Multiomics and proteomic insights into Alzheimer’s disease biology in Down syndrome

Abstract

Introduction:

Down syndrome (DS) confers a high risk of Alzheimer’s disease (AD) and is a genetically determined form of AD. As such, DS provides a uniquely informative biological context in which to investigate AD initiation and progression. Defining the molecular mechanisms that link trisomy 21 to neurodegeneration has broad implications for AD biology and neurotherapeutic development

Areas covered:

This review summarizes findings from brain, cerebrospinal fluid, and blood-based proteomic studies, integrated with transcriptomic and multiomics analyses, to characterize molecular pathways underlying AD in DS. The literature was identified through iterative PubMed/MEDLINE searches and manual review of reference lists, considering studies available through June 2026 with no limitation to publication dates.

Expert opinion:

Brain, lesion-specific, cerebrospinal fluid, and blood-based proteomics, interpreted alongside transcriptomic and complementary omics data, position DSAD as a network-level disorder in which amyloid and tau pathology interact with immune, vascular, metabolic, synaptic, and proteostasis pathways. This integrated proteomic framework helps define shared and subtype-specific mechanisms across DSAD, sporadic AD, and autosomal dominant AD, while supporting biological staging, patient stratification, and therapeutic target discovery.

1. Introduction

1.1. Down syndrome prevalence and characteristics

Down syndrome (DS) is the most common human chromosomal disorder and results from partial or complete triplication of human chromosome 21 (Hsa21), occurring in approximately 1 in 700–1,000 live births worldwide [1–3]. Most of the DS cases arise from meiotic nondisjunction (∼96%), with mosaicism and translocation accounting for the remainder [4–6].

Trisomy 21 produces widespread biological effects, including intellectual disability and altered brain development [7]. Neuroimaging and neuropathological studies show delayed maturation and reduced brain volume in DS, particularly in frontal and temporal regions, as well as accelerated age-related neuronal loss [8–10]. Although DS phenotypes vary widely, trisomy 21 consistently induces genome-wide transcriptional and proteomic changes that extend beyond Hsa21 itself [11]. Multiomics studies demonstrate that gene-dosage imbalance reshapes immune signaling, endosomal trafficking, extracellular matrix organization, and metabolic pathways across the lifespan, creating a biological landscape in which neurodegenerative processes later emerge.

1.2. The relationship between Down syndrome and Alzheimer’s disease

Medical advances have greatly increased life expectancy in DS to over 60 years [12–14], revealing a striking association with Alzheimer’s disease (AD). Nearly all individuals with DS develop AD neuropathology by ~40 years of age, and the majority progress to dementia later in life, making AD the leading cause of death in this population [15,16]. Although the age at dementia onset is often described as highly variable in DS [17–21], dementia rarely occurs before age 40 despite the shared genetic predisposition. The timing and progression of clinical symptoms resemble those observed in autosomal dominant AD (ADAD) [22]. As in ADAD studies, estimated years to onset (EYO) is commonly used to align individuals by disease stage rather than chronological age [23,24].

The strong association between DS and AD is driven by triplication of the amyloid precursor protein (APP) gene on Hsa21 [25–27]. Evidence from rare cases with partial trisomy 21 carrying only two copies of APP and absent AD pathology even in late adulthood, indicates that APP triplication is necessary for early-onset AD development in DS [20,25,28]. Accordingly, DS is often regarded as a genetically determined form of AD [20,29] (Figure 1(A)). However, trisomy 21 perturbs multiple biological systems across the lifespan, with alterations in immune signaling, endosomal trafficking, and extracellular matrix regulation emerging long before amyloid accumulation, indicating that AD pathology in DS develops within a broader trisomy-driven molecular framework.

Figure 1.

Down Syndrome as a Genetically Defined form of Alzheimer’s Disease. A. Schematic illustrating trisomy 21. Key chromosome 21 genes implicated in Alzheimer’s disease pathology are highlighted, including APP, DYRK1A, RCAN1, SOD1, BACE2, SYNJ1, and S100β. Their increased dosage reflects genetic mechanisms by which trisomy 21 contributes to Alzheimer’s disease – related molecular pathways in Down syndrome. B. Temporal and regional progression of Alzheimer’s disease neuropathology in Down syndrome. Diffuse Aβ_1–42 plaques emerge early, precede neuritic plaque formation, and show region-specific accumulation, with increasing post-translational modification over time. Tau pathology follows, beginning in the hippocampus and progressing to widespread neurofibrillary tangle formation, with greater overall Tau burden compared with non-Down syndrome Alzheimer’s disease. The long interval between early Aβ_1–42 deposition and typical dementia onset emphasizes that amyloid accumulation is an early and necessary component of DSAD biology, while clinical expression emerges over time through subsequent Tau, synaptic, inflammatory, vascular, metabolic, and white matter changes. C. Horizontal gradient bars depict the approximate temporal onset and progressive severity of key pathobiological processes across the DS lifespan: (1) trisomy 21 as a lifelong genomic context, with triplication of AD-relevant genes. (2) Neuroinflammation, evolving from mixed M1/M2b microglial phenotypes in younger DS to predominant M2b profiles in DS-associated AD; (3) endo-lysosomal dysfunction, driven by APP β-CTF – mediated endosomal hyperactivation and lysosomal deacidification; (4) white matter and oligodendrocyte pathology, progressing from commissural and limbic tract abnormalities to corticocortical involvement; (5) proteostasis disruption and oxidative stress, including oxidative modification of protein quality control machinery (UCHL1, GRP78, cathepsin D) and reduced proteasome activity; and (6) synaptic and mitochondrial decline, including loss of synaptic markers (SYP, GAP43) and reduced mitochondrial oxidative phosphorylation and NRF2 antioxidant responses. Vertical bars represent severity of molecular changes with age and disease progression, increasing from left to right. Figure created in Biorender.com.

Within this framework, APP triplication is best understood as a major initiating axis rather than a complete explanatory model for DSAD. Although lifelong Aβ overproduction establishes a genetically anchored amyloidogenic state, the progression from trisomy 21 to dementia likely reflects a broader molecular vulnerability landscape shaped by altered neurodevelopment, baseline immune activation, endolysosomal dysfunction, oxidative and mitochondrial stress, vascular and extracellular matrix remodeling, and white matter vulnerability. From this perspective, DSAD may be conceptualized as a systems-level network disorder in which amyloid pathology is one major component of a temporally evolving interaction among gene dosage, cell-type-specific responses, and tissue-level dysfunction [20,29–32].

1.3. The neuropathology of Alzheimer’s disease in Down syndrome

The neuropathological hallmarks of AD in DS closely resemble those observed in sporadic AD and ADAD, both in morphology and regional distribution [20]. Diffuse amyloid plaques composed predominantly of Aβ_1–42 are detectable in DS before 20 years of age [33], precede cortical neuritic plaque formation, and remain more abundant than Aβ_1–40 plaques across the lifespan [33,34]. By approximately 30 years of age, diffuse Aβ_1–42 deposits are evident in the cerebellum and striatum, although fibrillar plaques remain scarce in these regions even in older individuals, supporting region-specific progression of amyloid pathology [35,36] (Figure 1(B)). In parallel, Aβ undergoes extensive post-translational modification, including isomerization, oxidation, and formation of pyroglutamate variants, which increase with age and disease stage [37–39].

The early appearance of diffuse Aβ_1–42 deposits also highlights an important issue in AD biology: amyloid accumulation and dementia are temporally related but not simultaneous. In DS, APP triplication establishes a highly penetrant amyloidogenic trajectory, and clinical and biomarker studies aligned by estimated years to onset indicate that AD pathophysiology follows a relatively predictable sequence across the DS lifespan. However, Aβ deposition begins decades before typical dementia onset, suggesting a long biological latency between the initiation of amyloid pathology and the emergence of clinical symptoms. This temporal gap suggests that amyloid-driven disease becomes clinically expressed over time through downstream processes, including tau propagation, synaptic failure, neuroinflammation, vascular dysfunction, mitochondrial stress, white matter injury, and impaired proteostasis. From this perspective, DSAD may resemble other genetically driven forms of AD, with differences in apparent predictability partly reflecting lifespan and the ability to align individuals by estimated years to onset [20,23,29,31,33,35].

