Tenascin-C Is Associated with Cored Amyloid-β Plaques in Alzheimer Disease and Pathology Burdened Cognitively Normal Elderly

Abstract

Tenascin-C (TN-C) is an extracellular matrix glycoprotein linked to inflammatory processes in pathological conditions including Alzheimer disease (AD). We examined the distribution of TN-C immunoreactivity (ir) in relation to amyloid-β (Aβ) plaques and vascular Aβ deposits in autopsy brain tissues from 14 patients with clinical and neuropathological AD and 10 aged-matched controls with no cognitive impairment; 5 of the controls had Aβ plaques and 5 did not. TN-C ir was abundant in cortical white matter and subpial cerebral gray matter in all cases, whereas TN-C ir was weak in blood vessels. In all cases with Aβ plaques but not in plaque-free controls, TN-C ir was detected as large (>100 µm in diameter) diffuse extracellular deposits in cortical grey matter. TN-C plaques completely overlapped and surrounded cored Aβ plaques labeled with X-34, a fluorescent derivative of Congo red, and they were associated with reactive astrocytes astrocytes, microglia and phosphorylated tau-containing dystrophic neurites. Diffuse Aβ plaques lacking amyloid cores, reactive glia or dystrophic neurites showed no TN-C ir. In cases with cerebral amyloid angiopathy, TN-C ir in vessel walls did not spread into the surrounding neuropil. These results suggest a role for TN-C in Aβ plaque pathogenesis and its potential as a biomarker and therapy target.

INTRODUCTION

Alzheimer disease (AD) is a chronic neurodegenerative disorder characterized histopathologically by extracellular amyloid plaques and intracellular neurofibrillary tangles. The principal component of amyloid plaques is the amyloid β (Aβ) peptide, a 39–43 amino acid fragment proteolytically produced from a transmembrane Aβ precursor protein (APP) ( 1–3 ). Numerous extracellular proteins bind to or colocalize with Aβ in AD plaques. These include inflammatory molecules (acute phase proteins, cytokines, chemokines, complement proteins) ( 4 ), amyloidogenic molecules (apolipoproteins and the non-Aβ component of AD amyloid, NAC) ( 5 ), lysosomal proteinases and ubiquitin ( 5 ) and numerous extracellular matrix (ECM) proteins including proteoglycans (heparan, chondroitin, keratin and dermatan sulphate proteoglycans) ( 6–10 ), reelin, agrin, collagen XVIII, collagen-like Alzheimer amyloid plaque component, ECM modulator lysyl oxidase, tissue inhibitor of matrix metalloproteinases and Goodpasture antigen-binding protein/ceramide transporter ( 11–16 ). The role of plaque-associated proteins in initiation or dissolution of Aβ aggregates in AD is poorly understood.

Tenascin-C (TN-C) is an ECM glycoprotein that regulates the differentiation and proliferation of astrocytes during embryogenesis and CNS development ( 17–19 ). TN-C expression is downregulated in the adult brain but it can be increased due to inflammatory processes and injuries to the nervous system ( 20 , 21 ). A recent study suggested that TN-C is involved in the pathogenesis of AD: TN-C gene transcription increased in microglia after they were exposed to Aβ, and brains of transgenic mice overexpressing APP had increased TN-C expression whereas Aβ plaque burden was reduced in TN-C-deficient APP mice ( 22 ). These associations warrant detailed investigations of brains from AD patients compared with normal elderly subjects. In this study, we examined localization of TN-C immunoreactivity (ir) in relation to Aβ pathology burden as well as inflammatory markers and phosphorylated tau in postmortem brain tissue from AD cases and age-matched cognitively normal controls.

MATERIALS AND METHODS

Protocol Approvals and Patient Consent

The study was approved by Rush University and the University of Pittsburgh’s Committee for Oversight of Research and Clinical Training Involving Decedents. Written informed consent for research and autopsy was obtained for all subjects in the study.

