Characterization of neurodegenerative pathologies in adult and pediatric subjects with Down syndrome

Abstract

Background

Down syndrome (DS) is frequently associated with Alzheimer's disease neuropathologic change (ADNC). However, studies assessing the full spectrum of neurodegenerative pathologies using modern consensus and staging criteria remain limited.

Objective

We aimed to elucidate the progression of neurodegenerative pathologies in DS and to explore the prevalence of comorbid pathologies across a broad age range (0–76 years), using comprehensive neuropathological assessments.

Methods

We conducted a two-phase analysis. First, we investigated an institutional dataset, followed by a pooled analysis incorporating data from the National Alzheimer's Coordinating Center and four published studies. Pathologies assessed included amyloid-β (Aβ), tau, α-synuclein, TDP-43, cerebral amyloid angiopathy (CAA), other cerebrovascular diseases (CVD), hippocampal sclerosis (HS), and basal ganglia mineralizations (BGM).

Results

Diffuse Aβ plaques appeared by age 11, with neuritic plaques emerging in the mid-thirties. Mild tau pathology, including pre-tangles and neuropil threads, first emerged in the second decade, with neurofibrillary tangles fully present in the fourth decade, always concurrent with Aβ plaques. All individuals over 30 exhibited ADNC. α-Synuclein pathology was observed in 27% of cases, while aging-related tau astrogliopathy (ARTAG), HS, and limbic-predominant age-related TDP-43 encephalopathy neuropathologic change (LATE-NC) were rare. CVD was present in approximately 60%, CAA was nearly universal (98%) after age 50, and 18% had BGM. Brain weight was consistently below the 25th percentile, even in younger individuals without ADNC.

Conclusions

DS shows a distinct neurodegenerative trajectory with early Aβ deposition. CAA, arteriolosclerosis, BGM, and α-synuclein pathology were highly prevalent, while ARTAG and LATE-NC were infrequently observed.

Introduction

Down syndrome (DS) is the most frequent chromosomal disorder linked to intellectual disabilities. It affects multiple organ systems, particularly musculoskeletal, cardiovascular, and nervous systems. In the United States, as of 2010, DS birth prevalence was estimated at 1 in 792 ¹ and population prevalence was estimated at 1 in 1499. ² DS is associated with the presence of an extra copy of chromosome 21 (i.e., trisomy 21). This anomaly can arise from nondisjunction, leading to a total of 47 chromosomes, or from the translocation of an extra chromosome 21 onto a different chromosome. Regardless of the underlying mechanism, the clinical manifestations of trisomy 21 remain consistent. ³ Several genes, both on chromosome 21 and in other genomic locations, including variations in the DS cell adhesion molecule and the amyloid precursor protein (APP) gene, play a role in the diverse clinical presentations such as memory impairment. ³

Early onset of amnestic symptoms in patients with DS was first reported in the mid-twentieth century. ⁴ Patients with DS typically develop clinical dementia at a mean age of 55 years,^5,6 with over 85% experiencing dementia by the age of 65 years.^6,7 This onset is remarkably earlier than that of sporadic Alzheimer's disease (AD) dementia, which characteristically presents with a much later onset (80–90 years).^8,9 The presence of an extra copy of the APP gene on chromosome 21 and the associated increase in amyloid-β (Aβ) peptide production, have long been accepted as the basis for the consistent development of AD pathology and the elevated risk of AD dementia in individuals with DS. ¹⁰ Investigations of DS along with familial AD (characterized by PSEN1, PSEN2, or APP mutations) have provided some of the most robust evidence supporting the amyloid cascade hypothesis, which proposes that the pathological accumulation of Aβ peptide is the primary cause of AD pathogenesis. ¹¹

Postmortem neuropathological examinations of DS brains consistently show evidence of Alzheimer's disease neuropathologic change (ADNC), with increased frequency in patients older than 30 years of age.^4,12 ADNC is characterized by the accumulation of neuritic plaques containing extracellular Aβ and intracellular neurofibrillary tangles (NFTs) composed of hyperphosphorylated tau that follow distinct regional progressions across brain regions. ¹³

Widely accepted neuropathological criteria for the diagnosis of AD were first established by the National Institute on Aging in 1985 ¹⁴ and went through a revision by the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) in 1991. ¹⁵ Initially, these criteria focused primarily on Aβ plaques. Subsequently, in 1997, the National Institute on Aging and the Reagan Institute integrated the CERAD neuritic plaque (NP) scoring with Braak and Braak NFT staging (hereinafter referred to as Braak stage) to ascertain whether clinical dementia had a high, intermediate, or low likelihood of being attributed to AD. ¹⁶ The 2012 National Institute on Aging-Alzheimer's Association (NIA-AA) guidelines provided the most recent criteria for the neuropathological evaluation of AD.^13,17 They utilize a semi-quantitative method for histopathological assessments of Aβ deposition using Thal phase, neurofibrillary degeneration using the Braak stage, and frequency of neuritic plaques using the CERAD plaque score. These most recent consensus criteria focus solely on neuropathological changes and do not consider the patient's cognitive status. The 2012 guidelines also incorporated protocols for the neuropathologic assessment of Lewy body disease pathology, vascular brain injury, hippocampal sclerosis (HS), and transactive response DNA binding protein of 43 kDa (TDP-43) inclusions, which are frequent comorbidities in subjects with ADNC. ¹⁸

To date, only a few studies have utilized the NIA-AA guidelines to assess neurodegenerative pathology in DS brains.^19–23 Moreover, only one study included pediatric subjects, ¹⁹ and none of them evaluated race. Notably, it has been shown that racial/ethnic disparities could affect the clinical and neuropathological expression of neurodegenerative disorders.^24,25 Our goal was to expand the knowledge base of DS neuropathology by evaluating the neuropathological features of DS individuals encompassing diverse age and race characteristics, with specific attention to the presence of comorbidities commonly associated with ADNC in non-DS subjects. We also aimed to enhance our analyses by combining our data with those from the four of the aforementioned studies,^19–22 where individual data were available, as well as with the DS cases from the National Alzheimer's Coordinating Center (NACC).

Methods

UTSW sample

We studied 34 brains from subjects with DS who underwent autopsies at our institution, UT Southwestern Medical Center (UTSW), between 1986 and 2023, using the same tissue collection and processing protocols. All macroscopic brain images, original hematoxylin and eosin (H&E) slides, available immunohistochemical (IHC) and special stains, and original neuropathology reports were reviewed. When existing H&E slides or immunostains were unavailable or difficult to interpret, formalin-fixed paraffin-embedded (FFPE) tissue blocks were retrieved and re-sectioned. New stains and appropriate IHC studies were performed to conduct a comprehensive neuropathological evaluation of neurodegenerative diseases, based on current diagnostic criteria. These evaluations included an assessment of NFTs and NP burden using fluorescent thioflavine-S staining and Aβ IHC, and more thorough assessment and staging of tau deposition using AT8 IHC. Immunohistochemistry assessment of phospho-alpha-synuclein and phospho-TDP-43 pathology was conducted only for subjects over the age of 30 years. The characteristics of individual cases from this dataset are presented in Table 1.

Table 1.

Sociodemographic and neuropathological characteristics of the UTSW cohort.

