Abstract
Objective: To determine whether hippocampal neurons and choroidal epithelial cells demonstrate a mitochondrial enzyme deficiency in AD more frequently than in normal aging.
Background: High levels of mutant mitochondrial DNA (mtDNA) cause a deficiency in cytochrome c oxidase (COX) (complex IV activity) because three of its 13 subunits are encoded for by mtDNA. In contrast, succinate dehydrogenase (SDH) (complex II activity) remains intact because all of its subunits are nuclear encoded. The histologic hallmark of cells containing high levels of mtDNA mutation in both primary mtDNA disorders and normal aging muscle is the presence of COX-deficient SDH-positive cells.
Methods: The authors applied a sequential histochemical method for COX and SDH to hippocampal sections in 17 AD and 17 age-matched control brains. This confers the advantages of both looking at individual cells in situ and measuring the actual mitochondrial complex activity rather than simply the complex quantity.
Results: COX-deficient SDH-positive hippocampal neurons and choroidal epithelial cells are more prevalent in patients with AD than in controls. In addition the COX-deficient SDH-positive choroidal cells are associated with an enlargement in size.
Conclusion: This increase in number of COX-deficient SDH-positive hippocampal pyramidal neurons and choroid epithelial cells provides strong evidence that a substantial mitochondrial enzyme activity defect occurs in individual cells more frequently in AD than in normal aging and that mitochondria may play a significant role in the pathogenesis of AD.
Mitochondrial abnormalities, either inherited or acquired, have been proposed to contribute to neurodegeneration in AD.1 Mitochondrial DNA (mtDNA) mutations result in a defect of oxidative phosphorylation so that cellular function is consequently compromised in metabolically active cells such as neurons. The accumulation of mtDNA mutations is accelerated by oxidative damage to mtDNA, which has been reported to be higher than normal in AD.2,3⇓
One of the central observations in patients with both inherited and acquired mtDNA defects is the presence of cells with decreased cytochrome c oxidase (COX) activity, but normal activity of succinate dehydrogenase (SDH). COX (complex IV of the respiratory chain) is composed of 13 subunits; the three main catalytic units (I to III) are encoded by mtDNA, and the remaining 10 subunits (IV to XIII) by nuclear DNA. SDH (complex II) is entirely encoded for by nuclear DNA. The sequential histochemical demonstration of COX activity, and then SDH activity in a single tissue section, allows the identification of COX-deficient SDH-positive cells. Loss of COX activity, but preservation of SDH activity, strongly suggests the involvement of mtDNA in the pathogenesis of the enzyme defect. This technique has the advantage of allowing us to study individual cells rather than tissue homogenates, which present major difficulties if the defect is not uniform. We have previously used this technique to investigate the CNS of a patient with an mtDNA disorder,4 age-related changes in the CNS5 and muscle6 of control subjects, and in ventral horn motor neurons in ALS.7 We have now applied this procedure to a series of AD patients and compared them to a series of age-matched controls, examining both the hippocampal cornus ammonis (CA) pyramidal neurons and the choroid plexus epithelial cells, both of which are known to accumulate pathologic changes in AD.
Methods and patients.
Neuropathology.
The brains of AD patients (12 women and 5 men), age 67 to 93 years (mean age = 82 years), and control brains (from 8 women, 9 men), age 66 to 87 years (mean age = 78 years), were selected from the Newcastle Brain Tissue Bank. All tissues were obtained after informed consent. The cerebral hemispheres were separated, half were frozen in dichloro,difluoromethane (Arcton-12, I.C.I., London, UK) cooled to −150 °C using liquid nitrogen, then stored at −80 °C. The other hemisphere was formalin fixed for neuropathologic studies. Four patients (three control subjects, one AD patient) lacked choroid plexus and four others (three control subjects, one AD patient) lacked hippocampi because of previous studies.
Classification was by case-note review and neuropathology, characterized using a panel of conventional stains and immunocytochemistry. All cases were confirmed to have an absence of other neurodegenerative disease. All the AD patients fulfilled the Consortium to Establish a Registry for Alzheimer’s Disease criteria8 for moderate to severe plaque and tangle pathology and had sporadic disease; genetic studies excluded presenilin or APP mutations.
