Astrocyte- and Microglia-Specific Mitochondrial DNA Deletions Levels in Sporadic Alzheimer’s Disease

Abstract

Oxidative stress is implicated in the pathogenesis of neurodegenerative diseases, including sporadic Alzheimer’s disease (AD). Mitochondrial DNA (mtDNA) deletions are markers of oxidative damage with an age-dependent accumulation. In a previous study, we analyzed mtDNA levels in diverse neuronal cell types in order to unravel the impact of oxidative stress in brains of AD patients. The aim of this study was to identify possible correlations between mtDNA deletion levels of selected astrocytes and microglia from three brain regions with different vulnerability to AD pathology and different stages of disease compared to controls. Our results reflect a higher vulnerability of hippocampal astrocytes and microglia to oxidative stress compared to other brain regions, such as cerebellum and brainstem.

Keywords

Alzheimer’s disease astrocytes brainstem cerebellum hippocampus microglia mitochondrial DNA oxidative stress selective vulnerability

INTRODUCTION

The most common neurodegenerative disease, sporadic Alzheimer’s disease (AD), progresses in an anatomically stereotyped pattern [1]. Some brain regions are highly vulnerable to typical pathological changes, such as amyloid-β (Aβ) plaques and neurofibrillary tangles (NFTs), whereas other regions remain intact even in advanced disease stages. The cerebellum seems to be less affected by AD-related changes, with only few Aβ deposits found in the molecular layer without detection of NFTs, while on the contrary, the hippocampus demonstrates the highest vulnerability [1 –3]. The involvement of the brainstem in AD is still discussed controversially. In early stages of the disease, NFTs occur in the dorsal raphe nucleus and, moreover, hyperphosphorylated tau proteins (pretangles) appear in the locus coeruleus in individuals younger than 30 years, in the absence of NFTs or Aβ [4 –6]. These findings may suggest that AD originates in some nuclei of the brainstem before spreading to cortical brain regions. A clinical method for early detection of AD, which is based on electric stimulation of the auricular branch of the vagus nerve, supports these findings [7]. Sensitive fibers of vagus nerve project among others in the nucleus tractus spinalis nervi trigemini of brainstem [8]. Vagus somatosensory evoked potentials in clinical trials showed longer latencies in AD individuals compared with age-matched controls [9].

In the last decades, many theories to the pathogenesis of AD were formulated (e.g., the amyloid cascade hypothesis, tau hypothesis, hypothesis of prion-like transmission, synapse pathology), but none of these hypotheses were able to explain entirely the pathogenesis of AD [10 –12]. Another theory postulates an age-dependent decline of organ function as a consequence of oxidative damage caused by reactive oxygen species arising from oxidative phosphorylation in mitochondria [13 –15]. The mitochondrial DNA (mtDNA) is particularly susceptible to oxidative damage because it is located in close proximity to the respiratory chain, and displays reduced repair capacity relative to nuclear DNA, resulting in an accumulation of point mutations and deletions [16, 17], including the 4977 base pairs deletion which has attracted the attention of the aging research community because of its age-dependent accumulation in postmitotic cells [18]. In various neurodegenerative diseases, oxidative damage has been suggested as a common process to explain selective vulnerability of certain brain regions [19]. Quantification of mtDNA deletions during the course of AD showed different findings in previous studies. Some authors found a distinct accumulation in brains of AD patients younger than 75 and a decrease in AD individuals above the age of 75, compared with controls [20]. Other authors found no difference in mtDNA levels between AD individuals and aged-matched controls [21]. These data were exclusively collected on tissue-homogenates containing all cell types of brain parenchyma, whereas cell-specific analyses based on laser microdissection are very scarce and merely restricted to neurons. In our previous study, we could not show a coherent pattern of neuronal mtDNA deletion levels correlating with the degree of selective vulnerability to AD [22]. In the last decades, the importance of astrocytes in the pathogenesis of neurodegenerative diseases has increased and the classical view of them as simple passive cellular support to neuronal function has begun to change. Astrocytes are now considered as full-fledged participants in brain circuitry and processing, and display a large spectrum of functions at the cell level, such as the formation, maturation, and elimination of synapses, ionic homeostasis, clearance of neurotransmitters, regulation of extracellular space volume, and modulation of synaptic activity and plasticity [23].