Case #	Age at death	Sex	Race	Brain weight (g)	Brain weight (percentile)	Thal Phase (0–5)	Braak Stage (0–6)	A score	B score	C score	ADNC level	AD pathology other than conventional NFT and NP	Lewy pathology	LATE-NC stage (0–3)	ARTAG	HS	Mineralization in BG	CAA	Other CVD
1	2	M	Hispanic	1030	N/A	0	0	0	0	0	None	None	N/A	N/A	No	No	No	No	No
2	2	M	Black	880	N/A	0	0	0	0	0	None	None	N/A	N/A	No	No	No	No	No
3	3	F	Black	770	N/A	0	0	0	0	0	None	None	N/A	N/A	No	No	No	No	No
4	3	F	White	1000	N/A	0	0	0	0	0	None	None	N/A	N/A	No	No	N/Av	No	No
5	5	F	Black	N/Av	N/A	0	0	0	0	0	None	None	N/A	N/A	No	No	No	No	No
6	11	F	White	1150	N/A	1	0	1	0	0	Low	DP	N/A	N/A	No	No	No	No	No
7	14	F	Black	1070	N/A	0	0	0	0	0	None	Rare PT	N/A	N/A	No	No	No	No	No
8	19	M	Hispanic	1050	<10	0	0	0	0	0	None	Rare NT	N/A	N/A	No	No	No	No	Yes
9	20	M	White	1255	<10	1	0	1	0	0	Low	DP	N/A	N/A	No	No	No	No	Yes
10	20	M	Black	1125	<10	3	0	2	0	0	Low	DP	N/A	N/A	No	No	No	No	Yes
11	21	F	N/Av	N/Av	N/Av	0	0	0	0	0	None	None	N/A	N/A	No	No	No	No	No
12	21	M	Black	1130	<10	4	0	3	0	0	Low	DP + rare PT	N/A	N/A	No	No	No	No	No
13	22	F	White	1195	10–25	2	0	1	0	0	Low	DP + rare NT	N/A	N/A	No	No	No	No	No
14	22	F	Hispanic	1190	10–25	5	0	3	0	0	Low	DP + rare NT	N/A	N/A	No	No	N/Av	No	No
15	25	F	White	N/Av	N/Av	2–4*	0	1–3	0	0	Low	DP + rare NT	N/A	N/A	No	No	N/Av	No	Yes
16	31	M	Black	1110	<10	2	0	1	0	0	Low	DP	None	0	No	No	No	Yes	Yes
17	34	M	White	1260	10	5	5	3	3	2	High	N/A	None††	0	No	No	No	Yes	Yes
18	35	M	White	N/Av	N/Av	5	2–3**	3	1–2	2	≥Intermediate	N/A	None†††	≤1‡	No	No	N/Av	No	No
19	40	M	Hispanic	N/Av	N/Av	2–3***	1	1–2	1	2	Low	N/A	None†††	≤1‡	No	No	N/Av	No	No
20	40	F	White	1280	10–25	2	1	1	1	1	Low	N/A	None††	0	No	No	No	Yes	Yes
21	44	F	White	1175	<10	5	6	3	3	3	High	N/A	None†††	≤1‡	No	No	No	Yes	Yes
22	47	F	Hispanic	855	<10	5	2	3	1	2	Intermediate	N/A	None††	0	No	No	Yes	Yes	Yes
23	48	F	Black	1030	<10	4	4	3	2	3	Intermediate	N/A	None†††	≤1‡	No	No	N/Av	Yes	Yes
24	48	M	White	1180	<10	5	6	3	3	3	High	N/A	None††	0	No	No	No	Yes	Yes
25	49	F	N/Av	N/Av	N/Av	3	5–6†	2	3	2	Intermediate	N/A	None††	0	No	No	No	Yes	Yes
26	49	M	White	1210	<10	5	5	3	3	3	High	N/A	None	0	No	No	No	Yes	Yes
27	54	M	White	1335	10–25	5	5–6†	3	3	3	High	N/A	None††	0	Yes	No	No	Yes	Yes
28	56	M	White	970	<10	5	6	3	3	3	High	N/A	None††	0	No	No	No	Yes	Yes
29	56	F	White	990	<10	5	5–6†	3	3	3	High	N/A	None††	0	No	No	No	Yes	Yes
30	59	M	White	1010	<10	5	5	3	3	3	High	N/A	None††	0	Yes	No	Yes	Yes	Yes
31	61	F	White	840	<10	5	6	3	3	3	High	N/A	None††	0	No	No	Yes	Yes	Yes
32	66	M	White	925	<10	5	6	3	3	3	High	N/A	None†††	≤1‡	No	No	No	Yes	Yes
33	70	F	N/Av	660	<10	5	6	3	3	3	High	N/A	Amygdala	0	Yes	Yes	Yes	Yes	Yes
34	70	F	White	840	<10	5	5–6†	3	3	3	High	N/A	None†††	≤1‡	No	No	Yes	Yes	Yes

^*At least 2, but less than 5 (a cerebellum section is negative for amyloid deposits; basal ganglia and midbrain sections are not available).

**At least 2, but less than 4 (Hippocampus incomplete. Fragments of hippocampus are positive for neurofibrillary tangles and neuritic plaques; neocortical sections are negative for tau pathology).

^***At least 2, but less than 4 (midbrain section is negative for diffuse plaques; basal ganglia section is not available).

^†At least 5 (calcarine cortex section is not available). ^††Olfactory bulbs not assessed; stage “olfactory bulb-only” Lewy pathology cannot be excluded. ^†††Olfactory bulbs and amygdala not assessed; “olfactory bulb only” or “amygdala-predominant” Lewy pathology cannot be excluded. ^‡Amygdala not assessed; stage 1 LATE pathology cannot be excluded.

AD: Alzheimer's disease; ADNC: Alzheimer's disease neuropathologic change; ARTAG: aging-related tau astrogliopathy; BG: basal ganglia; CAA: cerebral amyloid angiopathy; CVD: cerebrovascular disease (arteriolosclerosis and/or arteriosclerosis, infarcts, hemorrhage, atherosclerosis); DP: diffuse plaques; HS: hippocampal sclerosis; LATE-NC: Limbic-predominant age-related TDP-43 encephalopathy neuropathologic change; N/A: not applicable; N/Av: not available; NT: neuropil threads; NFT: neurofibrillary tangles; NP: neuritic plaques; PT: pretangles.

Diagnostic criteria for neurodegenerative disorders

Comprehensive anatomical mapping and staging were applied for all commonly assessed protein aggregates using the 2012 NIA-AA guidelines. ¹³ More specifically, IHC for Aβ (6E10) was used to assess the Thal phase (scale, 0–5), IHC for p-tau (AT8) was used to establish the Braak stage (scale, 0-VI), and thioflavine-S staining was used to determine CERAD neuritic plaque score (scale, 0–3).^15,26,27 Thal phase quantifies the topographic progression of Aβ deposition (A score), Braak stage reflects the spread of tau pathology (B score), and CERAD evaluates the density of neuritic plaques (C score). These individual scores are then converted into a unified ABC score (ranging from 0 to 3), which is the basis for grading the ADNC as none, low, intermediate, or high. ¹³ For statistical analysis, we assigned the ADNC level as: none = 0, low = 1, intermediate = 2, and high = 3.

All sections stained with AT8 were additionally evaluated for the presence of aging-related tau astrogliopathy (ARTAG) using a standardized assessment method. ²⁸

The brainstem, amygdala, and olfactory bulb sections, when available, were screened for the presence of Lewy pathology, using the most recent consortium staging protocol to categorize cases into “no Lewy pathology”, “olfactory bulb-only”, “amygdala-predominant”, “brainstem-predominant”, “limbic (transitional)”, and “neocortical (diffuse)” stages. ²⁹

Limbic-predominant age-related TDP-43 encephalopathy neuropathologic change (LATE-NC) was assessed based on LATE-NC working group recommendations. ³⁰ Cases were either TDP-43 negative (Stage 0), TDP-43 positive in the amygdala only (Stage 1), TDP-43 positive in the amygdala and hippocampus (Stage 2), or TDP-43 positive in the amygdala, hippocampus, and middle frontal gyrus (Stage 3).

The HS was assessed based on the presence of severe neuronal loss and gliosis involving predominantly the subiculum and CA1 subfield of hippocampus. ³¹

Brain weight percentiles

Percentiles for brain weights by sex and age were determined for subjects aged 18 years or older (n = 27), based on a recent study that screened over 4000 autopsies performed on adults. ³² However, a reliable study on the percentiles for the brain weight of subjects younger than 18 years old (n = 7) could not be found; thus, percentiles for brain weights were not calculated for this subgroup.

Immunohistochemistry

Adjacent four-micrometer thick sections of formalin-fixed, paraffin-embedded tissue underwent IHC staining with either an antibody to phospho-tau (AT8 [pSer202/Thr205] ThermoScientific) diluted 1:200, an antibody to amyloid-β (6E10, Covance) diluted 1:3000, an antibody to phospho-α-synuclein [MJF-R13(8-8), Abcam] diluted 1:4000, and an antibody to phospho-TDP-43 (Cosmo Bio Co.) diluted 1:2000. IHC was performed on Leica Bond III automated immunostaining platform (Leica Biosystems).

Additional neurofilament-specific antibodies (NF-L and SMI-31) were used to stain selected sections of the basal ganglia that harbored exuberant mineralizations. An (NF-L antibody (Dako, Agilent) diluted 1:2000 detects intermediate filaments in neuronal cell bodies, dendrites, and axons, while an SMI-31 antibody (Covance) diluted 1:4000 recognizes phosphorylated epitopes on the heavy (NF-H) and medium (NF-M) neurofilament subunits and yields stronger staining in axon shafts compared to neuronal cell bodies.

Special histologic stains

Select sections of the basal ganglia that featured significant mineralizations on the H&E stains (cases #30 and #31) were subsequently stained with Prussian blue for ferric iron, Von Kossa for calcium, and periodic acid-schiff (PAS) for glycogen.