Cases

The brains of 24 cases were evaluated ( Table 1 ); these included 19 from the Rush Religious Order Study, a longitudinal clinical-pathologic study of aging and AD in older Catholic nuns, priests, and brothers who died with a premortem clinical diagnosis of “no cognitive impairment” ([NCI], n = 10) or “mild-moderate AD” (n = 9) and received a neuropathological evaluation postmortem ( 23 ). An additional 5 cases with moderate-severe AD were participants in the University of Pittsburgh Alzheimer’s Disease Research Center. Details of clinical evaluation and diagnostic criteria were published previously ( 23 , 24 ). A consensus conference of neurologists and neuropsychologists reviewed clinical data, medical records and interviews with family members and assigned a final clinical diagnosis. Mini-Mental State Examination ( 25 ) scores in the AD group ranged from 6 to 25 ( Table 1 ). Neuropathologic diagnoses were based on the National Institute on Aging (NIA)–Reagan Institute criteria (NIA-RI Working Group) ( 26 ), recommendations of the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) ( 27 ) and Braak staging of neurofibrillary tangles ( 28 ). Application of the new National Institute on Aging-Alzheimer’s Association guidelines ( 29 ) is currently ongoing in the examined cohorts. Cases with strokes, Lewy body disease, Parkinson disease or hippocampal sclerosis were excluded. All cases were de-identified and randomly assigned a unique identifier prior to the study.

TABLE 1.

Demographic, Clinical and Neuropathological Characteristics by Clinical Diagnosis Category

Clinical diagnosis
	NCI (n = 10)	AD (n = 14)	Total (n = 24)	p value
Age (years) at death:
Mean ± SD (Range)	84.3 ± 5.3 (74.0–93.0)	85.1 ± 6.6 (69.0–94.0)	84.8 ± 5.9 (69.0–94.0)	0.60 m
Number (%) of males:	6 (50.0%)	6 (50.0%)	12 (50.0%)	0.68 f
MMSE: Mean ± SD (Range)	27.0 ± 1.5 (25-30)	16.9 ± 7.8 (6-25)	21.1 ± 7.8 (6-30)	0.0001 m
PMI (h):
Mean ± SD (Range)	6.7 ± 3.2 (2.2–11.0)	6.6 ± 2.8 (3.0–10.7)	6.6 ± 2.9 (2.2–11.0)	0.80 m
Braak scores:
0	2	0	2
I/II	2	0	2	0.0001 m
III/IV	6	4	10
V/VI	0	10	10
CERAD diagnosis:
No AD	5	0	5
Possible	5	0	5
Probable	0	5	5	0.001 m
Definite	0	9	9
NIA-RI diagnosis
No AD	2	0	2
Low	4	0	4
Intermediate	4	6	10	0.0001 m
High	0	8	8

NCI, no cognitive impairment; AD, Alzheimer disease; PMI, postmortem interval; MMSE, Mini-Mental State Examination; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease; NIA-RI, National Institute on Aging-Reagan Institute; ^f Fisher’s exact test; ^m Mann-Whitney U test.

Tissue Preparation

At autopsy (mean postmortem interval 6.6 hours, range 2.2–11 hours; Table 1 ), superior frontal cortex (Brodmann area 9), caudate nucleus, and cerebellum were dissected and immersed in 4% paraformaldehyde in sodium phosphate buffer ([PB], pH 7.4), for 48 hours, cryoprotected in glycol solution ( 30 ), and sectioned at 40 μm on a freezing sliding microtome. Sections were then stored in cryoprotection solution ( 31 ) at –20 °C until processing.

Immunohistochemistry and Histology

Chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated. Table 2 lists all histological compounds and antibodies used in the study.

TABLE 2.

Histological Compounds and Antibodies

Compound	Affinity	Concentration	Source
X-34	β-Pleated sheet	100 μM	Synthesized ( 69 )
6-CN-PiB	Amyloid-beta fibrils	10 μM	Synthesized ( 33 )
Antibody (host)	Epitope or immunogen	Dilution	Source
TN-C (mouse)	Human TN-C	1:500	Generated by R. Chiquet-Ehrismann ( 32 )
Aβx-42 (rabbit)	Aβx-42	1:200	Millipore
Aβx-40 (rabbit)	Aβx-40	1:200	Millipore
Iba1(rabbit)	A synthetic peptide corresponding to C-terminus of Iba1	1:500	Wako
GFAP (chick)	Bovine GFAP	1:1000	Abcam
Fibronectin (mouse)	Extra domain A	1:500	Millipore
Agrin (rabbit)	Human agrin	1:50	Generated by G. J. Cole ( 12 )
Collagen IV (rabbit)	Human and bovine placenta	1:1000	Rockland
Phosphorylated tau (rabbit)	Serine 404	1:1000	Abcam