The other brain-specific cell type, namely microglia, migrate from the yolk bag into the brain during embryogenesis, act as endogenous neuronal immune system, and play important roles in neurogenesis, synaptic plasticity, and neuronal death [24]. Furthermore, microglia seem to play an important role in the pathogenesis of AD: microglia secrete proteolytic enzymes that degrade Aβ, and express receptors that promote the clearance and phagocytosis of Aβ. But microglia are also activated by Aβ and, once activated, release inflammatory mediators that could promote AD pathology [25]. Other studies support the idea that progressive, aging-related microglial degeneration and loss of microglial neuroprotection, rather than induction of microglial activation, contributes to the onset of sporadic AD [26, 27].

The influence of oxidative stress on astrocytes and microglia in terms of mtDNA deletions was not yet considered in cell-specific analyses; therefore, the aim of the present study was to identify possible correlations between cell-specific mtDNA deletion levels of astrocytes and microglia from three brain regions with different vulnerability to AD pathology and different stages of disease compared to controls.

MATERIALS AND METHODS

Tissue samples

Frozen brain tissue from three different brain regions (brainstem, cerebellum, and hippocampus) from neuropathological characterized AD with different Braak and Braak stages and age-matched controls were collected from Brain Bank Center Wuerzburg, a member of the BrainNet Europe Brain Bank Consortium Network (http://www.brainnet-europe.org/). Human brains were obtained with the consent of the next of kin and according to the guidelines of the National and Local Ethics Committees. The study was also approved by the Local Ethics Committee of the University of Wuerzburg (internal application number 99/11) and was performed in consent with the ethical standards described in the most recent version of the Declaration of Helsinki. Three groups were created: control cases (n = 10; Braak and Braak stage 0– I; no other neuropsychiatric disease), intermediate AD cases (n = 13; Braak and Braak stages III– IV), and late AD cases (n = 9; Braak and Braak stages V– VI). The postmortem intervals (PMIs) varied from 2 to 72 h, with no significant difference between groups: the mean±standard deviation (SD) for PMIs was 28.4±22.5 h in the control group, 33.8±18.3 h in the intermediate AD group, and 26.6±12.8 h in the late AD group. The tissue was uninterrupted stored at – 80°C between 3 months and 10 years. The age ranged from 62 to 80 years in the control group (69.4±5.0 years), from 62 to 92 years in the intermediate group (78.7±8.6 years), and from 56 to 86 years in the late AD group (72.2±10.0). There was a significant difference in mean age between the control group and the intermediate group (p < 0.05). The male/female ratio in the control group was 7/3, in the intermediate group 6/7, and in the late 3/6. Demographical data of the cases are shown in Table 1.

Table 1

Demographical data

No.	Group	Age (y)	PMI (h)	Sex	Braak Stage
1	Control	80	23	m	I
2	Control	70	40	f	I
3	Control	62	48	f	I
4	Control	67	15	m	I
5	Control	64	9	m	I
6	Control	68	11	m	I
7	Control	70	2	m	I
8	Control	73	16	f	I
9	Control	69	72	m	0
10	Control	69	48	m	I
11	Interm.AD	77	24	f	III
12	Interm.AD	78	38	f	IV
13	Interm.AD	75	20	f	III
14	Interm.AD	85	46	f	III
15	Interm.AD	64	24	m	III
16	Interm.AD	86	61	f	III
17	Interm.AD	80	60	m	III
18	Interm.AD	92	20	f	III
19	Interm.AD	86	63	m	IV
20	Interm.AD	85	25	m	IV
21	Interm.AD	76	6	m	III
22	Interm.AD	77	28	f	III
23	Interm.AD	62	24	m	III
24	Late AD	74	26	f	V
25	Late AD	65	24	f	V
26	Late AD	80	24	f	VI
27	Late AD	83	22	f	VI
28	Late AD	70	39	m	VI
29	Late AD	62	n.a.	m	VI
30	Late AD	86	21	f	VI
31	Late AD	56	7	f	VI
32	Late AD	74	50	m	VI

AD, Alzheimer’s disease; f, female; m, male; n.a., not available; PMI; postmortem interval; Interm., Intermediate.