Combined dataset

To enhance our analyses, we combined our institutional data with those from three similar published studies^19–21 and included data for DS subjects from the National Alzheimer's Coordinating Center (NACC), funded by the National Institute on Aging (U24 AG072122). NACC data, downloaded on 03/05/2024 (Data Request Number 12040), included the “Neuropathology Dataset” from September 2005 to December 2023, collected from 37 Alzheimer's Disease Research Centers (ADRCs). Out of 7801 unique cases with neuropathological data, only six had a diagnosis of DS and were included in the combined dataset (form versions v1.2, v2.0, and v3.0/3.2). The combined dataset included 56 cases from Davidson et al., ¹⁹ 11 cases from Ichimata et al., ²⁰ 33 cases from Wegiel et al., ²¹ 20 cases from Martá-Ariza et al., ²² and six NACC cases. The characteristics of all 126 individual cases from these datasets are presented in Table 2. With the addition of 34 cases from our institutional dataset (UTSW cohort), the combined dataset comprised a total of 160 cases. Although all included studies used current NIA-AA guidelines to assess ADNC pathology in DS brains, their methodologies differed, and not all provided detailed protocols. The diagnostic staging criteria and antibody clones used by each study are summarized in Supplemental Table 2.

Table 2.

Characteristics of the cases from previously published studies and NACC dataset (n = 126).

Data group	Age at death	Sex	Race	Genetically proven dx	APOE genotype	Brain weight (g)	Brain weight (percentile)	A score	B score	C score	ADNC dx	ADNC level	AD pathology	Lewy pathology†	LATE_NC stage (0–3)	CAA	ARTAG	Hippocampal sclerosis
1	0	M	N/Av	No	N/Av	N/Av	N/A	0	0	0	No	0	None	0***	0**	No	No	No
1	0	F	N/Av	No	N/Av	N/Av	N/A	0	0	0	No	0	None	0***	0**	No	No	No
1	2	M	N/Av	No	N/Av	N/Av	N/A	0	0	0	No	0	None	0***	0**	No	No	No
1	2	M	N/Av	No	N/Av	N/Av	N/A	0	0	0	No	0	None	0***	0**	No	No	No
1	3	F	N/Av	No	N/Av	N/Av	N/A	0	0	0	No	0	None	0***	0**	No	No	No
1	11	M	N/Av	No	N/Av	N/Av	N/A	0	0	0	No	0	None	0***	0**	No	No	No
1	13	F	N/Av	No	N/Av	N/Av	N/A	0	0	0	No	0	None	0***	0**	No	No	No
1	13	M	N/Av	Yes	3/3	N/Av	N/A	1	0	N/Av	Yes	1	DP only	0***	0**	No	No	No
1	23	M	N/Av	No	N/Av	N/Av	N/Av	0	0	0	No	0	None	0*	0**	No	No	No
1	27	F	N/Av	No	N/Av	N/Av	N/Av	1	0	N/Av	Yes	1	DP only	0***	0**	No	No	No
1	35	M	N/Av	No	N/Av	N/Av	N/Av	1	0	N/Av	Yes	1	DP only	0***	0**	No	No	No
1	36	M	N/Av	No	N/Av	N/Av	N/Av	3	1	N/Av	Yes	1	NFTs + NPs	0***	0**	No	No	No
1	37	F	N/Av	Yes	3/4	N/Av	N/Av	3	1	N/Av	Yes	1	NFTs + NPs	0***	0**	No	No	No
1	39	F	N/Av	No	N/Av	N/Av	N/Av	1	0	N/Av	Yes	1	DP + PT/NT	0***	0**	No	No	No
1	42	M	N/Av	No	N/Av	N/Av	N/A	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	47	F	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	47	M	N/Av	No	N/Av	N/Av	N/Av	3	2	N/Av	Yes	2	NFTs + NPs	0***	0**	Yes	No	No
1	49	F	N/Av	No	3/4	N/Av	N/Av	3	2	N/Av	Yes	2	NFTs + NPs	0***	0**	Yes	No	No
1	50	M	N/Av	No	3/4	N/Av	N/Av	2	1	N/Av	Yes	1	NFTs + NPs	0***	0**	Yes	No	No
1	50	F	N/Av	No	N/Av	N/Av	N/Av	1	0	N/Av	Yes	1	DP + PT/NT	0***	0**	No	No	No
1	51	M	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	5	0**	Yes	No	No
1	53	M	N/Av	Yes	3/3	N/Av	N/Av	3	2	N/Av	Yes	2	NFTs + NPs	0***	0**	Yes	No	No
1	53	F	N/Av	No	3/4	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	54	M	N/Av	No	2/4	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0	2**	Yes	No	No
1	55	M	N/Av	No	3/4	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	55	M	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	56	F	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	56	F	N/Av	No	3/4	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	57	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	57	M	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	58	F	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	58	M	N/Av	No	3/4	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	58	F	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	58	F	N/Av	No	2/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	59	M	N/Av	No	2/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	59	M	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	2**	Yes	No	No
1	59	M	N/Av	No	2/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	59	M	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	2**	Yes	No	No
1	60	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	60	M	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	60	F	N/Av	No	N/Av	N/Av	N/Av	3	0	N/Av	No	1	DP only	0***	0**	No	No	No
1	60	M	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	60	M	N/Av	No	3/3	N/Av	N/Av	3	2	N/Av	Yes	2	NFTs + NPs	0***	0**	Yes	No	No
1	61	F	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	2**	Yes	No	Yes
1	61	F	N/Av	No	3/4	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	5	2**	Yes	No	Yes
1	62	F	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	62	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	5	0**	Yes	No	No
1	62	F	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	62	M	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	Yes
1	62	M	N/Av	No	2/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	5	2**	Yes	No	No
1	63	M	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	5	0**	Yes	No	No
1	64	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	64	F	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	5	0**	Yes	No	No
1	65	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	65	M	N/Av	No	3/3	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	66	M	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	69	F	N/Av	No	N/Av	N/Av	N/Av	3	3	N/Av	Yes	≥2	NFTs + NPs	0***	0**	Yes	No	No
1	76	M	N/Av	Yes	2/3	N/Av	N/Av	3	2	N/Av	Yes	2	NFTs + NPs	0***	0**	Yes	No	No
2	38	M	N/Av	No	N/Av	1010	<10	2	1	1	Yes	1	NFTs + NPs	0	0	Yes	No	No
2	42	F	N/Av	No	N/Av	900	<10	3	3	3	Yes	3	NFTs + NPs	0	0	Yes	No	No
2	49	M	N/Av	No	N/Av	1100	<10	3	3	3	Yes	3	NFTs + NPs	3	0	Yes	No	No
2	51	F	N/Av	No	N/Av	750	<10	3	3	3	Yes	3	NFTs + NPs	5	0	Yes	No	No
2	54	M	N/Av	No	N/Av	1130	<10	3	2	3	Yes	2	NFTs + NPs	0	0	Yes	No	No
2	57	F	N/Av	No	N/Av	756	<10	3	3	3	Yes	3	NFTs + NPs	0	0	Yes	No	No
2	58	M	N/Av	No	N/Av	1100	<10	3	3	3	Yes	3	NFTs + NPs	4	0	Yes	Yes	No
2	61	M	N/Av	No	N/Av	905	<10	3	3	3	Yes	3	NFTs + NPs	0	0	Yes	Yes	No
2	61	F	N/Av	No	N/Av	900	<10	3	3	3	Yes	3	NFTs + NPs	0	0	Yes	No	No
2	65	F	N/Av	No	N/Av	880	<10	3	3	3	Yes	3	NFTs + NPs	3	1	Yes	No	No
2	66	F	N/Av	No	N/Av	760	<10	3	3	3	Yes	3	NFTs + NPs	0	2	Yes	No	No
3	26	M	N/Av	No	N/Av	1160	<10	0	1	N/Av	No	0	NFTs	NR	0	No	N/Av	N/Av
3	28	M	N/Av	No	N/Av	1320	10–25	3	1	N/Av	Yes	1	NFTs + NPs	NR	0	No	N/Av	N/Av
3	41	M	N/Av	No	N/Av	1130	<10	3	1	N/Av	Yes	1	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	41	F	N/Av	No	N/Av	926	<10	3	2	N/Av	Yes	2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	43	F	N/Av	Yes	3/4	990	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	48	M	N/Av	No	N/Av	940	<10	3	2	N/Av	Yes	2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	48	M	N/Av	Yes	4/4	920	<10	3	2	N/Av	Yes	2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	49	M	N/Av	Yes	N/Av	1100	<10	3	2	N/Av	Yes	2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	51	M	N/Av	Yes	N/Av	840	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	51	M	N/Av	Yes	N/Av	954	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	52	M	N/Av	Yes	N/Av	1194	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	54	F	N/Av	Yes	3/3	900	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	54	M	N/Av	Yes	N/Av	1060	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	54	M	N/Av	Yes	3/3	774	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	56	F	N/Av	Yes	N/Av	656	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	56	F	N/Av	Yes	N/Av	822	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	56	M	N/Av	Yes	3/3	762	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	56	M	N/Av	Yes	4/4	1114	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	57	F	N/Av	Yes	3/3	960	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	57	F	N/Av	Yes	3/4	696	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	57	M	N/Av	Yes	3/3	1000	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	57	F	N/Av	Yes	3/3	790	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	59	F	N/Av	Yes	3/4	740	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
3	59	M	N/Av	Yes	N/Av	1060	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
3	59	M	N/Av	Yes	3/4	1028	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
3	61	M	N/Av	Yes	N/Av	1028	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
3	61	F	N/Av	Yes	2/3	1102	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
3	63	M	N/Av	Yes	N/Av	700	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	65	F	N/Av	Yes	4/4	830	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
3	65	M	N/Av	Yes	3/4	900	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
3	67	M	N/Av	Yes	N/Av	700	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	71	M	N/Av	Yes	3/3	960	<10	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
3	72	F	N/Av	Yes	N/Av	1100	10–25	3	3	N/Av	Yes	≥2	NFTs + NPs	N/R	0	Yes	N/Av	N/Av
4	49	M	White	No	4/4	N/Av	N/Av	N/Av	3	3	Yes	N/Av	NFTs + NPs	0	N/Av	Yes	N/Av	No
4	49	M	White	No	3/3	N/Av	N/Av	N/Av	3	2	Yes	N/Av	NFTs + NPs	0	N/Av	Yes	N/Av	No
4	53	M	White	No	3/4	1063	<10	3	3	3	Yes	3	NFTs + NPs	2	0	Yes	No	N/Av
4	60	F	White	No	3/3	N/Av	N/Av	N/Av	3	3	Yes	N/Av	NFTs + NPs	0	N/Av	No	N/Av	No
4	61	F	White	No	3/4	696	<10	3	3	3	Yes	3	NFTs + NPs	2	1	Yes	N/Av	N/Av
4	63	M	White	No	2/3	1272	25–50	3	3	3	Yes	3	NFTs + NPs	2	0	Yes	N/Av	N/Av
5	48	F	N/Av	Yes	3/3	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	50	F	N/Av	Yes	3/4	N/Av	N/Av	N/Av	3	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	54	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	57	F	N/Av	Yes	3/3	N/Av	N/Av	3	2	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	58	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	58	F	N/Av	Yes	3/3	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	58	M	N/Av	Yes	N/Av	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	59	F	N/Av	Yes	3/3	N/Av	N/Av	N/Av	3	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	59	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
5	60	M	N/Av	Yes	4/4	N/Av	N/Av	N/Av	N/Av	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	61	M	N/Av	Yes	4/4	N/Av	N/Av	N/Av	N/Av	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	61	M	N/Av	Yes	3/3	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
5	62	M	N/Av	Yes	3/4	N/Av	N/Av	N/Av	3	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	62	M	N/Av	Yes	3/4	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	63	M	N/Av	Yes	3/3	N/Av	N/Av	N/Av	N/Av	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	63	F	N/Av	Yes	2/4	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	N/R	N/Av	N/Av
5	64	F	N/Av	Yes	3/3	N/Av	N/Av	N/Av	3	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	64	M	N/Av	Yes	3/3	N/Av	N/Av	N/Av	3	N/Av	Yes	N/Av	N/Av	N/R	N/R	N/R	N/Av	N/Av
5	64	M	N/Av	Yes	2/4	N/Av	N/Av	3	3	3	Yes	3	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av
5	71	M	N/Av	Yes	2/4	N/Av	N/Av	3	2	3	Yes	2	NFTs + NPs	N/R	N/R	Yes	N/Av	N/Av

