Belbin et al. Molecular Neurodegeneration (2020) 15:46 https://doi.org/10.1186/s13024-020-00398-0 RESEARCH ARTICLE Open Access Cerebrospinal fluid profile of NPTX2 supports role of Alzheimer’s disease-related inhibitory circuit dysfunction in adults with Down syndrome Olivia Belbin1,2* , Mei-Fang Xiao3, Desheng Xu3, Maria Carmona-Iragui1,2,4, Jordi Pegueroles1,2, Bessy Benejam1,4, Laura Videla1,2,4, Susana Fernández4, Isabel Barroeta1,2,4, Raúl Nuñez-Llaves1,2, Victor Montal1,2, Eduard Vilaplana1,2, Miren Altuna1,2,4, Jordi Clarimón1,2, Daniel Alcolea1,2, Rafael Blesa1,2, Alberto Lleó1,2, Paul F. Worley3,5 and Juan Fortea1,2,4 Abstract Background: Alzheimer’s disease (AD) is the major cause of death in adults with Down syndrome (DS). There is an urgent need for objective markers of AD in the DS population to improve early diagnosis and monitor disease progression. NPTX2 has recently emerged as a promising cerebrospinal fluid (CSF) biomarker of Alzheimer-related inhibitory circuit dysfunction in sporadic AD patients. The objective of this study was to evaluate NPTX2 in the CSF of adults with DS and to explore the relationship of NPTX2 to CSF levels of the PV interneuron receptor, GluA4, and existing AD biomarkers (CSF and neuroimaging). Methods: This is a cross-sectional, retrospective study of adults with DS with asymptomatic AD (aDS, n = 49), prodromal AD (pDS, n = 18) and AD dementia (dDS, n = 27). Non-trisomic controls (n = 34) and patients with sporadic AD dementia (sAD, n = 40) were included for comparison. We compared group differences in CSF NPTX2 according to clinical diagnosis and degree of intellectual disability. We determined the relationship of CSF NPTX2 levels to age, cognitive performance (CAMCOG, free and cued selective reminding, semantic verbal fluency), CSF levels of a PV-interneuron marker (GluA4) and core AD biomarkers; CSF Aβ1–42, CSF t-tau, cortical atrophy (magnetic resonance imaging) and glucose metabolism ([18F]-fluorodeoxyglucose positron emission tomography). Results: Compared to controls, mean CSF NPTX2 levels were lower in DS at all AD stages; aDS (0.6-fold, adj.p < 0.0001), pDS (0.5-fold, adj.p < 0.0001) and dDS (0.3-fold, adj.p < 0.0001). This reduction was similar to that observed in sporadic AD (0.5-fold, adj.p < 0.0001). CSF NPTX2 levels were not associated with age (p = 0.6), intellectual disability (p = 0.7) or cognitive performance (all p > 0.07). Low CSF NPTX2 levels were associated with low GluA4 in (Continued on next page) * Correspondence: obelbin@santpau.cat 1 Memory Unit and Biomedical Research Institute Sant Pau (IIB Sant Pau), Neurology Department, Hospital de la Santa Creu i Sant Pau, 08025 Barcelona, Spain 2 Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas (CIBERNED), 28031 Madrid, Spain Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 2 of 10 (Continued from previous page) all clinical groups; controls (r2 = 0.2, p = 0.003), adults with DS (r2 = 0.4, p < 0.0001) and sporadic AD (r2 = 0.4, p < 0.0001). In adults with DS, low CSF NPTX2 levels were associated with low CSF Aβ1–42 (r2 > 0.3, p < 0.006), low CSF t- tau (r2 > 0.3, p < 0.001), increased cortical atrophy (p < 0.05) and reduced glucose metabolism (p < 0.05). Conclusions: Low levels of CSF NPTX2, a protein implicated in inhibitory circuit function, is common to sporadic and genetic forms of AD. CSF NPTX2 represents a promising CSF surrogate marker of early AD-related changes in adults with DS. Keywords: Neuronal Pentraxin-2, Alzheimer’s disease, Down syndrome, Inhibitory circuits, Cerebrospinal fluid, Biomarker, GluA4, Cortical atrophy, Glucose metabolism Background homeostatic maintenance of excitation/inhibition balance Down syndrome (DS) has been conceptualized by the and for circuit rhythmicity essential for memory [13, 14]. International Working Group for Alzheimer’s disease GluA4 is also down-regulated in AD brains in a coordi- (AD) as a genetically determined form of AD [1] due to nated manner with NPTX2 [12], consistent with a disrup- the triplication of the gene encoding amyloid precursor tion of PV interneuron function. The reduction in brain protein (APP), among others [2]. AD in DS has a long NPTX2 expression is also reflected in the CSF and a preclinical phase and a predictable sequence of changes recent study has shown that the ratio of CSF NPTX2 to in biomarkers that is strikingly similar (both in order CSF t-tau (neurodegeneration marker) was the best CSF and temporality) to that described in autosomal domin- predictor of cognitive decline in sporadic AD [15]. ant AD [3, 4]. The similar patterns of amyloid depos- Based on the aforementioned data, we hypothesized ition, atrophy and hypometabolism indicate that AD in that, as a genetic form of AD, DS CSF will have reduced DS targets the same cortical regions as those affected in NPTX2 and correlated GluA4 levels, reflecting PV inter- the sporadic and autosomal dominant forms [5]. This is neuron dysfunction, thus representing a potential surro- supported by neuropathological studies, which show that gate marker of early AD pathophysiology in adults with DS brains from donors who arrive to autopsy in the 4th DS. To test this hypothesis, we have evaluated CSF levels decade present AD pathology [6]. The cumulative inci- of NPTX2 as a surrogate marker of inhibitory circuit dence of AD dementia exceeds 90% in the seventh dec- dysfunction using the Down Alzheimer Barcelona Neu- ade with a median age-of-onset of 55 years [7, 8], and as roimaging Initiative (DABNI) cohort, the largest collec- such, AD is now the leading cause of death in adults tion worldwide of CSF and neuroimaging data collected with DS [9]. While neuropsychological tests are able to from adults with DS (n = 94). We have included non- detect cognitive decline in this population, the intellec- trisomic population controls (n = 34) and sporadic AD tual disability associated with the syndrome can compli- patients (n = 40) from the Sant Pau Initiative for Neuro- cate an AD diagnosis. Thus, there is an urgent need for degeneration (SPIN) cohort for comparison. objective markers of AD in adults with DS to improve early diagnosis, monitor disease progression and assess Methods response to treatment. Study design Accumulating evidence suggests that early dysfunction This is a single-center cross-sectional study of cogni- of inhibitory interneuronal circuits contributes to cogni- tively normal non-trisomic individuals, adults with DS tive abnormalities in AD decades before clinical disease and sporadic AD patients. Controls and sporadic AD onset [10]. It follows that a surrogate marker of inhibi- patients were retrospectively selected from the Sant Pau tory circuit dysfunction could be a valuable biomarker of Initiative for Neurodegeneration (SPIN) cohort (Memory early AD pathogenesis. In support of this, recent studies Unit, Hospital de la Santa Creu i Sant Pau, Barcelona, have shown that neuronal pentraxin-2 (NPTX2), a protein Spain, [16]). Controls were without cognitive or neuro- that accumulates at excitatory synapses onto GABA-ergic logical disorders and normal levels of core AD CSF parvalbumin (PV) interneurons [11], is down-regulated in biomarkers (Aβ1–42, t-tau and p-tau), determined pathologically confirmed sporadic AD and adult trisomy according to our internal cut-offs [17]. AD patients were 21 brains [12]. At the excitatory synapse on PV interneu- clinically diagnosed with dementia due to AD [18]. rons, NPTX2 has been shown to mediate activity- Adults with DS were recruited from the Down Alzhei- dependent increases of the AMPA receptor subunit, mer Barcelona Neuroimaging Initiative, a prospective GluA4, which is selectively expressed in PV interneurons longitudinal cohort, linked to a population-based health [11]. This action of NPTX2 enhances PV interneuron ac- plan in Catalonia, Spain, led by the Catalan Foundation tivity and thereby inhibitory circuit function important for for Down Syndrome and Hospital de la Santa Creu i Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 3 of 10 Sant Pau, Barcelona, Spain [19]. Inclusion criteria for Aβ1–42 (Lumipulse Aβ1–42, Fujirebio-Europe, Belgium) total participation in the study required that participants were tau (Lumipulse G Total Tau, Fujirebio-Europe) and tau over 18 years of age, that they received a comprehensive phosphorylated at threonine residue 181 (Lumipulse G neurological and neuropsychological evaluation and PTAU 181, Fujirebio-Europe) in the DABNI cohort and underwent a lumbar puncture to assess AD biomarkers SPIN controls [17]. An in-house enzyme-linked immuno- levels (CSF Aβ1–42, t-tau and p-tau) [17]. As in previous sorbent assay targeting the N-terminal fragment encoding studies [19–21], participants with DS were classified by amino acid 1 to 201 of NPTX2 was performed in the Balti- neurologists and neuropsychologists, masked to bio- more Laboratory as previously described [12]. marker data in a consensus meeting into the following groups: asymptomatic AD (aDS; no clinical or neuro- Targeted liquid chromatography mass spectrometry psychological suspicion of symptomatic AD), prodromal We monitored a set of 3 proteotypic peptides AD (pDS; suspicion of cognitive decline due to AD, but corresponding to GluA4 by Selected Reaction Monitoring symptoms did not fulfill criteria for dementia) and AD (SRM) as previously described [28]. Briefly, we digested in- dementia (dDS; full blown dementia). dividual CSF samples overnight and spiked isotopically- labeled peptides (Pepotech SRM custom peptides, grade 2, Neuropsychological assessment Thermo Fisher Scientific) into each sample. We analyzed The level of intellectual disability was categorized accord- an equivalent of 5ul of each sample in a randomized order ing to the Diagnostic and Statistical Manual of Mental over a 120-min gradient (0–35% ACN + 0.1%FA) in SRM Disorders, Fifth Edition as mild, moderate, severe, or mode using a triple quadrupole-Qtrap mass spectrometer profound intellectual disability, based on carers’ reports of (5500 QTrap, Sciex, Massachussetts) coupled to a nano- the individuals’ best-ever level of functioning and the LC chromatography column (300 ul/min, 25-cm C18 Kaufmann Brief Intelligence Test [22]. Neurological and column, 75 um I.d., 2 um particle size). We ran BSA neuropsychological examination of individuals with DS, as technical controls between each sample. We visualized previously described [15], included a semi-structured and analyzed transitions using Skyline 3.5. Injection of a health questionnaire (Cambridge Examination for Mental pool of all the samples over the duration of the mass Disorders of Older People with DS and others with intel- spectrometric measurements and monitoring the peak lectual disabilities) [23] and a neuropsychological battery area of the standard peptides confirmed the stability of all including the Cambridge Cognition Examination (CAM- peptides over the course of the experiment. We processed COG) adapted for intellectual disabilities in individuals the SRM transitions using the dataProcess function of with DS limited to mild or moderate intellectual disability. MSstats v3.5 package in R [29] and removed transitions Declarative verbal memory was evaluated by the Free and with between run interference (betweenRunInterference- Cued Selective Reminding