Immunohistochemical stainings

Frozen sections (10μm thickness) from each of the three brain regions were cut using a CM1950 cryostat (LEICA, Wetzlar, Germany), mounted on polyester membranes (Leica, Wetzlar, Germany), and fixed in ice-chilled acetone. Sections were processed for immunohistochemical single-staining using primary antibodies, glial fibrillary acidic protein (GFAP) (mouse monoclonal, 1:5; DAKO, Glostrup, Denmark) for detecting astrocytes, and CD68 (mouse monoclonal, 1:25; DAKO, Glostrup, Denmark) for detecting microglia, respectively soluted in a phosphate buffered saline with 0.1 % Diethylpyrocarbonat (PBS-DEPC, ph 7.4) including 0.1% bovine serum albumin. For visualizing the primary antibody binding sites, the sections were treated with Envision+ System-Horse Radish Peroxidase (HRP) (DAKO, Glostrup, Denmark) and HistoGreen HRP-Substrate Kit (Linaris, Dossenheim, Germany). Between the incubation steps, the sections were rinsed in PBS-DEPC. The last rinse step consisted of merely 0.1% DEPC-H₂O. The polyester membranes dried completely in a desiccator for 45 min and were used for laser microdissection directly after staining.

Laser microdissection and DNA extraction

The membranes were microdissected using the LMD6000 laser microscope (Leica, Wetzlar, Germany). 150 astrocytes and 150 microglia cells respectively were microdissected within the following localization: hippocampal gyrus dentatus, brainstem in the vicinity of the nucleus tractus spinalis nervi trigemini, and in the cerebellar white matter, exemplary depicted in (Fig. 1). The dissected cells were transferred in RLT buffer (Qiagen, Hilden, Germany) and stored at – 80°C until DNA extraction. The DNA was isolated using the All Prep DNA/RNA Micro Kit according to the manufacturer’s protocol (Qiagen, Hilden, Germany). The DNA was eluted twice with 30μl elution buffer to obtain a final volume of 60μl. The purity of the extracted DNA was proved by measuring the 280/260 and 280/230 ratio using a NanoDrop Spectrophotometer (ThermoFisher Scientific, Erlangen, Germany). Only samples with a ratio close to 1.5–1.8 and 1.8–2.0, respectively, were used for this study.

Fig. 1.

GFAP staining for astrocytes (upper) and CD68 staining for microglia (lower) before and after laser microdissection, respectively, exemplary depicted for the hippocampus. Scale bar: 100μm.

Real time quantitative polymerase chain reaction

A duplex quantitative real time PCR (qPCR) was carried out to quantify the mtDNA deletion levels, as previously described by Bender et al. [28, 29]. According to this, ND1 can be used as an internal control, whereas a drop in ND4 quantity serves as a measure for the abundance of deletions, as the majority of mtDNA deletions comprises the ND4 region and spares the ND1. The qPCR was performed using the TaqMan Gene Expression Master-mix (Life Technologies, Darmstadt, Germany), and the following primers and probes (a) ND4: forward primer nucleotide (nt) 12087–12109, reverse primer nt 12170–12140 (Eurofins MWG, Elsberg, Germany); VIC-labeled probe nt 12111–12138 (Life Technologies, Darmstadt, Germany); and (b) ND1: forward primer nt 3485–3504, reverse primer nt 3553–3532, FAM-labeled probe nt 3506–3529 (Eurofins MWG, Elsberg, Germany). The primers and probes specifications are summarized in Table 2. They were used in a final concentration of 900 nM for primers and 250 nM for probes. ND4 served as the target gene while ND1 officiated as the reference gene which is not affected by the 4977 bp deletion. As described by Elstner et al. [30], this method shows high correlation with mtDNA deletion quantification by Southern blot and also with the original method [31]. Reactions were carried out using a CFX C1000^TM real time system (BioRad, Muenchen, Germany) in a total volume of 13μl, using 2μl of template DNA. Run conditions consisted as follows: initial phase of activation for 2 min at 50°C, followed by 10 min at 95°C; 50 cycles including a 15 s denaturation step at 95°C, and primer annealing and elongation for 1 min at 60°C. Each sample was analyzed in triplicate with mean values calculated at the end of the run. Non-template controls exhibited no Cq values or values distant from those of the samples. The individual efficiency of each primer pair was determined using LinReg [32]; efficiencies for astrocytes (2) were 1.783 for ND4 and 1.698 for ND1; efficiencies for microglia (3) were 1.765 for ND4 and 1.697 for ND1. Based on Pfaffl [33] the efficiencies-corrected model was used to quantify the mtDNA deletion levels:

Table 2

Primers and probes specifications

Description	Sequence (5^′ ⟶ 3^′)	Label	Manufacturer
ND1 FP	CCC TAA AAC CCG CCA CAT CT		Eurofins MWG Operon¹
ND1 RP	GAG CGA TGG TGA GAG CTA AGG T		Eurofins MWG Operon¹
ND1 probe	CCA TCA CCC TCT ACA TCA CCG CCC	5^′FAM	Eurofins MWG Operon¹
ND4 FP	CCA TTC TCC TCC TAT CCC TCA AC		Eurofins MWG Operon¹
ND4 RP	CAC AAT CTG ATG TTT TGG TTA AAC TAT ATT T		Eurofins MWG Operon¹
ND4 probe	CCG ACA TCA TTA CCG GGT TTT CCT CTT G	5^′VIC	Life Technologies²

FP, forward primer; RP, reverse primer; ¹Ebersberg, Germany; ²Darmstadt, Germany.

$R = \frac{1, 783_{ND 4}^{CqND 4 (Kalibrator - Preobe)}}{1, 698_{ND 1}^{CqND 1 (Kalibrator - Preobe)}}$ (1) $R = \frac{1, 765_{ND 4}^{CqND 4 (Kalibrator - Preobe)}}{1, 697_{ND 1}^{CqND 1 (Kalibrator - Preobe)}}$ (2)

Calibration was performed as described previously [22]. Deletion levels were calculated by Microsoft Excel 2007.

Statistical analysis

Statistical analysis of ND4 deletion levels in the different subject groups and brain regions was performed using the none-parametric Kruskal-Wallis and the post-hoc Mann-Whitney test. For analysis of covariance we used the one-way ANCOVA (StatView5.0). Significance was set at p < 0.05. Values are shown as mean±SD.

RESULTS

Astrocyte-specific mtDNA deletion levels

Hippocampal astrocytes showed a non-significant decrease of the mtDNA deletions level from control (38.0% ±6.46) to intermediate AD group (35.86% ±9.9), followed by an increase in the late AD group (40.57% ±9.45) (H(2) = 0.647, p = 0.724). In the nucleus tractus spinalis nervi trigemini, there was a non-significant increase from control group (28.01% ±8.65), to intermediate AD group- (33.38% ±13.3), followed by a decrease of 5.9% in the late AD group compared to the intermediate AD group (H(2) = 1.587, p = 0.452). The cerebellar astrocytes showed non-significant increase from controls (30.84% ±9.59) to intermediate AD cases (35.18% ±8.84) as well as in the late AD group with 36.45% ±11.97 (H(2) = 0.980, p = 0.613) (Fig. 2).

Fig. 2.

mtDNA deletion levels (%) in astrocytes from hippocampus (HC), nucleus tractus spinalis nervi trigemini of the brainstem (BS), and cerebellum (CB) in control, intermediate, and late AD. Error bars indicate standard deviations. Statistical analysis was performed using Kruskal-Wallis and the post-hoc Mann-Whitney test. ^*p < 0.05, ⁺0.1 > p > 0.05.

In the control group, a significant difference between the three brain regions was observed (H(2) = 6.006, p = 0.0496), in which the hippocampus showed statistical significant higher deletion levels compared to brainstem (U = 17, p = 0.0222), and nominal significance higher levels compared to the cerebellum (U = 26, p = 0.0696).

Regarding the brain regions independent of AD progression, significantly higher rates of mtDNA deletion were observed between astrocytes of the hippocampus and nucleus tractus spinalis nervi trigemini (U = 209, p = 0.0046) (Fig. 3). For this purpose, the DNA from all cases was pooled region-specific and the deletions levels were calculated as described.

