CSF Samples Research Articles

Longitudinal AD biomarkers in monozygotic twins: genetic contribution to AD biomarker concentrations over time

AbstractBackgroundUnderstanding the earliest pathophysiological changes in Alzheimer’s disease (AD) is important for development of therapeutic intervention strategies. Studying genetically identical twins provides unique information on genetic and environmental influences on development of AD. Here we studied change in AD biomarkers amyloid β1‐42 (aβ42), amyloid β1‐40 (aβ40), phosphorylated tau181 (p‐tau) and total tau (t‐tau) in CSF over time in a cohort of older cognitively normal monozygotic twins and tested the genetic contribution to biomarker concentrations at baseline and to change of biomarker concentrations over time.MethodWe included 103 twins (npairs = 42) from the EMIF‐AD preclinAD study with 2‐4 repeated CSF sampling available across a time period of 4.47±2 years (mean±SD age: 68.8±7 years, 54 (52%) female). We measured aβ42, aβ40, p‐tau and t‐tau using the Lumipulse platform (Fujirebio Diagnostics, Inc.). We used linear mixed models to investigate change in biomarker concentrations over time. We used twin‐pair intraclass correlation (ICC) analysis to estimate the genetic contribution to biomarker concentrations at baseline and at repeated measurements, and to annual change in biomarker concentrations.ResultAt baseline, 25 (24%) individuals had an abnormal aβ42/aβ40 ratio, 27 (26%) abnormal p‐tau and 35 (34%) abnormal t‐tau, and at last measurement respectively 34 (33%), 38 (37%) and 45 (44%). Across the group, aβ42/aβ40 decreased (β±SE ‐0.001±0, p‐value<0.001), and p‐tau and t‐tau increased over time (respectively 1.37±0.3, p‐value<0.001; 6.58±2.3, p‐value = 0.004). Increases in p‐tau levels were most pronounced in amyloid positive individuals (2.72±0.38, p‐value<0.001) compared to amyloid negative individuals (0.65±0.24, p‐value = 0.007; p‐value interaction<0.001). At baseline, the genetic contribution to biomarker concentrations was lowest for t‐tau (ICC = 0.5, p<0.001) and highest for aβ40 (ICC = 0.79, p<0.001). ICC’s remained similar at repeated measurements (table 1). Annual changes of aβ40 and p‐tau showed significant twin‐pair correlations (resp. ICC’s: 0.58 and 0.42; p‐values <0.01; Figure 1), indicating a genetic contribution to change in these concentrations over time.ConclusionIn this sample of older cognitively normal monozygotic twins, we found that AD biomarkers became increasingly abnormal over time. Our unique twin design further indicates that these very early pathophysiological changes are under moderate to high genetic influence, which also indicates a role of environmental factors.

Longitudinal white matter integrity alterations as a function of Preclinical Alzheimer’s Pathology: findings from the EPAD non‐demented cohort

AbstractBackgroundCross‐sectional diffusion tensor imaging (DTI) studies report white matter (WM) microstructural alterations, mainly involving fornix and corpus callosum (CC), in symptomatic stages of Alzheimer’s Disease (AD). Exploring these changes longitudinally in relation to preclinical and prodromal AD pathology is essential to track disease progression. We investigated trajectories of WM alterations in non‐demented individuals from the European Prevention of Alzheimer’s Dementia (EPAD) cohort and their relationship with AD biomarkers.MethodEPAD inclusion criteria were age>50 years and Clinical Dementia Rating = 0.5. We selected participants with baseline CSF samples, and longitudinal DTI scans. After defining AT groups(Ingala et al. 2021), A‐T+ participants were excluded (final N = 283). MRI acquisition and preprocessing was previously described(Lorenzini et al. 2022). We extracted fractional anisotropy (FA) in 14 regions(Mori et al. 2005) from DTI as a marker of WM integrity. Site harmonization of MRI‐derived data was performed with ComBat toolbox(Fortin et al. 2018). Age‐ and sex‐corrected linear mixed models with random intercepts on participants were used to estimate time‐effect on MRI‐derived phenotypes, and their interaction with AT status. Results were corrected for multiple testing.ResultAt baseline, A+T‐ participants showed lower FA in body/columns (ß = ‐0.04, p<0.01) and crus of fornix (ß = ‐0.01, p = 0.04), and in splenium (ß = ‐0.007, p = 0.04) and body of CC (ß = ‐0.01, p = 0.03) compared to A‐T‐. In the whole group, FA decreased over time in splenium (ß = ‐0.003, p<0.01), body (ß = ‐0.002, p<0.01) and genu of CC (ß = ‐0.009, p<0.01), body/columns (ß = ‐0.004, p<0.01) and crus of fornix (ß = ‐0.005, p<0.01), and inferior longitudinal fasciculus (ß = ‐0.007, p<0.01). Compared to A‐T‐, A+T+ participants showed steeper FA reduction in superior (Interaction:ß = ‐0.004, p = 0.03) and anterior corona radiata (Interaction:ß = ‐0.006, p = 0.02).ConclusionOur results suggest that WM integrity in key AD regions may degenerate early in the AD continuum, showing greater impairment over short‐time intervals in individuals with higher AD‐specific pathological burden.

Tissue inhibitor of matrix metalloproteinases 4: a novel marker associated with CAA