TN-C Immunofluorescence. Tissue sections were washed in PB, incubated for 1 hour in PB containing 0.3% Triton X-100, 3% normal goat serum and 2% bovine serum albumin, followed by an overnight incubation at 4 °C in monoclonal antibody (mab) B28-13 (diluted 1:500, provided by Dr. Ruth Chiquet-Ehrismann, University of Basel, Basel, Switzerland) raised against human TN-C, with its binding site localized to the last 3 fibronectin type III repeats, and its specificity confirmed previously by Western blotting and immunocytochemistry ( 32 ). Tissue sections were subsequently incubated for 2 hours with affinity purified goat anti-mouse IgG (Cy3; 1:500 dilution; Jackson ImmunoResearch Laboratories, West Grove, PA). Control sections processed in the absence of TN-C antibody resulted in no immunoreactivity.

Combined Immunofluorescence and Histofluorescence. To assess co-distribution of ECM proteins with fibrillar Aβ deposits, tissue sections were first processed as described above using antibodies generated against TN-C (mab B28-13; diluted 1:500), collagen IV (600-401-106-0.5; rabbit polyclonal; diluted 1: 1000; Rockland, Gilbertsville, PA), fibronectin (mab 1940; diluted 1: 500; EMD Millipore, Billerica, MA), or agrin (rabbit polyclonal; 1:50; gift from Dr. Gregory Cole, North Carolina Central University, Durham, NC) ( 12 ), followed by an affinity purified species-appropriate secondary IgG (Cy2 and Cy3; 1:500 dilution; Jackson ImmunoResearch Laboratories). Sections were then processed for X-34 (100 µM, a highly fluorescent derivative of Congo red) or 6-CN-Pittsburgh Compound B (6-CN-PiB) (10 µM, a highly fluorescent derivative of PiB), as previously described in ( 33 ).

Multiple-Labeling Immunofluorescence. Sections were rinsed in PB, treated with 85% formic acid for 2 min, and incubated in a cocktail of TN-C antibody (mab B28-13; diluted 1:500) and rabbit polyclonal antibodies against Aβ40 (AB5078P; 1:200; Millipore) and Aβ42 (AB5074P; 1:200; Millipore) overnight at 4 °C. Sections were then incubated in a cocktail of affinity purified goat anti-mouse IgG (Cy3; 1:500 dilution; Jackson ImmunoResearch Laboratories) and goat-anti-rabbit IgG (Alexa488; 1:500 dilution; ThermoFisher Scientific, Grand Island, NY) secondary antibodies for 3 hours at room temperature. To assess co-distribution of TN-C with phosphorylated tau, astrocytes, and microglia, sections were processed as described above (with the omission of formic acid pretreatment) using a cocktail of TN-C antibody (mab B28-13; diluted 1:500) and polyclonal antibodies against either tau phosphorylated at Serine 404 (ab30666, diluted 1:500; Abcam, Cambridge, MA), glial fibrillary acidic protein ([GFAP], ab4674; diluted 1:500; Abcam), or ionized calcium binding adaptor molecule 1 ([Iba1], 091-19741; diluted 1:500; Wako, Richmond, VA). Sections were then processed in a cocktail of species-appropriate secondary antibodies: goat anti-mouse IgG (Cy3; 1:500 dilution; Jackson ImmunoResearch Laboratories) for TN-C; goat-anti-rabbit IgG (Alexa488; 1:500 dilution; ThermoFisher Scientific) for phosphorylated tau and Iba1; goat-anti-chicken (Alexa488; 1:500 dilution; ThermoFisher Scientific) for GFAP.

Quantitative Analyses of TN-C and Aβ-Immunoreactive Plaques in the Frontal Cortex

Analyses of TN-C and Aβ Plaque Count. Numbers of total and co-labeled TN-C and Aβ40/42-immunoreactive (-ir) plaques were counted in three randomly selected, non-overlapping 10× microscopic fields (2.35 mm ² field of view), in 3 sections per case. Results are expressed as mean ± SD.