Test (FCSRT) as previously Score< 0.8). Censored missing values were samples with described [24]. log base-2 endogenous intensities under the cut-off desig- nated by the MSstats package (10.35). We used the Equali- Karyotyping zeMedians function to normalize the transitions and DNA samples extracted from blood were subjected to Tukey’s Median Polish to generate a summarized value. high-density single nucleotide polymorphism genotyping with the Illumina Infinium Global Screening Array (Illu- Magnetic resonance imaging (MRI) acquisition and mina, San Diego, CA, USA). Visualization and confirm- analysis ation of chromosome 21 aneuploidy was done through A high-resolution three-dimensional structural dataset the Genome Viewer tool within Genome Studio version was acquired in a 3 T MRI scanner (Philips 3.T X Series 2.0 (Illumina). A shift towards 66 and 33% in the B allele ACHIEVA) with the following parameters: T1-weighted frequency, and an increase from approximately 0 to 0.25 magnetization-prepared rapid gradient-echo, repetition in the log R ratio across all single nucleotide polymor- time 8.1 msec, echo time 3.7 msec, 160 slices, matrix size phisms contained within chromosome 21 was used as an 240 × 234; slice thickness 1 mm, voxel size 0.94 × 0.94 × indicator of trisomy [25]. All adults with DS included in 1 mm. Cortical thickness reconstruction was performed the study had confirmed trisomy of chromosome 21. with Freesurfer package v6 (http://surfer.nmr.mgh.har vard.edu) using a procedure that has been described in CSF collection and biomarker assessment detail elsewhere [30], as previously reported [28, 29]. A CSF samples were collected following international consen- Gaussian kernel of 15 mm full width at half maximum sus recommendations [26] as previously described [27]. (FWHM) was applied to the subjects’ cortical thickness Samples had been previously stored at − 80 °C and had not maps before further analyses as it is customary in been thawed prior to analysis. Commercially available auto- surface-based analyses. From the 75 adults with DS who mated immunoassays were used to determine levels of CSF had available MRI, 14 were excluded due to suboptimal Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 4 of 10 image quality, which included subtle movement artifacts, [35]. All resulting images were visually inspected to poor signal to noise ratio, and gradient artifacts. The check for possible registration errors. breakdown of diagnosis in these individuals was as follows: aDS n = 38, pDS n = 15, dDS n = 8. Statistical analysis Statistical analyses were performed in R version 3.4.3 [36]. FDG-PET acquisition and analysis Group differences were compared using One-way Analysis From the remaining 61 adults with DS with MRI, 42 of Variance (ANOVA) with post hoc Tukey tests adjusted underwent [18F]-fluorodeoxyglucose (FDG) positron for multiple testing using the Bonferonni-Hochberg emission tomography (PET) scan on an integrated method. Robust logistic regressions were performed using PET-TC system (Philips Gemini TF). The breakdown the lm function included in the “robustbase” package that of diagnosis in these individuals was as follows: aDS uses MM-type regression estimators to compute robust- n = 28, pDS n = 10, dDS n = 4. Overnight fasting ness weights in the fitting process. All reported r2 values subjects’ plasma glucose was measured and confirmed were adjusted for the number of predictors in the model. to be < 180 mg/dL prior to starting the scanning Where residuals deviated from a Gaussian distribution procedure. Data were acquired 50 min after the injec- (Shapiro-Wilk p < 0.05), square root transformed values tion of 259 MBq/ml (7 mCi) of [18F]-FDG. After were used. Outliers were excluded using the 3 × IQR rule. computerized tomography data was obtained, brain For the neuroimaging analyses we performed Monte- PET was acquired. The reconstruction method was Carlo simulation with 10,000 repeats (family wise error iterative (LOR RAMBLA, 3 iterations and 33 subsets) [FWE] correction at p < 0.05). Due to the small size of with a 128 × 128 image size, 2 mm pixel size and 2 mm each clinical group with available neuroimaging, analyses pixel slice thickness. [18F]-FDG-PET images were were performed on the complete DS dataset only. The intensity-scaled by the reference pons-vermis region figures show only those results that survived FWE correc- [31], spatially normalized to the Freesurfer anatomical tion. Alpha was set at 0.05 for global analyses and 0.01 for space using a PET surfer approach [32] and then post-hoc analyses. smoothed with a 10 mm FWHM kernel in order to ob- tain similar relative smoothing FDG maps as compared Results with cortical thickness maps [33, 34]. With this The Table 1 shows the demographic and clinical data surface-based approach, the reliability of PET effects for the samples included in the study. The mean age was improves and the inter-individual variance is reduced 52 years (standard deviation 15) across the whole study Table 1 Demographics and clinical data for study participants Group Control sAD aDS pDS dDS N 34 40 49 18 27 Mean age-at-collection (years) 51 70 37 52 52 SD (range) 13 (24–74) 7 (52–83) 10 (22–57) 4 (45–60) 5 (42–62) Sex (% Female) 68% 65% 47% 44% 42% Intellectual Disabilitya None None 80%;20% 78%;22% 67%;33% FCSRT Immediate Total Recall 34 13 35 19 15 SD (range), n 3 (34–48), 40 10 (0–36), 30 2 (30–36), 33 11 (0–36), 11 7 (0–32), 13 Semantic Verbal Fluency Test 22 9 10 8 7 SD (range), n 5 (14–33), 34 4 (3–15), 30 3 (3–18), 37 2 (4–13), 12 4 (1–12), 15 CAMCOG N/A N/A 79 67 54 SD (range), n N/A N/A 12 (51–96), 36 12 (41–80), 10 22 (2–87), 11 Mean CSF Aβ1–42 (pg/ml) 1359 492 1089 489 504 SD (range) 284 (1000–2034) 107 (275–690) 474 (302–2335) 146 (235–865) 160 (235–765) Mean CSF t-tau (pg/ml) 249 781 342 972 1095 SD (range) 63 (157–407) 423 (229–2110) 248 (86–1394) 658 (118–2565) 653 (212–3150) Mean CSF p-tau (pg/ml) 37 137 44 173 173 SD (range) 9 (22–55) 62 (55–351) 47 (10–277) 128 (22–564) 92 (31–354) a Proportion of cases with mild/moderate vs severe/profound intellectual disability Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 5 of 10 (n = 168). Compared to controls the mean age was higher in the sporadic AD group (+ 19 years, 95% CI 13 to 25, adj.p < 0.0001), lower in the aDS group (− 14 years, 95% CI 8 to 19; adj.p < 0.0001) and comparable in pDS and dDS (both adj.p > 0.9). As would be expected for an age-related disease such as AD, mean age was lower in aDS than pDS (− 14 years, 95% CI − 7 to − 21; adj.p < 0.0001) and dDS (− 15 years, 95% CI − 9 to − 21; adj.p < 0.0001). The proportion of males:females included in the study was 46:54 and was comparable across diagnostic groups (p = 0.09). The proportion of individuals with severe/profound intellectual disability was comparable across AD stages (p = 0.7). All cogni- tive measures were lower in sAD compared to con- trols (all adj.p < 0.0001) and in dDS compared to aDS (all adj.p < 0.0001). CSF biomarker levels were as pre- viously reported [17]. Briefly, compared to controls mean Aβ1–42 levels were lower in all groups (all adj.p < 0.0001) and p-tau and t-tau levels were higher in sAD, pDS and dDS (all adj.p < 0.0001). CSF NPTX2 levels are similarly reduced in DS and sporadic AD Mean CSF NPTX2 levels were associated with clinical diag- nosis (Fig. 1a; n = 168, p < 0.0001). Specifically, post hoc analyses showed that compared to controls, mean NPTX2 levels were 0.5-fold in sporadic AD (− 491 pg/ml, 95% CI −280 to − 702; adj.p < 0.0001), 0.6-fold in aDS (− 400 pg/ml, 95% CI − 196 to − 604; adj.p < 0.0001), 0.5-fold in pDS (− 472, 95% CI − 208 to − 736; adj.p < 0.0001) and 0.3-fold in dDS (− 595, 95% CI − 361 to − 828; adj.p < 0.0001). Mean NPTX2 levels were comparable between sporadic AD and DS (all AD stages adj.p > 0.3) and across AD stages in DS (all adj.p > 0.07). In adults with DS, CSF NPTX2 levels were not associ- Fig. 1 NPTX2 levels in control, sporadic AD and DS CSF. Box and ated with level of intellectual disability (Fig. 1b; n = 94, whisker plots and individual points of the CSF levels of NPTX2 (pg/ p = 0.7) and were not associated with any measure of ml) are grouped (a) by clinical diagnosis and (b) by degree of cognitive performance in DS (n = 57–64, all p > 0.07). intellectual disability (DS population only). P-values resulting from Mean NPTX2 levels were not associated with age overall one-way ANOVA on square root transformed NPTX2 levels are (p = 0.6) or at any AD stage (all p > 0.1). shown in the top right corner. P-values resulting from Bonferroni- adjusted post-hoc Tukey test are shown where p < 0.05 Low CSF NPTX2 levels correlate with low CSF levels of a PV-interneuron specific receptor We next sought to determine whether the PV- Low CSF NPTX2 levels correlate with CSF markers of AD interneuron specific marker, GluA4, shows a coordi- pathology in adults with DS nated reduction with NPTX2 in the CSF. While CSF We next sought to determine the relationship between CSF GluA4 levels were not associated with clinical diagnosis NPTX2 levels and core CSF AD biomarkers in DS. Low (Fig. 2a; p = 0.2), low NPTX2 levels correlated with low NPTX2 levels were associated with low CSF Aβ1–42 levels GluA4 levels in controls (n = 27 r2 = 0.2, p = 0.003), sAD (Fig. 3a) in all diagnostic groups; aDS (r2 = 0.5, p < 0.0001), (n = 20, r2 = 0.4, p < 0.0001) and in DS (n = 81, r2 = 0.4, pDS (r2 = 0.3, p = 0.0003) and dDS (r2 = 0.6, p = 0.006). Low p < 0.0001): The association in DS was apparent at all NPTX2 levels were also associated with p-tau (Fig. 3b) in AD stages; aDS (n = 36, r2 = 0.5, p = 0.04), pDS (n = 18, all diagnostic groups; aDS (r2 = 0.4, p < 0.0001), pDS (r2 = r2 = 0.5, p = 0.005) and dDS (n = 21, r2 = 0.4, p = 0.02). 0.3, p = 0.0004) and dDS (r2 = 0.3, p = 0.001). Similar Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 6 of 10 Fig. 3 Correlation between CSF NPTX2 levels and CSF markers of AD pathology. CSF NPTX2 levels are plotted against (a) CSF Aβ1–42 and Fig. 2 GluA4 levels in control, sporadic AD and DS CSF. a Box and (b) CSF p-tau with associated linear regression lines shown for each whisker plots and individual points of the CSF levels of GluA4 (log2 clinical diagnosis (see legend) intensity) grouped by clinical diagnosis. P-value resulting from one- way ANOVA on square root transformed GluA4 levels is shown in the top right corner. b CSF NPTX2 levels are plotted against CSF widespread regions of the temporal and parietal cortices GluA4 levels with associated linear regression lines shown for controls, sporadic AD and DS groups (see legend) (bottom; n = 44, FWE < 0.05). Discussion correlations were observed for CSF t-tau (r2 = 0.3 to 0.4, Neuronal Pentraxin Receptor 2 (NPTX2) has recently p < 0.003). been identified as a novel CSF biomarker of inhibitory circuit dysfunction in sporadic AD [12, 15]. Here we Low CSF NPTX2 levels in adults with DS are associated report that CSF levels of NPTX2 are also reduced in with cortical atrophy and neuronal dysfunction in CSF of adults with DS who are highly susceptible to signature AD brain regions developing AD due to triplication at chromosome 21. We also analyzed the relationship between CSF NPTX2 The reduction in CSF NPTX2 compared to non- levels and structural and functional brain