AbstractBackgroundWe have previously shown that matrix metalloproteinase 9 (MMP9) and tissue inhibitor of proteinases 3 (TIMP3) are specifically expressed in the vasculature of patients with cerebral amyloid angiopathy (CAA). Moreover, we have demonstrated increased MMP9 and decreased TIMP3 levels, as well as an altered MMP9:TIMP3 ratio in the vasculature of CAA patients that suffered from intracerebral haemorrhages (CAA‐ICH) compared to CAA patients without ICH (CAA‐nonICH). We now extended our studies to investigate whether TIMP4 may have a similar role in CAA, as this has not been studied before.MethodUsing immunohistochemistry, we studied occipital lobe tissue of 44 controls (mean age 78.4 years; 50% female) and 58 CAA patients (77.5 years; 55% female; 40 CAA‐nonICH and 18 CAA‐ICH). Vascular TIMP4 staining was scored in a semiquantitative manner. Using ELISA, we measured TIMP4 CSF levels of 33 CAA patients (mean age 72.2 years; 47% female) and 35 controls (mean age 71.2 years; 51% female). For a subset of CSF samples (11 CAA samples and 14 controls), the levels of MMP2, MMP9, and MMP14 have been determined previously, and now their ratios to TIMP4 were calculated. Differences between CAA patients and controls were assessed using (ordinal) linear regression analysis.ResultIn brain tissue, CAA cases had increased vascular TIMP4 expression when compared to controls (p = 0.00002, figure 1). TIMP4 expression did not differ between CAA‐nonICH and CAA‐ICH cases. In the absence of vascular TIMP4 staining, we often observed perivascular staining. Decreased levels of perivascular TIMP4 staining were observed in CAA‐ICH cases when compared to CAA‐nonICH cases (p = 0.045) and to controls (p = 0.035). CSF levels of TIMP4 were decreased in CAA patients compared to controls (p = 0.03, figure 2). In addition, the ratios of MMP2 (p = 0.038), MMP9 (p = 0.005), and MMP14 (p = 0.025) to TIMP4 were increased in CAA patients compared to controls.ConclusionWe demonstrated increased TIMP4 expression in vessel walls of CAA patients compared to controls. In contrast, TIMP4 CSF levels were decreased in CAA patients compared to controls, whereas MMP:TIMP4 ratios were increased, indicating a disbalance of MMP/TIMP4 levels in CAA pathophysiology.Funding: BIONIC project (no. 733050822, ZonMW) and CAFÉ project (no. 5R01NS104147‐02, NIH).

Association of novel CSF biomarker candidates with cortical thickness in genetic frontotemporal dementia

AbstractBackgroundA novel panel of 14 proteins measured in the CSF could separate individuals with genetic frontotemporal dementia (FTD) from controls, with most significant findings observed for neurofilament medium (NEFM), neuronal pentraxin 2 (NPTX2), neurosecretory protein VGF (VGF) and aquaporin 4 (AQP4) [1]. However, it is currently unknown whether these altered protein levels in the CSF reflect neurodegenerative changes in the brain. The aim of this study was to explore the cross‐sectional associations between the previously identified CSF biomarker candidates and cortical thickness in presymptomatic and symptomatic mutation carriers, and whether those associations differ by FTD mutation type.MethodWe have analyzed T1 MRI scans alongside concurrent CSF samples from 202 individuals from the GENFI cohort, belonging to families that carry FTD mutations in either C9orf72, GRN or MAPT genes. The study sample included symptomatic mutation carriers, presymptomatic mutation carriers and non‐carrier controls. Cortical thickness was estimated with FreeSurfer and CSF protein levels were measured via a multiplexed antibody‐based suspension bead array. The correlations between regional cortical thickness and protein levels were calculated via linear regression.ResultAltered levels of NEFM, AQP4, APOA1, PTPRN2, CTSS, SERPINA3, C4, AMPH and CD14 were all correlated with increased atrophy of at least one cortical region. Some effects were mutation specific, but NEFM, AQP4 and APOA1 correlated with atrophy in all mutation groups. We also observed mutation specific effects for 10 of the proteins. CTSS levels were only correlated with cortical atrophy in C9orf72 mutation carriers while NPTX2, VGF and PTPRN2 correlated with atrophy in GRN mutation carriers. In MAPT mutation carriers, 6 different proteins correlated with atrophy in the right temporal pole.ConclusionThe proposed fluid biomarker candidates continue to show promise and further longitudinal studies will contribute to elucidate their relationship to cortical atrophy and their prognostic value in genetic FTD.[1] Bergström et al. Mol Neurodegener. 2021; 16(1):79

Safety, Pharmacokinetics and Immunogenicity of Single and Multiple Ascending Doses of the Anti‐Tau Therapeutic Antibody E2814: A Phase 1, First‐In‐Human (FIH) Study in Healthy Subjects

AbstractBackgroundE2814 is an anti‐tau therapeutic monoclonal antibody (mAb) which is being developed as a potential disease‐modifying therapy for Alzheimer’s disease (AD). E2814 inhibits the propagation of pathological tau species by binding to the microtubule binding region (MTBR). This study evaluated safety, pharmacokinetics (PK), and immunogenicity of single and multiple doses of E2814.MethodsThis randomized, placebo‐controlled study evaluated 5 single ascending doses (SAD) and 4 multiple ascending doses (MAD) in healthy subjects. Subjects (n = 8/cohort) received single or multiple (every 4 weeks) 1‐hour IV infusions of E2814 (or placebo). Safety was evaluated through a review of adverse events (AE), clinical laboratory test results, physical examination, electrocardiogram results, and vital signs. Serial blood and CSF samples for measurement of serum and CSF E2814 concentrations by a validated electrochemiluminescence (ECL) assay were obtained at prespecified time‐points. The serum PK parameters were estimated using non‐compartmental analysis. Anti‐drug antibodies (ADA) in serum were measured by a validated ECL assay.ResultsE2814 has an adequate single and multiple dose safety profile. No treatment‐emergent serious AEs, severe AEs, infusion‐related reactions, or dose limiting events were observed. There was a dose‐related increase in serum and CSF E2814 exposures. The median time to maximum E2814 concentrations in serum was 1.5 to 2 h. The half‐life of E2814 was 23 days which is typical of mAb. In SAD, serum AUC and Cmax were approximately dose proportional at lower doses and greater than dose‐proportional at higher doses. In MAD, accumulation based on Cmax and AUC was ∼1.2 and 1.3‐fold, respectively by the third dose in all dose group. The serum‐to‐CSF concentration ratio ranged between 0.1 to 0.3%. The immunogenicity to E2814 was low. In SAD, 5 subjects had transient low level ADA titers by last study day. In MAD, 1 subject had a confirmatory positive ADA at baseline prior to dose on Day 1. PK in the ADA positive subjects was comparable to that of ADA negative subjects.ConclusionE2814 presented an adequate safety and immunogenicity profile in healthy adults. E2814 PK was comparable to that with other mAbs. These results support further development of E2814 as a potential disease‐modifying therapy for AD.

Clinical Manifestations.