Data groups:

#1- Davidson YS, Robinson A, Prasher VP, et al. The age of onset and evolution of Braak tangle stage and Thal amyloid pathology of Alzheimer's disease in individuals with Down syndrome. Acta Neuropathol Commun 2018; 6: 56.

#2- Ichimata S, Yoshida K, Visanji NP, et al. Patterns of mixed pathologies in Down syndrome. J Alzheimers Dis 2022; 87: 595–607.

#3- Wegiel J, Flory M, Kuchna I, et al. Developmental deficits and staging of dynamics of age associated Alzheimer's disease neurodegeneration and ikbneuronal loss in subjects with Down syndrome. Acta Neuropathol Commun 2022; 10: 2.

#4- NACC dataset

#5- Martá-Ariza M, Leitner DF, Kanshin E, et al. Comparison of the amyloid plaque proteome in Down syndrome, early-onset Alzheimer's disease, and late-onset Alzheimer's disease. Acta Neuropathol 2025; 149: 9.

^*Olfactory bulbs not assessed.

^**Amygdala not assessed.

^***Olfactory bulbs and amygdala not assessed.

^† Lewy pathology: 0-None, 1-Olfactory only, 2-Amygdala predominant, 3-Brainstem predominant, 4-Limbic, 5-Neocortical. ADNC level: 0-None, 1-Low, 2-Intermediate, 3-high.

AD: Alzheimer's disease; ADNC: Alzheimer's disease neuropathologic change; ARTAG: aging-related tau astrogliopathy; CAA: cerebral amyloid angiopathy; DP: diffuse plaques; dx: diagnosis; LATE: Limbic-predominant age-related TDP-43 encephalopathy neuropathologic change; N/A: not applicable; N/Av: not available; NFT: neurofibrillary tangle; N/R: not reported; NP: neuritic plaque; PT/NT: Pretangles and/or neuropil threads.

Statistical analysis

Statistical analyses were performed using SPSS 29 (IBM SPSS Statistics for Windows, Version 29.0.0; IBM Corp., Armonk, NY). Continuous variables were checked for normal distribution with the Shapiro-Wilks test for goodness-of-fit. None of the continuous variables except for age and brain weight were normally distributed; therefore, the Mann–Whitney U test was used to assess between-group differences (e.g., ADNC-negative versus ADNC-positive) and Spearman's correlation coefficient was used to analyze associations between continuous variables (e.g., age, ABC scores). The χ2 test was used to compare frequency distributions of categorized variables between groups (e.g., ADNC-positive versus ARTAG-positive). Fisher's exact test was used when expected cell values were less than five. Statistical significance was set at a p-value of less than 0.05.

Results

UTSW cohort

The characteristics of each case included in the study are described in Table 1. The ages at death ranged from 2 to 70 years, with a slight female predominance (n = 18, 53%). One fifth (20.1%) of the subjects were under the age of 18. The youngest subject with ADNC diagnosis (referred to as “ADNC+” hereafter) was a 11-year-old white female. The oldest subject without a diagnosis of ADNC (referred to as “ADNC-” hereafter) was a 21-year-old female. The youngest ADNC+ subject with high level ADNC was a 34-year-old white female.

Among the ADNC+ subjects (n = 26), 10 (41.7%) had a low level of ADNC, three (12.5%) had an intermediate level, and 12 (50%) had a high level, while one case could not be assigned a definite ADNC level due to incomplete Braak staging. However, having an A score of 3, this case had at least an intermediate level of ADNC. Among the ADNC+ subjects with low-level ADNC (n = 10), four (40%) harbored only diffuse plaques (DPs), four (40%) showed both DPs and pretangles/neuropil threads (PT/NTs), and two (20%) exhibited well-formed neuritic plaques and neurofibrillary tangles (Figure 1).

Figure 1.

Neuropathological changes in DS. (A) In case #6, aged 11 years, amyloid deposits are present in the frontal cortex in the absence of any tau pathology. (B) The area marked in A is shown at a higher magnification. (C) In case #7, aged 14 years, the hippocampus is mostly devoid of tau pathology. (D) At a higher magnification, the area marked in C is shown to exhibit very mild tau pathology in the form of a few pretangles and neuropil threads. This case did not exhibit any amyloid deposits. (E) In case #28, aged 56 years, Thal phase 5 (A score of 3) is evident, characterized by numerous amyloid deposits in the cerebellar cortex. (F) In the same case, Braak stage VI (B score of 3) is evident, showing abundant tau accumulation with tangles and neuropil threads in the primary visual cortex. (G) In case #34, aged 70 years, a section of the cerebellum reveals multiple vessels with intramural amyloid-β accumulation, indicating cerebral amyloid angiopathy (CAA). (H) The amygdala of case #33, aged 70 years, exhibits notable α-synuclein accumulation. This case also had α-synuclein-positive Lewy bodies in bilateral olfactory tracts/bulbs (not shown).

Among the ADNC- subjects (n = 8), 6 (75%) exhibited no evidence of neurodegenerative pathology and 2 (25%), harbored only rare hippocampal PT/NTs, and thus did not meet the criteria for a diagnosis of ADNC.

The average A score of the total sample was zero in the first decade. As depicted in Figure 2, the average A score then gradually increased from the second decade onwards. By the sixth decade of life, virtually all cases showed the highest possible A score (A3). Similarly, the mean B and C scores were zero in the first three decades, with a gradual increase from the fourth decade onwards. By the sixth decade of life, virtually all cases showed the highest possible B and C scores (B3 and C3, respectively).

Figure 2.

Average ABC scores and ADNC level by age decade in UTSW cohort (n = 34). ADNC level: 0-none, 1-low, 2-intermediate, 3-high (subjects in first decade of life not shown due to lack of pathological findings).

