Apolipoprotein A1 in Cerebrospinal Fluid and Plasma and Progression to Alzheimer’s Disease in Non-Demented Elderly

Abstract

Background: HDL-cholesterol transporter Apolipoprotein A1 (ApoA1) holds neuroprotective properties, such as inhibition of amyloid-β aggregation. Low plasma ApoA1 concentrations are associated with Alzheimer’s disease (AD). Little is known about ApoA1 levels in the pre-dementia stages of AD.

Objective: To investigate associations between cerebrospinal fluid (CSF) and plasma ApoA1 levels and clinical progression toward AD in non-demented elderly.

Methods: From the Amsterdam Dementia Cohort, we included 429 non-demented elderly with subjective cognitive decline (SCD; n = 206, 61±9 years, Mini-Mental State Exam (MMSE) 28±2) and mild cognitive impairment (MCI; n = 223, 67±8 years, MMSE 27±2), with a mean follow-up of 2.5±1.6 years. We used Cox proportional hazard models to investigate relations between CSF and plasma ApoA1 concentrations and clinical progression, defined as progression to MCI or AD for SCD, and progression to AD for MCI. Analyses were adjusted for age, gender, MMSE, and plasma cholesterol levels. Analyses were stratified for diagnosis and APOE ɛ4 carriership.

Results: 117 patients (27%) showed clinical progression. One standard deviation increase of CSF ApoA1 was associated with a 30% increased risk of clinical progression (hazard ratio (HR) (95% CI) = 1.3(1.0–1.6)). The effect appeared to be attributable to the APOE ɛ4 carriers with SCD (HR 3.3(1.0–10.9)). Lower plasma ApoA1 levels were associated with an increased risk of clinical progression in APOE ɛ4 carriers with SCD (HR 5.0(1.3–18.9)).

Conclusion: Higher CSF and lower plasma ApoA1 levels were associated with an increased risk of clinical progression in APOE ɛ4 carriers with SCD; suggesting that ApoA1 may be involved in the earliest stages of AD.

Keywords

Alzheimer’s disease Apolipoprotein-A1 Apolipoprotein E mild cognitive impairment subjective cognitive decline

INTRODUCTION

The ɛ4 allele of the Apolipoprotein E gene (APOE) is a major genetic risk factor for Alzheimer’s disease (AD) [1 –3], and seems to influence, besides Apolipoprotein E (ApoE) concentrations, the expression of other apolipoproteins as well [4, 5]. Recent proteomic studies have identified Apolipoprotein A1 (ApoA1) as being related to AD pathology [6, 7]. ApoA1 is mainly involved in reverse cholesterol transport, preventing atherosclerosis by transporting excessive cholesterol back to the liver [8]. ApoA1 is, next to ApoE, the second most abundant apolipoprotein in cerebrospinal fluid (CSF), present in HDL-like particles and maintaining cholesterol homeostasis in the brain [8 –10]. ApoA1 is probably transported over the blood-brain barrier by transcytosis, facilitated by the Scavenger Receptor class B type 1 [11], but it is also expressed by brain capillary endothelial cells [12 –14]. ApoA1 has been shown to inhibit amyloid-β (Aβ) aggregation and prevent Aβ induced neurotoxicity in vitro [15 –17]. ApoA1 deficiency in AD mouse models lowered plasma cholesterol and increased cerebral amyloid angiopathy (CAA) and cognitive deficits, but it did not alter parenchymal amyloid deposition [17, 18]; whereas overexpression of ApoA1 in APP/PS1 mice did not prevent amyloid deposition, but preserved cognitive function and attenuated CAA [19].

In clinical studies, reduced plasma ApoA1 levels have been observed in AD patients compared with controls [20 –22]. In mild cognitive impairment (MCI), low plasma ApoA1 was the strongest predictor for cognitive decline out of several apolipoproteins [5]. In a community based study, the combination of APOE ɛ4 genotype and low plasma ApoA1 levels was associated with an increased risk of AD [4]. A few previous studies have reported on ApoA1 levels in CSF, comparing AD patients and controls, and results were not consistent [23 –29]. Moreover, no data on CSF ApoA1 in pre-dementia stages of AD were available.