In DS, amyloid also accumulates prominently within cerebral blood vessels, with vascular deposition occurring decades earlier than in late-onset AD (LOAD) [40]. Early work by Glenner and Wong isolated Aβ from DS cerebral vasculature and demonstrated its biochemical similarity to Aβ in LOAD, establishing a direct connection between chromosome 21 and AD pathology [21]. Subsequent identification of APP mutations on Hsa21 that selectively increase Aβ_1–42 production in ADAD further supported a shared mechanism of Aβ dysregulation across LOAD, ADAD, and DSAD [41,42].

Modified Aβ species contribute substantially to DS neuropathology. Older individuals with DS exhibit elevated plasma levels of pyroglutamate-3 Aβ, and DS brains contain abundant pyroglutamate-11 Aβ within plaque cores and vascular deposits [43–45]. These phosphorylated and pyroglutamate-modified Aβ species are more prominent in DS plaques than in early-onset AD (EOAD) [46]. Despite a low prevalence of systemic vascular risk factors, cerebral amyloid angiopathy (CAA) is more frequent and severe in DS than in EOAD or sporadic AD and appears largely independent of APOE ε4 status, unlikely in EOAD or late-onset AD [18,47,48].

Tau pathology also develops early in DS. Abnormal tau is first detected in the appear in the hippocampal molecular layer between 30 and 40 years of age, followed by neurofibrillary tangle (NFT) formation in CA1 and the subiculum and neuronal loss in the entorhinal cortex [49]. Although the anatomical distribution parallels non-DS AD, tau burden is consistently greater in DS, indicating a more severe neurodegenerative process [50,51].

Recent spatially resolved proteomic analyses revealed that amyloid deposits in DS contain complex networks of immune, endo-lysosomal, and extracellular matrix–related proteins. These molecular signatures largely overlap with those observed in ADAD and sporadic AD, suggesting that amyloid recruits a conserved pathological microenvironment despite distinct genetic origins [46,52].

Together, these features establish DS as a uniquely informative context for AD research. This biological framework has prompted increasing use of multiomics and proteomic approaches to define molecular pathways underlying disease initiation and progression. The near-universal development of AD pathology and the ability to identify trisomy 21 early in life provide an unparalleled opportunity to study preclinical disease stages, validate biomarkers linked to cognitive decline, and inform preventive and early-intervention strategies relevant across AD subtypes [40,53].

1.4. Literature search methodology

This article was prepared as a narrative expert review. PubMed/MEDLINE was used as the primary bibliographic database, supplemented by manual screening of reference lists from relevant primary research articles and reviews. Searches were conducted iteratively using combinations of terms related to Down syndrome, Alzheimer’s disease, DSAD, trisomy 21, APP, amyloid, tau, proteomics, multiomics, transcriptomics, cerebrospinal fluid biomarkers, plasma biomarkers, neuroinflammation, endo-lysosomal dysfunction, mitochondrial dysfunction, oligodendrocytes, white matter, cerebral amyloid angiopathy, blood–brain barrier, neurovascular unit, and clinical trials. Literature available through June 2026 was considered, with no formal lower publication-date limit. Recent studies from the past 10 to 20 years were prioritized when available, particularly for omics technologies, biomarkers, imaging, and therapeutic development. Older studies were included when they represented foundational neuropathological, genetic, biochemical, or mechanistic evidence directly relevant to DSAD biology. Articles were selected based on relevance to the biological mechanisms linking trisomy 21 to AD pathogenesis, contribution to proteomic or multiomics interpretation, methodological quality, and importance for the conceptual framework of the review.

2. Review

2.1. Alzheimer’s disease pathogenesis in Down syndrome

Beyond amyloid and tau pathology, individuals with Down syndrome-associated Alzheimer’s disease (DSAD) exhibit prominent neuroinflammatory, endo-lysosomal, and white matter alterations that overlap with mechanisms implicated in sporadic and early-onset AD [54–61] (Figure 1C). These processes contribute to disease progression and provide biological context for multi-system and omics-based studies discussed in the following sections.

2.1.1. Genes of interest for Alzheimer’s disease in Hsa21

The strongest genetic link between DS and AD is triplication of APP. However, Hsa21 also contains multiple genes whose overexpression affects AD-relevant pathways, including tau phosphorylation, oxidative stress, endosomal trafficking, and immune signaling [21,62–65] (Figure 1(A)).

Several Hsa21 genes converge on tau dysregulation. DYRK1A phosphorylates tau directly and facilitates further phosphorylation by GSK3β, while also altering tau splicing to increase the neurodegeneration-associated 3R:4R tau ratio [64]. Accordingly, individuals with DS aged 30–40 years and older show increased DYRK1A-positive and 3R Tau-positive NFTs than those with sporadic AD [64]. RCAN1 inhibits calcineurin and may further promote tau phosphorylation by reducing phosphatase activity and increasing GSK3β signaling [66,67]. Aβ_1–42 may amplify this cascade by upregulating both RCAN1 and DYRK1A [68,69].

Oxidative stress pathways are also affected by Hsa21 dosage. Triplication of SOD1, a key antioxidant enzyme, disrupts redox balance by increasing hydrogen peroxide accumulation [70,71]. S100β, an astrocyte-derived neurite growth factor increased in neural progenitor cells in DS [69,70], is markedly increased around amyloid plaques and reflects astrocyte activation rather than developmental effects alone [72,73].

Additional Hsa21 genes influence amyloid processing and trafficking. BACE2 can generate Aβ through β-site cleavage of APP but may also reduce Aβ formation via alternative α- or θ-site cleavage, indicating a context-dependent role in amyloid metabolism [74–77]. SYNJ1, which regulates endocytosis and endosomal trafficking, is upregulated in DS brains and correlates with Aβ levels, implicating endosomal dysfunction in DS-related neurodegeneration [78,79]. Finally, triplication of interferon receptor genes (IFNAR1, IFNAR2, IFNGR2 and IL10RB) results in chronic interferon signaling that may promote microglial activation and neurotoxicity [80].

Although DSAD differs from autosomal dominant AD caused by APP, PSEN1, or PSEN2 variants, recent case reports involving dual genetic risk support the broader concept that AD phenotypes are shaped by interacting molecular mechanisms rather than by single-gene effects alone. Ogg et al. reported a rare case of an individual with both DS and the PSEN2 N141I variant, in whom compounded genetic risk was associated with increased total Aβ burden across multiple brain regions compared with individuals carrying either risk context alone, but without corresponding increases in hyperphosphorylated Tau or neuroinflammatory markers [81]. Similarly, Bisceglia and collaborators described an individual with MCI carrying two PSEN1 variants, K311R and E318G, but without abnormal amyloid accumulation on CSF or amyloid PET measures, highlighting the complexity of variant interpretation and genotype-phenotype relationships in AD [82]. In other case report, Chang and colleagues described a 33-year-old patient with EOAD carrying both PSEN2 V214L and IDE R261Q variants, with reduced CSF Aβ42 and mild cortical atrophy. The case is mechanistically relevant because it combines a rare PSEN2 variant potentially affecting γ-secretase-mediated Aβ production with a previously unreported IDE variant in a gene involved in Aβ degradation. However, the patient’s sister carried the same two variants without clinical manifestations, leading the authors to conclude that these variants may contribute to EOAD through incomplete penetrance, synergistic effects, or interactions with additional genetic, environmental, or lifestyle modifiers, rather than acting as fully deterministic mutations [83]. Thus, compared with the Ogg et al. DS plus PSEN2 case, which suggests additive effects on Aβ burden, and the Bisceglia et al. PSEN1 double-variant case, which illustrates uncertainty in variant interpretation when amyloid biomarkers are negative, the Chang et al. report highlights the importance of amyloid production-clearance balance and modifier burden in shaping clinical expression. Together, these reports should be interpreted cautiously because they are single-case studies, but they reinforce the relevance of genetic interaction, amyloid production and clearance balance, and modifier burden when comparing DSAD with other genetically determined forms of AD.

2.1.2. Neuroinflammation in DSAD

Neuroinflammation is a core feature of DSAD and contributes to disease progression alongside amyloid and Tau pathology [84,85]. In sporadic AD, neuroinflammation is increasingly viewed as a dynamic process rather than a uniformly detrimental response. Early immune activation may be partially adaptive, supporting Aβ containment, debris clearance, synaptic remodeling, and tissue repair. However, with persistent amyloid and tau pathology, this response can transition toward a chronic maladaptive state characterized by sustained microglial and astrocytic activation, cytokine release, complement signaling, impaired phagocytosis, and progressive synaptic and neuronal injury [84–86].

