Power Outage—Damaged Mitochondria In PET Changes, Early Aging

Mitochondria get more than their fair share of the blame for neurodegeneration, being linked to Alzheimer, Parkinson, Huntington diseases, and even normal aging. Recent papers offer little respite for these energy-generating organelles. In this week’s PNAS online, scientists report that throughout the AD brain, reduced expression of neuronal, nuclear-encoded mitochondrial genes mimics patterns of declining glucose utilization that precede the disease. The finding suggests that compromised mitochondria may be one of the earliest manifestations of AD. Another paper, published March 2 in Nature Genetics online, reveals that deletions in mitochondrial DNA drive premature aging in mice. Together, the papers suggest that whether from within or without, dysfunctional mitochondrial genes present a problem for successful aging.

For the first study, researchers led by Dietrich Stephan at the Translational Genomics Research Institute, Phoenix, Arizona, used laser-capture microscopy to isolate neurons from various regions of the brain, then applied microarray analysis to determine expression of mitochondrial genes. They focused on genes that keep the mitochondrial electron transport chain flowing. The study was designed to address an important question raised by prior fluorodeoxyglucose positron emission tomography (FDG-PET), that is, whether decreased glucose metabolism seen long before the onset of symptoms in people at increased risk for AD is due to problems in neurons, glia, or both (see, e.g., Reiman et al., 2005 and ARF related news story). The new findings indicate that neurons are at least partly to blame.

FDG-PET scans show that the earliest detectable losses in the cerebral metabolic rate for glucose (CMRgl) occur in the posterior cingulate cortex (PCC). The visual cortex and the superior frontal gyrus are relatively unaffected. First author Winnie Liang and colleagues found that deficient expression of electron transport proteins follows a similar pattern in AD patients. The researchers examined transcript levels of 80 nuclear-encoded electron transport chain (ETC) components, as well as 17 genes that code for inner and outer mitochondrial membrane translocases (the transporters that ferry proteins into the mitochondria). They compared expression patterns in AD samples with those from age-matched normal controls.

Liang and colleagues found that expression of 70 percent of the ETC genes was significantly lower in the PCC of AD patients compared to controls. In contrast, only 5 and 16 percent of the genes were underexpressed in the superior frontal gyrus (SFG) and visual cortex (VC), respectively. Reductions were intermediate in other brain areas, including the CA1 of the hippocampus (61 percent), the medial temporal gyrus (65 percent), and the entorhinal cortex (23 percent). In at-risk individuals, these regions show FDG-PET signal changes between those in the PCC and visual cortex. Some genes were overexpressed in the AD brain, though the fraction was small. Only 4 percent of the genes were overexpressed in the PCC, slightly more in the hippocampus (8 percent) and medial temporal gyrus (13 percent).

Expression of the translocase genes did not follow the same pattern as the ETC genes—fewer translocase genes were underexpressed in the PCC (35 percent) than in the medial temporal gyrus (47 percent) or the hippocampus (41 percent), for example, suggesting that only expression of the metabolic genes follows the pattern of FDG-PET signal loss. Similarly, in post-hoc fashion the researchers looked at expression of all genes, not just mitochondrial. They found significant reductions in all areas of the brain, but again the pattern did not match the CMRgl rates seen in PET studies. “These findings suggest that although metabolism genes may not be disproportionately affected in AD, the regional pattern of underexpressed metabolic genes (PCC>SFG and VC) more closely reflects the pattern of metabolic reductions observed in PET studies,” write the authors.

The findings help address the question of which cells are responsible for the loss in CMRgl seen in at-risk individuals, suggesting that neurons are at least partly to blame, though it leaves the potential contribution of glia to be decided. The authors claim that combined with previous PET studies, their current findings “raise the possibility that molecular processes involved in neuronal energy metabolism may be involved in the earliest pathogenesis of AD.” But as they note, it remains to be clarified if that loss in energy metabolism precipitates reductions in the activity of neurons or whether it is the other way around.