Analyses of TN-C and Aβ Plaque Loads (Percent Area). Percent areal coverages for TN-C and Aβ40/42-ir plaques were determined in the same 10× microscopic fields used for plaque count analyses. Percent area values were obtained by dividing immunolabeled area by total area sampled. Fluorescence illumination intensity was held constant throughout the sampling procedure and images were analyzed using the public domain ImageJ program (available at: http://rsb.info.nih.gov/nihimage/ ). A predetermined threshold value was held constant throughout the analysis.

Statistical Analysis

Statistical comparisons of the demographic, cognitive, and neuropathologic data between AD and NCI groups were performed using the Fisher exact test (gender) and Mann Whitney U test (age, postmortem interval, MMSE, NIA-RI diagnoses of AD, CERAD and Braak scores). TN-C and Aβ40/42 plaque counts and load were compared in AD and NCI cases using the Student t -test. Correlation analyses were performed using Spearman correlation. Statistical significance was set at 0.05 (two-sided).

RESULTS

Table 1 shows demographic, cognitive, and neuropathologic characteristics of the subjects grouped by clinical diagnosis. The NCI and AD groups did not differ significantly by age, gender, or postmortem interval. However, the 2 groups differed significantly by MMSE, Braak scores, CERAD and NIA-RI diagnosis of AD, although the NCI group included Aβ plaque pathology burdened (Aβ-positive NCI, n = 5) and Aβ pathology-free (Aβ-negative NCI, n = 5) cases.

In the frontal cortex from all AD and NCI cases, intense homogeneous TN-C ir was observed in the white matter (WM) and subpial portion of gray matter ( Fig. 1A, B ), whereas weaker TN-C ir was seen in cortical blood vessels ( Fig. 1C ). In AD and Aβ-positive NCI, TN-C-ir plaques were observed in the cortical gray matter; they ranged in diameter from 80 to 200 µm (average 120 ± 20 µm) and were more frequent in deep cortical layers, consistent with known distribution of classic neuritic plaques in the frontal cortex ( 34 ). TN-C plaques were not detected in Aβ-negative NCI cases. Quantitative analyses showed that compared with the Aβ-positive NCI, AD cases had significantly higher numbers of TN-C and Aβ plaques in the frontal cortex (TN-C count: p = 0.004, Aβ plaque count: p = 0.0001) and higher plaque loads by percent area (TN-C percent area: p = 0.001; Aβ percent area: p = 0.029). TN-C plaques completely surrounded cored Aβ plaques ( Fig. 1C–F ), which co-labeled with X-34, a highly fluorescent derivative of Congo red ( Fig. 2A ). There was a significant correlation (R ² =  0.98; p < 0.001) between numbers of TN-C plaques and cored Aβ plaques. The majority of TN-C plaques (73%) surrounded cored Aβ plaques, whereas approximately 80% of cored Aβ plaques were observed inside TN-C plaques.

FIGURE 1.

TN-C immunofluorescence in the frontal cortex from representative Aβ plaque burdened NCI (A) and AD (B) cases. Dense TN-C ir is present in the WM and subpial layer (arrows) in both cases, whereas TN-C deposits in the gray matter are more abundant in AD. (C) Section of frontal cortex from an AD case processed for dual immunofluorescence with antibodies against TN-C (red) and Aβ (green) showing that the majority of large TN-C deposits completely surround cored Aβ plaques. TN-C-labeled blood vessels are also seen (arrow). (D-F) TN-C/Aβ dual-immunolabeling in the area boxed in C . TN-C ir (D, red) surrounds the cored Aβ plaque (arrow in E , green) while the diffuse Aβ plaque (open arrow in E , green) is not associated with TN-C ir. Merged image is shown in (F) . Scale bars: bar in A  = 500 µm (for A and B ); bar in C  =50 µm; bar in D  = 50 µm (for D - F ).

FIGURE 2.

High-magnification fluorescence microscopy images of frontal cortex tissue sections from an AD case, processed for dual labeling with X-34 and extracellular matrix proteins. (A) Typical X-34-labeled cored amyloid plaque (blue) is completely surrounded by larger diffuse TN-C deposit (red). (B) Collagen IV is detected exclusively in blood vessels (red) and is not seen in the X-34-labeled cored plaque (blue). (C–E) A proteoglycan agrin (red) is present in blood vessels (arrows) and in the periphery, but not the central core, of an X-34-positive Aβ plaque (blue). Scale bar: A – E  = 30 µm.

