As mentioned in Chapter 1, people with Down syndrome (DS) develop Alzheimer’s disease (AD) neuropathology by 40 years of age (Chapter 2), and in most, dementia develops in their 50s (Chapter 15). In this context, diagnostic and prognostic biomarkers of AD continue to be the focus of many research groups and could contribute significantly to the diagnosis and prognosis of dementia. Because the lifespan of this population has more than doubled in the past few decades due to better medical treatment of comorbidities  (see Chapter 11), AD and other age-related conditions have now become more prevalent in those with DS. Some investigators even consider DS a “segmental” form of progeria or accelerated aging , where multiple tissues are affected by aneuploidy with more striking effects manifesting in the brain and heart. In contrast to DS, other segmental progeroid syndromes, like Hutchinson-Gilford progeria and Werner syndrome, have significantly lower rates of brain aging relative to other organ systems.
The brain appears to age faster than other organs in DS, presumably due to overexpression of several genes located on Chr. 21 that code for proteins that may enhance neurodegeneration and contribute to the early incidence of AD . For example, APP is the substrate of amyloid-beta peptide (Aβ) production and is sufficient for the clinical, biochemical, and neuropathological manifestation of AD in DS . Research has revealed other key proteins, including the dual specificity tyrosine phosphorylation regulated kinase 1A (DYRK1A), which plays a role in the development of microtubule-associated protein Tau (MAPT) pathology in DS [4, 5], the regulator of calcineurin 1 (RCAN1) and superoxide dismutase 1 (SOD1), which have been associated with the increased oxidative stress seen in DS [6–8]. In addition, immune-related genes on Chr. 21, which contribute to an exacerbated immune dysregulation in DS, have been identified to play a role in DS-AD . Thus several genes, as well as gene-gene relationships, may drive the pathology related to AD observed in the brain of those with DS. Investigating proteins that are involved in AD pathology and their expression in body fluids also allows examination of the series of events during the lifespan that may create the “perfect storm” that accelerates AD in DS.
Clinical studies and interventions targeting the DS population are likely to inform clinicians of early changes occurring in brain pathways that may lead to the development of AD. Regrettably, despite the importance of this population for preventative studies, individuals with DS are often excluded from clinical trials of AD drugs  (see Chapter 16). With access to better biomarkers, which may lead to earlier recognition of transition to AD, or monitoring of treatment effects, people with DS may also get access to new clinical trials that can improve their lives and slow the progression of dementia. Since the course of AD in DS resembles that of familial and sporadic AD [11–13], at least in some aspects of the disease, such knowledge could also be translated and benefit those with AD in the general population.
This chapter provides an overview of the current status of fluid biomarkers in DS, focusing on core AD biomarkers and markers of neurodegeneration in plasma and CSF. The chapter also includes recent and novel work on other biomarker candidates such as inflammatory cytokines, oxidative stress, and miRNA, as well as investigations of biomarkers in brain-derived extracellular vesicles.
One of the earliest biomarkers described for AD in brain, plasma and CSF is Aβ [14–17]; for a recent review, see Ref. . Aβ peptides are the main aggregation-prone component of the neuritic plaques found in the brains of people with AD (Fig. 1) and can occur in various lengths of 36–43 amino acids [19, 20] (also see Chapter 2). The two most studied peptides are Aβ40 and Aβ42. They share similar sequences, except for an extra alanine and isoleucine at the C-terminal domain of Aβ42 .
Adults with DS (> 18 years), regardless of clinical classification, exhibit higher concentrations of Aβ in plasma compared to euploid individuals [13, 22–30]. Significant increases in both plasma Aβ40 and Aβ42 have been reported using either conventional ELISA immunoassays, electro-chemiluminescent-based immunoassays from MesoScale Discovery (MSD), bead-based immunoassays from Luminex xMAP technology (LMX), and ultrasensitive single-molecule detection arrays (Simoa) from Quanterix [31, 32]. The degree of reported increases in Aβ levels ranges between 1.5 and 2.8 fold for Aβ40 and 1.5 and 2 fold for Aβ42 at a group level, and in part reflects the expected overproduction of APP due to Chr. 21 trisomy, although other factors such as APOE genotype, Aβ clearance efficiency, or increased dosage of other Chr. 21 genes have been proposed to play a role [22, 23, 27]. The increased levels of plasma Aβ seen in adults with DS were also found to occur in the plasma of newborns and children with DS [30, 33], but not in fetuses .
