Metallothionein‐I/II expression associates with the astrocyte DNA damage response and not Alzheimer‐type pathology in the aging brain

Oxidative stress and oxidative DNA damage are early features of mild cognitive impairment and Alzheimer's disease (AD), occurring before the formation of classical AD neuropathology, and resulting from an imbalance between pro‐ and anti‐oxidants. Astrocytes play a major neuroprotective role, producing high levels of anti‐oxidants including metallothionein‐I and ‐II (MT‐I/II). In the present study we characterized the immunoreactive profile of MT‐I/II in the temporal cortex of the Cognitive Function and Ageing Study (CFAS) aging population‐representative neuropathology cohort, and examined H2O2‐modulation of MT transcription by human astrocytes. MT‐I/II is primarily expressed by astrocytes in the aging brain, but is also associated with pyramidal neurons in a small proportion of cases. Astrocyte expression of MT‐I/II does not correlate with Alzheimer‐type pathology (Aβ plaques and neurofibrillary tangles) but does relate to astrocyte oxidative DNA damage (rs = .312, p = .006) and the astrocyte response to oxidative DNA damage in vivo (rs = .238, p = .04), and MT gene expression is significantly induced in human astrocytes response to oxidative stress in vitro (p = .01). In contrast, neuronal MT‐I/II does not relate to oxidative DNA damage or the neuronal DNA damage response, but is significantly higher in cases with high levels of local tangle pathology (p = .007). As MT‐I/II is neuroprotective against oxidative stress, modulation of MT‐I/II expression is a potential therapeutic target to treat the onset and progression of cognitive impairment.

characterized by the activation of sensor kinases, including DNA-protein kinase catalytic subunit (DNA-PKcs), and regions of DNA damage incorporating phosphorylated histone γH2AX. We have previously shown an association between the neuronal DDR and cognitive impairment, independent of AD pathology in the ageing brain , and demonstrated that increased levels of oxidative stress and the astrocyte DDR are an early event and not simply a secondary response to established AD pathology (Simpson, Ince, Haynes, et al., 2010).
Astrocytes play a major neuroprotective role, producing high levels of anti-oxidants (Ben Haim, Carrillo-de Sauvage, Ceyzeriat, & Escartin, 2015;Liddell, Robinson, Dringen, & Bishop, 2010). Metallothioneins (MT) are low molecular weight cysteine-rich proteins with a high metal content that can protect against reactive oxygen species (ROS; Vasak & Meloni, 2011). They comprise four subfamilies: MT-I and MT-II are present in most tissues, with MT-III and MT-IV expression confined to the central nervous system (CNS) and stratified squamous epithelium, respectively. While MT-I and -II have a similar structure, differing only by one amino acid and their ability to bind different metal ions, they are generally expressed in response to the same stimuli and perform the same function, so they are commonly grouped together and referred to as MT-I/II (Artells, Palacios, Capdevila, & Atrian, 2013). Within the CNS, MT-I/II is predominantly expressed by astrocytes, but has also been detected in neurons, microglia and endothelial cells (Pedersen et al., 2009;West, Hidalgo, Eddins, Levin, & Aschner, 2008), and is highly inducible in response to a variety of stimuli, including oxidative stress, neuroinflammation and toxic levels of metal ions (Andrews, 2000;Hidalgo et al., 2006;Manso, Adlard, Carrasco, Vasak, & Hidalgo, 2011). MT-I/II gene expression is regulated by antioxidant response elements, and reduced expression of MT-I/II in cell models or knock out animals results in heightened sensitivity to oxidative stress (Pitt et al., 1997;Qu, Pi, & Waalkes, 2013). MT-I/II is upregulated in AD (Adlard, West, & Vickers, 1998;Zambenedetti, Giordano, & Zatta, 1998) and is primarily associated with Aβ plaques in the hippocampus of several AD animal models Hidalgo et al., 2006).
Given the major role oxidative stress plays in the pathogenesis of neurodegenerative pathologies the present study evaluated MT-I/II expression in the temporal cortex of the Cognitive Function and Ageing Study (CFAS) neuropathology cohort, a well characterized large, ageing population-representative cohort which facilitates unbiased assessment of relationships between pathology and dementia, with the aim of assessing the relationship between MT-I/II and Alzheimertype pathology and the association with the astrocyte DDR in the aging human brain. In addition, the direct effect of oxidative stress on MT-I/II expression by human astrocytes was also assessed.