changes in trisomic cognitively normal controls was observed in DS adults with DS. As shown in Fig. 4, low CSF NPTX2 across all AD stages and was similar to the reduction ob- levels were associated with reduced cortical thickness in served in sporadic AD patientsboth here and in previous the temporal and parietal cortices (top; n = 63, FWE < studies [12, 15]. This finding is consistent with studies 0.05) and with reduced glucose metabolism in more that report a similar reduction in NPTX2 expression in Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 7 of 10 Fig. 4 Structural and functional neuroimaging correlates of CSF NPTX2 levels in DS. Correlation of CSF NPTX2 levels and cortical thickness (top) and glucose metabolism (bottom) in all DS participants. Regions in red represent significant correlations (p < 0.05). Only the clusters that survived family-wise error correction p < 0.05 are shown. Scatterplots with linear regression line (black) and standard error (gray) show the correlation between CSF NPTX2 and neuroimaging biomarkers by diagnosis in a representative cluster that survived multiple comparisons, indicated with an asterisk. Regression lines are shown for each clinical group for informative purposes (see legend) postmortem sporadic AD and DS brains [12] and indi- The lack of association of NPTX2 levels with intellec- cates that the mechanism underlying these changes is tual disability suggests that the reduction in NPTX2 common to both sporadic and genetic forms of AD and levels is not neurodevelopmental. This is supported by that these changes may be reflected in the CSF. As these the finding that reduced CSF NPTX2 levels were associ- changes were apparent across all AD stages, CSF NPTX2 ated with increased cortical atrophy and reduced glucose may be a useful marker of AD-related changes in adults metabolism in temporal and parietal regions of the with DS even prior to symptom onset. brain, regions that are particularly susceptible to AD DS cohorts with available CSF are scarce and current pathology. Thus, CSF NPTX2 levels inversely correlated biofluid markers for AD in the DS population are with the extent of both neuronal loss and neuronal restricted to surrogate markers of amyloidosis (CSF dysfunction in signature AD regions. Taken together, we Aβ1–42 or amyloid PET tracers) and tau-mediated propose that changes in CSF NPTX2 levels in adults neurodegeneration (CSF p-tau and t-tau) combined with with DS are related to AD pathophysiology and not to neuropsychological assessment, which can be con- intellectual disability. founded by substantial inter-individual variation in intel- Low CSF NPTX2 levels in adults with DS were associated lectual disability. NPTX2 represents a new addition to with low CSF GluA4 and low CSF Aβ1–42, a surrogate CSF the DS biomarker arsenal that may be used to detect marker of increased aggregation of brain Aβ1–42. This is the early signs of AD in this relatively under-studied popula- first evidence that a synergized increase in brain Aβ1–42 ag- tion. Furthermore, this work opens the door to future gregates and reduced NPTX2 can be measured indirectly in studies exploring the potential of CSF NPTX2 as a prog- the CSF of adults with DS. NPTX2 is expressed by pyram- nostic marker for late-onset myoclonic epilepsy in Down idal neurons and secreted by axon terminals to mediate syndrome and sporadic AD. activity-dependent strengthening of pyramidal neuron Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 8 of 10 excitatory synapses on GABA-ergic PV interneurons, a disability in the DS population can complicate cognitive mechanism that is dependent upon the AMPA receptor assessment and therefore these neuropsychological tests subunit, GluA4 [11]. Histological studies have shown re- could only be performed in individuals with mild or mod- duced NPTX2 and GluA4 expression in signature AD re- erate intellectual disability, which resulted in a limited gions of AD and DS brains [12], regions that also show a sample size. Thus, the potential relationship between CSF reduced number of PV interneurons [37]. This is consistent NPTX2 levels and cognitive performance in the DS popu- with the accumulating evidence for reduced inhibitory lation warrants further study in a larger population. On interneuronal activity in the AD brain [10]. Studies in mice the other hand, there is evidence that the temporal onset [37–39] have shown that low NPTX2 expression at excita- of cognitive decline may differ in DS compared to auto- tory synapses of PV interneurons combined with brain somal dominant AD [4], thus underscoring the import- amyloidosis is critical for inhibitory circuit dysfunction ance of considering the neurodevelopmental differences in [12].Thus, a hypothesis emerges whereby a loss of NPTX2- the DS population when interpreting the cognitive results. GluA4 synapses in regions affected by amyloidosis may lead The association of increased CSF NPTX2 levels with in- to inhibitory circuit dysfunction in AD. While extrapolation creased CSF p-tau levels in the DS population could indi- of data from mouse to human brain should be treated with cate that widespread neuronal loss at later disease stages caution, this could be one plausible explanation for the low may lead to increased NPTX2 clearance into the CSF and CSF NPTX2 levels we report in sporadic AD patients and may mask a reduction in NPTX2 synapses. Nevertheless, adults with DS. the inverse relationship we report between CSF NPTX2 Unlike NPTX2, GluA4 levels were not significantly levels and a neuroimaging marker of cortical atrophy, sug- reduced in CSF of adults with DS compared to controls, gests that