Fig. 3.

mtDNA deletion levels (%) in astrocytes from hippocampus (HC), nucleus tractus spinalis nervi trigemini of the brainstem (BS) and cerebellum (CB). Error bars indicate standard deviations. Statistical analysis was performed using Kruskal-Wallis and the post-hoc Mann-Whitney test., ^*p < 0.05. A one-way ANCOVA was conducted to determine a statistically significance difference between brain regions (HS, BS, CB) on the mtDNA deletion levels (%) in astrocytes controlling for age and PMI, F(1,85) = 42.705, p < 0.0001. The data present pooled results of controls and AD.

Microglia-specific mtDNA deletion levels

Hippocampal microglia deletion levels showed a non-significant decrease from control group (33.98% ±9.22) over intermediate AD group (29.63% ±9.63) to late AD group (28.12% ±11.77) (H(2) = 1.327, p = 0.515). In the nucleus tractus spinalis nervi trigemini, there was a negligible decrease from control (22.87% ±9.65), to intermediate AD group (21.23% ±10.56), followed by an increase of 5.16% in the late AD group compared to the intermediate AD group (H(2) = 0.357, p = 0.837). The cerebellar microglia showed a low-level of deletions with moderate increase from controls (22.01% ±10.83) over intermediate AD cases (22.33% ±15.28) to late AD group (25.89% ±8.09) (H(2) = 1.697, p = 0.428) (Fig. 4). Similarly to the astrocytes results, in the control group, a significant difference between the three brain regions was observed (H(2) = 6.513, p = 0.0385), in which the hippocampus showed statistical significant higher deletion levels compared to brainstem (U = 14, p = 0.0209), as well as a significance higher levels compared to the cerebellum (U = 14, p = 0.0343).

Fig. 4.

mtDNA deletion levels (%) in microglia from hippocampus (HC), nucleus tractus spinalis nervi trigemini of the brainstem (BS) and cerebellum (CB) in control, intermediate and late AD. Error bars indicate standard deviations. Statistical analysis was performed using Kruskal-Wallis and the post-hoc Mann-Whitney test. ^*p < 0.05.

Regarding the brain regions independent of AD progression, significant differences were observed between microglia of the hippocampus and nucleus tractus spinalis nervi trigemini on the one hand (U = 161, p = 0.0033) and microglia of the hippocampus and cerebellum on the other hand (U = 203, p = 0.0138). For this purpose, the DNA from all cases was pooled region-specific and the deletions levels were calculated as described. The results are summarized in (Fig. 5).

Fig. 5.

mtDNA deletion levels (%) in microglia from hippocampus (HC), nucleus tractus spinalis nervi trigemini of the brainstem (BS) and cerebellum (CB). Error bars indicate standard deviations. Statistical analysis was performed using Kruskal-Wallis and the post-hoc Mann-Whitney test, ^*p < 0.05. A one-way ANCOVA was conducted to determine a statistically significance difference between brain regions (HS, BS, CB) on the mtDNA deletion levels (%) in astrocytes controlling for age and PMI, F(1,77) = 5.539, p < 0.021. The data present pooled results of controls and AD.

DISCUSSION

The frequency and importance of mtDNA deletions in patients with neurodegenerative diseases, especially in those with AD, has not been fully established. Deletions of mtDNA are observed most frequently in aging postmitotic cells like neurons, whereas in dividing cells such as astrocytes and microglia this phenomenon was not yet studied. The aim of this study was to identify possible correlations between mtDNA deletion levels of selected astrocytes and microglia from three brain regions with different vulnerability to AD pathology and different stages of disease compared to controls.

We found a tendency to an increased level of deletions in astrocytes of the hippocampus in patients with advanced disease compared to controls. As a major homeostatic cell type in the central nervous system, astrocytes have been implicated in AD pathology by their capacity to undergo hypertrophy and to accumulate small amounts of Aβ with the level of accumulation correlating to the severity of AD pathology [34]. Oxidative stress, a hallmark of AD, was linked to amyloid toxicity, and the astrocytes seem to take part in this process, since amyloid stimulates reactive oxygen species production in these cells [35]. It is therefore to be assumed that the increased oxidative stress leads to damage in mtDNA and to an accumulation of deletions especially in hypertrophic astrocytes near the amyloid plaques. In contrast, astrocytes further away from amyloid plaques tend to get atrophic especially in advanced disease, as shown in a transgenic mouse model [36]. This could explain our findings in brainstem, which showed a slight decrease in mtDNA deletions level in late AD stages compared to intermediate stages.