A recent meta-analysis of nine studies examining plasma Aβ biomarkers in DS concluded that higher concentrations in plasma Aβ40 and lower ratios of Aβ42/Aβ40 distinguished individuals with DS-AD from those without dementia symptoms. However, findings between different studies included in the meta-analysis were heterogeneous . Indeed, there has been discrepancy regarding differences in plasma Aβ between individuals with DS-AD and those who were asymptomatic to dementia symptoms, as briefly discussed later.
When comparing diagnostic groups, most research studies have reported no differences in single time point measures of plasma Aβ in people with DS with or without dementia [22, 25, 26, 36–38]. However, some have found that individuals with DS and dementia showed higher levels of both Aβ40 and Aβ42 compared to asymptomatic individuals with DS , while others have found only differences in either Aβ40 alone [27, 39], Aβ42 alone , or in the Aβ42/Aβ40 ratio .
The levels of Aβ40 and Aβ42 in plasma have predicted cognitive decline and incidence of dementia in both cross-sectional [25, 41] and longitudinal investigations [26, 39, 40, 42], although the direction of change (increase/decrease) differed when single or multiple time point measures of Aβ were assessed. For example, higher baseline concentrations of plasma Aβ42 were associated with worse semantic verbal fluency, poorer communication skills, and a higher dementia rating , as well as with poorer performance on the Test of Severe Impairment, which assesses language, memory, executive function, and motor performance . Higher baseline concentrations of plasma Aβ42 have also been associated with increased AD risk in DS [39, 40]. However, when longitudinal changes in plasma Aβ (across two time points) were examined, declining plasma Aβ42 levels were associated with poorer cognitive performance  and with increased AD risk in adults with DS .
The apparent opposing outcomes of plasma Aβ changes across studies could be in part related to the use of different methodologies with different sensitivities in a complex matrix (e.g., ELISA [22, 23, 25, 36]; xMAP ; MesoScale  and Simoa ). Likewise, differences in age and dementia duration between cohorts, as well as the timing of Aβ assessment with respect to AD stage, could also explain some of these discrepancies . Indeed, according to a natural history study of AD in DS, which evaluated the order and onset of biomarker changes in this population, plasma Aβ42 decreased with age until AD symptom onset [13, 30], while rising thereafter . These results indicate that plasma Aβ changes in DS might differ according to the stage of AD progression, offering a new context to interpret the apparent opposing findings from previous studies, not necessarily as contradictory.
Regardless of the direction of change of plasma Aβ levels in DS, the findings from a large biomarker study employing Simoa do not support the utility of plasma Aβ measures as a reliable biomarker of AD in DS with the technologies discussed so far . In that study it was reported that the diagnostic performance of both plasma Aβ40 and Aβ42 peptides was poor in DS, with an AUC of 0.54 and 0.56, respectively, to distinguish DS-AD from asymptomatic individuals with DS. Moreover, it showed that plasma Aβ levels did not correlate with levels in CSF, as previously demonstrated in participants with AD in the general population , further suggesting limited use of this biomarker with this technology, at least when measured in plasma.