While reductions in ETC gene expression are related to regulation of nuclear-encoded genes, there is ample evidence that mitochondrial DNA exerts its own influence on aging and disease. Work from Nils-Goran Larsson and colleagues, at Karolinska University in Stockholm, Sweden, showed that impaired mitochondrial DNA proofreading shortens the lifespan of mice (see ARF related news story), a finding supported by work from Thomas Prolla’s lab at the University of Wisconsin. Prolla and colleagues found that shortened lifespan due to shoddy mitochondrial DNA proofreading goes hand-in-hand with excessive apoptosis and rampant mutations in mitochondrial DNA promoter regions (see ARF related news story). The latter is of particularly interest to AD researchers since similar mutations have been found in AD brain (see ARF related news story). Now Prolla, in collaboration with Lawrence Loeb at the University of Washington, Seattle, report that it is not just point mutations that drive premature aging in proofreading-deficient mice, but wholesale deletions. Interestingly, it is these deletions, rather than point mutations, that most dramatically affect the brain.

First author Marc Vermulst and colleagues came to this conclusion by studying polymerase gamma (Polga)-deficient mice, which most labs have used to address the mitochondria-age connection. Polga is the proofreading DNA (mtDNA) polymerase of the mitochondria. When it is mutated, as in Polga “mutator” mice, mtDNA point mutation rates grow three- to eightfold and the animals die prematurely. But last year Vermulst and colleagues showed that homozygous Polga mutator mice do not die by point mutation alone (see Vermulst et al., 2007); something else contributes to their early demise. To get to the bottom of this, the researchers looked at mtDNA deletions. They found that while these deletions occur at the same rate in heterozygous Polga mutator mice as in wild-type, there is a rapid acceleration of deletions in homozygous mutant mice. By 20 months, homozygotes have about five times as many mtDNA deletions. Specifically, these deletions are mostly between non-homologous sequences, indicating that Polga suppresses rearrangements between dissimilar stretches of DNA (most mtDNA deletions are between homologous sequences in WT and Polga heterozygotes). In fact, the authors found that there is almost a 100-fold increase in these “non-homologous” deletions in the homozygous mutator mice.

Given this new finding, the authors conclude that both deletions and point mutations must contribute to premature aging in homozygous animals. “However, as an increase in mtDNA deletions is associated exclusively with prematurely aging mice, and as mtDNA deletions affect thousands of base pairs per event, mtDNA deletions seem to be the most important force behind the shortened lifespan of Polgamut/mut mice,” write the authors.

To test whether this mattered functionally, the researchers studied expression of mitochondrial encoded cytochrome oxidase (COX). They found that there are no COX-negative cells in the duodenum, heart, or brain of wild-type animals at 15 months. In the duodenum of heterozygous mutator mice, about 20 percent of cells are COX-negative, but despite a documented 100-fold increase in DNA point mutations in these animals, there are no cells in the brain or heart that are COX-negative. This leads the authors to conclude that “a considerable difference exists in the rate at which mtDNA point mutations reach phenotypic expression between cell types.” The finding suggests that the brain may be relatively resistant to mtDNA point mutations.

The brain does not fare so well when it comes to deletions, however. In heart, duodenum, and brain, the researchers found many COX-negative cells in homozygous mutator mice. Interestingly, a high number of mtDNA deletions was recently linked to Parkinson disease (see ARF related news story), suggesting that for both aging and neurodegeneration, mtDNA deletions may be the mutations to avoid.—Tom Fagan

Comments

  1. Which Came First?
    For over a decade, imaging studies of cerebral metabolism have determined the cingulate gyrus to be a region of interest in Alzheimer disease (e.g., Alexander et al., 1997; Imamura et al., 1997), and in some of these papers the focus was turned to the posterior cingulate gyrus (e.g., Reiman et al., 1996; Salmon et al., 2000). Thus, it is not surprising that the present manuscript should focus on energy metabolism in the posterior cingulate gyrus. However, these prior imaging studies have been unable to answer the question, Was the decreased metabolism of the posterior cingulate in AD attributable to neurons, to glia, or other cellular elements within the local region? To the great credit of the authors of the present paper, they have effectively addressed this major limitation of more global approaches by laser capturing single neurons of the posterior cingulate gyrus (as well as five other regions) and analyzing the expression of 80 nuclear genes related to energy metabolism. The essence of their findings is that 72 percent of metabolism-related genes showed reduced expression in the posterior cingulate gyrus. In middle temporal gyrus and hippocampus, 69 percent and 64 percent, respectively, of metabolism genes showed reduced expression. Surprisingly, in the entorhinal cortex, a major affected area in AD, only 23 percent did. This cannot be attributed to neuron loss since the data were derived from neurons remaining. Might it relate to normal aging versus AD? Other, traditionally less affected areas also showed a lower percentage of genes reduced.