The ALLFTD (ARTFL-LEFFTDS Longitudinal Frontemporal Lobar Degeneration) study is an NIH-funded effort to prepare for clinical trials in sporadic (s-FTLD) and familial (f-FTLD) FTLD syndromes by characterizing cohorts, developing new clinical trial outcome measures, and evaluating disease progression. ALLFTD represents the merger and continuation of two prior studies: Advancing Research and Treatment of Frontotemporal Lobar Degeneration (ARTFL) and Longitudinal Evaluation of Familial Frontotemporal Dementia Subjects (LEFFTDS). ALLFTD recruits participants with FTLD spectrum disorders (bvFTD, svPPA, nfvPPA, FTD-ALS, CBS or PSP), with strong family histories of FTLD, or known FTLD-associated genetic variants within the family. 26 sites in North America enroll participants for clinical and neuropsychological evaluations, MR Imaging, blood draws, and CSF collection in willing participants; participants are requested to return annually for follow-up evaluation. All participants are genotyped for dementia-associated mutations. Participants also enroll in the FTD Disorders Registry for follow-up surveys. A subset of participants undergo remote assessment via the ALLFTD app. Visits conducted under ARTFL and LEFFTDS are included in the ALLFTD dataset. ALLFTD has been successful in developing methods to conduct remote or hybrid visits and in actively evaluating participants. Since ALLFTD began in January 2020, 1034 participants (736 new; 298 previously enrolled in ARTFL or LEFFTDS) have been evaluated in the longitudinal arm; an additional 119 participants have been participated in a shortened, biofluid-focused visit. 349 have returned for follow-up visits. Including prior participants, the ALLFTD dataset contains 3871 visits from 2343 individuals. 1277 people have sporadic syndromes; 1045 were enrolled as familial. Among s-FTLD, bvFTD is the most common phenotype (32.8%) followed by PSP (15.9%). 61.4% of f-FTLD participants were clinically normal at first visit; bvFTD is the most common syndrome in symptomatic f-FTLD (40%). 477 are confirmed carriers of FTLD-associated genetic mutations. MRIs were obtained for 2594 visits; 3492 visits have associated blood biospecimens banked; 978 visits have associated CSF samples. The ALLFTD consortium is actively evaluating participants across North America to better characterize FTLD syndromes and support the planning and development of FTLD clinical trials. Longitudinal datasets including clinical, genetic, and imaging data are available by request.

Two novel beta synuclein specific assays for the detection of Alzheimer’s disease

AbstractBackgroundRecent studies reported the potential use of elevated CSF β‐synuclein levels as an AD‐specific synaptic biomarker. These results were obtained with a sandwich ELISA where one of the antibodies recognizes both alpha‐and β‐synuclein1. Blood contamination can affect alpha‐synuclein concentrations in CSF, which is not observed for β‐synuclein2. Therefore, we aimed to develop a β‐synuclein‐specific assay to avoid possible interference with alpha‐synuclein.MethodsTwo novel sandwich CSF ELISA assays were established using only β‐specific synuclein monoclonal antibodies: a mid‐region and C‐terminal assay. Both assays use a monoclonal capture antibody that maps to C‐terminal. The mid‐region assay involves a detection antibody mapping to an epitope that overlaps with the SRM peptide used for CSF mass‐spectrometry quantification3. The C‐terminal assay includes a detection antibody targeting a less well‐defined C‐terminal epitope. Forty‐four remnant routine CSF samples were analyzed in duplicate (22 individuals with AD biomarker profile based on pTau181 and amyloid‐beta(1‐42), age70±7 years, 38%F and 22 control profiles, age70±7, 38%F).ResultsThe mid‐region assay showed a precision of 4%CV in clinical samples and a sensitivity of 1.25 pg/ml compared to 10%CV and 4.22 pg/ml with the C‐terminal assay (Fig. 1). CSF β‐synuclein levels were significantly elevated in patients with an AD CSF‐profile compared to individuals with negative‐AD CSF‐profiles, with AUCs of 0.77 [95%CI: 0.62‐0.90] for the C‐terminal assay and 0.71 [0.55‐0.87] for the mid‐region assay (Fig. 2). The assays showed comparable (Delong’s P = 0.53) clinical performance and their results correlated (r = 0.61, P = <0.001)(Fig. 2). Both C‐terminal and mid‐region β‐synuclein associated with CSF pTau181 (Spearman’s rho = 0.58 (p<0.001), rho = 0.51 (P = <0.001), respectively), and did not correlate with CSF Abeta(1‐42) (Fig. 3). Both assays are specific for recombinant beta‐synuclein, but the C‐terminal assay showed higher selectivity for β‐synuclein compared to the mid‐region assay, as increasing signals were observed with the latter with increasing concentrations of spiked recombinant alpha/gamma‐synuclein.ConclusionThe two novel β‐synuclein assays show clinically relevant changes in relation to AD CSF biomarker positivity. Both assays are specific to β‐synuclein. The assays will be transferred to SIMOA, which provides a wider dynamic range, and thus we will further assess the selectivity in plasma/serum.

Synucleinopathy in Alzheimer’s disease: role of the alpha‐synuclein seed amplification assays