| Immunohistochemistry
Immunohistochemistry was performed using a standard avidin-biotin complex (ABC) method. Sections were deparaffinized, rehydrated to water and endogenous peroxidase activity quenched by placing the sections in 0.3% H 2 O 2 /methanol for 20 min at room temperature (RT). Sections were subjected to antigen retrieval (Access Super pH9.5, pressure cooker [Menarini Diagnostics, UK]). Following incubation with 1.5% normal serum for 30 min at RT, the sections were incubated for 60 min at RT with the well characterized, commercially available mouse monoclonal antibody against MT-I/II (1:100; DakoCytomation, UK). To visualize antibody binding, the horse-radish peroxidase avidin biotin complex (ABC-HRP) was used (Vectastain Elite kit, Vector Laboratories, UK) with 3,3 0 -diaminobenzidine (DAB) as the chromogen (Vector Laboratories, UK; brown). Negative controls, either omission of the primary antibody or a mouse isotype control, were included in every run.
To determine if MT-I/II + astrocytes colocalized with Aβ plaques, NFT or the astrocyte DDR, sequential dual labeling was performed with MT-I/II and either Aβ, AT8 or γH2AX. Briefly, amyloid plaques (mouse monoclonal Aβ 1:100; DakoCytomation, UK), tau tangles (mouse monoclonal AT8 1:400; Endogen, UK) or the astrocyte DNA damage response (rabbit polyclonal γH2AX 1: 1,000, R&D Systems, UK) were first visualized using the standard ABC-HRP protocol with DAB as substrate described above. For Aβ immunohistochemistry, antigen retrieval was carried out following pre-treatment with formic acid for 40 min at RT. Sections were then incubated with the avidin-biotin blocking kit (Vector Laboratories, UK), according to the manufacturer's instructions. The tissue was incubated overnight at 4 C with anti-MT-I/II (1:100), followed by biotinylated secondary antibody, and visualized with the alkaline-phosphatase-conjugated avidin-biotin complex and alkaline phosphatase substrate 1 (Vector Laboratories, UK, red). Every immuno-run included a single-labeled section, which showed the same pattern and intensity of immunoreactivity as seen in the dual-labeling experiments.

| Quantitative analysis of MT-I/II
Assessment of MT-I/II-specific immunoreactivity was performed by capturing bright-field microscopic images in three adjacent 350 μmwide cortical ribbons, consisting of contiguous fields to cover the total cortical thickness through the apex of the gyrus, using a 20 × objective (Nikon Eclipse Ni-U microscope, Nikon, UK) and analyzed using the Analysis^D software (Olympus Biosystems, Watford, UK). The image was thresholded and the immunoreactive area of the field determined per total area examined.

| Immunocytochemistry
Cells were fixed in 4% paraformaldehyde in PBS for 5 min at 37 C, then permeabilized (0.3% Triton X-100 in PBS) and blocked with 3% bovine serum albumin for 20 min at RT. Cells were incubated with MT-I/II (1:100) for 60 min at RT, followed by biotinylated secondary antibody, and visualized with streptavidin alexa fluor 488 (1:500).
Nuclei were counterstained with Hoechst 33342, and the staining visualized using a Nikon Eclipse Ni-U microscope (Nikon, Kingston upon Thames, UK).

| Statistical analysis
As the immunohistochemistry data was skewed, median and interquartile range (IQR) was used for descriptive analyses. Spearman's correlation coefficient (r s ) was calculated to verify the strength of correlations between continuous variables, and Kruskal-Wallis (KW) to test the correlation with parametric data. All tests were two-tailed.
Student's paired t test was used to analyze the qPCR data. Statistical analyses were performed using statistical package IBM SPSS Statistics version 23 and GraphPad Prism version 7.