    The choice of Affymetrix arrays (U133 Plus 2.0) dictated other aspects of the study, most notably the need to sum RNA collected from about 100 neurons (exact number not provided) from each brain region. Thus, although different cell types are not confounded, and although picked cells were characterized as lacking thioflavin S staining, summing expression of ~100 cells must confound cells in different states of AD pathology. For example, although negative thioflavin S staining demonstrates absence of frank neurofibrillary tangles, variation in other states of tau was admissible.

    Unfortunately, the sensitivity of current AFFY arrays is limited so that summation of material from large numbers of cells is required, even after two rounds of amplification, as was done here. From a statistical point of view, such summation discards information about one source of variance. cDNA arrays have, however, been demonstrated to have sufficient sensitivity that material from individual cells can provide reliable results (e.g., Chow et al., 1998). The ability to profile transcripts of single cells (of any specified type) in addition to allowing immunohistochemical (or otherwise) characterization also provides the option of borrowing from single unit electrophysiology the concept of population reconstruction of neuronal responses, which then could be correlated with psychophysical function (e.g., Mountcastle et al., 1967). Of course, the dimensionality of data in the case of single unit electrophysiology is much simpler than in the case of array data, where one may be dealing with hundreds or thousands of variables—a daunting statistical task that may require excessively large numbers of cells.

    Reviewers generally require validation of array data. Validation may be considered as addressing either technical variability or biological variability. Quantitative RT-PCR of the same samples as those used to obtain array data addresses technical variability, but not biological variability. If different samples are used, both technical and biological variability are addressed. By validating with Western blotting, Liang et al. have added another aspect to validation that deals with potential non-linearity of the relationship between message and protein expression. By apparently using the same samples for Western blotting as were used to obtain array data, the authors miss a chance to simultaneously address biological (sample-to-sample) variability.

    The authors appropriately point out the possibility that the reduced expression of transcripts related to metabolism may reflect reduced neuronal activity related to loss of synapses. So, the question remains whether reduced synaptic activity precedes reduced energy metabolism or whether reduced energy metabolism precedes reduced synaptic activity. Which is cause and which is effect? It is still a chicken-or-egg question.

    References:

    Alexander GE, Furey ML, Grady CL, Pietrini P, Brady DR, Mentis MJ, Schapiro MB. Association of premorbid intellectual function with cerebral metabolism in Alzheimer's disease: implications for the cognitive reserve hypothesis. Am J Psychiatry. 1997 Feb;154(2):165-72. PubMed.

    Chow N, Cox C, Callahan LM, Weimer JM, Guo L, Coleman PD. Expression profiles of multiple genes in single neurons of Alzheimer's disease. Proc Natl Acad Sci U S A. 1998 Aug 4;95(16):9620-5. PubMed.

    Imamura T, Ishii K, Sasaki M, Kitagaki H, Yamaji S, Hirono N, Shimomura T, Hashimoto M, Tanimukai S, Kazui H, Mori E. Regional cerebral glucose metabolism in dementia with Lewy bodies and Alzheimer's disease: a comparative study using positron emission tomography. Neurosci Lett. 1997 Oct 10;235(1-2):49-52. PubMed.

    Mountcastle VB, Talbot WH, Darian-Smith I, Kornhuber HH. Neural basis of the sense of flutter-vibration. Science. 1967 Feb 3;155(3762):597-600. PubMed.

    Reiman EM, Caselli RJ, Yun LS, Chen K, Bandy D, Minoshima S, Thibodeau SN, Osborne D. Preclinical evidence of Alzheimer's disease in persons homozygous for the epsilon 4 allele for apolipoprotein E. N Engl J Med. 1996 Mar 21;334(12):752-8. PubMed.

    Salmon E, Collette F, Degueldre C, Lemaire C, Franck G. Voxel-based analysis of confounding effects of age and dementia severity on cerebral metabolism in Alzheimer's disease. Hum Brain Mapp. 2000 May;10(1):39-48. PubMed.