The cerebellum, a region with low degree of vulnerability regarding tau and Aβ pathology, exhibited in our study unexpected a slight stage-dependent increase of the mtDNA deletions level. Since some authors [37] reported a gradual increase of astrogliosis in human cerebellar tissue from AD patients during the progression of the disease, it is assumable that the reactive hypertrophic astrocytes accumulate mtDNA damages due to the oxidative stress reflecting in higher levels of deletions.

Regarding the analyzed brain regions independent of AD stage, we found significant higher deletion levels in hippocampus compared to brainstem that was particularly significant in the healthy control astrocytes. This result could be interpreted as an indicator for the overall high vulnerability of this region to various pathological conditions, not only to AD changes but also to hypoxia, which is generally associated with aged brain and prolonged agonal state.

The microglia showed a different pattern of mtDNA deletion levels compared to astrocytes. In hippocampus, we observed a continuous decrease of the deletion levels during the disease progression, with lowest levels in late stages. It is known that microglia show altered morphology and reduced arborization in human brain during aging and AD. Davies et al. [38] demonstrated that microglial cell processes were reduced in length, showed less branching and reduced arborized area with aging and that this process is even stronger in AD compared to aged-matched controls. It is possible that these alterations on microglial cells correlate with a reduction of the number of mitochondria and also with a decrease of the mtDNA deletion levels during disease progression.

In cerebellum, by contrast, we observed a continuous but slight increase of deletion levels during the course of disease. These results could be explained by the possibility that, in a region with low vulnerability to AD changes such as cerebellum, the microglia suffer a process of activation rather than senescence and these changes may be also triggered by premortem hypoxia which is often augmented in cerebellum [39]. Based on highest brainstem accumulation of mtDNA deletions in late stage of AD, this phenomenon seems do not represent an early alteration in AD. Similarly to the astrocytes, we found overall significant higher deletions levels in hippocampus compared to brainstem and cerebellum regardless to AD stage. This was observed to be most pronounced in the control microglia.

Based on our results, it is supposed that accumulation of mtDNA deletions do not correlate with AD, rather they could be related to the aging process.

Conclusion

According to the literature, the hippocampus, especially the CA1 sector, shows an overall higher vulnerability of neuronal cells toward different pathologies, but the possible causes and the pathogenesis of this selective neuronal loss is still not fully elucidated despite intensive efforts of scientists in the last decades. Furthermore, even less is known about the behavior and the role of astrocytes and microglia in human brains during the course of AD, or about their mtDNA alterations. Our results reflect a generally higher vulnerability of hippocampal astrocytes and microglia to accumulate mtDNA deletions compared with other brain regions, such as cerebellum and brainstem. These findings do not explain the differing vulnerabilities of the analyzed brain regions toward AD pathology and seems rather to be related to the aging process.

Astrocytes and microglia gain a growing importance in the pathogenesis of neurodegenerative disease, especially of AD, and many studies to date are focused on the clarification of their role in neurodegeneration. This could be the key for more effective therapies for AD in near future. Our study is to our knowledge the first cell-specific analysis to assess the mtDNA deletion levels of astrocytes and microglia in human AD brains. To obtain more representative results regarding the significance, distribution,and level of mtDNA deletions over the course of AD progression, further analyses on larger patient cohort and additional brain regions are needed.

Footnotes

ACKNOWLEDGMENTS

This work was supported by a grant from the Interdisciplinary Center for Clinical Research (IZKF), University of Wuerzburg, Germany, and by a grant from Edda-Neele-Stiftung Frankfurt/Main, Germany.

Authors’ disclosures available online ().

References

Braak

, Braak

(1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82, 239–259.

Larner

(1997) The cerebellum in Alzheimer’s disease. Dement Geriatr Cogn Disord 8, 203–209.

Mavroudis

, Fotiou

, Adipepe

, Manani

, Njau

, Psaroulis

, Costa

, Baloyannis

(2010) Morphological changes of the human purkinje cells and deposition of neuritic plaques and neurofibrillary tangles on the cerebellar cortex of Alzheimer’s disease. Am J Alzheimers Dis Other Demen 25, 585–591.