The non-white subjects (n = 13), with an average age of 21.1 years (range: 2–48 years), were notably younger compared to the white subjects (n = 18), whose average age was 41.8 years (range: 3–70 years) (z-score = −4.002, p = <0.001). The proportion of ADNC+ cases among non-white subjects (7/13, 53.8%) was significantly lower than that among white subjects (17/18, 94.4%) (χ² = 7.17, p = 0.012). In the ADNC+ group, non-white (n = 7) and white (n = 17) subjects did not differ in terms of average A-score (z-score = 0.501, p = 0.617) and C-score (z-score = −1.715, p = 0.087). However, non-white subjects had significantly lower average B scores (z-score = −2.238, p = 0.025) and ADNC levels (z-score = −2.219, p = 0.026) than the white subjects. As expected, ADNC- subjects were younger than ADNC+ subjects (Table 3).

Table 3.

Comparison of ADNC- and ADNC + DS subjects in UTSW cohort (n = 34).

	ADNC− (n = 8)	ADNC + (n = 26)	z-score*	p
Age at death, years, mean (range)	8.7 (2–21)	42.2 (11–70)	−202.5	<0.001
Brain weight, mean, grams (range)	966.7 (770–1070)	1077.9 (660–1335)	−93.5	0.124
			χ2 value	p
Female, n (%)	5 (63%)	9 (50%)	0.693	0.384**
White, n (%)	1 (14%)	13 (81%)	7.117	0.012**
CAA, n (%)	0 (0%)	15 (83%)	−	−
ARTAG, n (%)	0 (0%)	3 (12%)	−	−
HS, n (%)	0 (0%)	1 (4%)	−	−
Mineralization in basal ganglia, n (%)	0 (0%)	5 (24%)	−	−

^*Mann-Whitney U test.

^**Fisher's exact test.

ARTAG: aging-related tau astrogliopathy; CAA: cerebral amyloid angiopathy; HS: hippocampal sclerosis.

The majority (17/26, 65.4%) of ADNC+ subjects also exhibited cerebral amyloid angiopathy (CAA) (Figure 1G), while none of the ADNC- subjects did (Table 3). The youngest subject to show evidence of CAA was a 31-year-old black male.

Almost two-thirds of the cases (21/34, 61.8%) exhibited cerebrovascular disease (CVD) other than CAA (Table 1 and Supplemental Table 3). All of these 21 cases had arteriolosclerosis and/or arteriosclerosis (9 mild [42.9%], 10 moderate [47.6%], 2 severe [9.5%]). No large infarcts, microhemorrhages, or macrohemorrhages were identified. The youngest individual with mild arteriolosclerosis was a 19-year-old Hispanic male. All subjects older than 40 years had at least mild arteriolosclerosis and/or arteriosclerosis. Four cases (11.8%) showed mild atherosclerosis, and one case (2.9%) exhibited a microinfarct; all of these also had arteriolosclerosis and/or arteriosclerosis (Supplemental Table 3 and Supplemental Figure 1).

A small subset (3/26, 11.5%) of ADNC+ subjects had comorbid ARTAG, while none of the ADNC- subjects had evidence of ARTAG (Table 3). The youngest subject to display ARTAG was a 54-year-old male, and the oldest was a 70-year-old female.

As expected, all ADNC components correlated with each other in both the total sample and the ADNC+ subgroup. Both in the total sample (Figure 3) and in the ADNC+ subgroup (Figure 4), ABC scores and ADNC level were positively and strongly correlated with age. Age was negatively correlated with brain weight in the ADNC+ group; however, no correlation was observed in the total sample or the ADNC- group (Figure 5).

Figure 3.

Scatterplot: age at death versus ABC scores and ADNC level in UTSW cohort (N = 34).

Figure 4.

Scatterplot: age at death versus ABC scores and ADNC level in ADNC+ subjects in UTSW cohort (n = 26).

Figure 5.

Scatterplot: age at death versus brain weight in UTSW cohort (N = 34).

None of the assessed subjects over the age of 30 (n = 19) exhibited LATE-NC proteinopathy. However, in six cases (31.6%), an amygdala section was unavailable, precluding the assessment of low-level TDP-43 pathology (stage 1 LATE-NC). Only one subject (2.9%), a 70-year-old female, had HS, and this case did not demonstrate concomitant LATE-NC.

Only one subject out of 19 over the age of 30 (5.3%), who was the same single case showing HS, exhibited Lewy pathology in bilateral olfactory bulbs and unilateral amygdala sections (Figure 1). However, 16 out of 19 cases (84.2%) lacked olfactory bulb sections and six cases (31.6%) also lacked an amygdala section. Thus, “olfactory bulb only” or “amygdala-predominant” Lewy pathology could not be ruled out in these cases.

Among the subjects with available basal ganglia (BG) sections (n = 28), five cases (17.9%) exhibited remarkable mineralizations, particularly in the globus pallidus (Figure 6). All five of these cases also had concomitant ADNC. The youngest subject harboring BG mineralizations was a 47-year-old Hispanic female. To further characterize these mineralizations, BG sections from cases #30 and #31 were stained with Aβ, neurofilament, SMI-31, Prussian blue, Von Kossa, and PAS. The mineralizations were predominantly positive for iron and variably for calcium, while the remaining stains yielded negative results (Figure 6). Among the 8 ADNC- subjects and 26 ADNC+ subjects, one (12.5%) and five (19.2%), respectively, lacked BG sections.

Figure 6.

Mineralizations in the basal ganglia. (A) Low magnification view depicting mineralizations in the basal ganglia, particularly in the globus pallidus of case #30, aged 59 years. (B) High magnification view of A showing a focus with striking mineralizations. (C) The mineralizations in the same case (#30) are positive for ferric iron. (D) In case #31, aged 61 years, the mineralizations also exhibit positivity for calcium.

Among the subjects over 18 years of age with available brain weight data (n = 22), five subjects (22.7%) fell within the 10–25 percentile range, while 17 subjects’ (77.3%) brain weight was below the 10th percentile. Brain weights were unavailable for five adult subjects (14.3%).

Combined dataset

A combination of 6 different datasets including our sample yielded a total of 160 subjects with a mean age at death of 47 years (range 0–76 years). Supplemental Table 2 describes the characteristics of each case included in the combined dataset. The subjects consisted of slightly more males (n = 93, 58.1%) and only a small subset of cases (n = 15, 9.4%) was under the age of 18. The information on race was available for 36 subjects (22.5%), all of whom were sourced from our institutional dataset and the NACC data. Forty percent of the total sample (n = 64) had a genetically confirmed diagnosis of DS, while the remaining cases were diagnosed based on clinical features.

The majority (143 of 160, 89.4%) of the subjects had ADNC. Eleven (7.7%) of the 143 ADNC+ cases could not be assigned a definite ADNC level due to incomplete staging. Among the remaining ADNC+ subjects (n = 132), 22 (16.7%) had a low level of ADNC, 15 (11.4%) had an intermediate level, and 34 (25.8%) had a high level. Although 61 cases could not be assigned a definite ADNC level due to the lack of a C score, they were reliably assigned at least an intermediate level ADNC based on their A and B scores. Thus, at least 110 (76.9%) out of 143 ADNC+ subjects had intermediate or high level ADNC.

The youngest ADNC+ subject was an 11-year-old white female from our institutional dataset, while the oldest ADNC- subject was a 26-year-old male from Wegiel et al. ²¹ Almost all subjects over 40 years had intermediate or high level ADNC (n = 108/112, 96.4%). Among those older than 40 years, only four individuals (3.6%)—aged 41, 50, 50, and 60—had low level ADNC; the one aged 60 had only DPs with an A score of 3.

Only two of the 35 subjects younger than 40 years (5.7%) had intermediate or high level ADNC (aged 34 and 35, respectively). Among low-level ADNC+ subjects aged 40 years or younger (18/35, 51.4%), seven (38.8%) had only DPs, five (27.7%) had both DPs and PT/NTs, and six (33.3%) had both neuritic plaques (NPs) and neurofibrillary tangles (NFTs). Two subjects, aged 14 and 19, exhibited only PT/NTs, which did not meet the criteria for an ADNC diagnosis. Only one subject, a 26-year-old male, had NFTs in the transentorhinal cortex without cortical NPs, corresponding to a Braak stage of I and a B score of 1, with a Thal phase of 0, an A score of 0, and an overall diagnosis of PART. ³³

The average A score was zero in the first decade, gradually increasing from the second decade onwards (Figure 7). Similarly, the mean C score was zero in the first three decades, gradually increasing from the fourth decade onwards. By the seventh decade, virtually all cases had the highest possible A and C scores (A3 and C3, respectively). The average B score was zero in the first two decades, gradually increasing from the third decade onwards, closely approached the highest possible score (B3) in the sixth decade and finally peaked at B3 in the seventh and eighth decades.

Figure 7.

Average ABC scores and ADNC level by age decade in the combined dataset. ADNC level: 0-none, 1-low, 2-intermediate, 3-high (subjects in first decade of life not shown due to lack of pathological findings).