Neurodegenerative changes, eventually leading to dementia due to AD, begin to accumulate at least twenty years before clinical symptoms appear [30, 31]. In the search for the underlying mechanism of the disease, the field is gradually moving forward in the disease process. MCI refers to the stage where patients experience memory impairment, but perform within normal limits on tests for global cognition and function independently at home [32, 33]. Subjective cognitive decline (SCD) refers to individuals who experience cognitive deterioration despite normal functioning on cognitive testing. This entity has recently been suggested as a potential first symptomatic expression of AD [34, 35].

We aimed to investigate the association between ApoA1 levels in CSF and plasma and clinical progression in non-demented patients with MCI and SCD. In addition, we evaluated whether these associations were modulated by APOE ɛ4 carriership. We found that higher CSF and lower plasma ApoA1 were associated with an increased risk of clinical progression in pre-dementia patients, particularly in APOE ɛ4 carriers with subjective cognitive decline.

MATERIALS AND METHODS

Subjects

From the Amsterdam Dementia Cohort, we included 452 non-demented patients with a baseline diagnosis of SCD or MCI, with available CSF and plasma and at least one year follow up. All patients underwent a standardized dementia screening including neuropsychological, physical, and neurologic examination as well as laboratory tests, electro-encephalography, and brain magnetic resonance imaging [36]. Diagnoses were made in a multidisciplinary consensus meeting. All MCI patients fulfilled NIA-AA core clinical criteria for MCI [32, 33]. Patients were labeled as having SCD when they presented with memory complaints, but cognitive functioning was normal and criteria for MCI, dementia, or any other neurological or psychiatric disorder known to cause cognitive decline were not met [35]. At the yearly follow-up visit in the memory clinic patient history, cognitive tests and physical and neurologic examination were repeated and diagnoses were re-evaluated. All AD patients fulfilled NIA-AA core clinical criteria for dementia due to AD [37]. The main outcome measure was clinical progression. In MCI, clinical progression was defined as progression to dementia due to AD. In patients with SCD, clinical progression was defined as progression to MCI or dementia due to AD. If patients with SCD first progressed to MCI and then to AD, the moment of progression to MCI was taken as time of clinical progression. Twenty-two patients that progressed to another form of dementia than AD were excluded. Furthermore, we excluded one patient with an ApoA1 CSF concentration of six standard deviations (SD) above the mean. This resulted in 206 patients with SCD and 223 with MCI.

Ethics, consent, and permissions

The local medical ethics committee of the VU University Medical Center approved collection of data and biomaterial from patients for research purposes. All patients gave written informed consent for the use of their data in research. All research was conducted in accordance with the Helsinki Declaration of 1975.

APOE genotyping

APOE genotyping was performed after automated genomic DNA isolation from 7–10 mL EDTA blood. It was subjected to PCR, checked for size and quantity using a QIAxcel DNA Fast Analysis kit (Qiagen, Venlo, The Netherlands) and sequenced using Sanger sequencing on an ABI130XL. Subjects with at least one ɛ4 allele were classified as APOE ɛ4 carriers, patients without an ɛ4 allele were considered asnon-carriers.

CSF and plasma biomarker analyses

CSF and plasma analyses were performed at the Neurochemistry Laboratory at the department of Clinical Chemistry of the VU University Medical Center Amsterdam. CSF and plasma were collected from non-fasted subjects. CSF was obtained by lumbar puncture between the L3/L4 or L4/L5 intervertebral space by a 25-gauge needle and collected in polypropylene tubes.

The first 2.5 mL CSF was collected in a separate tube for cell (erythrocyte) counting (expressed as number of erythrocytes/3 μL CSF) to determine possible blood contamination of the CSF. 87% of the available CSF samples had erythrocyte counts lower than 1500, corresponding to— based on average ApoA1 CSF (3 μg/mL) and plasma (1 mg/ml) levels— a contribution of plasma apoA1 to the CSF apoA1 levels of 3.8% at the maximum. This CSF sample was also used to determine AD biomarker levels (Aβ_1–42, Tau, and phosphorylated Tau 181 (pTau)), after centrifugation at 1,800×g, using ELISA (Innotest, Fujirebio, Ghent, Belgium) [38]. The interassay CVs obtained were 11.3% (4.9) for Aβ₄₂, 9.3% (1.5%) for Tau, and 9.4% (2.5%) for pTau. Staff involved in AD-biomarker analysis was blinded for clinical diagnosis. EDTA plasma was collected in 7 mL tubes. CSF and plasma for biobanking were centrifuged, aliquotted into 0.5 mL vials and stored at –80°C until further analysis. A maximum of 2 h was allowed between collection and freezing [39].