AbstractBackgroundAlzheimer’s disease (AD) and Parkinson’s disease (PD) are the most frequent proteinopathies of the central nervous system. Alpha‐synuclein seed amplification assays (αS‐SAA) are the most promising techniques for the in vivo detection of synucleinopathies, ie, PD and DLB. In neuropathological studies, Lewy Bodies (LBs) have been found in 50‐60% of AD patients. In AD, concomitant LB pathology could lead to parkinsonism and to a more aggressive course of the disease. Whether concomitant LBs could be related to some clinical aspects of AD, i.e., clinical variants or clinical progression, remains to be assessed. In our work, we applied αS‐SAA to determine whether the rate of assay positivity differs in AD vs controls (CTRL), and among AD clinical variants.MethodWe used αS‐SAA to test 457 CSF samples from a cohort of AD, CTRL, and patients with PD and DLB. Classical AD CSF biomarkers were used to select CTRL and AD patients according to the A/T/(N) system. All CTRL showed a CSF A‐/T‐/N‐ profile while all AD patients were A+T+. Details of the subjects included are reported in Table 1. Fisher’s Exact test for count data and logistic regression (by assuming age and gender as covariates) were used to evaluate the significance of the different prevalences observed.ResultWe observed significantly higher αS‐SAA positivity in PD/DLB (86%) and AD patients (30%) with respect to CTRL subjects (13%, 7% in CTRL CN and 16% in CTRL MCI subjects) (Fig‐1A). Among the AD clinical variants, patients with posterior variant exhibited significantly higher αS‐SAA positivity (Fig‐1C). Interestingly, αS‐SAA positivity increased in the diagnostic groups as they progressed to later stages of AD, with the highest percentage in the AD at dementia phase (AD‐dem) group (36%) (Fig‐1B).ConclusionOur results show that a large portion of AD cases have concomitant synucleinopathy and αS‐SAA could allow stratification of AD cases, as assay positivity was associated with AD subtypes and progression. Further evaluation of cross‐sectional and longitudinal clinical data is needed to determine potential prognostic value of αS‐SAA in AD. Early detection of synucleinopathy in AD patients will be increasingly important for design of personalized disease‐modifying therapies.

Longitudinal white matter integrity alterations as a function of Preclinical Alzheimer’s Pathology: findings from the EPAD non‐demented cohort

AbstractBackgroundCross‐sectional diffusion tensor imaging (DTI) studies report white matter (WM) microstructural alterations, mainly involving fornix and corpus callosum (CC), in symptomatic stages of Alzheimer’s Disease (AD). Exploring these changes longitudinally in relation to preclinical and prodromal AD pathology is essential to track disease progression.We investigated trajectories of WM alterations in non‐demented individuals from the European Prevention of Alzheimer’s Dementia (EPAD) cohort and their relationship with AD biomarkers.MethodEPAD inclusion criteria were age>50 years and Clinical Dementia Rating≤0.5. We selected participants with baseline CSF samples, and longitudinal DTI scans. After defining AT groups(Ingala et al. 2021), A‐T+ participants were excluded (final N = 283).MRI acquisition and preprocessing was previously described(Lorenzini et al. 2022). We extracted fractional anisotropy (FA) in 14 regions(Mori et al. 2005) from DTI as a marker of WM integrity. Site harmonization of MRI‐derived data was performed with ComBat toolbox(Fortin et al. 2018).Age‐ and sex‐corrected linear mixed models with random intercepts on participants were used to estimate time‐effect on MRI‐derived phenotypes, and their interaction with AT status. Results were corrected for multiple testing.ResultAt baseline, A+T‐ participants showed lower FA in body/columns (ß = ‐0.04, p<0.01) and crus of fornix (ß = ‐0.01, p = 0.04), and in splenium (ß = ‐0.007, p = 0.04) and body of CC (ß = ‐0.01, p = 0.03) compared to A‐T‐. In the whole group, FA decreased over time in splenium (ß = ‐0.003, p<0.01), body (ß = ‐0.002, p<0.01) and genu of CC (ß = ‐0.009, p<0.01), body/columns (ß = ‐0.004, p<0.01) and crus of fornix (ß = ‐0.005, p<0.01), and inferior longitudinal fasciculus (ß = ‐0.007, p<0.01). Compared to A‐T‐, A+T+ participants showed steeper FA reduction in superior (Interaction:ß = ‐0.004, p = 0.03) and anterior corona radiata (Interaction:ß = ‐0.006, p = 0.02).ConclusionOur results suggest that WM integrity in key AD regions may degenerate early in the AD continuum, showing greater impairment over short‐time intervals in individuals with higher AD‐specific pathological burden.

Validation of Cerebrospinal Fluid Alpha‐synuclein Assay against Autopsy in Non‐Parkinson’s Disease Neurodegenerative Syndromes

AbstractBackgroundAlpha‐synuclein (αSyn) is the underlying neuropathology in both Parkinson’s Disease (PD) and non‐PD neurodegenerative syndromes, e.g. dementia with Lewy bodies (DLB). Importantly, αSyn is also a common co‐pathology, especially in Alzheimer’s disease (AD), where autopsy studies reveal comorbid αSyn in >50% of cases. A seeded aggregation amplification (SAA) assay (SynTap) has been shown to be highly sensitive and specific for pathogenic αSyn seeds in patients with PD and DLB, with lower performance in amygdalar‐predominant Lewy Body disease (LBD). Here we evaluated the antemortem CSF performance of SynTap in patients with non‐PD syndromes, with specific attention to performance in brainstem‐predominant αSyn and patients with atypical PD, e.g. progressive supranuclear palsy Richardson’s syndrome (PSP‐RS, N = 6) and corticobasal syndrome (CBS, N = 11).MethodPatients were evaluated through observational studies at the UCSF Alzheimer’s Disease Research Center and neuropathologic evaluation was performed by the Neurodegenerative Disease Brain Bank. Antemortem CSF samples were tested with SynTap. Operators were blinded to diagnosis, and all samples were run as a single batch. Sensitivity and specificity were compared, and clinical correlations were examined across pathological groups and clinical diagnoses, using antemortem clinical records.ResultOverall, in 56 patients, SynTap was 59% sensitive and 96% specific for autopsy‐confirmed αSyn pathology of any type. SynTap had perfect performance for diffuse (4/4) and transitional (5/5) LBD, and correctly identified 3/3 cases where LBD was primary. Consistent with prior reports, sensitivity was low in amygdalar‐predominant LBD (6/14, 43%), and lower in brainstem‐predominant LBD (1/6, 17%). A false positive result was present in 1/27 cases (4%). In the PSP‐RS cohort, SynTap positively identified 1 case caused by diffuse LBD and AD, and was true negative in 5/5. Within CBS, SynTap positively identified 1/1 case with diffuse LBD co‐pathology, 1/2 cases with brainstem‐predominant LBD, but missed 1 case with amygdalar‐predominant LBD.ConclusionSynTap best identifies αSyn pathology in patients with diffuse and transitional LBD, though specificity is high in the entire cohort. The etiology of low sensitivity in amygdalar‐ and brainstem‐predominant LBD co‐pathology requires further elucidation. The test may aid diagnostically in atypical PD syndromes, but results should be carefully considered, given the possibility of co‐pathology.