About three-fourths of the total sample (107/144, 74.3%) exhibited CAA (16 subjects did not have available information on the presence of CAA). The youngest subject to exhibit CAA was a 31-year-old black male from the UTSW cohort. Among the subjects who were 50 years old or older (n = 103, mean age 59.5 ± 5 years, range 50–76 years), only two subjects did not show CAA (1.9%), and they were the same subjects with low level ADNC (aged 50 and 60 years).

ARTAG was not evaluated in 58 cases (36.3%). Among the remaining 102 subjects, only five (4.9%), aged 54 to 70 years, exhibited ARTAG; all were also ADNC+.

Limbic-predominant age-related TDP-43 encephalopathy neuropathologic change was not systematically assessed in 38 cases (23.8%). Among the remaining 122 cases, an amygdala section was unavailable in 62 (50.8%), precluding the assessment of stage 1 LATE-NC. Although amygdala was not assessed in one of the studies, ¹⁹ five out of 56 cases examined by that study were found to have stage 2 LATE-NC, based on the available hippocampus sections. Among the subjects with available sections (n = 65) in the combined cohort, five (7.7%) had stage 1, and six (9.2%) had stage 2 LATE-NC.

HS was not assessed in 56 cases (35%). Among the remaining 104 subjects, only four (3.8%) had HS (aged between 61 and 70 years), two of which (both aged 61 years) also demonstrated concomitant LATE-NC (stage 2).

Alpha-synuclein proteinopathy was not assessed in 15 cases (10.7%) from our study sample because these individuals were younger than 30 years. Additionally, one study ²¹ assessed only the amygdala for α-synuclein pathology and did not report individual results. However, they found α-synuclein pathology involving the amygdala in 20 of 33 DS subjects. Thus, among a total of 125 subjects who were evaluated for α-synuclein pathology in the combined dataset, 24 subjects (19.2%) exhibited “amygdala-predominant,” two (1.6%) “brainstem-predominant,” one (0.8%) “limbic,” and seven (5.6%) “neocortical” Lewy pathology. Two cases with neocortical Lewy pathology were 51 years old, and the other five were all in their sixties. All cases with any level of α-synuclein pathology also exhibited concomitant ADNC of at least an intermediate level. Thirteen cases had no α-synuclein pathology, including in the olfactory bulbs and amygdala. However, 65 out of 125 cases (52%) lacked an olfactory bulb section and 54 cases (43.2%) also lacked an amygdala section. Thus, “olfactory-bulb only” or “amygdala-predominant” Lewy pathology could not be ruled out in about half of the subjects.

In the combined sample, age at death was negatively correlated with brain weight (r = −0.456, p < 0.001) and positively correlated with A score (r = 0.673, p < 0.001), B score (r = 0.716, p < 0.001), C score (r = 0.878, p = <0.001), ADNC level (r = 0.833, p < 0.001), and LATE-NC stage (r = 0.319, p = 0.020). Brain weight was negatively correlated with A score (r = −0.250, p = 0.031), B score (r = −0.371, p = 0.001), and LATE-NC stage (r = 0.319, p = 0.020); but not correlated with C score (r = −0.290, p = 0.063) and ADNC level (r = −0.258, p = 0.073).

Brain weight information was unavailable for 85 subjects (53.1%). Among the subjects over 18 years of age with available brain weight data (n = 69), one subject (1.4%) fell within 25–50 percentile range, seven subjects (10.1%) fell within the 10–25 percentile range, while 61 subjects (88.4%) fell below the 10th percentile.

A comparison of UTSW cohort with a combination of NACC cases and previously published data revealed that subjects in our cohort were significantly younger at death and had higher brain weights (Table 4). Our dataset also had lower ABC scores, while overall ADNC level was not statistically different between groups. The proportion of ADNC+ cases and the prevalence of CAA were significantly lower in our cohort, likely explained by the high number of younger subjects. Over 40% of our sample consisted of non-white subjects, whereas race information was available only for the NACC data (six white subjects) in the combined dataset. There were no significant differences in sex distribution, ARTAG prevalence, or HS.

Table 4.

Comparison of UTSW cohort (n = 34) and non-UTSW data (n = 126).

	UTSW institutional data	Others’ cohorts	z-score*	p
Age at death, years, mean (range)	34.3 (2–70)	52.5 (0–76)	−4.773	<0.001
Brain weight, mean, grams (range)	1054.1 (660–1335)	944.2 (656–1320)	2.728	0.006
A score (0–3), mean	1.9	2.7	−3.976	<0.001
B score (0–3), mean	1.3	2.4	−4.261	<0.001
C score (0–3), mean	1.4	2.2	−2.930	0.003
ADNC level (0–3), mean	1.6	1.9	−1.104	0.311
			χ2 value	p
Female, n (%)	18 (53%)	49 (53%)	2.172	0.171
Non-white, n (%)	13 (42%)	0 (0%)	−	−
ADNC	26 (76%)	117 (93%)	7.571	0.011
CAA, n (%)	16 (47%)	93 (83%)	12.961	<0.001
ARTAG, n (%)	3 (9%)	2 (3%)	1.682	0.330**
HS, n (%)	1 (3%)	3 (4%)	0.112	1.000**

^*Mann-Whitney U test.

^**Fisher's exact test.

ARTAG: aging-related tau astrogliopathy; CAA: cerebral amyloid angiopathy; HS: hippocampal sclerosis.

Apolipoprotein E (APOE) genotype was available for 75 cases (46.9%), which included cases from the NACC dataset (n = 6), Davidson et al. ¹⁹ (n = 34), Wegiel et al. ²¹ (n = 16), and Martá-Ariza et al. ²² (n = 19). Of these, 40 cases (53.3%) had the APOE3/3 genotype, 18 (24%) had APOE3/4, seven (9.3%) had APOE2/3, six (8%) had APOE4/4, and four cases (5.3%) had APOE2/4. Twenty-eight of the 75 cases (37.3%) carried at least one APOE4 high-risk allele for the development of ADNC. ³⁴ These APOE4 carriers had a slightly lower mean age at death (56.6 years) compared to non-carriers (n = 47, mean age = 58.5 years), although this difference did not reach statistical significance (z score = −1.225, p = 0.220). Additionally, APOE4 carriers did not differ from non-carriers in terms of A score (χ² = 0.734, p = 1.000), B score (χ² = 2.281, p = 0.520), C score (χ² = 0.580, p = 1.000), ADNC level (χ² = 1.33, p = 0.710), and presence of CAA (χ² = 0.354, p = 0.615).

Discussion

ADNC pathology

In the UTSW cohort and the combined dataset, we observed that the youngest individual to display Aβ pathology in the form of neocortical DPs was an 11-year-old female. Previous reports on DS patients have indicated the presence of DPs in the neocortex as early as eight years old, and in the hippocampus as early as nine years old. ³⁵ Our findings demonstrate an overall gradual age-related progression of Aβ pathology among DS subjects; however, DPs do not give way to neuritic plaques (NPs) until subjects are in their mid-thirties. This finding corroborates prior research indicating that DS patients frequently exhibit DPs during their second decade of life, with A scores gradually rising with age until the appearance of NPs in their thirties.^{19,21,35–37} These findings are not surprising, as individuals with DS possess an additional copy of chromosome 21 housing the APP gene, long recognized as the origin of the fibrillogenic Aβ peptide that accumulates and forms the DPs in the brain. ³⁸ Mutations in the APP gene are deemed adequate to induce early-onset AD. Consequently, DS is recognized as a genetically predisposed form of AD. ³⁹

ADNC is universally present in DS patients older than 30 years, with the majority exhibiting at least an intermediate level. Only four cases over the age of 40 (range:41–60) had low level ADNC. None of these four cases had genetic confirmation of a DS diagnosis. One possible explanation for our finding is that these four cases might have been erroneously diagnosed with DS. Notably, in a UK study, ⁴⁰ about 30% of individuals with clinical findings suggestive of DS, such as epicanthal folds, protruding tongue, and simian crease, were shown to have normal karyotypes. On the other hand, although the majority of DS patients have complete trisomy 21, mosaicism for trisomy 21 occurs in about 2% of cases, and partial trisomy 21 occurs in an even smaller subset. ⁴¹ Case reports of DS patients with partial trisomy 21 have documented individuals who lived into their 70 s without clinical dementia and showed no evidence of ADNC at autopsy.^42,43 Therefore, it is plausible that the four patients over the age of 40 with lower-than-average levels of ADNC in the combined dataset might have an undetected partial trisomy 21. Alternatively, this small subgroup of DS patients in our study might have been resistant to the effects of an extra copy of chromosome 21 through an unknown mechanism. ⁴⁴