neurodegeneration has a discernable but limited making NPTX2 a potentially more valuable CSF bio- confounding effect on CSF NPTX2 levels. marker in DS. Based on these findings, CSF NPTX2 can A limitation of this study is the cross-sectional design; be added to the growing list of biomarkers that show longitudinal studies are needed to fully establish the prog- changes similar, both in direction and magnitude, to nostic value of NPTX2 in both sporadic AD and in DS. those described in sporadic or autosomal dominant AD Furthermore, these findings warrant replication in an in- [4, 20, 40]. The potential of CSF NPTX2 as a novel dependent DS cohort. Exploration of the potential of biomarker of interneuronal dysfunction can now be NPTX2 as a blood biomarker would also be valuable to extended to DS, a genetic form of AD that targets the minimize the use of lumbar puncture in people with DS. same cortical regions as the sporadic and autosomal dominant forms [5]. Conclusions CSF Aβ1–42, t-tau and p-tau are excellent diagnostic Here we report that CSF NPTX2 levels are reduced in markers in sporadic AD and in the DS population and we adults with DS compared to cognitively normal, non- do not expect NPTX2 to improve upon those markers. trisomic controls even prior to AD onset. Similar reduc- Rather, this study supports the increasing literature that tions were observed in the CSF of sporadic AD patients suggests CSF NPTX2 as a surrogate marker of underlying and have been previously reported in postmortem AD and interneuronal circuit dysfunction in sporadic and, now DS brains [12, 15]. Substantial evidence from animal studies genetic, forms of AD even at the preclinical stage. Further suggests that such changes may be associated with reduced studies in longitudinal cohorts are needed to explore the inhibitory inter-neuronal activity. While further functional full potential of CSF NPTX2. The capacity of NPTX2 as studies are needed, this could explain the correlation we re- an early prognostic marker would be one potential use port between low CSF NPTX2 and low CSF GluA4, the worth exploring for example. Alternatively, as tau and Aβ predominant AMPA receptor subunit in PV interneurons. are common drug targets in AD clinical trials, a surrogate In conclusion, this study shows for the first time that measure of synaptic dysfunction that does not directly NPTX2 could be a valuable CSF biomarker of AD-related measure the drug target levels would be a useful addition inhibitory neuronal circuit dysfunction in adults with DS to the biomarker arsenal. even prior to symptom onset. In previous studies in sporadic AD, CSF NPTX2 levels correlated better with cognitive performance than the core Abbreviations AD biomarkers [12]. Furthermore, in genetic frontotem- AD: Alzheimer’s disease; aDS: Asymptomatic AD in the context of DS; poral dementia, low CSF NPTX2 levels predicted subse- CAMCOG: Cambridge Cognition Examination; CSF: Cerebrospinal fluid; DABNI: Down Alzheimer Barcelona Neuroimaging Initiative; dDS: AD quent decline in phonemic verbal fluency and Clinical dementia in the context of DS; DS: Down syndrome; FDG: [18F]- Dementia Rating scale [41]. In contrast, here we report fluorodeoxyglucose; FWE: Family wise error; FWHM: Full width at half that reduced CSF NPTX2 levels were not associated with maximum; GABA: Gamma-Aminobutyric acid; MRI: Magnetic resonance imaging; NPTX2: Neuronal pentraxin-2; pDS: Prodromal AD in the context of any measure of cognitive performance in adults with DS. DS; PET: Positron emission tompgraphy; PV: Parvalbumin; SPIN: Sant Pau That being said, the high variability in intellectual Initiative for Neurodegeneration; SRM: Selected Reaction Monitoring Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 9 of 10 Acknowledgements diagnostic criteria for Alzheimer’s disease: the IWG-2 criteria. Lancet Neurol. We thank the participants and their families for their generosity. We also 2014;13:614–29. acknowledge the Fundació Catalana Síndrome de Down for global support; 2. Wiseman FK, Pulford LJ, Barkus C, Liao F, Portelius E, Webb R, Chavez- Soraya Torres, and Laia Muñoz for laboratory and sample handling; Sílvia Gutierrez L, Cleverley K, Noy S, Sheppard O, et al. Trisomy of human Valldeneu for administrative support; and Lidia García Losada for nursing chromosome 21 enhances amyloid-beta deposition independently of an management. extra copy of APP. Brain. 2018;141:2457–74. 3. Bateman RJ, Xiong C, Benzinger TL, Fagan AM, Goate A, Fox NC, Marcus DS, Authors’ contributions Cairns NJ, Xie X, Blazey TM, et al. Clinical and biomarker changes in OB, MX, DX, MC-I, JP, BB, SF, LV, IB, RNL, VM, EV, JC, DA and RB contributed dominantly inherited Alzheimer’s disease. N Engl J Med. 2012;367:795–804. to the acquisition and analysis of data. OB, MX, MC-I, JP, BB, JC, AL, PW and 4. Fortea J, Vilaplana E, Carmona-Iragui M, Benejam B, Videla L, Barroeta I, JF drafted a significant portion of the manuscript and figures. All authors Fernandez S, Altuna M, Pegueroles J, Montal V, et al. Clinical and biomarker read and approved the final manuscript. changes of Alzheimer’s disease in adults with Down syndrome: a cross- sectional study. Lancet. 