Before discussing technological improvements that might change the future landscape and utility of plasma Aβ biomarkers in DS, it should be noted that the lack of correlation between plasma and CSF Aβ measures also highlights the effects that other organs outside the brain may have on the plasma Aβ pool (e.g., heart, kidneys, and platelets) [44–46]: an aspect which has not been investigated in depth in people with DS. Indeed, evidence suggests that 30%–50% of plasma Aβ originates from the CNS  and it is possible that individuals with DS exhibit differences in Aβ production and clearance in peripheral organs compared to people with AD in the general population. Although these factors would most certainly affect plasma levels of Aβ, they have yet to be studied in detail. More studies are therefore needed to examine the complex relationship between Aβ depositions in brain tissue with plasma levels of Aβ in the population with DS.
Despite the limited use of plasma Aβ as a biomarker of AD in DS even when measured with ultrasensitive immunoassays, novel techniques involving mass spectrometry (MS) are emerging, and their promising results in the general population, together with their use and applicability by independent research groups, warrant future studies in DS. Using these techniques, it was reported that plasma Aβ42 or the ratio Aβ42/Aβ40 correlated well with CSF Aβ42 levels and with amyloid PET [48–50], suggesting that these assays could be used for the screening of brain amyloidosis in individuals at risk. Indeed the diagnostic accuracy to distinguish AD and controls or cerebral amyloid deposition was found to be superior to most plasma amyloid assays. Therefore future avenues of research in plasma Aβ biomarkers should include assessing the performance of these techniques in the DS population.
The MAPT gene codes for a structural protein (Tau) that is the major pathological component in abnormal intraneuronal neurofibrillary tangles (NFTs) observed in AD  and in DS (see Chapter 2). Tau is mainly expressed in neurons of the central nervous system, where it maintains axonal architecture and transport functions. The rationale for Tau as a biomarker of neurodegenerative diseases is based on the fact that its increased expression levels, aggregation, and modification status reflect neuronal distress and degeneration (Fig. 1). Tau has approximately 85 phosphorylation sites and 71 of these sites are known to be differentially altered during pathological conditions . The exact mechanisms that regulate Tau phosphorylation are complex, as more than 20 protein kinases can phosphorylate Tau. However, the most prominent change in phosphorylation of Tau that is associated with AD occurs on epitopes localized at positions 181, 199, 231, 396, and 404 .
Tau protein is in relatively low concentrations in plasma making it difficult to detect; this is why studies of plasma tau are mostly recent, as a result of the development of ultrasensitive technologies, which opened new opportunities to explore this biomarker. A total of seven studies have reported Tau determinations in plasma of DS from population-based cohorts in Europe (Spain, United Kingdom), Asia (Japan, Taiwan), and the United States, as discussed later.
Despite finding agreement in the literature that mean total Tau concentrations in plasma are higher in DS with respect to non-DS controls, particularly in the older age brackets (40–50 + years) [27, 30, 54, 55], as well as between DS-AD and the asymptomatic group , the overlap between individual Tau values is substantial; and no differences between DS groups have even been reported in some studies [13, 28]. Indeed, the diagnostic performance of this biomarker was found to be poor to distinguish between DS-AD and asymptomatic individuals with DS (AUC = 0.74) and weak correlations were found between plasma Tau and its levels in CSF , in agreement with reports from patients with AD in the general population . These results indicate that the diagnostic usefulness of this biomarker is limited.
Beyond these arguments, an additional disadvantage of total Tau as a biomarker is that its levels have been found to rise in other neurological conditions not associated with AD, such as Creutzfeldt-Jakob disease, traumatic brain injury, and stroke [57, 58]. In contrast, certain phosphorylation sites of the Tau protein have been established as reliable markers of neurofibrillary tangle accumulation in AD .
Indeed, promise is emerging on the use of phosphorylated Tau at threonine 181 (p-Tau T181) as a sensitive and accurate blood biomarker of AD in the general population. Using data from three independent clinic-based cohorts, Karikari and colleagues recently showed that plasma p-Tau T181 levels increased along the AD continuum and that it exhibited an excellent diagnostic performance to distinguish people with AD from Aβ-negative cognitively unimpaired older adults (AUC = 0.90–0.98 across cohorts), and from people with other neurological disorders, including frontotemporal dementia (AUC = 0.83–1 across cohorts), vascular dementia (AUC = 0.92), progressive supranuclear palsy or corticobasal syndrome (AUC = 0.89), and Parkinson’s disease or multiple systems atrophy (AUC = 0.82) . When validated in a primary care setting, plasma p-Tau T181 further showed its robustness to distinguish individuals with AD from healthy older adults (AUC = 0.84), although the discriminative performance of this biomarker was not sufficient to distinguish AD from the MCI group (AUC = 0.55) .