ApoA1 levels in CSF (n = 401) and EDTA plasma (n = 411) were measured using a commercial sandwich ELISA^PRO kit for human ApoA1 (Mabtech, Nacka Strand, Sweden). This assay utilizes ELISA strips pre-coated with capture monoclonal antibody (HDL110), to which samples are added. Captured ApoA1 is detected by adding another, biotinylated ApoA1 specific monoclonal antibody (HDL44). Concentration in the sample is determined by comparison to a serial dilution of purified human ApoA1, resulting in a standard range of 0.32–31.6 ng/ml. Pools of surplus routine plasma samples selected to have either high or low ApoA1 concentrations were run as a quality control. Plasma and CSF samples were tested at 1 : 100.000 and 1 : 1000 dilutions, respectively. Intra-assay CVs for ApoA1 results were on average 2.3% for plasma and 4.0% for CSF samples. Inter-assay CVs (26 plates) were 13.0% and 9.1% for the low and high ApoA1 plasma controls, respectively.

Plasma lipid levels were measured with a Modular P system (Roche, Mannheim, Germany). For total cholesterol, HDL-cholesterol, and triglycerides, the following reagents were used: CHOD-PAP, HDL-C plus, and GPO-PAP, respectively (Roche, Mannheim, Germany). The inter-assay CVs were all less than 7%. LDL-cholesterol was calculated using the Friedewald formula [39]. Cholesterol measurements were available for 411 of 429 patients.

Neuropsychological assessment

We used standardized measurements to assess cognitive functioning. Of the standardized test battery we used total immediate recall and delayed recall of the Dutch version of the Rey auditory verbal learning task (RAVLT) to evaluate memory function [40]. To evaluate executive function we used Trail Making Test (TMT) B, and for global cognition we used the Mini-Mental State Examination (MMSE) [41, 42].

Statistical analysis

Data were analyzed using SPSS for Macintosh, version 20 (IBM, Armonk, NY). Groups were compared using t-tests and chi-squared tests as appropriate. All biomarkers were log-transformed, because they did not have a normal distribution. Subsequently, we transformed values to z-scores. We used Cox proportional hazard models to assess associations between CSF and plasma ApoA1 concentrations and clinical progression to MCI or dementia due to AD. CSF and plasma ApoA1 were used as independent variables (separate models), and clinical progression was used as outcome measure. Plasma ApoA1 values were inverted prior to analysis. Results are presented as hazard ratio (HR) (95% confidence interval (CI)). HRs represent the risk of clinical progression associated with one standard deviation (SD) increase in CSF ApoA1 concentration or one SD decrease in plasma ApoA1. We cumulatively adjusted for age, sex, MMSE (model 1), and total cholesterol, HDL-cholesterol, LDL-cholesterol, and triglycerides (model 2). To assess the effect of the combination of CSF and plasma ApoA1 on clinical progression, we performed Cox proportional hazard models with the CSF/plasma ApoA1 ratio. The CSF/plasma ApoA1 ratio was calculated based on raw ApoA1 scores, followed by transformation to z-scores. Additionally, all analyses were repeated for SCD and MCI separately, and we re-ran all analyses stratified for APOE ɛ4 carriership.

To assess the effects ApoA1 on clinical progression in confirmed amyloid positive patients, we repeated Cox proportional hazards models in non-demented individuals with CSF Aβ₄₂ levels below a cut-off of 550 ng/L, based on previous studies of our center [38].

We used multivariate regression analyses with forward selection to identify the combination of biomarkers most predictive of cognitive function. Predictive variables were: Aβ₄₂, tau, pTau, plasma ApoA1, and CSF ApoA1. We adjusted analyses for age and gender age, gender. Dependent variables were MMSE, immediate and delayed recall (RAVLT), and TMT-A and B (in separate models). We reported associations between the biomarkers and cognitive markers as standardized beta.

Correlations between CSF and plasma ApoA1 levels and plasma cholesterol levels, CSF Aβ₄₂, Tau, and pTau levels and age were assessed using Pearson’s correlation coefficient. p < 0.05 was considered significant.