Histochemical analysis.
Transverse frozen sections of anterior hippocampus and adjacent choroid plexus were cut at 20 μm at −15 °C, mounted on gelatin coated slides, air dried for 1 hour, then stored at −80 °C. Unfixed sections were used to preserve SDH activity. The sequential demonstration of COX9 and SDH10 activities was performed after air-drying at room temperature for 1 hour. The COX-SDH histochemistry was as previously described,10 the only modification being an increased incubation time for the COX reaction of 1 hour.5
Cell counting and morphometry.
Adjacent fields (0.45 × 0.66 mm) were counted manually using a ×20 objective lens, to determine the proportion of COX-deficient SDH-positive choroidal epithelial cells and hippocampal pyramidal neurons. Triplicate nonserial sections were examined in a blinded manner by two independent observers.
Random choroid plexus fields from five elderly patients were captured (×40 objective lens; Image Grabber PCI by Neotech) using a Zeiss (Jena, Germany) Axioplan-2 microscope and a JVC KY-F55B camera. The two-dimensional size of cells was calculated using Image Tool version 2.00 by UTHSCSA.
Statistics.
We used correlation as measured by linear regression, or the unpaired Student’s t-test for comparison between two groups using the software Graphpad (San Diego, CA) prism 2.01. Goodness of fit was measured by r2; a p value of <0.05 is considered significant.
Results.
Analysis of all patients for factors influencing the incidence of cytochrome c oxidase–deficient cells.
Variability in the postmortem delay in our patients ranged from 5 to 30 hours (median = 25 hours; mean = 23.47 hours). There was no difference (p = 0.87) between the postmortem delays of the AD patients (mean = 24.06 hours) and the control subjects (mean = 23.65 hours). When all the cases were examined there was no correlation between postmortem delay and percentage of COX-deficient SDH-positive CA pyramidal neurons (n = 30, r2 = 0.007, p = 0.67) or choroid plexus epithelial cells (n = 30, r2 = 0.00002, p = 0.98). With gender there was no correlation for percentage of COX-deficient CA hippocampal pyramidal neurons (men = 11, women = 19, p = 0.95) or percentage of COX-deficient choroid plexus epithelial cells (men = 11, women = 19, p = 0.99). There was also no correlation with age and percentage of COX-deficient CA neurons (r2 = 0.07, n = 30, p = 0.173), yet there was a correlation between age and percentage of COX-deficient choroid epithelial cells (r2 = 0.47, n = 27, p = 0.011). The AD and control cohorts, were, however, age-matched.
Cytochrome c oxidase–deficient hippocampal pyramidal neurons.
The mean age for the AD patients was 82.4 years (n = 16) and for control subjects, 78.4 years (n = 14). The AD patients had higher percentages of COX-deficient SDH-positive CA pyramidal neurons (figure 1, a and b) than the controls (p = 0.014) (figure 2A). There is, however, a considerable variation within the AD group, with some of the AD patients overlapping with the control subject percentages of COX-deficient CA neurons (figure 2A). When the regional distribution of the COX-deficient neurons in the four different CA regions was compared (see figure 2B), CA2 had significantly greater numbers (table). The morphology of the COX-deficient neurons appeared unchanged from that of the COX-positive neurons.
Figure 1. Sequential demonstration of cytochrome c oxidase (COX) and succinate dehydrogenase (SDH) activities in hippocampal CA neurons and choroid plexus epithelial cells. (a) Hippocampal CA2 region of a control subject, age 86 years. Two COX-deficient SDH-positive (blue) pyramidal neurons are shown against COX-positive (brown) neurons. (b) Hippocampal CA2 region of an AD patient age 83 years. Numerous COX-deficient pyramidal neurons are shown. (c) High-power magnification of a COX-deficient pyramidal neuron. (d) Choroid plexus from an 87-year-old control subject showing a few COX-deficient epithelial cells. (e) Choroid plexus from an AD patient age 83 years showing numerous enlarged COX-deficient cells. (f) High-power magnification of a COX-deficient choroid epithelial cell. Scale bars, 200 μm in a, b, d, and e and 50 μm in c and f.