The increase in plasma p-Tau T181 along the AD continuum is seen in other independent studies [61–63] and its diagnostic performance was likewise found to be robust in distinguishing between AD and frontotemporal lobar degeneration (AUC 0.89) or other non-AD neurodegenerative diseases (AUC 0.94–0.98) [61, 63]. Janelidze and colleagues also found that higher p-Tau T181 levels were associated with prospective dementia onset both in cognitively normal individuals as well as in those with MCI , indicating a potential prognostic value of this marker that warrants replication in future studies. Moreover, as opposed to total Tau, plasma levels of p-Tau T181 correlated with p-Tau T181 levels in CSF and with Tau pathology in brain, as assessed with flortaucipir PET [61, 63], strengthening the evidence that this analyte could be a more reliable and noninvasive diagnostic and prognostic biomarker of Tau pathology associated with AD compared to total Tau. Along this line, Janelidze and colleagues mention that an additional explanation to account for the differences between these two biomarkers could be due to their different turnover in plasma. While total Tau appears to degrade quickly in plasma (hours), as opposed to CSF where it can be detectable for weeks [58, 64], it has been hypothesized that the Tau fragment measured by the p-Tau T181 immunoassay could be resistant to this degradation . Another promising development is the use of plasma p-Tau T217 as biomarker, which showed high accuracy to distinguish AD from other dementias (AUC = 0.96) and to discriminate normal from abnormal tau PET scans .
In the DS population, so far there has been only one study in which investigators examined levels of p-Tau T181 in plasma with Simoa, revealing higher p-Tau T181 levels in DS when compared to non-DS controls at a group level . However, individually, there was variability in p-Tau T181 concentrations, with some people with DS showing little to undetectable levels, while others (mainly those above 40 years) showed very high levels (> 1.00 pg/mL). The age range for the group with DS was 19–57 years and it is possible that part of the variability in p-Tau T181 levels found in this study is a reflection of the dynamic evolution of Tau pathology in DS across these ages and AD stages (see Chapter 2). Future studies on p-Tau T181 are warranted in DS, particularly to evaluate both its diagnostic and prognostic performance, based on the promising results reported in the general population. A new study from the DABNI cohort (Down Alzheimer Barcelona Neuroimaging Initiative) in collaboration with researchers from the University of Gothenburg, Sweden reported that plasma p-Tau T181 showed an excellent diagnostic performance to detect AD in DS , in line with results from the general population. No studies on plasma p-Tau-T217 have been yet performed in DS.
Another useful biomarker that is gaining momentum in neurology is the neurofilament light chain protein (NfL), which is a neuronal cytoplasmic protein highly expressed in myelinated axons . The rationale for serum or plasma NfL as a biomarker of neurodegenerative diseases relies on the fact that NfL levels are relatively low in peripheral fluids under normal physiological conditions. However, in response to axonal damage due to inflammation and neurodegeneration, the release of NfL sharply increases (Fig. 1).