RESULTS

Table 1 shows the baseline characteristics. On average, patients were 64±9 years of age, 180 (42%) were female, and they had a mean MMSE of 27±2. 213 (50%) of the patients were APOE ɛ4 carriers. There was a mean follow up was 2.5±1.6 years. During follow-up, 117 patients (27%; 26 SCD and 91 MCI) showed clinical progression. Patients with clinical progression were older, had lower MMSE scores, and were more frequently APOE ɛ4 carriers. Patients showing clinical progression had lower CSF Aβ₄₂ and higher CSF Tau and pTau concentrations at baseline, than those who remained stable. There were no differences in CSF or plasma ApoA1 levels between patients with clinical progression versus patients that remained stable in the total group. In the SCD subgroup CSF ApoA1 and the CSF/plasma ApoA1 ratio was higher in patients that showed clinical progression than stable individuals, seeTable 2.

We used Cox proportional hazard models to investigate relationships between CSF and plasma ApoA1 levels and the risk of clinical progression. Table 2 shows the HR (95% CI). Adjusted for age, sex, and MMSE, CSF ApoA1 was associated with an increased risk of clinical progression in the total group; as an increase of one SD in CSF ApoA1 was associated with a 30% increased risk of clinical progression (HR (95% CI) = 1.3 (1.0–1.6)). Additional adjustment of the models for HDL-, LDL-, and total cholesterol did not change the effect. Stratification for APOE ɛ4 genotype revealed that the effect was attributable to APOE ɛ4 carriers (HR (95% CI) = 1.4 (1.1–18), compared to APOE ɛ4 non-carriers HR 1.1 (0.7–1.6)). There was no relation between plasma ApoA1 and the risk of clinical progression in the total group. We then repeated the Cox proportional hazard models with the CSF/plasma ApoA1 ratio to assess the relation with clinical progression. A higher CSF/plasma ApoA1 ratio was associated with an increased risk of clinical progression in non-demented elderly (HR (95% CI) = 1.14 (1.04–1.24), adjusted for age, sex, and MMSE).

Subsequently, we re-ran all models in SCD and MCI patients separately (Table 2). We found that the observed effects were mostly attributable to SCD patients. In the SCD group, higher CSF ApoA1 was associated with an increased risk of clinical progression (HR (95% CI) = 1.5 (0.9–2.4)), although not significant. After stratification for APOE genotype, the effect of CSF ApoA1 on progression was most prominent in the APOE ɛ4 carriers (HR (95% CI) = 2.7 (1.1–6.5), adjusted for age, sex, and MMSE; and HR (95% CI) = 3.3 (1.0–10.9), additionally adjusted for cholesterol concentrations). Lower plasma ApoA1 was associated with an increased risk of clinical progression in APOE ɛ4 carriers with SCD (HR (95% CI) = 5.0 (CI 1.3–18.9), after adjustment for age, sex, MMSE, and cholesterol concentrations; plasma ApoA1 scores were inverted prior to Cox proportional hazards analyses). For the CSF/plasma ApoA1 ratios, effects were most prominent in APOE ɛ4 positive individuals with SCD (HR (95% CI) = 1.63 (1.11–2.40)).

In MCI, we found no significant associations between either CSF or plasma ApoA1 and clinical progression.

To assess the effects of CSF and plasma ApoA1 on clinical progression in amyloid positive patients (CSF Aβ_1–42 value below 550 ng/L), we repeated Cox proportional hazards models in non-demented Amyloid positive individuals (n = 143, 29% of total group), but results were not significant anymore.

Correlations between CSF and plasma ApoA1 concentrations, the CSF/plasma ApoA1 ratio, plasma cholesterol values, and CSF AD biomarkers are summarized in Table 3. We found no correlation between CSF and plasma ApoA1 (r = 0.03, p = 0.594) for the total group, and also when cases with clinical progression toward the AD trajectory at follow up (n = 117; r = –0.189, p = 0.057) and cases that remained clinically stable (n = 312; r = 0.097, p = 0.104) were considered separately (Fig. 1). Plasma ApoA1 was strongly related to HDL cholesterol, and moderately with total cholesterol and triglycerides. CSF and plasma ApoA1 did not correlate with CSF Aβ_1–42, but there was a weak correlation between plasma ApoA1 and CSF Tau, largely attributable to the MCI group. There was a modest positive correlation between CSF ApoA1 and age.