2020;395:1988–97. Funding 5. Neale N, Padilla C, Fonseca LM, Holland T, Zaman S. Neuroimaging and This work was supported by the Fundació Catalana Síndrome de Down and other modalities to assess Alzheimer’s disease in Down syndrome. Fundació Víctor Grífols i Lucas and research grants from Institute of Health Neuroimage Clin. 2018;17:263–71. Carlos III (ISCIII), Spain (PI18/00327 to OB, PI17/01019 to JF; PI16/01825 to RB; 6. Schweber MS. Alzheimer’s disease and Down syndrome. Prog Clin Biol Res. PI17/01895 to AL and PI18/0335 to MC-I and CIBERNED Program 1, Alzheimer 1989;317:247–67. Disease to AL and SIGNAL study, www.signal.es) and the National Institutes 7. McCarron M, McCallion P, Reilly E, Mulryan N. A prospective 14-year of Health, US (NIA grants R35NS097966 to PFW, R01AG056850-01A1, longitudinal follow-up of dementia in persons with Down syndrome. J R21AG056974 and R01AG061566 to JF). This work has also been supported Intellect Disabil Res. 2014;58:61–70. by the Fundació La Marató de TV3 (20141210 to JF and 044412 to RB), the 8. Sinai A, Mokrysz C, Bernal J, Bohnen I, Bonell S, Courtenay K, Dodd K, Fundació Bancaria Obra Social La Caixa (DABNI project to RB), Fundación Gazizova D, Hassiotis A, Hillier R, et al. Predictors of age of diagnosis and BBVA (to AL) and by the Department of Health Generalitat de Catalunya the survival of Alzheimer’s disease in Down syndrome. J Alzheimers Dis. 2018; Department of Health, Generalitat de Catalunya PERIS and SGR programs 61:717–28. (SLT006/17/125 to DA, SLT002/16/00099 to RNL, SLT006/17/00119 to JF 2017 9. Hithersay R, Startin CM, Hamburg S, Mok KY, Hardy J, Fisher EMC, and SGR 2017 00547 to the “Grup de Recerca en Demències: Sant Pau".) and Tybulewicz VLJ, Nizetic D, Strydom A. Association of Dementia with partly funded by Fondo Europeo de Desarrollo Regional (FEDER), Unión Eur- Mortality among Adults with Down syndrome Older than 35 years. JAMA opea, “Una manera de hacer Europa”. OB is contractually supported by the Neurol. 2018;1;76(2):152–160. Miguel Servet Type II fellowship (CP18/00011; ISCIII, FEDER). The funders of 10. Palop JJ, Mucke L. Network abnormalities and interneuron dysfunction in this study played no role in the design of the study, collection, analysis or in- Alzheimer disease. Nat Rev Neurosci. 2016;17:777–92. terpretation of data or writing the manuscript. 11. Chang MC, Park JM, Pelkey KA, Grabenstatter HL, Xu D, Linden DJ, Sutula TP, McBain CJ, Worley PF. Narp regulates homeostatic scaling of excitatory Availability of data and materials synapses on parvalbumin-expressing interneurons. Nat Neurosci. 2010;13: The data analyzed during the current study are available from the 1090–7. corresponding author on reasonable request. 12. Xiao MF, Xu D, Craig MT, Pelkey KA, Chien CC, Shi Y, Zhang J, Resnick S, Pletnikova O, Salmon D, et al. NPTX2 and cognitive dysfunction in Ethics approval and consent to participate Alzheimer’s disease. Elife. 2017;6:e23798. The study was approved by the Sant Pau Ethics Committee following the 13. Fuchs EC, Zivkovic AR, Cunningham MO, Middleton S, Lebeau FE, standards for medical research in humans recommended by the Declaration Bannerman DM, Rozov A, Whittington MA, Traub RD, Rawlins JN, Monyer H. of Helsinki and reported to the Minister of Justice according to the Spanish Recruitment of parvalbumin-positive interneurons determines hippocampal law for research in people with intellectual disabilities. All participants gave function and associated behavior. Neuron. 2007;53:591–604. signed, informed consent for the interventions described, inclusion of 14. Pelkey KA, Barksdale E, Craig MT, Yuan X, Sukumaran M, Vargish GA, Mitchell samples, neuroimaging and demographic data. RM, Wyeth MS, Petralia RS, Chittajallu R, et al. Pentraxins coordinate excitatory synapse maturation and circuit integration of parvalbumin Consent for publication interneurons. Neuron. 2015;85:1257–72. Not applicable. 15. Galasko DR, Xiao M, Xu D, Smirnov D, Salmon DP, Dewit N, Vanbrabant J, Jacobs D, Vanderstichele H, Vanmechelen E, Worley P. Synaptic biomarkers Competing interests in CSF aid in diagnosis, correlate with cognition and predict progression in OB, AL, JF and DA declare a filed patent application (pending) to the MCI and Alzheimer’s disease. Alzheimer’s Dement. 2019;5:871–82. European Patent Office for Markers of synaptopathy neurodegenerative 16. Alcolea D, Clarimon J, Carmona-Iragui M, Illan-Gala I, Morenas-Rodriguez E, disease (does not relate to data included in this manuscript). MX, DX and PW Barroeta I, Ribosa-Nogue R, Sala I, Sanchez-Saudinos MB, Videla L, et al. The are co-founders of CogNext. Sant Pau initiative on Neurodegeneration (SPIN) cohort: a data set for biomarker discovery and validation in neurodegenerative disorders. Author details Alzheimers Dement (N Y). 2019;5:597–609. 1 Memory Unit and Biomedical Research Institute Sant Pau (IIB Sant Pau), 17. Alcolea D, Pegueroles J, Muñoz L, Camacho V, López-Mora D, Fernández- Neurology Department, Hospital de la Santa Creu i Sant Pau, 08025 León A, Bastard NL, Huyck E, Nadal A, Olmedo V, et al. Agreement between Barcelona, Spain. 2Centro de Investigación Biomédica en Red sobre 18F-Florbetapir PET imaging and cerebrospinal fluid Abeta1–42, Abeta1–40, Enfermedades Neurodegenerativas (CIBERNED), 28031 Madrid, Spain. tTau and pTau measured on the LUMIPULSE G fully automated platform. 3 Solomon H. Snyder Department of Neuroscience, Johns Hopkins University bioRxiv. 2018. https://doi.org/10.1101/476937. School of Medicine, Baltimore, MD 21205, USA. 4Barcelona Down Medical 18. McKhann GM, Knopman DS, Chertkow H, Hyman BT, Jack CR Jr, Kawas CH, Center, Fundació Catalana Síndrome de Down, Barcelona, Spain. Klunk WE, Koroshetz WJ, Manly JJ, Mayeux R, et al. The diagnosis of 5 Department of Neurology, Johns Hopkins University School of Medicine, dementia due to Alzheimer’s disease: recommendations from the National Baltimore, MD 21205, USA. Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement. 