DSAD appears to share this broad temporal organization, but with important differences in baseline immune tone and disease trajectory. Trisomy 21 includes several immune-related genes, including interferon receptor genes, and consistently activates interferon signaling, creating a chronically primed inflammatory state that may precede clinical symptoms and classical AD pathology [80,87]. This suggests that the early inflammatory phase in DS may not be equivalent to that observed in sporadic AD. Rather than emerging primarily in response to aging, amyloid, or Tau accumulation, immune activation in DS is embedded within lifelong gene-dosage effects and may lower the threshold for later maladaptive neuroinflammation [80,84,88].

Neuropathological studies support a stage-dependent inflammatory profile in DSAD. Comparative analyses of DS, DSAD, and sporadic AD show that younger individuals with DS tend to exhibit mixed M1/M2b immune phenotypes, whereas older individuals with DSAD display a predominant M2b response, a phenotype associated with immune complexes and rarely observed in sporadic AD [87]. This pattern indicates both overlap and divergence from the inflammatory trajectory of sporadic AD. As in other forms of AD, glial activation in DSAD likely participates in Aβ and Tau-associated tissue responses, but the DSAD inflammatory environment is shaped by trisomy 21-associated immune response predisposition, altered interferon signaling, and disease-stage-specific microglial phenotypes [80,84,87,88].

Microglial alterations further illustrate this dual role. Across the DS lifespan, microglia show evolving activation states that become more pronounced with age and AD pathology [88]. In DSAD, microglia display amoeboid and rod-like morphologies associated with Tau pathology, axonal injury, and increased expression of activation and phagocytosis markers such as CD64 and CD86 [87–91]. These findings suggest that early immune responses may initially contribute to surveillance and clearance, but sustained activation may later promote synaptic dysfunction, white matter injury, vascular alterations, and neurodegeneration. Thus, neuroinflammation in DSAD should be understood as a temporally evolving process in which an initially adaptive or compensatory response occurs within a trisomy 21-primed immune environment and may transition earlier or more intensely into chronic maladaptive inflammation than in sporadic AD.

2.1.3. Endo-lysosomal pathways in DSAD

Disruption of endosomal and lysosomal pathways is an early and persistent feature of DSAD. Brains from individuals with DS show morphological abnormalities of late endosomes, including clustering not observed in normal aging [92,93]. Key regulators of endosomal maturation, including Rab5 and Rab7, are upregulated, indicating impaired coordination of endosomal trafficking [94].

A central contributor to this dysfunction is the APP β-CTF fragment generated during amyloidogenic processing in endosomes, which drives Rab5 hyperactivation, slows endosomal transport, and impairs neuronal support [55,57,95]. Lysosomal de-acidification further compromises proteolysis and intracellular trafficking, and similar disruptions are observed in DS fibroblasts and mouse models with modest APP overexpression [57,96–98]. Consistent with these observations, multiple AD risk genes identified by GWAS (PICALM, BIN1, SORL1, CD2AP) directly regulate endo-lysosomal pathways [56].

2.1.4. Mitochondrial dysfunction, oxidative stress, and energetic failure in DSAD

Mitochondrial dysfunction and oxidative stress represent central mechanisms linking trisomy 21 to AD vulnerability. Several Hsa21 genes influence redox homeostasis directly or indirectly. Triplication of SOD1 can increase hydrogen peroxide production when not matched by proportional increases in downstream antioxidant enzymes, while APP overexpression, RCAN1, DYRK1A, and BACH1/Nrf2 pathway imbalance may further disrupt mitochondrial function, calcium signaling, stress responses, and antioxidant defense [62,99–101]. These gene-dosage effects occur in a brain that already has high energetic demands, making neurons, synapses, and oligodendrocytes particularly vulnerable to cumulative oxidative and metabolic stress.

Evidence from cellular, biochemical, and proteomic studies supports early mitochondrial impairment in DS. Mitochondrial abnormalities in DS include altered morphology and dynamics, reduced oxidative phosphorylation, decreased ATP production, increased reactive oxygen species, impaired mitophagy, and reduced capacity to maintain redox balance [99,101,102]. These alterations are not restricted to symptomatic dementia stages. Redox proteomic analyses of young DS frontal cortex, before extensive AD neuropathology, showed oxidative modification of proteins involved in protein quality control and degradation, including UCH-L1, GRP78, cathepsin D, V-ATPase subunits, and GFAP, together with reduced proteasome activity and impaired autophagic flux [103]. This indicates that oxidative stress and impaired proteostasis are early features of the DS brain and may predispose to later Aβ and Tau accumulation.

Recent brain proteomic studies further support a model in which mitochondrial and energetic deficits intensify with aging and AD progression. In young DS brains, differentially expressed proteins are enriched in mitochondrial metabolism, synaptic function, stress responses, and proteostasis, whereas DSAD brains show broader reductions in proteins related to mitochondrial oxidative phosphorylation, synaptic signaling, and NRF2-mediated antioxidant responses [104]. These changes connect mitochondrial dysfunction to two major downstream consequences: impaired protein clearance and synaptic failure. Oxidative modification of proteostasis proteins can reduce the efficiency of degradation pathways, increasing vulnerability to Aβ and tau aggregation, while energetic failure at synapses can impair neurotransmission, plasticity, and network integrity.

Mitochondrial dysfunction may also interact with amyloid burden and cognition in DSAD. Recent clinical work has linked mitochondrial dysfunction to brain Aβ deposition and poorer memory and executive function in adults with DS, supporting the translational relevance of bioenergetic failure along the AD continuum [105]. In parallel, early postnatal single-nucleus multi-omic studies in DS prefrontal cortex show metabolic and synaptic gene dysregulation across neuronal populations, together with glial and inflammatory changes, suggesting that bioenergetic vulnerability is embedded within trisomy 21 biology long before dementia [106]. Overall, mitochondrial dysfunction in DSAD should be viewed not as an isolated downstream consequence of amyloid pathology, but as an early and progressive systems-level mechanism that links gene dosage, oxidative damage, proteostasis impairment, synaptic dysfunction, oligodendrocyte vulnerability, and neurodegeneration.

2.1.5. Oligodendrocytes dysfunction and white matter alterations in DSAD

Oligodendrocyte dysfunction and white matter alterations in DSAD are best understood as the convergence of early developmental demyelination and later neurodegenerative injury. Developmental transcriptomic studies show that trisomy 21 disrupts genes involved in oligodendrocyte differentiation and myelination across multiple brain regions, with parallel evidence of hypomyelination, reduced myelinated fiber density, altered axonal architecture, and slower neocortical white matter conduction in DS models [59]. This developmental vulnerability is supported by recent single-nucleus multi-omic data from early postnatal DS prefrontal cortex, which identified impaired oligodendrocyte lineage progression, depletion of the oligodendrocyte progenitor cell pool, and reduced myelin-related transcription in mature oligodendrocytes, together with metabolic and synaptic transcriptional deficits [106]. Thus, white matter abnormalities in DS cannot be interpreted solely as secondary consequences of AD pathology, since altered oligodendrocyte maturation and myelination are present from early life.

This early developmental vulnerability is likely further shaped by aging and AD-related changes. Neuroimaging studies show that adults with DS exhibit alterations in major white matter tracts, including the corpus callosum, long-range association pathways, limbic tracts, projection fibers, and cerebellar white matter [60,107,108]. White matter hyper-intensities (WMH) are also increased in DS and appear closely linked to AD-relevant biology. Cross-sectional work showed that WMH burden increases approximately 10 years before expected AD symptom onset, is associated with NfL, and relates to gray matter volume in parieto-temporal regions [109]. Longitudinal data further indicate that posterior WMH change accelerates across the AD continuum in adults with DS, particularly around progression from MCI-DS to AD dementia [110]. These findings support a model in which developmental demyelination and atypical structural connectivity create a vulnerable white matter substrate that is later affected by amyloid, Tau, vascular injury, neuroinflammation, and axonal degeneration.