We used multivariate regression analyses with forward selection to identify the best combination of biomarkers associated with individual cognitive markers (MMSE, immediate and delayed recall (RAVLT), and TMT-B). Results are displayed in Table 4. The optimal model for global cognition (MMSE) included CSF Tau and CSF ApoA1. The optimal model for both immediate and delayed recall included CSF Tau, CSF Aβ_1–42, and plasma ApoA1, but after adjustment for age and gender plasma ApoA1 was excluded from the best fitted model. For executive functioning, the optimal model included CSF Tau and CSF Aβ_1–42.

DISCUSSION

The main finding of this study is that higher baseline CSF ApoA1 was associated with an increased risk of clinical progression in non-demented APOE ɛ4 carriers with SCD. In these patients, lower levels of plasma ApoA1 were also associated with increased risk of clinical progression. To further assess the coherence between these two biomarkers, we looked into the CSF/plasma ApoA1 ratio. This ratio was higher in patients that progressed toward dementia than in patients that remained stable, but only in SCD these differences were significant. The increased CSF/plasma ratio in progressors reflects the relatively higher CSF ApoA1 and lower plasma ApoA1 levels associated with an increased risk of progression in individual patients. This, and the percentage-wise larger change in CSF (Table 1), suggests that higher CSF ApoA1 contributes relatively more to progression than lower plasma ApoA1 in our population.

Former studies have shown reduced plasma ApoA1 levels in patients with AD compared to controls [20 –22]. Low plasma ApoA1 was also identified as the apolipoprotein with the strongest predictive value for cognitive decline in patients with MCI [5]. In the Honolulu aging study, decreased plasma ApoA1 was related to clinical progression in a community based sample, particularly in APOE ɛ4 carriers [4]. Our findings are in line with these former studies, as we found that low plasma ApoA1 concentrations were associated with an increased risk of clinical progression. In our study the effect was specific for subjects in the earliest stages of AD (i.e., SCD), and particularly the APOE ɛ4 carriers.

Few previous studies on CSF ApoA1 in relation to AD are available. Available studies mainly reported reduced ApoA1 levels in patients with AD compared to controls, but results were not consistent. Some studies showed decreased levels of CSF ApoA1 [26 –29], while others showed no effect [24] or increased CSF ApoA1 [25] in patients with AD versus healthy controls. In our sample of pre-dementia patients, we observed an association between increased concentrations of CSF ApoA1 and an increased risk of clinical progression toward AD. Differences in results for CSF ApoA1 between our study and previous studies may be due to methodological issues, as well as to differences in disease stage of subjects included in the different studies. Concerning methodological issues, results in former studies were not adjusted for age, sex, or other potential confounders, possibly because of their small sample sizes [25 –28]. In addition, in some studies CSF was obtained by lumbar puncture [24 –27], while in others ventricular CSF was obtained postmortem [28, 29], which makes the comparison of protein concentrations difficult [43]. Concerning differences in disease stage, previous studies on CSF ApoA1 compared only AD patients with healthy controls, while we studied pre-dementia patients [26 –29]. It is conceivable that levels increase in the earliest phase of AD, and decrease later in the disease process. Support for this notion comes from our findings that specifically in subjects with SCD, elevated CSF ApoA1 levels were associated with an increased risk of clinical progression, while in MCI we did not observe this effect.

Limitations of the study include the relative short mean follow up of 2.5 years in view of a disease process that may take at least 20 years. The number of subjects with cognitive decline at follow-up was relatively small in the group of patients with SCD, and as a result Cox models did not converge in some of the smaller strata (APOE ɛ4). Future studies should include longitudinal measurements of ApoA1 to investigate changes in CSF and plasma ApoA1 values over time in relation to cognitive decline. A previous study indicated that blood contamination of CSF samples, measured as CSF hemoglobin concentration, could influence reliability of CSF measurements [44]. In our study, we used available erythrocyte concentration to assess possible blood contamination of the CSF samples. The majority (>87%) of available CSF samples had erythrocyte levels lower than 1500/3 μl CSF, corresponding to a maximal contribution of plasma apoA1 to the CSF apoA1 level of 3.8% (0.135 μg/ml). Therefore, and because of possible overestimation of erythrocyte numbers as these were determined in the first 2.5 ml collected that was not used for storage in our biobank, we considered the influence of possible blood contamination on our CSF results a minor problem. Strengths of the current study are the longitudinal design and highly standardized clinical follow-up. To our knowledge this is the first study to assess such a large sample of paired CSF and plasma ApoA1 data, in combination with CSF AD biomarkers, in the pre-dementia stages of AD.