As of today, five studies have examined plasma NfL levels in DS; all of them using the Simoa ultrasensitive technology, as discussed later and further detailed in Table 1. Strydom and colleagues first reported significant increases in plasma NfL (approximately threefold) in individuals with DS-AD (N = 18) compared with those classified as asymptomatic (N = 76) at the time of NfL determination . Diagnostic performance was not assessed in this first study. Interestingly, they showed that individuals with higher baseline NfL levels had a higher likelihood of developing dementia at follow-up (time frame not specified).
|Reference||Subjects (N)||Sex (M/F)||Age (years ± SD)||Plasma NfL (median pg/mL)||Assay||Main result|
|Strydom et al. ||aDS: 76||aDS: 41/35||aDS: 39.7 ± 14.3||aDS: 19.96||Simoa (HD-1)||NfL correlates with age (r = .789, P < .001);|
|dDS: 18||dDS: 12/6||dDS: 55.2 ± 9.9||dDS: 63.76||NfL dDS > aDS (P < .001)|
|Fortea et al. ||CTL: 67||CTL: 20/47||CTL: 52.05*||CTL: 4.15||Simoa (HD-1)||aDS, pDS, dDS > CTL (P< .0001); pDS, dDS > aDS (P < .0001); pDS vs. dDS: n.s.|
|aDS: 194||aDS: 105/89||aDS: 37.05*||aDS: 5.88|
|pDS: 39||pDS: 21/18||pDS: 49.83*||pDS: 15.39|
|dDS: 49||dDS: 28/21||dDS: 54.88*||dDS: 23.04|
|Rafii et al. ||DS: 12||2/10||45.33 ± 8.53||5.79–71.73 (range)||Simoa (HD-1)||NfL correlates with Florbetapir PET (r = .73, P = .007), hippocampal volume (r = −.52, P = .084), CAMCOG (r = − .66, P = .022), CANTAB PAL errors (r = .68, P = .015), and inverse relation with FDG-PET 5 regionsa|
|Mengel et al. ||CTL: 100||CTL, DS: 62/38||CTL and DS: range 0.3–68 years||CTL vs. DS (mean ± SEM)||Simoa (HD-1)||NfL significantly elevated in DS compared to CTL in the older age brackets (41–50, 50 +)|
|DS: 100||0–10 years: 6.3 ± 0.6 vs. 7.5 ± 0.5|
|11–20 years: 3.6 ± 0.3 vs. 5.1 ± 0.4|
|21–30 years: 5.4 ± 0.9 vs. 6.7 ± 0.4|
|31–40 years: 6.8 ± 1.3 vs. 10.2 ± 1.6|
|41–50 years: 8.3 ± 1.0 vs. 15.9 ± 1.0|
|50 + years: 11.7 ± 1.1 vs. 31.0 ± 6.0|
|Fortea et al. ||CTL: 242||CTL: 80/162||CTL: 56.6*||CTL: 4.4||Simoa (HD-1, SR-X)||NfL significantly elevated in DS since age 30 (20 years before AD diagnosis)|
|aDS: 257||aDS: 144/113||aDS: 38.7*||aDS: 6.0|
|pDS: 48||pDS: 26/22||pDS: 50.2*||pDS: 13.0|
|dDS: 83||dDS: 44/39||dDS: 53.7*||dDS: 22.3|
Median plasma NfL concentrations are shown, unless otherwise indicated. Abbreviations: aDS, asymptomatic; CAMCOG, Cambridge Cognitive Examination; CANTAB, Cambridge Neuropsychological Test Automated Battery; CTL, control; dDS, dementia due to AD; DS, Down syndrome; F, female; FDG, fluorodeoxyglucose; M, male; N, number of subjects; NfL, neurofilament-light; PAL, paired associate learning; pDS, prodromal AD; SD, standard deviation.
a Inverse correlations between NfL and regional glucose metabolism in 5 of 6 regions examined: anterior cingulate (r = −.55, P = .067, Pa = .067), posterior cingulate (r = − .90, P < .001, Pa < .001), lateral temporal (r = − .78, P = .004, Pa = .012), frontal cortex (r = − .90, P < .001, P Pa < .001), parietal cortex (r = − .82, P = .002, Pa = .008), and precuneus (r = − .73, Pa = .010, Pa = .020); Pa, adjusted P value.