Fluid proteomics reinforces this interpretation. CSF proteomic studies in DS show early alterations in markers of white matter and axonal pathology, including decreased myelin-associated proteins such as MAG and MOG long before estimated symptom onset, with changes occurring before NfL, total and phosphorylated Tau alterations [31]. Plasma NfL is also associated with microstructural white matter changes in adults with DS, suggesting that axonal injury biomarkers may partly reflect selective white matter vulnerability before overt dementia [111]. Together, these findings suggest that myelin dysfunction may precede and contribute to later axonal degeneration rather than merely following neuronal loss.

Interactions among oligodendrocytes, microglia, and metabolic stress may further amplify this vulnerability. Oligodendrocytes require high energetic and lipid-metabolic support to maintain myelin and provide trophic support to axons. Poorly myelinated or demyelinated axons lose this support, increasing energetic demands and susceptibility to degeneration [112]. In DS, early microglial activation, astrocyte reactivity, cytokine dysregulation, and vascular-glial-immune signaling may impair oligodendrocyte lineage maturation, myelin maintenance, and repair capacity [106]. Overall, white matter abnormalities in DSAD likely arise from both developmental demyelination and later neurodegenerative injury, highlighting oligodendrocyte–microglial–metabolic interactions as an important part of the broader systems-level vulnerability landscape in DSAD.

2.1.6. Neurovascular unit dysfunction, cerebral amyloid angiopathy, and blood–brain barrier alterations in DSAD

Neurovascular dysfunction is an important component of DSAD pathogenesis and provides a framework for integrating cerebral amyloid angiopathy (CAA), blood–brain barrier (BBB) disruption, extracellular matrix remodeling, impaired clearance, and white matter injury. Although adults with DS have a relatively low prevalence of conventional vascular risk factors such as hypertension and atherosclerosis, cerebrovascular abnormalities are common with aging and AD progression. This suggests that vascular injury in DSAD is not primarily driven by classical vascular comorbidity, but rather by interactions among APP triplication, lifelong Aβ overproduction, vascular amyloid deposition, inflammatory signaling, and neurovascular unit vulnerability [18,48,113,114].

CAA is the most prominent vascular manifestation of AD pathology in DS. Postmortem studies show that CAA is more frequent in DS than in sporadic AD and controls, whereas atherosclerosis and arteriolosclerosis are comparatively uncommon, supporting a vascular phenotype dominated by amyloid deposition rather than conventional small-vessel disease [48]. Comparative studies also indicate that CAA is more frequent in genetically determined forms of AD, including DS and ADAD, than in sporadic early-onset AD, although DS appears to differ from APP duplication in CAA severity, capillary involvement, and hemorrhagic risk [47,113,115]. This distinction is important because DS and APP duplication both involve three copies of APP, but individuals with DS appear to have fewer intracerebral hemorrhages and less severe CAA than expected from APP overexpression alone, suggesting that additional chromosome 21 genes, vascular regulatory mechanisms, or BBB-related protective factors may modify the vascular phenotype [18,113]. A recent systematic neuropathological analysis further showed that CAA in adult DS is frequent, severe, and extensive, with leptomeningeal-predominant neocortical involvement, capillary involvement, marked cerebellar CAA, and iron pathology suggestive of chronic hypoperfusion, despite a relative paucity of gross hemorrhagic lesions [116].

The neurovascular unit links vascular amyloid to BBB integrity and clearance biology. Endothelial cells, pericytes, vascular smooth muscle cells, astrocytic endfeet, microglia, basement membrane components, and extracellular matrix proteins collectively regulate BBB permeability, cerebral blood flow, immune trafficking, and interstitial solute clearance. Aβ deposited in cerebral vessels may derive from local production by neurovascular unit cells, drainage from the parenchyma through perivascular pathways, or peripheral sources, and BBB transport mechanisms such as LRP1-mediated Aβ clearance may influence whether Aβ accumulates in parenchymal plaques or vascular walls [113]. In DSAD, vascular Aβ deposition may impair vessel wall structure, reduce vascular reactivity, disrupt perivascular drainage, and increase vulnerability to microbleeds, white matter injury, and downstream neurodegeneration. MRI studies in adults with DS show that enlarged perivascular spaces and infarcts may emerge in the early 30s, whereas microbleeds, white matter hyperintensities, amyloid, and tau emerge in the mid-to-late 30s, before typical dementia onset [117]. Enlarged perivascular spaces may reflect impaired perivascular or glymphatic clearance, while white matter hyperintensities may capture downstream effects of small-vessel injury, BBB disruption, demyelination, chronic hypoperfusion, and inflammation [18,117].

Direct evidence for BBB permeability changes in DSAD remains limited, but current data support incorporating BBB and neurovascular unit dysfunction into a unified model of disease progression. Valay and Potier emphasize that few longitudinal studies have directly examined BBB integrity in DS and that the apparent dissociation among APP dosage, CAA severity, and hemorrhagic burden may reflect both vulnerability and resilience within the neurovascular unit [113]. This interpretation is consistent with fluid and lesion-level proteomic studies showing early immune, extracellular matrix, plasma-derived, glial, and endo-lysosomal protein signatures in DSAD, suggesting that amyloid accumulates within a broader vascular and stromal microenvironment rather than as an isolated parenchymal event [31,46,52]. Together, these findings indicate that neurovascular unit dysfunction should be viewed as a core component of DSAD biology. Integrating CAA-sensitive MRI markers, BBB-related fluid biomarkers, extracellular matrix remodeling, and vascular proteomic signatures into DSAD staging models may improve interpretation of disease progression and inform therapeutic trial design, particularly because high CAA burden and BBB alterations may influence the safety and biological response to anti-Aβ immunotherapy.

Although Down syndrome, autosomal dominant Alzheimer’s disease, and sporadic late-onset Alzheimer’s disease share the core neuropathological features of amyloid plaques, Tau pathology, synaptic dysfunction, and neurodegeneration, their upstream biological contexts differ substantially (Figure 2). Sporadic AD reflects the combined effects of aging, polygenic risk factors, environmental exposures, and vascular and metabolic comorbidities, whereas ADAD arises from pathogenic variants in APP, PSEN1, or PSEN2 that directly alter amyloid processing. DSAD represents a distinct genetically defined context in which trisomy 21 increases the dosage of APP and other chromosome 21 genes from conception, resulting in lifelong amyloid overproduction and broad disruption of neurodevelopmental, immune, endo-lysosomal, mitochondrial, vascular, and white matter pathways. Together, these features support a temporally organized, systems-level model of DSAD pathogenesis. In an early developmental phase, trisomy 21 alters brain maturation, oligodendrocyte differentiation, myelination, interferon signaling, mitochondrial homeostasis, endosomal trafficking, and extracellular matrix organization, thereby establishing a vulnerable biological substrate before classical AD pathology becomes prominent [32,59,80,103].

Figure 2.

Alzheimer’s disease forms and their genetic drivers. Sporadic late-onset Alzheimer’s disease arises from aging-related processes and multiple genetic risk factors, including APOE and TREM2. Autosomal dominant Alzheimer’s disease results from pathogenic variants in APP, PSEN1, or PSEN2 that directly alter amyloid processing and drive early amyloid accumulation. Down syndrome represents a genetically defined context in which trisomy 21 increases the dosage of APP and other chromosome 21 genes, leading to lifelong amyloid overproduction and near-universal Alzheimer-type pathology. Despite differing etiologies, all forms converge on shared downstream processes, including neuroinflammation, synaptic dysfunction, and disrupted protein homeostasis. Figure created in Biorender.com.

In a subsequent preclinical network dysregulation phase, immune, vascular, synaptic, myelin, and proteostasis alterations become detectable across brain and fluid proteomic studies, including changes that precede overt Aβ and Tau biomarker abnormalities [31,104]. Amyloid and Tau accumulation then emerge within this context, with amyloid plaques recruiting conserved immune, endo-lysosomal, glial, and extracellular matrix protein networks that overlap with those observed in early- and late-onset AD [30,46,52]. In vivo PET comparisons with ADAD further support this interpretation, showing broad similarity in the amyloid-to-Tau sequence but DS-specific differences in the spatial distribution, timing, and magnitude of Tau burden, which may influence clinical progression and the optimal timing of therapeutic intervention [118].

Finally, symptomatic disease can be viewed as a phase of systems-level neurodegeneration in which amyloid and tau pathology converge with synaptic failure, neuroinflammation, white matter injury, vascular dysfunction, impaired proteostasis, and energetic failure. This model does not diminish the central role of APP dosage in DSAD, but instead places amyloid within a broader and temporally dynamic vulnerability network that may help explain why similar amyloid and tau pathologies can lead to heterogeneous ages of onset, rates of decline, and therapeutic responses across individuals with DS.