Neurofibrillary tangles were not observed in DS brains until the mid-thirties except for a 26-year-old male in the combined dataset. Although not reported in the relevant study, ²¹ this case meets criteria for PART because it did not demonstrate any concomitant Aβ pathology. For the rest of the combined dataset, early forms of tau pathology (i.e., PT/NTs) emerged in the second decade of life. While DPs were often concurrent with PT/NTs in many subjects, two of them exhibited solely PT/NTs (both in our UTSW cohort), challenging the amyloid cascade hypothesis, which suggests that increased Aβ is the primary inciting event in AD that leads to the development of tau pathology. ⁴⁵

Braak's early study of 2332 nonselected brains from 1- to 100-year-old individuals, ⁴⁶ demonstrated that tau pathology emerged in the first decade of life in the form of pretangles in the noradrenergic projection neurons of the locus coeruleus and in the second decade in the form of neurofibrillary tangles in the hippocampus. The Aβ pathology was not observed until the fourth decade of life, except for a 17-year-old male with DS (the only DS subject in the cohort). Moreover, all the cases demonstrating Aβ pathology (n = 1301) also showed concomitant tau pathology. According to these findings, tau pathology is highly prevalent in the general population, even in younger subjects. Our observations may be in part explained by a sampling error, where a low level Aβ accumulation may preferentially affect certain cortical regions, or show patchy distribution that evaded sampling, or alternatively might be explained by the expected accumulation of pathologic tau protein in the brains of young individuals irrespective of the Aβ-driven process seen in DS. Because Aβ pathology, in contrast to tau pathology, appears to be the initial and relatively uniform pathology in DS, the pathogenesis of the development of ADNC in DS may be different from that of sporadic AD, and is more supportive of the amyloid cascade hypothesis in this patient group. Of note, the presence of tau pathology in the locus coeruleus was not assessed in the current study, which almost certainly has resulted in the overall underestimation of the tau pathology prevalence. While Aβ is considered the initiating molecule in the pathogenesis of both sporadic AD and DS-associated AD, the exact underlying mechanisms differ: in DS, Aβ accumulation arises primarily due to APP gene triplication and subsequent overproduction, whereas in sporadic AD, Aβ dysregulation is attributed to multifactorial processes. ⁴⁷

Down syndrome can show variations in prevalence, as well as in morphological and clinical features, among different ethnic and racial groups.^48–50 For the first time, this study assessed neurodegenerative pathology in non-white DS subjects, who comprised over 40% of our cohort. We demonstrated that the proportion of ADNC+ subjects was lower in non-white DS subjects, who also had significantly lower B scores and ADNC levels compared to white DS subjects. However, non-white subjects were significantly younger at death, and the ADNC+ subgroup consisted mostly of older white subjects (17 white versus 7 non-white). Thus, higher levels of ADNC are most likely associated with advanced age rather than with white race. Race information was available in the NACC dataset, where all subjects were white. However, none of the other four studies^19–22 in the combined dataset reported race as a characteristic, precluding us from making robust conclusions on the association between race and neurodegenerative pathology in DS subjects.

CAA

Only 2 out of 35 subjects younger than 40 years (5.7%) exhibited CAA in the combined dataset. However, nearly all subjects aged 50 years or older (98.1%) exhibited CAA, except for two, who also had low-level ADNC. This incidence is significantly higher than that found in early-onset sporadic ADNC cases (86%) (χ² = 10.16, p = 0.001), whose mean age at death was 66.4 years (range: 56–81 years). ⁵¹ Intriguingly, in a NACC dataset study, ⁵² among 436 ADNC+ subjects with an average age of 70.7 years, only 42.4% were found to have comorbid CAA. Moreover, a meta-analysis of 100 studies showed that the prevalence of CAA was 48% among 11,400 AD subjects with an average age of 79.8 to 80.6 years at death. ⁵³ Based on our findings, which corroborate those of others,^51,53,54 we conclude that CAA is more prevalent in early-onset AD compared to late-onset AD. This conclusion is also supported by imaging and neuropathological studies that demonstrated a negative correlation between the presence and severity of CAA and age at death in AD subjects.^54,55 Because the onset of ADNC in DS subjects is even earlier than that of sporadic early-onset AD, DS appears to be associated with a higher likelihood of CAA. ⁵⁶ There are complex interactions at multiple levels between CAA and ADNC: CAA can contribute to the development and acceleration of ADNC pathology, rather than being merely a comorbid pathology ⁵⁷ or can be a downstream manifestation of increased Aβ production and impaired vascular clearance.

Approximately two-thirds (62%) of the UTSW cohort exhibited arteriolosclerosis, a proportion notably lower than the 92% reported by Liou et al. ²³ in individuals with DS. However, the mean age in their study was 61 years, considerably older than our cohort's mean age of 34 years. Similarly, only 12% of our cohort showed mild atherosclerosis, compared to 22% in the DS cohort reported by Head et al., ⁵⁶ whose participants had a mean age of 55 years. These findings illustrate that arteriolosclerosis is relatively common even in younger individuals with DS and atherosclerosis is much less common in this younger DS cohort. Nonetheless, both arteriolosclerosis and atherosclerosis appear to increase in prevalence with advancing age.

Previous research has shown that individuals with DS have a higher risk of ischemic stroke, and a greater burden of atherosclerotic risk factors compared to age-matched non-DS controls. ⁵⁸ Due to the lack of clinical data in our cohort, we were unable to determine whether the presence of CVD was associated with specific cerebrovascular risk factors such as hypertension or diabetes.

ARTAG

ARTAG is a 4R-tauopathy commonly observed in individuals over the age of 60 and is characterized by thorn-shaped and granular/fuzzy astrocytes predominantly in subpial, subependymal, and perivascular regions. ²⁸ Only 12% of ADNC+ cases in our institutional dataset and 4.9% in the combined dataset (age range: 54–70 years) were found to have comorbid ARTAG. ARTAG is commonly observed in aging brains, but is not necessarily associated with ADNC independent of age. ⁵⁹ The prevalence of ARTAG in AD ranges from 38% to 66%, depending on whether AD is of early-onset or late-onset. ⁵¹ While ARTAG can co-occur with other tauopathies, such as ADNC and chronic traumatic encephalopathy (CTE), it possesses distinct pathological and anatomical features. ⁶⁰ The very low rates of ARTAG observed in our study support the idea that ARTAG is primarily an aging-related phenomenon and only incidentally associated with ADNC pathology.

Alpha-synuclein pathology

Alpha-synuclein pathology is one of the most common comorbid proteinopathies in patients with ADNC, with rates ranging from 33% to 53%, depending on age and whether low-level (olfactory bulb only and/or amygdala-predominant) pathology was evaluated.^51,61,62 These reported rates are much higher than the frequency that we observed in our series (5.3% in our dataset and 27.2% in the combined dataset). Among the α-synuclein-positive cases within the combined dataset, most exhibited amygdala-predominant Lewy-related pathology (24/34, 70.6%). The relatively low levels of α-synuclein pathology in our institutional sample may be explained by the relatively younger age at death compared to typical sporadic AD cases. However, these rates are still lower than those reported in early-onset AD subjects, with one study reporting 49% ⁵¹ and another 53% ⁶² in cases with mean age at death ranging from 53 to 64 years. Spina et al. ⁵¹ found that 44.7% of the early onset AD cases with α-synuclein pathology were in the category of amygdala-predominant Lewy-related pathology (versus 15% in the late-onset AD group). However, they did not assess the olfactory bulbs for the presence of α-synuclein pathology. Regardless, the stage of Lewy-related pathology in our sample appears to be lower than that observed in sporadic early-onset AD.

Notably, Sepulveda-Falla et al. ⁶² have shown that 57% to 70% of familial AD cases with the PSEN1 mutation exhibited comorbid α-synuclein pathology. Therefore, the rate of comorbid α-synuclein pathology in our sample is lower not only than that in sporadic AD cases of a similar age range, but also in cases with another genetic cause of AD, which is typically associated with a much younger onset of ADNC pathology.

In contrast to our findings, an early study detected comorbid α-synuclein pathology in 50% (8/16) of DS cases with ADNC (aged 53–70 years). ⁶³ Six of these cases had amygdala-predominant, one brainstem-predominant, and one neocortical stage Lewy-related pathology. However, the authors did not sample olfactory bulbs, which might have resulted in missing olfactory bulb only Lewy-related pathology. The discrepancy between our findings and those of Lippa et al. ⁶³ could be attributed to the absence of amygdala sections in nearly half of our study population.

Although the lack of olfactory bulbs and amygdala sections may explain the relatively low levels of α-synuclein pathology in our study, the fact that brainstem-predominant and neocortical stages of Lewy pathology (29.4%) were not observed as frequently as in sporadic early-onset AD (55.3%) ⁵¹ might indicate a different type of interaction between ADNC pathology and α-synuclein pathology in these patient groups.

Most of the cases with Lewy body pathology in the combined dataset (20 out of 34; 58.8%) were derived from the study by Wegiel et al., ²¹ which reported α-synuclein–positive Lewy bodies in the amygdala in 54.5% of individuals with DS aged 41–59 years and in 75% of those aged 61–72 years. These findings suggest that the presence of Lewy body pathology may be more closely associated with advanced age than with DS itself.