Figure 2. (A) Column scatter-graph with mean values indicated to compare the percentages of cytochrome c oxidase (COX)–deficient CA pyramidal neurons of the control subjects (n = 14, mean = 0.58, SD = 0.42) versus the AD patients (n = 16, mean = 1.51, SD = 1.26) (p = 0.014). (B) Column scatter-graph with mean values indicated of percentages of COX-deficient choroid epithelial cells of control subjects (n = 14, mean = 2.71, SD = 1.83) versus AD patients (n = 16, mean = 4.34, SD = 1.89) (p = 0.024). (C) Column scatter-graph with mean values indicated to show the percentages of each CA region of the hippocampus in each AD case. For CA1, n = 16, mean = 1.32, SD = 1.55. For CA2, n = 16, mean = 2.90, SD = 3.12. For CA3, n = 16, mean = 1.21, SD = 1.04. For CA4, n = 16, mean = 1.08, SD = 0.80.
Comparison of COX-deficient neurons in different hippocampal regions
Cytochrome c oxidase–deficient choroid epithelial cells.
The mean age for the AD patients was 81.8 years (n = 16) and for the controls, 78.4 years (n = 14). The AD patients had higher percentages of COX-deficient SDH-positive choroid epithelial cells (see figure 1, c and d) than the controls (p = 0.024) (see figure 2C).
There was an overlapping but bimodal distribution of cell area measurements for COX-positive and COX-deficient SDH-positive choroid plexus cells (p < 0.001; figure 3). The COX-deficient cells were larger (mean = 208 μm2, n = 333) than the COX-positive cells (mean = 87 μm2, n = 610), with cells larger than 240 μm2 almost exclusively being COX deficient.
Figure 3. Distribution of two-dimensional size of both cytochrome c oxidase (COX)–positive (filled columns, n = 610, mean = 87 μm2) and COX-deficient (open columns, n = 332, mean = 208 μm2) choroid plexus epithelial cells. Unpaired Student’s t-test with Welch’s correction is correlated to a difference in size between COX-positive and COX-deficient choroidal cells (p < 0.0001).
Analysis of AD cohort for factors influencing the incidence of cytochrome c oxidase–deficient cells.
APOE typing.
In the AD cohort, nine had no ε4 alleles, six had one, and one patient had two. There was no significant difference in the quantities of COX-deficient CA neurons or choroidal cells between those with no ε4 alleles and those with one or two.
Sex.
Gender had no effect on the percentage of COX-deficient cells in the AD patients either in the hippocampus (p = 0.72) or the choroid plexus (p = 0.77) using the unpaired Student’s t-test.
Age at death.
No correlation was found with linear regression for age at death and percentage of COX-deficient CA pyramidal neurons in the AD patients (r2 = 0.05, p = 0.40). However, a correlation was found with age at death and the percentage of choroidal cells that were COX deficient (r2 = 0.42, p = 0.007).
Postmortem delay.
Postmortem delay had no effect on the percentage of COX-deficient cells in the AD patients alone either in the hippocampus (n = 16, p = 0.46) or in the choroid plexus (n = 16, p = 0.83) using linear regression.
Discussion.
COX-deficient SDH-positive cells are a histologic hallmark of mtDNA disorders. Previous studies have identified expansion of a single mtDNA mutation to very high levels within these cells in mtDNA disorders11 and normal aging muscle.6 Our results establish that COX-deficient SDH-positive hippocampal pyramidal neurons (see figure 2A) and choroidal epithelial cells (see figure 2C) accumulate in larger quantities in AD patients than in age-matched controls.