A lack of association between specific biomarker concentrations in CSF and plasma has been reported for AD biomarkers, for example Aβ_1–42 [45]. CSF and plasma ApoA1 concentrations probably do not correlate, because ApoA1 is produced in the liver and small intestine, and is also expressed by brain capillary endothelial cells, but not in the CNS itself [13, 14]. The blood-brain barrier separates these compartments, and therefore concentrations in plasma and the brain might be independent from each other. On the other hand, it has also been suggested that ApoA1 is possibly transported over the blood-brain barrier by transcytosis, facilitated by the Scavenger Receptor class B type 1, suggesting an influence of plasma ApoA1 concentrations on ApoA1 in cerebro. Further research is needed to assess the impact of this transport over the blood-brain barrier on cerebral and CSF concentrations of ApoA1.

ApoA1 has been shown to inhibit Aβ aggregation and also to exert protective effects against Aβ mediated neurotoxicity in vitro [15 –17], as well as in animal models for AD [17, 19]. In ApoA1 deficient AD mice, ApoA1 levels in brain and CSF were reduced, and, possibly as a compensatory mechanism, plasma ApoE levels, but not CSF ApoE levels, were increased [18]. ApoA1 deficient AD mouse models exhibited memory deficits to a certain degree which paralleled cerebral vascular Aβ accumulation [17], whereas overexpression of ApoA1 in AD mice attenuated cognitive deficits and reduced the degree of CAA and neuroinflammation [19]. This protective effect of ApoA1 against neuroinflammation has also been previously described in Parkinson’s disease [46]. The beneficial effects of increased ApoA1 plasma levels on cerebral amyloid deposition [19] may be due to ApoA1 binding to Aβ, and preventing its toxic effects on vascular smooth muscle cells, since ApoA1, lipidated as well as non-lipidated, was found to protect brain vascular smooth muscle cells from Aβ in vitro [17]. In AD mouse models, ApoA1 levels were not related to Aβ accumulation in the brain parenchyma [17 –19]. Thus, although ApoA1 has been suggested to influence Aβ metabolism [15], its main effect seems to be on vascular Aβ accumulation. Cerebral amyloid deposition has been observed at neuropathological examination in 23–45% of the non-demented elderly [47]. It can be suggested that vascular amyloid deposition may be an early phenomenon in the pathophysiology of AD, which, as judged from the AD mouse model studies [17 –19], can be associated with ApoA1 expression levels. The early increase in CSF ApoA1 observed in our study in non-demented patients, may be a protective measure in the earliest stages of AD, in which ApoA1 can exert neuroprotective effects, especially when Aβ-mediated [15–17 , 19].

Because of the intimate involvement of Aβ in the pathogenesis of AD and CSF Aβ_1–42 levels being reduced in very early stages of AD, we next looked if ApoA1 levels were associated with AD biomarkers in CSF, but found no strong associations between CSF and plasma ApoA1, and the AD biomarkers CSF Aβ₄₂ and Tau (Table 3). The significance of the relation between CSF and plasma ApoA1 and clinical progression, was lost when Cox proportional hazard models for CSF and plasma ApoA1 were repeated after stratification for CSF Aβ_1–42 values below our center cut off of 550 ng/L, indicative of AD, probably due to the small group sizes.

Although associations between CSF ApoA1 concentrations and CSF Aβ₄₂ values have not been studied before, research on APOA1 polymorphisms showed that presence of the APOA1 -75bp A allele may be associated with lower CSF Aβ₄₂, and also with increased risk of AD [48, 49]. It would be of interest to further assess the contribution of APOA1 polymorphisms on ApoA1 concentrations in relation to Alzheimer biomarkers, within the light of other apolipoproteins and APOE ɛ4 carriership.

Further research on the possible influence of apolipoproteins on AD pathology is currently ongoing in our cohort using mediation analyses. We also investigated how CSF and plasma ApoA1, in combination with CSF Tau and Aβ₄₂, were aligned with cognitive measurements. The optimal model of biomarkers explaining global cognition included CSF Tau, Aβ₄₂, and CSF ApoA1. The optimal model explaining the memory domain included CSF Tau, CSF Aβ₄₂, and plasma ApoA1, but after adjustment for age and gender plasma ApoA1 was excluded. Lower plasma ApoA1 was related to worse cognition, which is in line with previous research indicating that lower plasma ApoA1 concentrations were associated with worse cognitive performance and severity of AD [22, 49].