2011;7:263–9. Received: 5 March 2020 Accepted: 10 August 2020 19. Fortea J, Carmona-Iragui M, Benejam B, Fernandez S, Videla L, Barroeta I, Alcolea D, Pegueroles J, Munoz L, Belbin O, et al. Plasma and CSF biomarkers for the diagnosis of Alzheimer’s disease in adults with Down References syndrome: a cross-sectional study. Lancet Neurol. 2018;17(10):860–9. 1. Dubois B, Feldman HH, Jacova C, Hampel H, Molinuevo JL, Blennow K, 20. Carmona-Iragui M, Balasa M, Benejam B, Alcolea D, Fernandez S, Videla L, DeKosky ST, Gauthier S, Selkoe D, Bateman R, et al. Advancing research Sala I, Sanchez-Saudinos MB, Morenas-Rodriguez E, Ribosa-Nogue R, et al. Belbin et al. Molecular Neurodegeneration (2020) 15:46 Page 10 of 10 Cerebral amyloid angiopathy in Down syndrome and sporadic and receptor 5 is a coreceptor for Alzheimer abeta oligomer bound to cellular autosomal-dominant Alzheimer’s disease. Alzheimers Dement. 2017;13: prion protein. Neuron. 2013;79:887–902. 1251–60. 40. Wisniewski KE, Wisniewski HM, Wen GY. Occurrence of neuropathological 21. Carmona-Iragui M, Santos T, Videla S, Fernandez S, Benejam B, Videla L, changes and dementia of Alzheimer’s disease in Down’s syndrome. Ann Alcolea D, Blennow K, Blesa R, Lleo A, Fortea J. Feasibility of lumbar Neurol. 1985;17:278–82. puncture in the study of cerebrospinal fluid biomarkers for Alzheimer’s 41. van der Ende EL, Xiao M, Xu D, Poos JM, Panman JL, Jiskoot LC, Meeter LH, disease in subjects with Down syndrome. J Alzheimers Dis. 2017;55:1489–96. Dopper EG, Papma JM, Heller C, et al. Neuronal pentraxin 2: a synapse- 22. Kaufman AS, Kaufman NL. Manual for the Kaufman Brief Intelligence Test. derived CSF biomarker in genetic frontotemporal dementia. J Neurol Circle Pines: American Guidance Service; 1990. Neurosurg Psychiatry. 2020;91:612–21. 23. Esteba-Castillo S, Dalmau-Bueno A, Ribas-Vidal N, Vila-Alsina M, Novell-Alsina R, Garcia-Alba J. Adaptation and validation of CAMDEX-DS (Cambridge examination for mental disorders of older people with Down’s syndrome Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in and others with intellectual disabilities) in Spanish population with published maps and institutional affiliations. intellectual disabilities. Rev Neurol. 2013;57:337–46. 24. Sala I, Illan-Gala I, Alcolea D, Sanchez-Saudinos MB, Salgado SA, Morenas- Rodriguez E, Subirana A, Videla L, Clarimon J, Carmona-Iragui M, et al. Diagnostic and prognostic value of the combination of two measures of verbal memory in mild cognitive impairment due to Alzheimer’s disease. J Alzheimers Dis. 2017;58:909–18. 25. Peiffer DA, Le JM, Steemers FJ, Chang W, Jenniges T, Garcia F, Haden K, Li J, Shaw CA, Belmont J, et al. High-resolution genomic profiling of chromosomal aberrations using Infinium whole-genome genotyping. Genome Res. 2006;16:1136–48. 26. Teunissen CE, Tumani H, Bennett JL, Berven FS, Brundin L, Comabella M, Franciotta D, Federiksen JL, Fleming JO, Furlan R, et al. Consensus guidelines for CSF and blood biobanking for CNS biomarker studies. Mult Scler Int. 2011;2011:246412. 27. Alcolea D, Martinez-Lage P, Sanchez-Juan P, Olazaran J, Antunez C, Izagirre A, Ecay-Torres M, Estanga A, Clerigue M, Guisasola MC, et al. Amyloid precursor protein metabolism and inflammation markers in preclinical Alzheimer disease. Neurology. 2015;85:626–33. 28. Lleo A, Nunez-Llaves R, Alcolea D, Chiva C, Balateu-Panos D, Colom-Cadena M, Gomez-Giro G, Munoz L, Querol-Vilaseca M, Pegueroles J, et al. Changes in synaptic proteins precede Neurodegeneration markers in preclinical Alzheimer’s disease cerebrospinal fluid. Mol Cell Proteomics. 2019;18:546–60. 29. Choi M, Chang CY, Clough T, Broudy D, Killeen T, MacLean B, Vitek O. MSstats: an R package for statistical analysis of quantitative mass spectrometry-based proteomic experiments. Bioinformatics. 2014;30:2524–6. 30. Fischl B, Dale AM. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci U S A. 2000;97:11050–5. 31. Landau SM, Harvey D, Madison CM, Koeppe RA, Reiman EM, Foster NL, Weiner MW, Jagust WJ. Associations between cognitive, functional, and FDG-PET measures of decline in AD and MCI. Neurobiol Aging. 2011;32: 1207–18. 32. Greve DN, Salat DH, Bowen SL, Izquierdo-Garcia D, Schultz AP, Catana C, Becker JA, Svarer C, Knudsen GM, Sperling RA, Johnson KA. Different partial volume correction methods lead to different conclusions: an (18)F-FDG-PET study of aging. Neuroimage. 2016;132:334–43. 33. La Joie R, Perrotin A, Barre L, Hommet C, Mezenge F, Ibazizene M, Camus V, Abbas A, Landeau B, Guilloteau D, et al. Region-specific hierarchy between atrophy, hypometabolism, and beta-amyloid (Abeta) load in Alzheimer’s disease dementia. J Neurosci. 2012;32:16265–73. 34. Bejanin A, Viard A, Chetelat G, Clarys D, Bernard F, Pelerin A, de La Sayette V, Eustache F, Desgranges B. When higher activations reflect lower deactivations: a PET study in Alzheimer’s disease during encoding and retrieval in episodic memory. Front Hum Neurosci. 2012;6:107. 35. Greve DN, Svarer C, Fisher PM, Feng L, Hansen AE, Baare W, Rosen B, Fischl B, Knudsen GM. Cortical surface-based analysis reduces bias and variance in kinetic modeling of brain PET data. Neuroimage. 2014;92:225–36. 36. R-Core-Team. R: a language and environment for statistical computing. In Book R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2018. 37. Brady DR, Mufson EJ. Parvalbumin-immunoreactive neurons in the hippocampal formation of Alzheimer’s diseased brain. Neuroscience. 1997; 80:1113–25. 38. Kamenetz F, Tomita T, Hsieh H, Seabrook G, Borchelt D, Iwatsubo T, Sisodia S, Malinow R. APP processing and synaptic function. Neuron. 2003;37:925–37. 39. Um JW, Kaufman AC, Kostylev M, Heiss JK, Stagi M, Takahashi H, Kerrisk ME, Vortmeyer A, Wisniewski T, Koleske AJ, et al. Metabotropic glutamate