Grinberg

, Rub

, Ferretti

, Nitrini

, Farfel

, Polichiso

, Gierga

, Jacob-Filho

, Heinsen

(2009) The dorsal raphe nucleus shows phospho-tau neurofibrillary changes before the transentorhinal region in Alzheimer’s disease. A precocious onset? Neuropathol Appl Neurobiol 35, 406–416.

Simic

, Stanic

, Mladinov

, Jovanov-Milosevic

, Kos-tovic

, Hof

(2009) Does Alzheimer’s disease begin in the brainstem? Neuropathol Appl Neurobiol 35, 532–554.

Braak

, Del Tredici

(2011) The pathological process underlying Alzheimer’s disease in individuals under thirty. Acta Neuropathol 121, 171–181.

Fallgatter

, Neuhauser

, Herrmann

, Ehlis

, Wagener

, Scheuerpflug

, Reiners

, Riederer

(2003) Far field potentials from the brain stem after transcutaneous vagus nerve stimulation. J Neural Transm 110, 1437–1443.

Nomura

, Mizuno

(1983) Central distribution of efferent and afferent components of the cervical branches of the vagus nerve. A HRP study in the cat. Anat Embryol (Berl) 166, 1–18.

Polak

, Ehlis

, Langer

, Plichta

, Metzger

, Ringel

, Fallgatter

(2007) Non-invasive measurement of vagus activity in the brainstem - a methodological progress towards earlier diagnosis of dementias? J Neural Transm 114, 613–619.

10.

Braak

, Del Tredici

(2011) Alzheimer’s pathogenesis: Is there neuron-to-neuron propagation? Acta Neuropathol 121, 589–595.

11.

Berridge

(2010) Calcium hypothesis of Alzheimer’s disease. Pflugers Arch 459, 441–449.

12.

Mudher

, Colin

, Dujardin

, Medina

, Dewachter

, Naini

SMA

, Mandelkow

, Buee

, Goedert

, Brion

(2017) What is the evidence that tau pathology spreads through prion-like propagation? Acta Neuropathol Commun 5, 99.

13.

Shigenaga

, Hagen

, Ames

(1994) Oxidative damage and mitochondrial decay in aging. Proc Natl Acad SciUSA 91, 10771–10778.

14.

Munch

, Schinzel

, Loske

, Wong

, Durany

, Li

, Vlassara

, Smith

, Perry

, Riederer

(1998) Alzheimer’s disease—synergistic effects of glucose deficit, oxidative stress and advanced glycation endproducts. JNeu-ral Transm (Vienna) 105, 439–461.

15.

Majd

, Power

JHT

(2018) Oxidative stress and decreased mitochondrial superoxide dismutase 2 and peroxiredoxin 1 and 4 based mechanism of concurrent activation of AMPK and mTOR in Alzheimer’s disease. Curr Alzheimer Res 8, 764–776.

16.

Moreira

, Nunomura

, Nakamura

, Takeda

, Shenk

, Aliev

, Smith

, Perry

(2008) Nucleic acid oxidation in Alzheimer disease. Free Radic Biol Med 44, 1493–1505.

17.

Meissner

, Bruse

, Mohamed

, Schulz

, Warnk

, Storm

, Oehmichen

(2008) The 4977 bp deletion of mitochondrial DNA in human skeletal muscle, heart and different areas of the brain: A useful biomarker or more? Exp Gerontol 43, 645–652.

18.

Meissner

, Bruse

, Oehmichen

(2006) Tissue-specific deletion patterns of the mitochondrial genome with advancing age. Exp Gerontol 41, 518–524.

19.

Lin

, Beal

(2006) Mitochondrial dysfunction and oxidative stress in neurodegenerative diseases. Nature 443, 787–795.

20.

Corral-Debrinski

, Horton

, Lott

, Shoffner

, McKee

, Beal

, Graham

, Wallace

(1994) Marked changes in mitochondrial DNA deletion levels in Alzheimer brains. Genomics 23, 471–476.

21.