DDR1 regulates CNS collagen IV expression and vascular fibrosis in AD to improve CNS vascular integrity

AbstractBackgroundBlood‐brain barrier breakdown, a feature associated with AD pathology and treatments, was examined as a potential therapeutic target in both genetic mouse models of AD and clinical trial data from human patients with AD treated with nilotinib.MethodsDDR1 knockout mice crossed with TgAPP mice were probed for levels of CNS collagen to examine the effects of DDR1 in vivo. CSF samples obtained from AD subjects treated with either nilotinib (n = 12) or placebo (n = 11) for 12 months were analyzed using whole genome RNA sequencing; significantly altered miRNA expression was examined for potential downstream effects regulating neurovascular integrity.ResultsCompared to TgAPP mice, staining of CNS sections for collagen IV in TgAPP/DDR1 knockout mice revealed decreased vascular collagen. CSF obtained from AD subjects treated with nilotinib revealed increased levels of miRNAs regulating collagen production, suggesting decreased vascular fibrosis and neuroinflammation, and improved neurovascular integrity.ConclusionDDR1 inhibition with nilotinib may improve neurovascular integrity in AD by reducing CNS collagen IV, vascular fibrosis, and neuroinflammation. Nilotinib should be considered as an adjunctive therapy to approved treatments ‐ including anti‐amyloid antibodies ‐ for individuals with AD.

REAL AD – A realistic screening approach for preclinical Alzheimer’s disease

AbstractBackgroundFor disease‐modifying treatments to be successful in Alzheimer’s disease (AD), it is essential to identify individuals in the early phases of the disease, characterised by first biomarker abnormalities. At present, most AD diagnoses in Sweden are made within primary care which is not equipped for detecting individuals who are at risk of developing AD dementia. Considering current treatment developments, it will be essential to have access to biomarkers to aid an informed diagnosis already in primary care. REAL AD will evaluate the feasibility of a screening approach for preclinical AD using remote cognitive testing and blood‐based biomarkers in a naturalistic setting.MethodREAL AD is a collaborative initiative between the Swedish Västra Götaland County Council (VGR), VGR Primary Care, the Sahlgrenska University Hospital, the University of Gothenburg and pharmaceutical industry partners. A targeted PR‐campaign will facilitate reaching a recruitment goal of a minimum of 3000 socioeconomically and ethnically diverse individuals aged 50‐80 years without AD dementia. All participants will perform remote cognitive testing at home and undergo blood sampling at a local primary care unit on two occasions. Based on baseline blood biomarker profiles, a subgroup of participants will be enrolled into a validation study and further undergo CSF sampling, neuroimaging and clinical and neuropsychological assessment, also on two occasions.ResultLongitudinal remote cognitive assessment and blood sampling will be conducted at baseline and after 18 months, the blood samples will be analysed for a range of established and novel biomarkers. Within the validation study, results of the screening approach will be validated against established neuropsychological, CSF, MRI and [18F]RO948 tau PET assessments, and the feasibility of implementing this approach in the existing primary care infrastructure will be evaluated. We plan to present first data on recruitment and implementation at AAIC.ConclusionREAL AD aims to evaluate the applicability of a population‐based screening method to detect preclinical AD in a naturalistic setting and to inform the implementation of recent biomarker developments in clinical routine.

Head‐to‐head comparison of aptamer‐ and antibody‐based proteomic platforms in CSF samples of a real‐world large cohort of a memory clinic

AbstractBackgroundHigh‐throughput proteomic platforms are an useful strategy to detect novel biomarkers and pathways involved in Alzheimer’s disease (AD) development. Due to the novelty of these techniques, an extensive quality control assessment is needed to ensure the reliability of those protein measurements. In this study, we aimed to assess the reproducibility and reliability of SomaScan and Olink proteomic platforms and select top‐performing proteins for further analysis.MethodTo assess CSF proteomic measurements, we used the SomaScan 7k platform (an aptamer‐based technique) measuring 7335 human proteins, and Olink proteomics platform (Explore panel, an antibody‐based technique), measuring 3072 proteins in a subset of 509 individuals from ACE Alzheimer Center Barcelona (Spain). To assess intra‐ and inter‐platform reproducibility and reliability, we evaluated 1) the correlations of aptamers of two studies carried out in the same samples with SomaScan platform and 2) the correlation between SomaScan vs Olink proteomics, respectively. Then, we created a metric of 6 categories of reproducibility combining both intra and inter‐ correlation analysis, considering good (r > 0.5), modest (0.2 > r ≥ 0.5), poor (r ≤ 0.2) correlation. We considered proteins in the top 3 categories as potential candidates to further analysis.ResultFrom the intra‐platform analysis, only 2402 proteins (35.45%) showed a Pearson’s correlation higher than 0.5. Regarding inter‐platform assessment, 2160 proteins were overlapping between both platforms. From these overlapping proteins, 712 proteins exhibited a Pearson’s r > 0.5 (26.67%). We considered potential candidate proteins those with good intra‐ and inter‐platform correlation (n = 687, 9.02%), only good intra‐platform correlation (n = 2015, 26.45%) and modest intra‐ and inter‐platform correlation (n = 5, 0.06%) for analysis.ConclusionThe validation of the reproducibility and reliability of the new proteomics platforms is essential for the correct analysis of the results. The classification of measured proteins in different categories could be useful to increase the confidence in these highly multiplexed proteomic panels.