    View all comments by Paul Coleman

  2. This microarray-based study is quite provocative. This research group employed findings gleaned from earlier positron emission tomography (PET) and cerebral metabolic rate for glucose (CMRgl) studies in vulnerable regions of cerebral cortex and hippocampus to drive a microarray study evaluating nuclear-encoded genes for mitochondrial/metabolic function. Specifically, principal neurons were microaspirated via laser capture microdissection (LCM) from well-characterized normal aged subjects and persons diagnosed with Alzheimer disease (AD). The genes were then subjected to Affymetrix gene chip analysis. Essentially, the group isolated relatively pure populations of pyramidal neurons that did not contain neurofibrillary tangle (NFT) pathology from six brain regions. This included areas known to be affected early in AD based upon PET/CMRgl studies such as posterior cingulate cortex (PCC), as well as relatively spared regions such as visual cortex. To my knowledge, this is one of the few microarray studies that evaluated single populations of neocortical neurons from cingulate and visual cortex (among other temporal and hippocampal areas) within postmortem human brain. The datasets generated from these cases have the potential to be extremely exciting and informative on many levels. Based upon the author’s hypothesis that metabolic changes occur early in AD pathogenesis, they concentrated on assessing approximately 80 of the nuclear genes encoding subunits of the mitochondrial electron transport chain pathway. The experimental design enabled a solid microarray analysis that was validated at the protein level via immunoblot analysis.

    Interestingly, the data provide molecular and cellular evidence for early metabolic dysfunction in the PCC that truly corroborates findings based on regional imaging techniques in living patients. Significant decrements were found in the class of nuclear-encoded mitochondrial genes in vulnerable PCC, middle temporal gyrus, and hippocampus. Lesser changes were found in these same transcripts within entorhinal cortex, visual cortex, and superior frontal gyrus. These data were validated in the PCC via immunoblot analysis for five electron transport chain subunits, indicating that gene changes are quite robust.
    Importantly, the LCM-based paradigm enabled the investigators to demonstrate that the downregulation was fairly neuron-specific, not a contamination effect from glial or vascular cells.

    In summary, this study illustrates the power of combining LCM-based high resolution in postmortem human brain samples with microarray analysis for quantitative analysis of relevant classes of transcripts that may have profound implications for the integration of functional imaging studies in living subjects with neuropathological investigations to understand the initiation and progression of AD. It would be highly desirable in future studies to combine microarray assessment of nuclear-encoded electron transport chain genes with antemortem functional imaging and assessment of cognitive performance across the progression of dementia (e.g., aged normal, mild cognitive impairment, and AD subjects). Moreover, an evaluation of cortical interneuron populations within several of the cortical regions evaluated in this manuscript (including the PCC and occipital/temporal cortices) would shed light on the contribution of inhibitory neurons to AD pathogenesis at the molecular and cellular level.

    View all comments by Stephen D. Ginsberg

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Functional Imaging Gives Early Glimpse of AD
  2. Poor Proofreading May Shorten Your Lifespan
  3. Cumulating Mitochondrial Mutations Hit Apoptosis Hardest
  4. Promoter Bashing—Mitochondrial Ones Damaged in AD Brain
  5. DNA Deletions Sap Mitochondria in Parkinson Neurons

Paper Citations

  1. Reiman EM, Chen K, Alexander GE, Caselli RJ, Bandy D, Osborne D, Saunders AM, Hardy J. Correlations between apolipoprotein E epsilon4 gene dose and brain-imaging measurements of regional hypometabolism. Proc Natl Acad Sci U S A. 2005 Jun 7;102(23):8299-302. PubMed.
  2. Vermulst M, Bielas JH, Kujoth GC, Ladiges WC, Rabinovitch PS, Prolla TA, Loeb LA. Mitochondrial point mutations do not limit the natural lifespan of mice. Nat Genet. 2007 Apr;39(4):540-3. PubMed.

Further Reading

Primary Papers

  1. Vermulst M, Wanagat J, Kujoth GC, Bielas JH, Rabinovitch PS, Prolla TA, Loeb LA. DNA deletions and clonal mutations drive premature aging in mitochondrial mutator mice. Nat Genet. 2008 Apr;40(4):392-4. PubMed.
  2. Liang WS, Reiman EM, Valla J, Dunckley T, Beach TG, Grover A, Niedzielko TL, Schneider LE, Mastroeni D, Caselli R, Kukull W, Morris JC, Hulette CM, Schmechel D, Rogers J, Stephan DA. Alzheimer's disease is associated with reduced expression of energy metabolism genes in posterior cingulate neurons. Proc Natl Acad Sci U S A. 2008 Mar 18;105(11):4441-6. Epub 2008 Mar 10 PubMed.

Annotate

To make an annotation you must Login or Register.