Blanchard

, Park

, Fripp

, Lerman

, Ingram

(1993) A mitochondrial DNA deletion in normally aging and in Alzheimer brain tissue. Neuroreport 4, 799–802.

22.

Gerschutz

, Heinsen

, Grunblatt

, Wagner

, Bartl

, Meissner

, Fallgatter

, Al-Sarraj

, Troakes

, Ferrer

, Arzberger

, Deckert

, Riederer

, Fischer

, Tatschner

, Monoranu

(2013) Neuron-specific mitochondrial DNA deletion levels in sporadic Alzheimer’s disease. Curr Alzheimer Res 10, 1041–1046.

23.

Vasile

, Dossi

, Rouach

(2017) Human astrocytes: Structure and functions in the healthy brain. Brain Struct Funct 222, 2017–2029.

24.

Southam

, Vincent

, Small

(2016) Do microglia default on network maintenance in Alzheimer’s disease? J Alzheimers Dis 51, 657–669.

25.

Katsumoto

, Lu

, Miranda

, Ransohoff

(2014) Ontogeny and functions of central nervous system macrophages. J Immunol 193, 2615–2621.

26.

Streit

, Braak

, Xue

, Bechmann

(2009) Dystrophic (senescent) rather than activated microglial cells are associated with tau pathology and likely precede neurodegeneration in Alzheimer’s disease. Acta Neuropathol 118, 475–485.

27.

Gutierrez

, Vitorica

(2018) Toward a new concept of Alzheimer’s disease models: A perspective from neuroinflammation. J Alzheimers Dis 64, 329–338.

28.

Bender

, Schwarzkopf

, McMillan

, Krishnan

, Rieder

, Neumann

, Elstner

, Turnbull

, Klop-stock

(2008) Dopaminergic midbrain neurons are the prime target for mitochondrial DNA deletions. J Neurol 255, 1231–1235.

29.

Bender

, Krishnan

, Morris

, Taylor

, Reeve

, Perry

, Jaros

, Hersheson

, Betts

, Klopstock

, Taylor

, Turnbull

(2006) High levels of mitochondrial DNA deletions in substantia nigra neurons in aging and Parkinson disease. Nat Genet 38, 515–517.

30.

Elstner

, Muller

, Leidolt

, Laub

, Krieg

, Schlau-draff

, Liss

, Morris

, Turnbull

, Masliah

, Prokisch

, Klopstock

, Bender

(2011) Neuromelanin, neuro-transmitter status and brainstem location determine the differential vulnerability of catecholaminergic neurons to mitochondrial DNA deletions. Mol Brain 4, 43.

31.

Krishnan

, Bender

, Taylor

, Turnbull

(2007) A multiplex real-time PCR method to detect and quantify mitochondrial DNA deletions in individual cells. Anal Biochem 370, 127–129.

32.

Ruijter

, Ramakers

, Hoogaars

, Karlen

, Bakker

, van den Hoff

, Moorman

(2009) Amplification efficiency: Linking baseline and bias in the analysis of quantitative PCR data. Nucleic Acids Res 37, e45.

33.

Pfaffl

(2001) A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29, e45.

34.

Filous

, Silver

(2016) Targeting astrocytes in CNS injury and disease: A translational research approach. Prog Neurobiol 144, 173–187.

35.

Abramov

, Canevari

, Duchen

(2004) Beta-amyloid peptides induce mitochondrial dysfunction and oxidative stress in astrocytes and death of neurons through activation of NADPH oxidase. J Neurosci 24, 565–575.

36.

Rodriguez

, Olabarria

, Chvatal

, Verkhratsky

(2009) Astroglia in dementia and Alzheimer’s disease. Cell Death Differ 16, 378–385.

37.

Sjobeck

, Englund

(2001) Alzheimer’s disease and the cerebellum: A morphologic study on neuronal and glial changes. Dement Geriatr Cogn Disord 12, 211–218.

38.

Davies

, Ma

, Jegathees

, Goldsbury

(2017) Microglia show altered morphology and reduced arborization in human brain during aging and Alzheimer’s disease. Brain Pathol 27, 795–808.

39.

Cervos-Navarro

, Diemer

(1991) Selective vulnerability in brain hypoxia. Crit Rev Neurobiol 6, 149–182.