2.2. Multiomics insights into Alzheimer’s disease biology in Down syndrome

The application of high-throughput omics technologies has substantially expanded the understanding of AD biology [119–122], including in the context of DS [32,123]. By integrating genomic, transcriptomic, proteomic, and metabolomic datasets, recent studies have begun to characterize the molecular networks that connect trisomy 21 to neurodegeneration, revealing systems-level disease mechanisms that extend beyond APP overexpression and affect diverse biological pathways [32,124]. Consistent with this perspective, large-scale proteomic analyses of human brain tissue have demonstrated that AD itself is molecularly heterogeneous, comprising reproducible subtypes defined by coordinated alterations in synaptic signaling, cytoskeletal organization, and blood–brain barrier–related processes [125]. Multilayer systems-level approaches further show that classical amyloid and tau pathologies are embedded within broader protein networks involving RNA splicing, axonal transport, and neuroinflammatory pathways, while prioritizing proteins such as MDK, NTN1, SFRP1, and GPNMB that are highly connected, aggregation-associated, or detectable in cerebrospinal fluid [125]. Together, these findings underscore the value of multiomics and network-based analyses for defining coordinated molecular pathways and prioritizing translationally relevant protein candidates in AD, providing a conceptual and analytical framework that is particularly relevant for DS, where genetically driven amyloid pathology enables investigation of early disease mechanisms.

While multiomics approaches have substantially advanced understanding of AD and DSAD, most current proteomic analyses quantify total protein abundance and therefore integrate multiple molecular species derived from a single gene. Proteoforms arise through alternative transcript usage, regulated translation, and post-translational modification, and can differ in stability, interaction profiles, and functional properties. As a result, bulk protein measurements may reflect shifts in proteoform composition rather than uniform changes in protein expression.

Protein diversity in the nervous system is further shaped by translational regulation, including non-canonical initiation and alternative reading frame usage, which remain poorly represented in standard proteomic workflows and complicate integration with transcriptomic data [126]. These limitations are particularly relevant in DS, where trisomy 21 induces widespread secondary molecular remodeling beyond direct gene dosage effects. Multiomics studies indicate that disease onset and progression are influenced by broader pathways involving proteostasis, immune signaling, and lipid metabolism rather than by uniform increases in chromosome 21 gene products alone [32].

In this context, a proteoform-aware perspective provides a useful framework for interpreting existing proteomics datasets without altering their core conclusions. Proteoform medicine emphasizes functional protein diversity as a key consideration in disease biology and offers guidance for future methodological refinement, while remaining compatible with current bulk proteomics approaches and network-based analyses [127]. With this framework in mind, transcriptomic and cell-resolved studies provide critical insight into how gene dosage and regulatory variation shape molecular states in the DS brain, while also highlighting the need for complementary protein-level interpretation.

2.2.1. Analytical frameworks for multiomics integration in DSAD

The value of multiomics studies in DSAD depends not only on the number of molecular layers measured, but also on how these layers are analytically integrated. Current approaches can be broadly organized into early, intermediate, late, and network-based integration strategies. Early integration combines features from multiple omics layers before modeling, which can improve discovery power but is vulnerable to dimensionality, batch effects, missing data, and overfitting. Intermediate integration methods, including latent-factor models, multi-view learning, and sparse partial least-squares frameworks, aim to identify shared molecular axes across transcriptomic, proteomic, metabolomic, epigenomic, and biomarker datasets while preserving layer-specific structure. Late integration combines results after separate analysis of each omics layer, such as pathway-level enrichment, molecular modules, or classifier outputs, and is useful when matched multiomics data are incomplete. Network-based approaches, including co-expression modules, protein-protein interaction networks, causal networks, knowledge graphs, and eQTL/pQTL-informed models, can further prioritize disease-relevant pathways and candidate proteins by linking genetic variation to downstream RNA, protein, and cellular phenotypes [124,128].

These analytical issues are particularly important in DSAD. Because trisomy 21 produces lifelong molecular remodeling, integration models must distinguish direct gene-dosage effects from secondary transcriptomic, proteomic, and cellular responses, as well as from AD-stage effects. Additional challenges include small cohort sizes, limited availability of matched tissue and fluid multiomics from the same individuals, differences in brain region and cell-type composition, postmortem interval, platform-specific batch effects, baseline intellectual disability, and the need to align individuals by age or estimated years to symptom onset rather than chronological age alone. Spatial and single-cell multiomics can help address some of these limitations by linking cell-type-specific transcriptional states to lesion-associated proteomic microenvironments, vascular and glial responses, and regional vulnerability. However, these approaches require careful validation across independent cohorts and orthogonal platforms before being used for patient stratification or therapeutic target selection [129,130]. Thus, future DSAD multiomics studies should prioritize harmonized longitudinal cohorts, transparent preprocessing pipelines, matched molecular layers when possible, and biologically interpretable models that connect trisomy 21 biology to AD progression.

2.2.2. Transcriptomic and single-cell studies of the Down syndrome brain

Transcriptomic studies in DS brain have provided early and enduring insight into the molecular basis by which trisomy 21 predisposes individuals to AD. Early bulk transcriptomic analyses of adult DS cortex demonstrated widespread gene expression dysregulation extending beyond chromosome 21, implicating synaptic, mitochondrial, and immune-related pathways in DS neuropathology [131]. Subsequent studies in developing DS brain tissue showed that trisomy 21 induces early and persistent alterations in neurodevelopmental gene programs, suggesting that deviations in neuronal maturation and circuit formation may establish a molecular landscape permissive to later AD pathogenesis [59,132]. More recent RNA-seq analyses have refined these findings by revealing cell-type–specific transcriptional changes in DS aging brains, including altered neuronal subtype composition and increased RNA isoform diversity, linking trisomy 21 to transcriptomic instability relevant to neurodegeneration [133].

Advances in single-cell and single-nucleus RNA sequencing have further expanded understanding of DS-associated AD by resolving transcriptional heterogeneity across brain cell populations. Studies of sporadic AD have identified disease-associated transcriptional states in microglia, astrocytes, oligodendrocytes, and vascular cells, highlighting coordinated glial responses and altered intercellular interactions as central features of AD biology [134]. Applying these approaches to DS-AD, spatial transcriptomic and single-nucleus analyses demonstrated both shared and distinct transcriptional signatures between DS-AD and sporadic AD, including region- and layer-specific gene expression changes linked to amyloid deposition, synaptic dysfunction, and neuroinflammatory signaling [30]. Notably, this work revealed that Hsa21 gene overexpression, including APP, varies substantially by cell type and brain region, and that DSAD is characterized by prominent glial and vascular transcriptional programs that overlap with known AD risk pathways.

While transcriptomic and genomic approaches have therefore been instrumental in defining genetic risk and cell-specific gene expression changes in DS-associated AD, they do not fully capture all molecular dimensions of disease biology. Particularly, transcript-level analyses are limited in their ability to reflect post-translational protein modifications, functional consequences of risk-associated genes, or regulatory variation residing in non-coding genomic regions. Consistent with this, proteomic studies have identified disease-relevant protein-level alterations that are not predictable from RNA abundance alone, highlighting the importance of complementary molecular layers when interpreting transcriptomic findings [135,136].

Integrating these perspectives, recent multi-modal studies combining proteomics with single-nucleus RNA sequencing have shown that trisomy 21 alters transcript abundance across astrocytes, endothelial cells, and pericytes, with elevated expression of non-chromosome 21 genes such as APOE. These findings emphasize that transcriptomic remodeling in DS-associated AD extends well beyond gene-dosage effects alone and reflects coordinated, system-level responses across cell types [137]. Collectively, transcriptomic and single-cell approaches provide a critical framework for understanding how early developmental perturbations, cell-type–specific responses, and network-level interactions converge to shape AD biology in the context of DS.

2.2.3. Brain proteomics in Down syndrome

Proteomic analyses of DS brain tissue have revealed early molecular alterations that precede and progressively converge with AD pathology. Across studies, DS is characterized by early disruption of proteostasis, redox homeostasis, and synaptic and metabolic pathways, followed by the appearance of proteomic signatures that closely resemble those observed in sporadic AD in both bulk tissue and neuropathological lesions.