The largest study to report differences in NfL plasma concentrations included 282 individuals with DS from the DABNI cohort . Regardless of clinical classification, adults with DS exhibited significantly higher levels of plasma NfL compared with nontrisomic controls (an overall increase of around twofold) . When comparing NfL levels across clinical categories (asymptomatic DS, prodromal DS-AD, and demented DS-AD), there were progressive increases in NfL, with significantly higher levels in the symptomatic groups compared to asymptomatic individuals with DS. Importantly, ROC analysis further revealed an excellent diagnostic performance of plasma NfL for distinguishing asymptomatic versus symptomatic individuals (0.88 for asymptomatic vs. prodromal and 0.95 for asymptomatic vs. dementia), which was greater to that of plasma Aβ and Tau biomarkers. Plasma and CSF levels of NfL were significantly correlated. Moreover, an ongoing longitudinal study involving six centers across Europe and the US has confirmed the clinical utility and excellent diagnostic performance of plasma NfL to distinguish AD in adults with DS (AUC 0.94). Importantly, this study showed for the first time that longitudinal NfL analyses also demonstrate good prognostic performance, supporting its use as a reliable and noninvasive biomarker that could be used in clinical trials .
Besides correlating with CSF NfL levels, a recent study with a small group of adults with DS (N = 12) from the DSBI pilot project (Down syndrome Biomarker Initiative) reported that plasma NfL correlated with neuroimaging PET biomarkers of Aβ and Tau load, as well as with hippocampal volume and cerebral glucose metabolism in specific AD-vulnerable regions . These results strengthen the evidence that NfL is a general biomarker of brain neurodegeneration; in fact, it is not specific to AD and is also altered in other dementias  as well as in concussed individuals . This aspect, however, does not represent a drawback for its use in the DS population, who due to their genetic condition are expected to uniformly develop AD pathology by middle age, and the clinical symptoms of dementia years later. Such a gap between the buildup of AD pathology and emergence of dementia symptoms represents a key window for prognostic indicators. Interestingly, the trajectory of plasma NfL in DS showed detectable increases 20 years before expected symptom onset [13, 30], underscoring the relevance and potential clinical applicability of this biomarker.
In sporadic and familial AD, CSF core AD biomarkers have been widely utilized in clinical research and increasingly incorporated into clinical practice [18,74]. In contrast, studies on CSF biomarkers remain scarce in DS since, in some countries, invasive procedures including lumbar punctures (LPs) are avoided in vulnerable populations such as those with DS. To address this, the Alzheimer Down Unit in Barcelona collected data from 80 adults with DS who underwent LP, and complications or adverse effects of the procedure were evaluated . The study reported that LP was safe and well tolerated (no adverse events in 90% of participants), with minimal and mild complications including headache, dizziness, and back pain, similar if not lower, to what is reported in the general population , demonstrating that individuals with DS are capable of undergoing such procedures.
The first study to measure CSF levels of Aβ species in DS determined Aβ1-40, Aβ1-42, Aβx-40, and Aβx-42 in 5 people with DS aged around 55 years using ELISA, and compared them with 34 nontrisomic individuals diagnosed with neurological diseases other than dementia . No significant differences were found in CSF Aβ1-40, Aβx-40, and Aβx-42; however, the concentration of Aβ1-42 was significantly lower in DS than in the control group: a finding that resembles that of sporadic and autosomal dominant AD populations [78,79].
Later, Tapiola et al. measured Aβ1-42 and total Tau (t-Tau) levels in the CSF of 12 participants with DS of a slightly younger age than the previous report (mean age 41 years), and in 19 nontrisomic controls without dementia (mean age 53 years) using ELISA. CSF Aβ1-42 was significantly lower in the group with DS with respect to controls but no differences were found in t-Tau . The authors further reported a negative correlation between age and CSF Aβ1-42, with the highest levels seen in the youngest participants. A positive correlation was revealed between age and CSF Tau .