Immunoreactivity for 2 other ECM proteins, collagen IV ( Fig. 2B ) and fibronectin (not shown), was restricted to blood vessels and did not colocalize with X-34 and Aβ-ir plaques. The ECM protein agrin ir localized to both blood vessels and cored Aβ plaques, but in contrast to TN-C it did not extend beyond the plaque halos ( Fig. 2C–E ). No ECM protein ir was found in diffuse (non-cored) Aβ plaques (not shown). Overlap of TN-C plaques with diffuse Aβ plaques was observed only when cored and diffuse Aβ plaques were in close proximity.

Cored Aβ plaques surrounded by TN-C plaques were also strongly labeled with 6-CN-PiB, a highly fluorescent derivative of the amyloid-binding compound PiB, and contained reactive glia, with GFAP-ir astrocytes restricted to the periphery ( Fig. 3A–D ), whereas Iba1-ir microglia clustered at the Aβ plaque core ( Fig. 3E–H ). TN-C/6-CN-PiB co-labeled plaques also contained phosphorylated-tau-ir dystrophic neurites ( Fig. 4 ).

FIGURE 3.

Triple fluorescent labeling in a single section of frontal cortex from an AD case. TN-C-ir deposits ( A and E , red) completely surround 6-CN-PiB-labeled cored amyloid plaques ( B and F , blue) with GFAP-ir astrocytes located near the periphery ( C , green) and Iba1-ir microglia at the center ( G , green) of the plaque. Panels D and H are merged fluorescence images of A – C and E – G , respectively. Scale bar: A – H  = 50 µm.

FIGURE 4.

Triple fluorescent labeling in a single section of frontal cortex from an AD case. Large TN-C deposits ( A , red) contain 6-CN-PiB-positive cored plaques ( B , blue) surrounded by clusters of phosphorylated-tau-ir dystrophic neurites ( C , green, arrows), whereas individual phosphorylated-tau-ir tangles show no TN-C immunofluorescence. Panel D shows merged fluorescence, with typical cored neuritic Aβ plaques surrounded by TN-C plaques (arrows in B – D ). Scale bar: A – D  = 100 µm.

As in the frontal cortex, TN-C ir was robust in the WM of the cerebellum and caudate nucleus where it labeled large bundles of myelinated fibers, while TN-C ir was not observed in diffuse Aβ plaques in these brain regions. Cerebral vascular localization of TN-C ir was less prominent in Aβ-burdened blood vessels (cerebral amyloid angiopathy [CAA]) observed in a subset of AD cases (not shown). In contrast to the association with cored plaques, TN-C did not form a halo or patch around CAA, but when cored Aβ plaques were positioned in close proximity to blood vessels TN-C ir was also seen around the blood vessels, regardless of whether the vessels were affected with CAA or not (not shown).

DISCUSSION

This report provides novel information regarding the localization of TN-C in the brains of AD and elderly people with NCI. We observed large TN-C-ir diffuse extracellular deposits in the frontal cortex of AD and Aβ pathology-burdened NCI, which is considered preclinical AD ( 35 , 36 ), while such deposits were absent from the NCI brains that were free of Aβ plaques. TN-C deposits colocalized with and surrounded cored neuritic Aβ plaques, but they did not associate with diffuse Aβ plaques lacking X-34 and a 6-CN-PiB-positive core, reactive astrocytes and microglia, and phosphorylated tau-containing dystrophic neurites. These results suggest a role for TN-C in the development of, or reaction to, classic senile plaques in AD. We also observed TN-C ir cerebral blood vessels in all cases regardless of clinical and pathology status; this is in agreement with previous reports of TN-C ir vasculature in brain ( 37 ) and other tissues ( 38 ). However, we found that when associated with CAA, which consists of fibrillar Aβ deposits similar to those in amyloid plaques, TN-C ir was weak and rarely extended into surrounding parenchyma. This might be due to lesser microgliosis associated with CAA-burdened vessels compared with cored plaques ( 39 , 40 ) and decreased ECM proteins; the latter may possibly relate to their susceptibility to rupture ( 41 ). Thus, reduced TN-C content in CAA relative to normal blood vessels might contribute to the progression of vascular pathology in AD.