In the hippocampus the levels of COX-deficient neurons are greatest in CA2, an area characteristically spared in AD by amyloid plaques, neurofibrillary tangles (NFT), or neuronal loss. This is consistent with the neuropathology observed in a multiple mtDNA deletion disorder.4 In this patient, COX-deficient neurons were found throughout the CNS, with levels often >50%. Paradoxically, regions with high neuronal dropout, axonal loss, and demyelination showed the lowest percentages of COX-deficient neurons. We hypothesize that certain neuronal populations are more prone to neuronal death via mtDNA dysfunction, perhaps via apoptosis, in which mitochondria are known to play a pivotal role.12 This may be due to a varying degree of dependence on oxidative phosphorylation or local levels of neurotoxic substances. Neurons in CA1, CA3, and CA4 may have a lower threshold for degeneration with COX deficiency than in CA2, contributing to the higher neuronal loss in these regions.
The histochemical technique used in this study identifies cells with high levels of mtDNA dysfunction via a biochemical deficiency in the oxidative phosphorylation pathway. Because neurons are heavily dependent on oxidative phosphorylation for the production of adenosine triphosphate (ATP),13 it is difficult to see how a dramatic fall in COX activity would not have an impact on neuronal adenosine triphosphate production. MtDNA dysfunction could lead to some of the observed neuropathology in AD by enhancing the amyloidogenic pathway of APP that increases with energy metabolism inhibition14 or by increasing the level of oxidative stress promoting the aggregation of amyloid β.15 Indeed cybrid cells made from mtDNA of AD subjects have shown an increase in amyloid β secretion and development of Congo red–positive amyloid β deposits.16 Alternatively mtDNA mutations may be a separate neuropathologic entity contributing to the neurodegeneration or a secondary phenomenon to the other established neuropathologic markers of AD.
As shown in figure 2A, the AD patients had percentages of COX-deficient CA neurons ranging from 0.17 to 4.04%, overlapping with the controls. We analyzed the AD cohort for gender, postmortem delay, and APOE genotype but found no association with any of these variables. Only age at death had a significant correlation and only in the choroid epithelial cells. This association is probably an aging phenomenon5 rather than a specific subset of AD.
As shown in figure 3, the mean size of COX-deficient cells (mean = 208 μm2) is greater than twice the size of the COX-positive cells (mean = 87 μm2). This corresponds to previous reports in mtDNA disorders that the choroid plexus cells undergo oncocytic transformation, becoming packed with abnormal-looking mitochondria.17,18⇓ The similarity of the changes in morphology between the choroid epithelial cells in patients with mtDNA disorders and those with AD suggests a similar mechanism. Using both excretion and secretion mechanisms, the choroid plexus controls the CSF concentration of various organic anions, metals, proteins, and pharmaceutical agents.19,20⇓ Consequently, impairment in these cells that contain numerous mitochondria21 could have profound effects on the integrity of the blood–CSF barrier and the CSF concentration of neuronal protective and toxic substances. Indeed in a number of mtDNA disorders abnormalities in the CSF composition occur, with high levels of protein and lactate.22
In aging and particularly AD, the choroidal cells undergo morphologic alterations with epithelial atrophy and basement membrane thickening23 and accumulate densely packed intracellular straight and paired helical filaments that closely resemble cortical NFT and neuropil threads known as Biondi inclusions.24 The similar accumulation of COX-deficient choroidal cells with aging and in greater quantities in AD may bear some relationship as to why certain cells accumulate these changes.
Acknowledgments
Supported by grants from the Wellcome Trust and the MRC. Glaxo Wellcome funded E.B.
Acknowledgment
The authors thank Jean Dawes, Dr. Chris Morris, Dr. Gill Borthwick, Dr. Patrick Chinnery, and the neuropathology staff at the Newcastle General Hospital for their advice and help.
- Received October 23, 2000.
- Accepted March 13, 2001.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Hastening the Diagnosis of Amyotrophic Lateral Sclerosis
Dr. Brian Callaghan and Dr. Kellen Quigg
► Watch
Topics Discussed
Alert Me
Recommended articles
Spinal muscular atrophy-like picture, cardiomyopathy, and cytochrome c oxidase deficiency
Single large-scale mitochondrial DNA deletion in a patient with mitochondrial myopathy associated with multiple symmetric lipomatosis
Mitochondrial abnormalities in inclusion-body myositis
Partial depletion and multiple deletions of muscle mtDNA in familial MNGIE syndrome