The higher levels of CSF Aβ in people with DS at young ages is also supported by a study in which investigators measured five species of Aβ, t-Tau, and p-Tau in longitudinal samples from a group of 24 infants with DS (measured at 8, 20–40, and 54 months). All of the Aβ species measured were found to increase with age, although only Aβ1-37 and Aβ1-38 were significant. The higher levels of Aβ in CSF at young ages are consistent with the minimal accumulation of Aβ in the brain and overexpression of APP. At older ages, CSF Aβ drops as Aβ accumulates in the brain in the form of plaques (see Chapter 2). With respect to T-Tau and p-Tau, their levels remained stable in CSF, suggesting that overt development of Tau pathology detectable in CSF may occur at later stages in DS , although a deficit in Tau phosphorylation processes is detected early at the fetal stage  and Tau biomarkers are increased in neuron-derived-exosomes in children with DS , as described later (see “Biomarkers in neuron-derived exosomes in Down syndrome” section).
Portelius and colleagues later expanded the findings of Tapiola et al. by investigating the same 12 subjects and including other CSF biomarkers for Aβ pathology (Aβ species Aβx-38, Aβx-40, Aβx-42; soluble APP fragments sAPPα, sAPPβ), in addition to the core AD biomarkers (Aβ1-40, Aβ1-42, t-Tau, p-Tau) measured with MSD arrays and ELISA . The authors confirmed the decrease of Aβ1-42 and the increase in t-Tau and p-Tau with age, but also described significantly increased levels of Aβx-40 and both soluble APP fragments in DS CSF compared to controls. Using hybrid immunoaffinity-mass spectrometry (MS) and ELISA, the same group further reported in the same DS cohort that the relative abundance of Aβ1-42 (based on the relative MS peak heights/areas of various CSF Aβ peptides) was decreased and negatively correlated with age . This result is also in line with the increased deposition into plaques and reduced clearance of Aβ that is present in the brains of people with DS.
Taken together, these small studies have pioneered biomarker characterization of AD in CSF in the DS population but have obvious limitations, mainly due to the small sample sizes and the lack of cognitive characterization of the participants with DS, and should therefore be cautiously interpreted . Because of these limitations in individual cohorts, several research groups in Europe and the United States have formed collaborations to combine many cohorts in the search for better biomarkers distinguishing the development of predementia processes in the brain of DS.
Strong evidence for the usefulness of CSF biomarker changes in DS across the AD continuum came from a cross-sectional study from the DABNI cohort in Spain, where it was shown that CSF Aβ42, t-Tau, p-Tau, and NfL exhibited a high diagnostic performance to differentiate prodromal AD and AD dementia in adults with DS . In contrast, CSF Aβ40 concentrations showed a poor diagnostic performance. In essence, this study demonstrated that the biological signature of AD can be robustly detected in CSF in people with DS, in agreement with previous reports in sporadic and familial AD. Moreover, it showed that while changes in CSF core AD biomarkers and NfL were similar (in both direction and magnitude) to those found in the general population and familial AD, they showed superior diagnostic performance in people with DS .
The timing of CSF biomarker changes with respect to expected symptom onset was described in a subsequent large multimodal study that characterized the course of AD in adults with DS. The study found decreased values of the CSF Aβ42/Aβ40 ratio starting in the second decade of life and being significantly different from controls since age 28 (22 years before prodromal AD diagnosis), reaching the lowest levels in the sixth decade . CSF levels of Tau, p-Tau, and NfL showed similar but opposite trajectories to amyloid CSF biomarkers, increasing since the third decade of life and being different from controls 10–14 years before prodromal AD diagnosis . In accordance with the conceptualization of DS as a form of genetically determined AD, biomarker changes in CSF begin decades before symptoms of AD dementia emerge, in a strikingly similar order and temporality to that described in familial AD . This finding further supports the consideration of DS as a suitable population who may benefit from AD clinical trials. Given the similarities between sporadic, familial AD and DS, such therapies could prove beneficial not only for people with DS but also for the other forms of AD. A summary of CSF studies in DS is presented in Table 2.