We found stronger effects for both CSF and plasma ApoA1 in APOE ɛ4 carriers than non-carriers, especially in SCD. APOE ɛ4 carriership has been suggested to influence the pathophysiological sequences leading to dementia due to AD [3, 50], and is associated with reduced CSF Aβ₄₂ concentrations in patients with AD [51, 52]. After stratification for ɛ4 carriership, effects of plasma ApoA1 were different in ɛ4-carriers versus non-carriers, which is similar to the results of two previous studies in which the effect of plasma ApoA1 on progression toward AD differed between APOE ɛ4 carriers and non-carriers with normal cognition [4, 5]. We now indicate that also for CSF ApoA1 concentrations, effects seem to differ between APOE ɛ4 carriers and non-carriers. How APOE ɛ4 carriership exerts its effects on ApoA1 remains to be elucidated.

In conclusion, both higher CSF ApoA1 and lower plasma ApoA1 are associated with an increased risk of clinical progression in the pre-dementia AD, especially in subjects with SCD and APOE ɛ4 carriership, suggesting a role for ApoA1 in the earliest stages of AD.

Footnotes

ACKNOWLEDGMENTS

The authors thank Mrs. M. van der Wal and Mr. J.A. Heijst from the Neurochemistry Laboratory of the VUmc for their expert technical support, and dr. S. Braesch-Andersen (MabTech AB) for sharing technical know-how on apolipoproteins. The VUmc Alzheimer Center is supported by Alzheimer Nederland (charity) and Stichting VUmc fonds (institutional support). Apolipoprotein measurements were funded with a grant from the Willem Meindert de Hoop Stichting (charity). The clinical database structure was developed with funding from Stichting Dioraphte (charity). Research of the VUmc Alzheimer Center is part of the neurodegeneration research program of the Neuroscience Campus Amsterdam.

Authors’ disclosures available online ().

References

Corder

, Saunders

, Strittmatter

, Schmechel

, Gaskell

, Small

, Roses

, Haines

, Pericak-Vance

(1993) Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Science 261, 921–923.

Strittmatter

, Saunders

ANNM

, Schmechel

, Pericak-vance

, Enghild

JAN

, Salvesen

, Roses

(1993) Apolipoprotein E: High-avidity binding to B-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease. Proc Nat Acad Sci U S A 90, 1977–1981.

(2009) Apolipoprotein E and its receptors in Alzheimer’s disease: Pathways, pathogenesis and therapy. Nat Rev Neurosci 10, 333–344.

Saczynski

, White

, Peila

, Rodriguez

, Launer

(2007) The relation between apolipoprotein A-I and dementia: The Honolulu-Asia aging study. Am J Epidemiol 165, 985–992.

Song

, Poljak

, Crawford

, Kochan

, Wen

, Cameron

, Lux

, Brodaty

, Mather

, Smythe

, Sachdev

(2012) Plasma apolipoprotein levels are associated with cognitive status and decline in a community cohort of older individuals. PLoS One 7, e34078.

Hye

, Riddoch-Contreras

, Baird

, Ashton

, Bazenet

, Leung

, Westman

, Simmons

, Dobson

, Sattlecker

, Lupton

, Lunnon

, Keohane

, Ward

, Pike

, Zucht

, Pepin

, Zheng

, Tunnicliffe

, Richardson

, Gauthier

, Soininen

, Kłoszewska

, Mecocci

, Tsolaki

, Vellas

, Lovestone

(2014) Plasma proteins predict conversion to dementia from prodromal disease. Alzheimers Dement 44, 1–9.

Sattlecker

, Kiddle

, Newhouse

, Proitsi

, Nelson

, Williams

, Johnston

, Killick

, Simmons

, Westman

, Hodges

, Soininen

, Kłoszewska

, Mecocci

, Tsolaki

, Vellas

, Lovestone

, Dobson

RJB

(2014) Alzheimer’s disease biomarker discovery using SOMAscan multiplexed protein technology. Alzheimers Dement 10, 724–734.

Dietschy

, Turley

(2001) Cholesterol metabolism in the brain. Curr Opin Lipidol 12, 105–112.

Roheim

, Careyt

, Fortet

, Vega

(1979) Apolipoproteins in human cerebrospinal fluid. Proc Nat Acad Sci U S A 76, 4646–4649.

10.