| RESULTS
3.1 | MT-I/II predominantly immunolabels astrocytes in the aging human temporal cortex MT-I/II was associated with astrocytes, predominantly staining the astrocyte cell body and their primary radiating processes throughout layers II-VI of the temporal cortex (Figure 1a). A punctate pattern of staining was observed throughout the parenchyma, likely reflecting immunolabeling of cross-sections of the astrocyte processes. The pattern and intensity of immunoreactivity varied greatly across the cohort (median 7.31%, IQR 5.77), with some cases showing reduced staining, particularly surrounding blood vessels (Figure 1b).
While MT-I/II predominantly labeled astrocytes, weak immunoreactivity was also associated with a subpopulation of neurons in 16.6% of cases (13 out of 78 cases, Figure 1c), but no microglial, oligodendrocyte, or endothelial staining was detected. The immunoreactivity associated with capillaries and blood vessels likely represents MT-I/II + astrocytes, whose foot processes line the blood brain barrier There were no differences in MT-I/II expression between Braak and Braak groups (Kruskal-Wallis p = .314) (Figure 2a). We also compared MT-I/II to measures of local AD neuropathology in temporal cortex.
For comparisons with CERAD scores of diffuse and neuritic plaques and neurofibrillary tangles, the moderate-severe score groups were combined. MT-I/II expression was not associated with neuritic plaques (Kruskal-Wallis p = .697), diffuse plaques (p = .871) or neurofibrillary tangles (p = .891) and there were no correlations with area immunoreactivity for phospho-tau (AT8; Spearman's rank, r s = .081, p = .498) or Aβ (r s = .001, p = .99). MT-I/II expression did not correlate with post-mortem delay (r s = .113, p = .37) or brain pH (r s = .038, p = .771). Furthermore, dual labeling demonstrated that while MT-I/ II + astrocytes were a prominent feature of the aging brain, they did not intimately associate with the majority of Aβ plaques (Figure 1e).