Determination of molecular chaperone levels in cerebrospinal fluids of Alzheimer’s patients

AbstractBackgroundMolecular chaperones are a component of the protein homeostasis network and regulate the biogenesis, folding, and trafficking of proteins, protecting the proteome against misfolding, aggregation, and proteotoxicity [1]. Therefore, they serve as the first line of defense against toxicity caused by misfolded proteins such as Aβ and tau [2]. Amyloid precursor protein (APP) interacts with chaperones from the moments of microtubule‐associated protein tau (MAPT) production. These properties keep the molecular chaperones in close contact with many proteins within the cell, including proteins associated with Alzheimer’s disease (AD), such as APP and tau [3]. According to the data we obtained using the web‐based FpClass protein‐protein association (PPI) prediction program, which investigates protein‐protein interactions, we determined that the AD‐related proteins MAPT and APP strongly interact with the chaperone proteins HSP90AA1, CHIP (STUB1) and HSPA4. In light of this information, we thought these chaperone proteins could interact with Aβ and tau and exit the CSF together, and these chaperones could be used as a biomarker.MethodFollowing this purpose, we determined the levels of HSP90AA1, STUB1, and HSPA4 chaperone proteins in CSF samples of AD, MCI, and FTD patients and individuals diagnosed with SCI using the ELISA method. Depending on the distribution of ApoE, any changes in CSF chaperone levels and co‐localization of these chaperones with MAPT and Aβ1‐42 were demonstrated by double‐immunofluorescence labeling in neuron‐differentiated human SH‐SY5Y cells. Within the scope of the study, these parameters included to be biomarkers that were investigated.ResultCSF HSP90AA1 protein levels were increased in sporadic AD and FTD patients compared to controls, CSF HSPA4 protein levels were increased in sporadic AD patients compared to family history of AD patients, and CSF STUB1 protein levels were decreased in AD, MCI, and FTD patients compared to controls.ConclusionConsidering the findings of our study, the presence of these chaperone proteins in CSF suggests that they may have an important place in the neuropathological processes of neurodegenerative diseases such as Alzheimer’s disease.The present work was supported by the Research Fund of Istanbul University‐Cerrahpasa. Project No. 33638

Oligoclonal Bands in CSF May be Associated with Chronic Graft Versus Host Disease after Haploidentical Hematopoietic Stem Cell Transplantation

Introduction Haploidentical hematopoietic stem cell transplantation (haplo-HSCT) is a straightforward and effective therapy for hematology patients, and chronic graft-versus-host disease (cGVHD) is recognized as a major and important complication of haplo-HSCT. cGVHD can affect a number of organs, such as the skin, eyes, oral mucosa, liver, and lung.In addition to classical manifestations, cGVHD can imitate almost any autoimmune disease, and its central nervous system (CNS) involvement is exceedingly rare. In our studies, IgG synthesis and anti-myelin oligodendrocyte glycoprotein antibody were risk factors for CNS demyelination ( Annals of Hematology, 2018), aGVHD and hypertension were associated with posterior reversible encephalopathy syndrome ( Bone Marrow Transplantation, 2020), and aGVHD and cGVHD were identified as factors for idiopathic inflammatory demyelinating disease ( Am J Hematol, 2021). However, there are limited data available on cGVHD in the CNS. Therefore, we performed a thorough assessment of cGVHD, including examining the CSF of patients with cGVHD. Methods We performed a retrospective nested case‒control study in patients receiving haplo-HSCT between January 2013 and June 2022 in our center. In total, 86 patients suffering from cGVHD who also received CSF examinations were analyzed, of whom 42 patients showed evidence of CNS-cGVHD. Written informed consent was, and ethical approval was obtained before sample collection by the Ethics Committee of Peking University People's Hospital. The CSF samples were collected by lumbar puncture (LP). At least 1 ml of CSF was obtained from all patients by nontraumatic LP. All CSF samples were routinely screened for infection with bacterial, fungal, and virological pathogens, examined by histopathology and analyzed by fluorescence-activated cell sorting to identify tumor cells. Immune biomarkers were also evaluated. Results The median time from transplantation to cGVHD was 236 days. In the 86 cGVHD patients, aGVHD (HR 0.431, 95% CI 0.229-0.812, P=0.009), CSF protein (HR 0.634, 95% CI 0.484-0.784, P=0.)and oligoclonal bands (OCB) (HR 0.527, 95% CI 0.312-0.890, P=0.017) were identified as risk factors for the onset of cGVHD after haplo-HSCT. Moreover, we combined the variables to establish a nomogram for the cumulative incidence of cGVHD. The test showed that the fitness of the multivariate logistic regression model was suitable (P = 0.9516). The area under the ROC curve was 0.741. There were no differences in overall survival (OS), disease-free survival (DFS), or the cumulative incidence of relapse (CIR) between the cGVHD group and the control group without cGVHD. Nonetheless, there was a significant difference in nonrelapse mortality (NRM) (P=0.009). The 2-year cumulative NRM was 48.2% (95% CI 45.8%-50.6%) versus 12.6% (95% CI 12.2%-13.0%) between the cGVHD and without cGVHD groups, respectively. Then, we performed multivariate analysis for data from 42 patients who were diagnosed with CNS-cGVHD. The underlying diseases (HR 1.334, 95% CI 1.027-1.733, P=0.030) and CSF proteins (HR 0.313, 95% CI 0.126-0.776, P=0.012) were independently associated with the occurrence of CNS-cGVHD. Conclusions In conclusion, CSF immunity may be associated with the onset of cGVHD after haplo-HSCT, especially OCB, and the identification of cGVHD increases NRM after haplo-HSCT. Further multicenter prospective studies are needed to confirm the results and establish the standard diagnostic criteria and therapy for CNS-cGVHD.

First Report of Phase 1 Studies of DZD8586, a BBB Penetrant LYN/BTK Dual Inhibitor, in Patients with B-Cell Non-Hodgkin Lymphoma (B-NHL)