Early evidence for this trajectory came from redox proteomic analyses of frontal cortex from young individuals with DS, prior to extensive amyloid plaque or neurofibrillary tangle deposition [103]. This study demonstrated selective oxidative modification of proteins involved in protein quality control and degradation, including UCH-L1, GRP78, cathepsin D, V-ATPase subunits, and GFAP, accompanied by reduced proteasome activity and impaired autophagic flux despite relatively modest changes in total protein abundance. Importantly, these alterations occurred alongside subtle increases in soluble Aβ_1–42 and phosphorylated tau, supporting the conclusion that impaired protein clearance mechanisms, driven by oxidative stress, constitute an early pathogenic event that predisposes the DS brain to subsequent AD-related protein aggregation and neurodegeneration [103].

A subsequent large-scale shotgun proteomic study of frontal cortex extended these findings across aging and disease stages by comparing young DS, DS-AD, and age-matched control brains [104]. In young DS individuals, most differentially expressed proteins were increased and spanned pathways related to mitochondrial metabolism, synaptic function, stress responses, and proteostasis, involving both trisomic and disomic genes. Only a small fraction of altered proteins mapped to Hsa21, indicating that trisomy induces widespread secondary remodeling of the proteome rather than simple gene dosage effects at the level of total protein abundance. With aging and AD progression, DS-AD brains showed a broad decline in protein abundance, particularly affecting mitochondrial oxidative phosphorylation, synaptic proteins, and NRF2-mediated antioxidant responses. Proteins such as GFAP, UCH-L1, TXN, VDACs, DNM1L, GAP43, and SYP were consistently altered across disease stages, indicating that AD-related molecular changes emerge early in DS and intensify over time. Notably, comparison with sporadic AD datasets revealed substantial overlap, reinforcing the interpretation of DS as a genetically driven form of AD with accelerated onset rather than a mechanistically distinct condition [104].

While bulk tissue proteomics highlighted global pathway perturbations, localized proteomic approaches provided lesion-specific insight into DS-associated AD pathology. Using laser-capture microdissection and label-free mass spectrometry, Drummond et al. [46] characterized the amyloid plaque proteome in DS-AD and sporadic early-onset AD (EOAD). Despite distinct etiologies, plaques in both conditions shared a core set of 48 highly enriched proteins, many of which were more strongly enriched than Aβ itself. Prominent plaque-associated proteins included COL25A1, SMOC1, MDK, NTN1, OLFML3, HTRA1, APOE, CLU, complement components, and lysosomal proteins, highlighting endo-lysosomal pathways, extracellular matrix remodeling, and immune activation as central features of plaque composition. Quantitative differences were observed between subtypes, with DS plaques enriched in phosphorylated and pyroglutamate-modified Aβ, whereas EOAD plaques contained higher levels of soluble oligomers, supporting the conclusion that plaque formation in DS-AD and EOAD arises through largely shared molecular mechanisms with subtype-specific modulation of Aβ states [46].

This lesion-level convergence was further reinforced by a larger comparative study encompassing DS, EOAD, and late-onset AD (LOAD) [52]. Across all three groups, 43 plaque-associated proteins were shared and showed strong correlation in abundance, with dominant functional themes related to APP/Aβ metabolism, lysosomal organization, immune response, and glial activation. Although the core plaque proteome was highly conserved, DS plaques displayed stronger enrichment of lysosomal and vesicular proteins, whereas EOAD plaques exhibited relatively greater immune-related signatures. In contrast, non-plaque tissue showed greater divergence, with DS cortex enriched for extracellular matrix and structural proteins, while EOAD and LOAD converged on chromatin remodeling and transcriptional regulatory processes. Importantly, mapping of plaque-associated proteins confirmed that chromosome 21 gene products were not preferentially enriched within plaques, indicating that trisomy influences amyloid pathology through network-level effects beyond APP, rather than direct accumulation of HSA21-encoded proteins [52] (Figure 3A).

Figure 3.

Convergent and distinct proteomic alterations across DS, ADAD and sporadic AD. Multi omics approaches have revealed both convergent and divergent pathways across Down syndrome, sporadic and autosomal dominant Alzheimer’s disease. A. Localized proteomics in brain tissue have elucidated a preferential alteration of proteins related to the extracellular matrix and mitochondrial function in DS, whereas sporadic and ADAD showed altered pathways related to gene expression, RNA processes and chromatin remodeling (Left). The right side of the panel shows convergent pathways and functions across DS and AD subtypes, including APP and tau processing, neuroinflammation, endo-lysosomal dysfunction, myelin and oligodendrocyte alterations, and synaptic dysfunction. B. Fluid proteomics highlights differences in the temporal sequence of protein alterations across DS and ADAD aligned to estimated years to symptom onset (EYO). In DS, early changes include neurofilament light, synaptic proteins, extracellular matrix components, and markers of cerebral amyloid angiopathy, preceding amyloid and tau alterations. In contrast, ADAD shows earlier changes in Tau-related proteins, including phosphorylated Tau and proteins such as SMOC1 and SPON1. Despite these differences, overlapping molecular signatures are observed between CSF and brain, underscoring convergent neurodegenerative processes with distinct temporal dynamics in DS. Figure created in Biorender.com.

Notably, the proteomic alterations observed in DS brain tissue often extend beyond direct chromosome 21 gene-dosage effects, pointing to regulatory and post-transcriptional mechanisms that reshape protein networks upstream of overt neuropathology. The strong convergence of plaque-associated proteomes across DS, EOAD, and LOAD, contrasted with greater divergence in non-plaque tissue, further suggests that system-level regulatory states and cellular context modulate molecular outcomes more than genotype alone. Together, these findings position proteomics as a critical integrative layer linking genetic risk to downstream molecular and pathological phenotypes, a theme that becomes particularly evident in fluid proteomic studies.

2.2.4. Fluid biomarkers in Down syndrome with Alzheimer’s disease

The near-universal development of AD neuropathology in adults with DS provides a unique opportunity for evaluating fluid biomarkers across the AD continuum. Because cognitive variability and baseline intellectual disability complicates clinical staging, biological markers in cerebrospinal fluid (CSF) and blood are particularly critical for disease detection, staging, and therapeutic trial design. Over the past decade, work in DS has progressed from validation of canonical Aβ and tau markers toward high-dimensional proteomic approaches that capture immune, vascular, and neurodevelopmental processes unique to trisomy 21, while integrating signal across heterogeneous protein states.

2.2.4.1. Cerebrospinal fluid biomarkers.

CSF biomarkers in adults with DS recapitulate the canonical Alzheimer’s disease AT(N) profile, with reductions in Aβ_1–42 and elevations in tau species and neurofilament light (NfL) that precede dementia onset by many years. In a seminal cross-sectional study, Fortea and colleagues demonstrated that individuals with DS showed marked decreases in CSF Aβ_1–42 and significant increases in total tau (t-tau), phosphorylated tau at threonine 181 (p-tau181), and NfL as they transitioned from asymptomatic to prodromal and dementia stages. These markers showed excellent diagnostic performance for symptomatic AD in DS, with CSF Aβ_1–42 reliably discriminating disease stage at the individual level [138].

Importantly, CSF Aβ_1–42 in DS behaves similarly to sporadic and autosomal-dominant AD, contrasting with plasma Aβ measures that remain elevated due to lifelong APP overexpression and show limited staging utility. This positions CSF Aβ_1–42 as a robust biological marker for early disease classification and trial stratification in DS [138].

Subsequent longitudinal and cross-sectional work has shown that CSF NfL rises early along the disease trajectory in DS, often before marked tau PET abnormalities, reflecting early axonal and white-matter vulnerability. This early prominence of neurodegeneration markers distinguishes DS-AD from sporadic AD and aligns with known developmental and myelination abnormalities associated with trisomy 21 [31].

Recent large-scale proteomic analyses of CSF have extended understanding beyond core AT(N) biomarkers. Montoliu-Gaya and collaborators performed untargeted mass-spectrometry-based proteomics in CSF from a large DS cohort spanning asymptomatic, prodromal, and dementia stages, and compared these trajectories with late-onset AD (LOAD) and autosomal-dominant AD (ADAD) [31]. While many protein changes overlapped across AD etiologies, DS was characterized by disproportionately severe alterations in immune-related proteins, extracellular matrix components, and plasma-derived proteins suggestive of early blood–brain barrier dysfunction [31].