| DISCUSSION
Oxidative stress and the associated oxidative-induced damage plays a key role in the initiation and progression of AD (Keller et al., 2005;Nunomura et al., 2006;Nunomura et al., 2001). In the current study we demonstrate that MT-I/II expression is predominantly associated with astrocytes in the ageing brain. Expression of this anti-oxidant does not correlate with Alzheimer-type pathology but does relate to the astrocyte response to oxidative DNA damage in vivo, and is induced in human astrocytes in response to oxidative stress in vitro. Within the aging temporal cortex MT-I/II is predominantly associated with astrocytes, supporting previous studies in AD (Adlard et al., 1998;Zambenedetti et al., 1998). In contrast to these studies which examined a small number of cases and report MT-I/II is exclusively associated with astrocytes, we demonstrate MT-I/II is also weakly associated with pyramidal neurons in cases with high levels of local tau pathology in a small proportion of a much larger cohort. MT-I/II is generally considered an intracellular protein which is primarily localized in the cytoplasm of astrocytes, however MT-I/II can be released into the extracellular matrix and internalized by neurons via low density lipoprotein receptors, decreasing phosphorylation of several intraneuronal kinases implicated in pro-inflammatory processes and apoptosis, and promoting neurite outgrowth and neuronal survival (Ambjorn et al., 2008;Asmussen, Von Sperling, & Penkowa, 2009;Chung et al., 2008;Fitzgerald et al., 2007). Furthermore, MT-I/II expression is induced in neurons in the hippocampus of animal models of multiple sclerosis (Grubic Kezele, Blagojevic Zagorac, Jakovac, Domitrovic, & Radosevic-Stasic, 2017), and in adult rats following systemic administration of kainic acid (Kim et al., 2003). Whether the immunopositive neurons detected in the current study are a result of neuronal uptake of MT-I/II or induction of MT-I/II expression is unknown.
In the current study we demonstrate MT-I/II expression is not a specific astrocyte response to local Alzheimer type pathology. This finding conflicts with rodent models of AD which show prominent upregulation of MT-I/II associated with Aβ plaques Hidalgo et al., 2006), but supports previous investigation of MT-I/II in AD which also demonstrated no spatial relationship between MT-I/II + astrocytes and Aβ plaques or tangle pathology (Adlard et al., 1998;Zambenedetti et al., 1998). However, in contrast to the study by Adlard et al, which reported a significant increase in MT-I/II expression in five MCI compared with five AD cases, we did not identify any association between MT-I/II and Braak stage in a large aging-representative neuropathology cohort.
Oxidative stress and metal ion dyshomeostasis are key features of AD pathology and may contribute to the pathogenesis of disease. Zinc plays a critical role in memory formation, regulating neuronal communication in the hippocampus (Pan et al., 2011;Wang et al., 2010), and is highly concentrated within Aβ plaques, with depleted levels in the surrounding parenchyma (Dong et al., 2003;Lovell, Robertson, Teesdale, Campbell, & Markesbery, 1998;Sayre et al., 2000;Suh et al., 2000).
Both zinc deficiency and zinc overload elicit oxidative stress that can lead to neuronal death (Sekler & Silverman, 2012;Wang et al., 2010).
MT play a neuroprotective role against ROS damage in AD, scavenging hydroxyl radicals, superoxide anions, hydrogen peroxide and reactive nitrogen radicals, and are generally considered more effective antioxidants than superoxide dismutase, catalase or glutathione peroxidase (Juarez-Rebollar, Rios, Nava-Ruiz, & Mendez-Armenta, 2017; Kumari, Hiramatsu, & Ebadi, 1998). In the current study expression of MT-I/II correlates with the astrocyte DDR and is induced in response to oxidative stress in vitro, suggesting that astrocytes play a neuroprotective role in response to oxidative stress and oxidative DNA damage, and not AD pathology, in the aging brain. In addition to their role in protecting against oxidative stress, astrocytes are actively involved in the neuroinflammatory response in AD (Gonzalez-Reyes, Nava-Mesa, Vargas-Sanchez, Ariza-Salamanca, & Mora-Munoz, 2017). Astrocyte expression of MT-I/II impacts the neuroimmune response in rodent models (Manso et al., 2011;Pankhurst, Bennett, Kirkcaldie, West, & Chung, 2011), suggesting that in addition to their anti-oxidant role, MTs may modify the microglial response in AD.
MT sequester heavy metals, regulating their availability to a range of enzymes and transcription factors. Under normal conditions, MT-I/II is located in the cytoplasm, however nuclear translocation of MT-I/II has been shown to occur in response to ultra-violet damage (Hanada, Tamai, Sawamura, Hashimoto, & Muramatsu, 1998;Jourdan et al., 2002) and cytokine activation (Spahl, Berendji-Grun, Suschek, Kolb-Bachofen, & Kroncke, 2003). In addition to increased expression of MT-I/II in the cytoplasmic compartment of human astrocytes in response to oxidative stress in vitro, we also demonstrate increased expression in the nuclear compartment, where it likely plays a role in the requirement of zinc for nuclear metalloenzymes and transcription factors.
While recent research has primarily focused on enhancing MT-III as a therapeutic strategy for AD (Aikins et al., 2017;Roy et al., 2017), the current study highlights the additional potential of targeting MT-I/ II expression. In summary we demonstrate that MT-I/II is predominantly expressed by astrocytes in the ageing brain, is associated with the astrocyte DDR and is upregulated in response to oxidative stress in vitro. Given the ability of MT-I/II to offer neuroprotection against oxidative stress, modulation of MT-I/II expression is a potential target to treat the onset and progression of cognitive impairment and dementia. tre NHS Foundation Trust, Liverpool. We would like to acknowledge the essential contribution of the liaison officers, the general practitioners, their staff, and nursing and residential home staff. We are grateful to our respondents and their families for their generous gift to medical research, which has made this study possible.