The source of TN-C surrounding cored Aβ plaques and the temporal sequence of its accumulation in relation to aggregated Aβ and phosphorylated tau are not clear. We observed robust TN-C ir in the cortical WM and layer I in all AD and NCI cases. This is in agreement with reports of greater TN-C expression in the WM compared with gray matter in normal human brain ( 42 ), and of TN-C synthesis by astrocytes and release into the extracellular environment of WM and layer I of the cortex ( 42 , 43 ). We also found that TN-C plaques surrounding cored Aβ plaques were permeated by reactive astrocytes and their processes whereas microglia were in closer contact with cored Aβ plaques. This distribution pattern of activated glia has been described in classic neuritic plaques of AD and in old APPV717F mice ( 44 , 45 ). Inflammatory cells, specifically interleukin-1 (IL-1)-expressing microglia and S100β-producing astrocytes, are believed to play major roles in the formation of neuritic plaques with tau-positive dystrophic neurites ( 4 , 46 , 47 ). In addition to activating the astrocytes, IL-1 upregulates APP and thereby stimulates Aβ production. Microglia can also stimulate astrocyte proliferation via TGF-β1 ( 48 , 49 ), and increased GFAP production ( 50 ) induces TN-C gene expression and TN-C protein production ( 51–53 ). Furthermore, TN-C is an endogenous activator of Toll-like receptor, which induces the expression of pro-inflammatory factors ( 54 ). Overall, these observations support a role for TN-C in brain inflammation ( 42 , 55–57 ), which is closely linked to Aβ deposition in AD. In mice overexpressing mutant human APP, higher levels of TN-C gene transcription were associated with more advanced Aβ plaque pathology and inflammatory reaction, while TN-C deficiency reduced pro-inflammatory activation and fibrillar Aβ concentration ( 22 ). Other studies provide additional evidence that TN-C influences the development or progression of neuritic Aβ plaques. TN-C stimulates the expression of 14-3-3 (a family of homologous proteins that consist of seven isoforms [β, γ, ϵ, η, ζ, σ, and τ/θ]) ( 58 , 59 ), which can bind tau ( 60 , 61 ) and regulate tau phosphorylation via glycogen synthase kinase-3 beta (GSK3β) ( 61 , 62 ). Furthermore, 14-3-3ζ can bridge the interaction of GSK3ζ with tau and facilitate GSK3β-mediated phosphorylation of tau ( 63 ). Additional studies using in vitro and appropriate transgenic AD models are needed to elucidate these pathological pathways.

Because TN-C deposits surrounded selectively cored neuritic Aβ plaques, our results also suggest a potential value for TN-C as an AD biomarker that may have diagnostic value in neuropathological evaluation of AD ( 27 , 64 ). Neuritic plaques are considered to be the prime target for amyloid PET imaging ( 65–68 ). While current PET radioligands cannot distinguish between cored/neuritic and diffuse plaques ( 69 , 70 ), or between parenchymal plaques and CAA ( 69 , 71 ), a TN-C PET tracer might show selectivity for cored Aβ plaques in the gray matter. However, the marked TN-C staining seen in the WM would present a considerable obstacle in PET imaging studies of brain amyloidosis. On the other hand, the use of TN-C as a CSF or blood biomarker for AD-associated brain amyloidosis and inflammation should be considered ( 72 ).

Some methodological considerations warrant further discussion. The observed 73% overlap of TN-C deposits with cored Aβ plaques is likely an underestimation because TN-C deposits are substantially larger and completely surround cored Aβ plaques, Thus, in some instances TN-C plaques with only the peripheral halo but not the core of Aβ plaques (or no Aβ plaque at all) can be present in the plane of section. It is also possible that our immunostaining assay is not sensitive enough to detect the TN-C around all cored Aβ plaques. A systematic evaluation of TN-C negative cored Aβ plaques in relation to markers of microglia and astrocytes as well as related inflammatory molecules will provide greater insight into this issue.

In summary, in AD and Aβ plaque-burdened cognitively normal elderly subjects, cortical TN-C deposits colocalized with and selectively surrounded cored Aβ plaques that labeled with PiB and X-34. The deposits were closely associated with inflammatory cells and phosphorylated tau containing dystrophic neurites. Collectively, our data suggest that TN-C plays a role in reactive processes involved either in the progression or clearance of amyloid plaques. These complex relationships could not be determined in the current cross-sectional analysis, thus the potential value of TN-C as a biomarker and target for therapy in AD remains to be investigated.