|Reference||Sample with available CSF||Age (years ± SD) a||Sex (M/F)||Assay||Aβ1-42 (median pg/mL)||Aβ1-40 (median pg/mL)||t-tau (median pg/mL)||p-tau (median pg/mL)||NfL (median pg/mL)||Main results|
|Tamaoka et al. ||5 DS||55.3 ± 3.4||ns||ELISA||181 ± 110a||1870 ± 1190a||ND||ND||ND||Aβ1-42 DS < CTL (P < .05)|
|34 CTL||67.9 ± 10.4||323 ± 165a||1520 ± 616a|
|Tapiola et al. ||12 DS (3 dDS)||41 ± 11||6/6||ELISA||572 ± 160 (< 40 years)||ND||144 ± 101 (> 40 years)b||ND||ND||Aβ1-42 DS < CTL (P < .01)|
|370 ± 105 (> 40 years)||500 ± 341 (> 40 years)b||No differences in t-Tau|
|19 CTL||53 ± 5||8/11||578 ± 129||246 ± 109b||Aβ1-42 and t-Tau correlate with age in DS (r = − .785, P < .005; r = 7.18, P < .05)|
|Englund et al. ||9 DS||8 months||ns||Western blot||1200||5900||ns||ns||ND||Aβ1-42 and Aβ1-40 increased with age (n.s.); t-Tau and p-Tau stable|
|11 DS||20–40 months||1800||7500|
|4 DS||54 months||1800||8800|
|Portelius et al. ||12 DS (3 dDS)||41 ± 11||6/6||ELISA||637 ± 201||ND||431 ± 369||52 ± 31||ND||Aβ1-42 similar in DS and CTL; Aβ1-42 and t-Tau correlate with age in DS (r = − .69, P = .0015; r = .68, P = .025); t-Tau DS (> 40 years) > DS (< 40 years) (P = .009); p-Tau DS (> 40 years) > DS (< 40 years) (P = .046)|
|20 CTL||40 ± 15||8/12||674 ± 145||ND||210 ± 87||34 ± 9|
|Portelius et al. ||12 DS (3 dDS)||41 ± 11||6/6||Mass spectrometry||Relative abundance||ND||ND||ND||ND||Aβ1-42 decreased in DS (rel. abund.); negative correlates with age (P < .001)|
|20 CTL||40 ± 15||8/12|
|Fortea et al. ||94 DS||54 aDS||37.2 ± 7.2||30/24||ELISA||753.7 ± 221||5559 ± 708||177 ± 72||36 ± 13||349 ± 140||Aβ1-42, t-Tau, p-Tau, NfL have a high diagnostic performance to detect symptomatic AD in DS|
|18 pDS||50.8 ± 1.1||11/7||418 ± 89||5239 ± 560||539 ± 288||81 ± 32||742 ± 217|
|22 dDS||54.1 ± 4.2||13/9||391 ± 57||5311 ± 694||853 ± 342||95 ± 24||1100 ± 21|
|67 CTL||52.1 ± 5||20/47||825 ± 122||5590 ± 1325||174 ± 42||37 ± 9||351 ± 54|
|Fortea et al. ||137 DS||73 aDS||47.7 ± 14||75/62||Lumipulse (Aβ1-42, tau, p-tau)||Aβ1-42/40:||0.078 ± 0.031||ns||40 ± 401||369 ± 289||Decrease in Aβ42/40 starts in the 2nd decade of life in DS, reaching the lowest levels in the 6th decade. Differs from controls 22 years before AD diagnosis. Tau, p-Tau, and NfL increase since 3rd decade and differ from CTL 14 years before AD diagnosis.|
|26 pDS||ELISA (NfL)||0.042 ± 0.011||110 ± 119||667 ± 530|
|38 dDS||0.046 ± 0.013||155 ± 92||1139 ± 679|
|137 CTL||57.1 ± 11.7||72/135||0.106 ± 0.010||33 ± 18||400 ± 206|