Notably, many of these immune and vascular proteomic changes were already present in young, cognitively asymptomatic adults with DS, prior to detectable abnormalities in CSF Aβ or tau. Network analyses revealed early perturbations in axonal, myelin, and synaptic protein modules, supporting the view that DS confers a biologically primed substrate for neurodegeneration that precedes classical AD pathology rather than merely resulting from amyloid accumulation [31] (Figure 3B).

2.2.4.2. Plasma biomarkers.

Plasma biomarkers offer a minimally invasive and scalable alternative to CSF, enabling repeated measurements and broader trial inclusion. Early studies established that plasma Aβ concentrations are chronically elevated in DS due to APP triplication but display poor diagnostic specificity for AD staging at the individual level [138]. More recent work has therefore focused on downstream markers of tau pathology and neurodegeneration.

In a large non-demented DS cohort, Schworer et al. showed that lower plasma Aβ_42/40 ratios and higher plasma total tau and NfL levels were associated with poorer episodic memory and visuospatial performance, even prior to clinical dementia. Although the associations were modest, they indicate that plasma biomarkers can track early cognitive vulnerability along the preclinical and prodromal AD continuum in DS [139].

Across studies, plasma NfL has emerged as the most sensitive plasma marker for disease progression in DS. Plasma NfL rises steeply with age, differentiates asymptomatic from symptomatic individuals, and correlates strongly with CSF NfL and neuroimaging measures of neurodegeneration. However, its lack of disease specificity underscores its role as a general marker of axonal injury rather than a direct indicator of amyloid or tau pathology [138,139].

Ultra-sensitive assays have enabled reliable measurement of phosphorylated tau species in plasma, particularly p-tau217. Wisch and colleagues examined longitudinal amyloid PET, tau PET, and plasma p-tau217 in more than 300 individuals with DS. Plasma p-tau217 increased after the emergence of widespread amyloid pathology and correlated with both amyloid and tau PET burden, confirming its biological relevance [140].

However, longitudinal modeling revealed a critical DS-specific insight: chronological age alone performed as well as plasma p-tau217 in detecting current pathology and predicting near-future amyloid and tau accumulation. This finding reflects the exceptionally tight coupling between age and AD pathology in DS and contrasts with sporadic AD, where plasma p-tau217 often adds substantial predictive value beyond age [140].

These results suggest that while plasma p-tau217 is a valid marker of disease stage in DS, its incremental diagnostic utility for early screening may be limited. Its greatest value may lie in confirming biological disease stage and serving as a pharmacodynamic marker in therapeutic trials.

High-dimensional plasma proteomics has revealed additional biological pathways involved in DS-AD. Using the Olink Explore platform, Wagemann and collaborators (2025) identified widespread dysregulation of immune, inflammatory, extracellular matrix, and synaptic proteins in plasma from adults with DS. Comparison of symptomatic and asymptomatic individuals revealed that none of the discriminating proteins were encoded on Hsa21, indicating that downstream pathological processes rather than gene dosage per se drive disease progression [141].

Machine-learning feature selection highlighted a panel of plasma proteins, including NfL, GFAP, IGFBP2, SPON1, and synaptic markers such as CBLN4 and SEPTIN3. This approach discriminated symptomatic from asymptomatic DS with good accuracy. The plasma proteomics findings complement CSF proteomic data and reinforce the importance of immune activation, vascular remodeling, and synaptic dysfunction in DS-AD beyond the traditional Aβ-tau framework [141].

2.2.4.3. Implications for staging and clinical trials.

Taken together, fluid biomarker studies demonstrate that AD-related biological changes in DS emerge early, follow a compressed and highly age-dependent trajectory, and involve neurodegenerative and immune pathways not fully captured by the classical AT(N) framework. CSF biomarkers remain the most specific indicators of amyloid and tau pathology, while plasma NfL provides a sensitive and scalable measure of neurodegeneration. Plasma p-tau and proteomic markers add mechanistic depth and may be particularly valuable for monitoring therapeutic engagement.

These findings underscore the central role of fluid biomarkers in the design of preventive and disease-modifying trials in DS, enabling biological staging, participant enrichment, and objective assessment of intervention efficacy in this genetically defined model of Alzheimer’s disease.

2.3. Clinical trials for DSAD and future perspectives

Individuals with DS have historically been excluded from the passive anti-Aβ immunotherapy trials that led to U.S. Food and Drug Administration (FDA) approval of disease-modifying treatments for AD. Given the need for DS-specific cognitive outcome measures and the higher prevalence of cerebral amyloid angiopathy (CAA) compared with sporadic AD, current recommendations advise against the use of approved anti-Aβ therapies in DSAD until DS-specific safety and efficacy data become available [142–145]. This caution reflects several biological and clinical challenges. Symptomatic DSAD may already represent an advanced disease stage after decades of APP overexpression, amyloid deposition, CAA, tau pathology, neuroinflammation, white matter injury, and synaptic dysfunction. In addition, the high burden of CAA raises concern for amyloid-related imaging abnormalities and hemorrhagic complications, particularly because lecanemab binds not only parenchymal plaques but also vascular amyloid in postmortem DS brain tissue [146]. Finally, amyloid lowering alone may not address broader DSAD mechanisms, including immune predisposition, mitochondrial dysfunction, endo-lysosomal impairment, BBB dysfunction, oligodendrocyte vulnerability, and synaptic failure. These limitations do not argue against amyloid-targeting trials, but they support earlier intervention, biomarker-confirmed staging, CAA-sensitive MRI surveillance, vascular risk assessment, DS-specific outcomes, and combination or sequential therapeutic strategies [142–144,147].

Ongoing trials within the Alzheimer’s Clinical Trials Consortium–Down Syndrome (ACTC–DS) are beginning to address this therapeutic gap. These include ABATE, a phase 1b/2 study of the active anti-Aβ immunotherapy ACI-24.060 in prodromal sporadic AD and DSAD; Hero, a phase 1b trial of the antisense oligonucleotide ION269 directed against the APP transcript; and ALADDIN, a phase 4 trial assessing the safety and tolerability of donanemab [145]. Future ACTC–DS studies linked to the Trial Ready Cohort–Down Syndrome are expected to investigate APP silencing using siRNA-based approaches as well as non-pharmacological interventions [145]. In parallel, immunomodulatory strategies are emerging as a complementary therapeutic direction. A phase II study of the JAK1/3 inhibitor tofacitinib in individuals with DS demonstrated an acceptable safety profile and preliminary efficacy, including improvement in autoimmune skin disease and reductions in interferon signaling, inflammatory cytokines, and autoantibody levels [148], supporting the potential relevance of targeting chronic immune activation in DSAD.

Non-pharmacological and neuromodulatory approaches should also be considered in future DSAD therapeutic development. Noninvasive brain stimulation, including repetitive transcranial magnetic stimulation and transcranial direct current stimulation, has shown potential biological and clinical effects in AD, although results remain heterogeneous across protocols, stimulation targets, disease stages, and outcome measures [149,150]. Recent work suggests that repetitive transcranial magnetic stimulation may modulate neuroplasticity, neurotransmission, cortical excitability, gamma oscillations, and neuroinflammatory pathways in AD [151]. Personalized stimulation of the precuneus, a hub of the default mode network, has also been associated with slower cognitive and functional decline over 24 and 52 weeks in mild-to-moderate AD [152,153]. These findings provide a rationale for exploring neuromodulation in DSAD, particularly because DS is characterized by altered neurodevelopmental circuitry, cerebellar hypoplasia, and AD-related network degeneration [154]. The cerebellum may also be relevant because cerebellar network dysfunction is increasingly considered in other neurodegenerative conditions, including frontotemporal dementia, where cerebellar noninvasive brain stimulation has been proposed as a potential diagnostic and therapeutic strategy [155]. However, these approaches should be framed as hypothesis-generating rather than established DSAD therapies, and DS-specific studies will need to establish feasibility, tolerability, individualized targets, seizure-related safety screening, sham-controlled designs, and outcome measures adapted to baseline intellectual disability and DSAD progression.

As therapeutic development in DSAD advances, multi-system and multiomics approaches are likely to play an increasingly important role [32]. Integrating transcriptomic, proteomic, imaging, and fluid-biomarker data across tissues and disease stages may help identify therapeutic targets beyond amyloid, clarify molecular mechanisms that underlie treatment response, and refine biological stratification for clinical trials. This approach will be essential for developing more precise, mechanism-informed intervention strategies in DSAD.