Introduction Resistance to covalent (ibrutinib, acalabrutinib and zanubrutinib) and non-covalent BTK inhibitors (pirtobrutinib) is an emerging clinical challenge. DZD8586 was designed as a full blood-brain barrier (BBB) penetrable LYN/BTK dual inhibitor. Preclinical study suggested DZD8586 has a good selectivity against other TEC family kinases (TEC, ITK, TXK, BMX). DZD8586 inhibits BTK-dependent and BTK-independent BCR signaling pathways by targeting LYN and BTK, which could potentially overcome resistance to both covalent and non-covalent BTK inhibitors for the treatment of CLL and MCL, as well as DLBCL. Two phase 1 clinical studies are ongoing in patients with relapsed or refractory (r/r) B-NHL to assess safety, pharmacokinetics (PK), pharmacodynamics (PD) and anti-tumor efficacy of DZD8586. Methods TAI-SHAN1 (NCT05844956; CTR20220558) and TAI-SHAN5 (NCT05824585) are phase 1 dose escalation and expansion studies in patients with r/r B-NHL. Both studies have similar dose escalation strategies. Patients who progressed following or intolerant of prior systemic therapy were enrolled. Pooled safety and efficacy analysis was performed. In addition, modulation of PD biomarkers was also evaluated. Tumor responses were investigator assessed according to Lugano 2014, IPCG 2005, or iwCLL 2018, as appropriate. Results As of June 15, 2023, 17 patients with r/r B-NHL have been enrolled and received DZD8586 at doses ranging from 10 to 100 mg once daily (QD) (10 mg， n=1；25 mg， n=4；50 mg， n=7；100 mg， n=5). The median age was 61 years, 71% were male, and 53% had an ECOG PS of 1. Histological subtypes included CLL, SCNSL, DLBCL, MCL, FL and MZL. The median number of prior therapies was 3 (range 2 - 8). All patients received prior chemotherapy and anti-CD20 therapy. Six patients, including 1 CLL, 2 MCL, 1 MZL, 1 SCNSL and 1 DLBCL had received prior treatment with covalent BTK inhibitors. DZD8586 was well tolerated across the doses investigated. Drug-related treatment emergent adverse events (TEAEs) reported in > 2 patients included thrombocytopenia (10/17), neutropenia (5/17), and upper respiratory tract infection (3/17). The most common ≥ grade 3 TEAE was thrombocytopenia (3/17, all at 100 mg). Other ≥ grade 3 TEAEs, including neutropenia, drug eruption and upper respiratory tract infection, were each reported by one patient. Bleeding, atrial fibrillation, and arthralgia were not reported. Grade 1 diarrhoea was reported in one patient. The maximum tolerated dose was defined as 100 mg QD. All these TEAEs were reversible after dose interruption and/or with supportive care. The longest duration on treatment was 10+ months. Among 13 patients treated at doses ≥ 25 mg QD with at least one tumor assessment, the objective response rate (ORR) was 9/13 (69.2%). Responses were observed across different histological subtypes (1/2 in SCNSL; 4/5 in DLBCL; 1/2 in FL; 1/1 in MCL, 1/2 in MZL and 1/1 CLL carrying C481S mutation). Among the 5 evaluable BTKi treated patients, tumor responses were observed in 3 patients. The longest duration of response (DoR) was 6.4+ months, and the patient was still ongoing with DZD8586 treatment and responding. PK was generally dose-proportional across the dose ranges of 10 to 100 mg. DZD8586 was absorbed with median t max 1.5 hours, and eliminated with mean terminal half-life of ~18 hours. Accumulation of about two folds in AUC was noted at steady state following once daily dosing. Dose dependent pBTK modulation in whole blood cells was detected, and at the doses ≥ 25 mg, sustained pBTK inhibition was observed. The unbound CSF-to-plasma concentration ratio (K puu,CSF) was 1.2 from one CNSL patient with available CSF samples, suggesting excellent CNS penetration of DZD8586 in humans. Conclusion In the preliminary pooled analysis of two ongoing phase 1 trials, DZD8586 has encouraging anti-tumor activity and safety in heavily pre-treated B-NHL patients. The combination of BBB penetration, dual inhibition of LYN/BTK and potential to overcome C481S mutant CLL and other covalent BTKi resistant lymphomas suggest this may be a useful agent. Updated data will be presented at the meeting.

1319. Microbiology and Epidemiology of Pediatric Ventriculoperitoneal Shunt Infections

Abstract Background Infections complicate approximately 10% of ventriculoperitoneal (VP) shunts in children, with considerable associated morbidity and mortality. Recent guidelines for empiric treatment of VP shunt infections recommend use of vancomycin combined with an anti-pseudomonal beta-lactam, but are based on low-quality evidence. We aimed to further characterize and update the microbiology of pediatric VP shunt infections using the U.S. Military Health System database. Methods Department of Defense beneficiaries less than 22 years-of-age admitted for VP shunt infection between Oct 1, 2008 and Sep 30, 2019 were included. VP shunt infection was defined using a combination of ICD-9-CM and ICD-10-CM codes, CPT codes for removal or replacement of a shunt (Table 1), and positive cultures from a normally sterile site (cerebrospinal fluid, blood, catheter tip, and peritoneum). Demographic data on age, sex, and year of admission were collected for each patient, as was microbiologic data including organism and specimen source. Results There were 32 qualifying admissions among 30 patients for VP shunt infection over the 10-year study period, with no discernable year-to-year trend identified. Half of all VP shunt infections occurred in patients less than 12 months of age (Table 2). Sixty-one unique positive cultures were identified among the 30 patients, 38 (62%) from CSF samples, and the remaining 23 (38%) from blood, peritoneum, and catheter tips (Figure 1). Coagulase-negative Staphylococcus (28%), Staphylococcus aureus (15%), and Enterococcus spp. (10%) were the leading causative pathogens (Figure 2). Enteric gram-negative rods accounted for 21% of positive cultures. Pseudomonas spp. were rarely isolated (3%), however, strict anaerobic pathogens made up 8% of positive cultures. Conclusion Empiric antibiotic treatment of VP shunt infections in children should cover gram-positive cocci and enteric gram-negative rods but need not cover Pseudomonas spp. Existing guidelines do not recommend anaerobic coverage, however, these data indicate empiric anaerobic coverage is prudent, especially when an intra-abdominal VP shunt infection is suspected. Disclosures All Authors: No reported disclosures