LATE-NC

LATE-NC is defined by the accumulation of TDP-43 protein in the limbic regions of the brain and can occur with or without HS. ³⁰ It is detected in up to 57% of ADNC cases, ⁶⁴ and is associated with more significant brain atrophy and greater memory loss in AD patients. ⁶⁵ None of the evaluated subjects in UTSW cohort and 16% of the subjects in the combined dataset exhibited LATE-NC. Approximately half of the positive cases were classified as stage 1, while the other half were classified as stage 2 LATE-NC. The prevalence observed in DS cases in our study is slightly higher than that found in early-onset ADNC cases (8%). ⁵¹ Since an amygdala section was not available for nearly half of the combined dataset, the exact prevalence of stage 1 LATE-NC is unknown and likely underestimated. However, none of the cases had stage 3 pathology and LATE-NC was positively correlated with age at death, suggesting that this condition is more closely associated with age than with a diagnosis of DS. Age being the primary driver of LATE-NC is further supported by Sepulveda-Falla et al., ⁶² who showed that LATE-NC was infrequent in young subjects with PSEN1 mutations.

Hippocampal sclerosis

HS is defined as severe neuronal loss and gliosis in predominantly the subiculum and CA1 subfield of the hippocampus. ³¹ It is associated with hippocampal atrophy, severe amnesia, and progressive dementia. ⁶⁶ HS is more common in older individuals, with a prevalence of 18% in those aged 90 years or older compared to 5–9% in those younger than 90 years.^67,68 Although pure cases of HS exist, it frequently co-occurs with other neurodegenerative pathologies. A study of 409 individuals aged 90 years or older showed co-occurrence rates of HS with ADNC at 77%, with CAA at 67%, and with LATE-NC at 65%. ³¹ In our institutional dataset, only one case, a 70-year-old individual, had HS. In the combined dataset, four cases (3.4%) exhibited HS, all of whom were older than 60 years of age. All HS-positive cases also had concomitant ADNC and CAA, with half of them also showing concomitant LATE-NC. Although this prevalence is somewhat lower than previously reported, our study population was younger at the time of death compared to those in other studies.^23,67,68 Additionally, HS is reported to be more common in the left hippocampus (82%) than in the right (25%), and bilateral involvement is observed in only 7% of affected subjects. ³¹ Because only unilateral sampling of the hippocampus is available in the current study, the actual prevalence of HS in DS cases might be higher than our findings show. However, it is unlikely that there is a relationship between HS and a diagnosis of DS independent of age, as all HS cases were in the oldest age decade.

Basal ganglia mineralization

Mineralization of the BG occurs as an incidental finding or can be associated with metabolic disorders, mitochondrial diseases, infections, trauma, and numerous hereditary syndromes. ⁶⁹ In a population-based computed tomography study, the prevalence of BG mineralization in the form of “punctate calcifications” was 11.8% in individuals younger than 65 years (n = 2894) and 28.3% in those aged 65 years or older (n = 2205), indicating that incidence correlates with age. However, “patchy calcifications” were much less frequent, occurring in 0.8% of the younger subgroup and 2.6% of the older subgroup. ⁷⁰ In our study, BG sections were considered positive for mineralizations only if they exhibited a remarkable amount (as referenced in Figure 6), likely corresponding to the “patchy calcifications” mentioned in the previous study. Since most of our study population was younger than 65 years, a prevalence of about 18% BG calcification in DS subjects appears considerably higher than that in the general population.

Two previous neuropathology studies have reported varying frequencies of BG mineralizations in DS patients, ranging from 45% ⁷¹ to 87.5%. ⁷² These studies, like ours, found that the globus pallidus region of the BG was particularly prone to mineralization. However, it should be noted that both studies used a relatively lower threshold to detect BG mineral deposition than our study; even “single globules in some fields” were deemed positive for mineralization. This might explain the relatively lower prevalence of BG mineralizations in our sample. Additionally, Mann et al. ⁷² found that ADNC+ cases without DS, who were older than 75 years at death, had a lower prevalence of BG calcifications (66.7%) compared to DS cases (87.5%), who were aged 13–71 years at death. These high rates of BG mineralization in DS cases can be attributed to “accelerated aging,” a phenomenon associated with the hypothesis that trisomy 21 increases the biological age of tissues, leading to a progeroid (i.e., clinical features of premature physiological aging) phenotype in DS patients.^73–76

Further evaluation of the BG mineralizations revealed that they were predominantly positive for iron and variably positive for calcium. This finding is consistent with a study that evaluated hippocampal mineralization and found that DS cases exhibited Von Kossa positivity (i.e., calcium) in the mineralization, which was comparable to ADNC subjects. ⁷⁷ However, the authors did not assess the presence of iron in the mineralization. Interestingly, a postmortem evaluation of two subjects who exhibited BG mineralization on imaging and had no history of neurodegenerative or neuroinflammatory diseases revealed that these mineralizations were predominantly positive for iron and also for calcium. ⁷⁸ Hence, the structural nature of BG mineralization in DS subjects does not seem to be disease-specific, despite its high prevalence.

Brain weight

Imaging studies have shown that children and adolescents with DS have smaller total brain volume, total gray matter, white matter, cortical lobar, hippocampal, and cerebellar volumes compared to age-matched controls.^79–81 In a review of the autopsy literature, Whalley ⁸² found that the brains of 50 DS patients, ranging in age from 12 to 65 years, exceeded 1200 grams in only 8 cases and had a weight less than 1000 grams in 18 cases, averaging 24% less than that of control adult brains. In our study, for the first time using robust percentile estimates from over 4000 adult autopsies, ³² we have shown that almost all DS subjects’ brain weights were below the 25th percentile for the subject's age and sex. In the combined dataset, ∼90% of cases had brain weights below the 10th percentile and only one subject, a 63-year-old male with a high level of ADNC, had a brain weight between the 25th and 50th percentiles. Even subjects aged 19–21 years at death with no ADNC had lighter than expected brains, indicating that atrophy or underdevelopment is an early event that contributes to a smaller average brain size of the DS subjects. It can be concluded that the lower brain weight in DS subjects is not necessarily a neurodegenerative process, but rather a neurodevelopmental one. However, ADNC can further contribute to this process, as brain weight was negatively correlated with A and B scores in the combined dataset.

Strengths and limitations

The study's strengths include a wide age range and comprehensive pathological assessment utilizing contemporary diagnostic approaches and consensus criteria. This study is the first to assess neurodegenerative pathologies in the brains of non-white DS subjects, who comprised over 40% of our cohort. By combining our institutional data with DS cases from the NACC database and four previously published studies,^19–22 we provide the largest sample size of DS subjects (N = 160) evaluated for the presence of neurodegenerative disorders. By including DS subjects from early childhood to older adulthood, we gained valuable insights into the early onset and progression of neurodegenerative pathologies. The systematic evaluation of multiple neuropathological features, such as Aβ, tau, and α-synuclein proteinopathies, as well as LATE-NC, HS, ARTAG, CAA, and CVD enhances our understanding of these processes in DS. Furthermore, the detailed age-specific data allow for a nuanced characterization of pathology progression, contributing significantly to our understanding of neurodegeneration in DS.

However, this study also has several limitations. The lack of clinical evaluation of dementia precludes correlating neuropathological findings with clinical symptoms. Additionally, the diagnosis of DS was not confirmed by karyotype analysis in most cases, raising the small possibility of inaccurate diagnosis. Because of the unavailability of olfactory bulb and amygdala sections in many cases, the assessment of α-synuclein and TDP-43 accumulation was limited, potentially underestimating the prevalence of these pathologies at lowest stages. While race was assessed in our study and the NACC dataset, other studies included in the combined dataset did not report race, limiting the robustness of evaluating race as an independent variable. Due to the lack of clinical data, the association between high rates of CVDs and clinical characteristics could not be assessed. Finally, only unilateral hippocampal sections were available, potentially underestimating the prevalence of HS and other pathologies, which may demonstrate differential degree of involvement impacted by laterality.

Conclusions

This study elucidates the distinct trajectory of neurodegenerative pathologies in individuals with DS, revealing an early onset of Aβ deposition and the progressive development of other neuropathologies with age. The findings support the notion that neurodegeneration in DS follows a characteristic progression pattern, somewhat distinct from sporadic ADNC, with significant implications for understanding the underlying mechanisms and optimizing potential interventions for this vulnerable population. In terms of comorbid pathologies, CAA, arteriolosclerosis, and BG mineralizations tend to be highly prevalent in DS patients. On the other hand, HS, ARTAG, LATE-NC, α-synuclein proteinopathy, and atherosclerosis do not appear to be specifically associated with a diagnosis of DS. Further research, including genetic confirmation and comprehensive clinical assessments, is essential to elucidate our understanding and improve diagnostic and therapeutic strategies for this population.