3. Conclusion

DS provides a uniquely informative biological context for understanding AD. The presence of trisomy 21 from conception creates a genetically defined environment in which amyloid pathology emerges predictably across the lifespan. Triplication of the APP gene establishes a strong mechanistic link between DS and AD, leading to early and nearly universal amyloid accumulation, followed by tau pathology, neurodegeneration, and dementia in later adulthood. At the neuropathological level, the distribution and composition of amyloid plaques, cerebral amyloid angiopathy, and neurofibrillary tangles in DS closely resemble those observed in autosomal dominant and sporadic forms of AD, although pathology often appears earlier and may reach greater severity. These shared features reinforce the concept of DS as a genetically defined form of AD while also highlighting biological processes that shape disease trajectory within the context of trisomy 21.

Beyond APP dosage, multiple chromosome 21 genes influence pathways central to AD biology, including tau phosphorylation, oxidative stress, immune signaling, and endosomal trafficking. As a result, individuals with DS exhibit early and persistent alterations in neuroinflammatory responses, endolysosomal function, and white matter integrity that precede overt neurodegeneration. These processes illustrate that AD in DS develops within a broader systems-level molecular environment rather than through amyloid dysregulation alone.

Multiomics approaches have significantly expanded understanding of this complex landscape. Transcriptomic and single-cell studies demonstrate widespread gene expression remodeling across neuronal, glial, and vascular cell populations, revealing cell-type-specific responses to trisomy 21 and amyloid deposition. Proteomic analyses further show that trisomy drives large-scale network changes in metabolism, proteostasis, synaptic function, and oxidative stress, many of which appear before extensive plaque or tangle formation. At the lesion level, spatial proteomics indicates that amyloid plaques in DS share a conserved molecular composition with plaques in early- and late-onset AD, enriched in proteins related to immune activation, extracellular matrix remodeling, and endo-lysosomal pathways. Fluid biomarker studies similarly demonstrate that adults with DS follow the canonical AD biomarker trajectory while also displaying molecular signatures that reflect trisomy-specific biology. Together, the findings summarized in this review establish DS as a powerful framework for examining early disease mechanisms, biomarker evolution, and therapeutic targets across the AD continuum.

The following expert opinion considers the translational implications of these findings and discusses how integrated proteomic and multiomics approaches may inform future neurodegenerative disease research and therapeutic development.

4. Expert opinion

The body of evidence reviewed here supports a shift in how neurodegenerative diseases are conceptualized and investigated. Insights from large-scale proteomics, studies of translational regulation, and multiomics integration indicate that gene-centric approaches and bulk protein measurements alone are insufficient to fully explain disease heterogeneity, progression, and therapeutic response in AD and related disorders. Disease risk and progression increasingly reflect coordinated changes in protein systems and molecular networks that vary across brain regions, peripheral compartments, and individuals.

Proteomic studies of human brain tissue have been particularly informative in defining these network-level alterations. Brain proteomics have revealed early disruption of proteostasis, mitochondrial function, synaptic signaling, and oxidative stress pathways that precede or accompany classical amyloid and tau pathology in both sporadic and genetically driven forms of AD [119]. In DS, proteomic analyses demonstrate that trisomy 21 induces widespread remodeling of brain protein networks beyond simple gene dosage effects, with strong convergence toward the molecular signatures observed in sporadic AD [52,137]. Lesion-specific analyses further show that amyloid plaque composition is remarkably conserved across DS, early-onset AD, and late-onset AD, highlighting shared pathogenic mechanisms at the tissue level while also revealing subtype-specific modulation [52]. These findings position brain proteomics as a critical layer linking genetic risk to downstream molecular and pathological phenotypes.

Complementing tissue-based studies, large cohort fluid proteomics analyses demonstrate that blood-based proteomic profiles can capture disease-relevant molecular signals at population scale. The Global Neurodegeneration Proteomics Consortium shows that plasma protein signatures are reproducible across cohorts and platforms and reflect both disease type and clinical impairment across AD, Parkinson’s disease, and frontotemporal dementia [156]. Pan-neurodegeneration analyses further indicate that molecular subtypes extend beyond traditional diagnostic boundaries and uncover shared and disease-specific biological pathways across disorders [125]. Together, brain and fluid proteomics underscore the value of protein-level measurements for biological staging and patient stratification.

Despite these advances, translation into clinical practice remains limited. High-dimensional proteomic analyses are not yet scalable in routine diagnostic settings, analytical pipelines lack regulatory standardization, and interpretation of complex molecular signatures remains challenging. While plasma proteomics offers a minimally invasive route for biomarker development, clinical integration will require longitudinal validation, linkage to clinical outcomes, and clear demonstration of added value over established imaging and cerebrospinal fluid markers. Integration across multiple omics further increases analytical complexity and currently confines most applications to specialized research environments [124].

At a mechanistic level, an important limitation of current approaches is incomplete capture and interpretation of protein diversity. Protein variation arises not only from genetic and post-translational processes but also from regulated translation, including non-canonical initiation and alternative reading frame usage, which are particularly active in the nervous system [126]. These features remain poorly represented in conventional proteomic workflows and help explain inconsistencies between transcriptomic and proteomic findings. Recognizing this biological complexity is essential for interpreting bulk tissue and fluid proteomics without overstating causal relationships.

DS-associated AD illustrates the importance of network-level biology in neurodegeneration. Multiomics studies show that trisomy 21 increases AD risk but does not fully determine disease onset or severity, which are influenced by pathways involving immune signaling, lipid metabolism, proteostasis, and epigenetic regulation beyond chromosome 21 [32]. These observations reinforce the view that neurodegeneration emerges from interacting molecular systems rather than isolated gene or protein changes.

In our view, the DSAD literature supports a model in which APP triplication and Aβ_1–42 accumulation are critical initiating and organizing events, but clinical dementia reflects the temporally structured convergence of amyloid pathology with broader molecular and cellular dysfunction. DS provides particularly strong evidence for this interpretation because APP triplication produces lifelong amyloid overproduction and a relatively predictable AD biomarker trajectory, yet clinical symptoms typically emerge only after downstream systems become progressively disrupted. Proteomic and multiomics studies indicate that immune activation, endo-lysosomal dysfunction, mitochondrial stress, vascular and BBB alterations, white matter injury, synaptic failure, and Tau pathology evolve alongside and downstream of amyloid. Thus, Aβ_1–42 accumulation in DSAD is part of a highly penetrant disease trajectory whose clinical expression depends on the cumulative failure of interacting biological networks.

Conceptual advances emphasizing functional protein diversity, including the paradigm of proteoform medicine, provide a useful perspective for integrating these observations and for guiding future methodological refinement [127]. Rather than replacing current approaches, such frameworks help contextualize bulk proteomics findings and motivate the development of more precise experimental and analytical strategies.

Looking forward, progress in neurodegeneration research is unlikely to converge on a single definitive endpoint. Disease processes are dynamic and multifactorial, and advances will be measured by improved predictive accuracy, mechanistic resolution, and translational relevance. Expansion of longitudinal cohorts, integration of brain and fluid proteomics with other molecular levels, and iterative validation in human-relevant models will be critical. Over the next five to ten years, research practice is likely to move toward integrated, network-based analyses supported by harmonized datasets. In this setting, proteomics will increasingly inform trial stratification, biomarker development, and mechanistic discovery. The key challenge will be managing growing methodological complexity while translating molecular insight into clinically actionable knowledge.

Article highlights.

• Down syndrome provides a powerful biological context for studying Alzheimer’s disease mechanisms, enabling investigation of early and genetically anchored disease processes.

• Brain proteomics reveals early disruption of proteostasis, mitochondrial function, synaptic signaling, and immune pathways that converge with sporadic Alzheimer’s disease pathology.

• Lesion-specific and tissue-level proteomic studies demonstrate strong molecular overlap across Alzheimer’s disease subtypes, indicating shared pathogenic mechanisms beyond gene dosage effects.

• Large-scale fluid proteomics captures reproducible disease-relevant signatures that support biological staging and patient stratification across neurodegenerative conditions.

• Integrated proteomic and multiomics approaches highlight network-level alterations and inform future neurotherapeutic research, while current challenges remain in scalability, standardization, and clinical translation.