2533. Cerebrospinal Fluid Penetration of Cefazolin and Flucloxacillin in Children

Abstract Background Effective treatment of methicillin-susceptible Staphylococcus aureus meningitis requires antimicrobial penetration into the CSF. Anti-staphylococcal penicillins, including flucloxacillin, are generally recommended as first-line treatment. However, some evidence suggests that the first-generation cephalosporin, cefazolin, may be an appropriate alternative. We investigated the CSF concentration of cefazolin and flucloxacillin in children. Methods CSF samples from children aged 0-18 years treated with cefazolin or flucloxacillin were collected when there was sample remaining after diagnostic tests were completed. Unbound antibiotic concentrations of cefazolin and flucloxacillin were measured. Results Overall, 222 CSF samples were collected from 163 children, median age 12 months (range 0.0-17.5 years). Of these, 157 were collected while the antibiotic was at steady state. For cefazolin, children ≤1 year had a higher median CSF concentration compared with those >1 year (1.8 versus 0.5 mg/L, p< 0.001), whereas for flucloxacillin the concentration was similar between the two age groups (0.2 versus 0.1 mg/L, p=0.9). In infants aged ≥ 1 month to 1 year, the CSF concentration of both antibiotics was higher in those with pleocytosis (WCC >5 x106/L) than those without (flucloxacillin 0.2 versus 0.065 mg/L, cefazolin 5.6 versus 1.06 mg/L). Notably, only those aged ≤1 year receiving cefazolin (7/12; 58%) achieved steady state CSF concentrations above the CLSI breakpoint for sensitive S. aureus (MIC ≤2 mg/L). Conclusion CSF penetration of cefazolin was higher in younger infants and for both antibiotics was higher in those with pleocytosis. In most children, the CSF concentrations in the CSF for either cefazolin or flucloxacillin were not above the S. aureus MIC. Disclosures All Authors: No reported disclosures

Temporal Profiles of P-Tau, T-Tau, and P-Tau:Tau Ratios in Cerebrospinal Fluid and Blood from Moderate-Severe Traumatic Brain Injury Patients and Relationship to 6-12 Month Global Outcomes.

Traumatic brain injury (TBI) can initiate progressive injury responses, which are linked to increased risk of neurodegenerative diseases known as "tauopathies." Increased post-TBI tau hyperphosphorylation has been reported in brain tissue and biofluids. Acute-to-chronic TBI total (T)-tau and phosphorylated (P)-tau temporal profiles in the cerebrospinal fluid (CSF) and serum and their relationship to global outcome is unknown. Our multi-site longitudinal study examines these concurrent profiles acutely (CSF and serum) and also characterizes the acute- to-chronic serum patterns. Serial serum and CSF samples from individuals with moderate-to-severe TBI were obtained from two cohorts (acute, subacute, and chronic samples from University of Pittsburgh [UPitt] [n = 286 unique subjects] and acute samples from Baylor College of Medicine [BCM] [n = 114 unique subjects]) and assayed for T-tau and P-tau using the Rolling Circle Amplification-Surround Optical Fiber ImmunoAssay platform. Biokinetic analyses described serum T-tau and P-tau temporal patterns. T-tau and P-tau levels are compared with those in healthy controls (n = 89 for both CSF and serum), and univariate/multivariable associations are made with global outcome, including the Disability Rating Scale (DRS) and the Glasgow Outcome Scale-Extended (GOS-E) scores at 3 and 6 months post-TBI (BCM cohort) and at 6 and 12 months post-TBI (UPitt cohort). For both the UPitt and BCM cohorts, temporal increases in median serum and CSF T-tau and P-tau levels occurred over the first 5 days post-injury, while the initial increases of P-tau:T-tau ratio plateaued by day 4 post-injury (UPitt: n = 99, BCM: n = 48). Biokinetic analyses with UPitt data showed novel findings that T-tau (n = 74) and P-tau (n = 87) reached delayed maximum levels at 4.5 and 5.1 days, while exhibiting long serum half-lives (152 and 123 days), respectively. The post-TBI rise in acute (days 2-6) serum P-tau (up to 276-fold) far outpaced that of T-tau (7.3-fold), leading to a P-tau:T-tau increase of up to 267-fold, suggesting a shift toward tau hyperphosphorylation. BCM analyses showed that days 0-6 mean CSF T-tau and P-tau levels and P-tau:T-tau ratios were associated with greater disability (DRS) (n = 48) and worse global outcome (GOS-E) (n = 48) 6 months post-injury. Days 0-6 mean serum T-tau, P-tau, and P-tau:T-tau ratio were not associated with outcome in either cohort (UPitt: n = 145 [DRS], n = 154 [GOS-E], BCM: n = 99 [DRS and GOS-E]). UPitt multivariate models showed that higher chronic (months 1-6) mean P-tau levels and P-tau:T-tau ratio, but not T-tau levels, are associated with greater disability (DRS: n = 119) and worse global outcomes (GOS-E: n = 117) 12 months post-injury. This work shows the potential importance of monitoring post-TBI T-tau and P-tau levels over time. This multi-site longitudinal study features concurrent acute TBI T-tau and P-tau profiles in CSF and serum, and also characterizes acute-to-chronic serum profiles. Longitudinal profiles, along with no temporal concordance between trajectory groups over time, imply a sustained post-TBI shift in tau phosphorylation dynamics that may favor tauopathy development chronically.

Metabolic characterization of neurogenetic disorders involving glutamatergic neurotransmission.

The study of inborn errors of neurotransmission has been mostly focused on monoamine disorders, GABAergic and glycinergic defects. The study of the glutamatergic synapse using the same approach than classic neurotransmitter disorders is challenging due to the lack of biomarkers in the CSF. A metabolomic approach can provide both insight into their molecular basis and outline novel therapeutic alternatives. We have performed a semi-targeted metabolomic analysis on CSF samples from 25 patients with neurogenetic disorders with an important expression in the glutamatergic synapse and 5 controls. Samples from patients diagnosed with MCP2, CDKL5-, GRINpathies and STXBP1-related encephalopathies were included. We have performed univariate (UVA) and multivariate statistical analysis (MVA), using Wilcoxon rank-sum test, principal component analysis (PCA), and OPLS-DA. By using the results of both analyses, we have identified the metabolites that were significantly altered and that were important in clustering the respective groups. On these, we performed pathway- and network-based analyses to define which metabolic pathways were possibly altered in each pathology. We have observed alterations in the tryptophan and branched-chain amino acid metabolism pathways, which interestingly converge on LAT1 transporter-dependency to cross the blood-brain barrier (BBB). Analysis of the expression of LAT1 transporter in brain samples from a mouse model of Rett syndrome (MECP2) revealed a decrease in the transporter expression, that was already noticeable at pre-symptomatic stages. The study of the glutamatergic synapse from this perspective advances the understanding of their pathophysiology, shining light on an understudied feature as is their metabolic signature.