Koch

, Donarski

, Goetze

, Kreckel

, Stuerenburg

, Buhmann

, Beisiegel

(2001) Characterization of four lipoprotein classes in human cerebrospinal fluid. J Lipid Res 42, 1143–1151.

11.

Balazs

, Panzenboeck

, Hammer

, Sovic

, Quehenberger

, Malle

, Sattler

(2004) Uptake and transport of high-density lipoprotein (HDL) and HDL-associated alpha-tocopherol by an in vitro blood-brain barrier model. J Neurochem 89, 939–950.

12.

Weiler-guttler

, Sommerfeldt

, Papandrikopoulou

, Mischek

, Bonitz

, Frey

, Grupe

, Scheerer

, Gassen

(1990) Synthesis of Apolipoprotein A-1 in pig brain microvascular endothelial cells. J Neurochem 54, 444–450.

13.

Möckel

, Zinke

, Flach

, Weiß

, Weiler-Güttler

, Gassen

(1994) Expression of Apolipoprotein A-I in porcine brain endothelium. J Neurochem 62, 788–798.

14.

Panzenboeck

, Balazs

, Sovic

, Hrzenjak

, Levak-Frank

, Wintersperger

, Malle

, Sattler

(2002) ABCA1 and scavenger receptor class B, type I, are modulators of reverse sterol transport at anblood-brain barrier constituted of porcine brain capillary endothelial cells. J Biol Chem 277, 42781–42789.

15.

Koldamova

, Lefterov

, Lefterova

, Lazo

(2001) Apolipoprotein A-I directly interacts with amyloid precursor protein and inhibits Abeta aggregation and toxicity. Biochemistry 40, 3553–3560.

16.

Paula-Lima

, Tricerri

, Brito-Moreira

, Bomfim

, Oliveira

, Magdesian

, Grinberg

, Panizzutti

, Ferreira

(2009) Human apolipoprotein A-I binds amyloid-beta and prevents Abeta-induced neurotoxicity. Int J Biochem Cell Biol 41, 1361–1370.

17.

Lefterov

, Fitz

, Cronican

, Fogg

, Lefterov

, Kodali

, Wetzel

, Koldamova

(2010) Apolipoprotein A-I deficiency increases cerebral amyloid angiopathy and cognitive deficits in APP/PS1DeltaE9 mice. J Biol Chem 285, 36945–36957.

18.

Fagan

, Christopher

, Taylor

, Parsadanian

, Spinner

, Watson

, Fryer

, Wahrle

, Bales

, Paul

, Holtzman

(2004) ApoAI deficiency results in marked reductions in plasma cholesterol but no alterations in amyloid-β pathology in a mouse model of Alzheimer’s disease-like cerebral amyloidosis. Am J Pathol 165, 1413–1422.

19.

Lewis

, Cao

, Lu

, Mans

, Su

, Jungbauer

, Linton

, Fazio

, LaDu

, Li

(2010) Overexpression of human apolipoprotein A-I preserves cognitive function and attenuates neuroinflammation and cerebral amyloid angiopathy in a mouse model of Alzheimer disease. J Biol Chem 285, 36958–36968.

20.

Kawano

(1995) Marked decrease of plasma apolipoprotein AI and AII in Japanese patients with late-onset non familial Alzheimer’s disease. Clin Chim Acta 239, 209–211.

21.

Merched

, Xia

, Visvikis

, Serot

, Siest

(2000) Decreased high-density lipoprotein cholesterol and serum apolipoprotein AI concentrations are highly correlated with the severity of Alzheimer’s disease. Neurobiol Aging 21, 27–30.

22.

Liu

H-C

, Hu

C-J

, Chang

J-G

, Sung

S-M

, Lee

L-S

, Yuan

R-Y

, Leu

S-J

(2006) Proteomic identification of lower apolipoprotein A-I in Alzheimer’s disease. Dement Geriatr Cogn Disord 21, 155–161.

23.

Harr

, Uint

, Hollister

, Hyman

, Mendez

(1996) Brain expression of apolipoproteins E, J, and A-I in Alzheimer’s disease. J Neurochem 66, 2429–2435.

24.

Song

, Saito

, Seishima

, Noma

, Nakashima

(1997) Cerebrospinal fluid apo E and apo A-I concentrations in early- and late-onset Alzheimer’s disease. Neurosci Lett 231, 175–178.

25.

Demeester

, Castro

, Desrumaux

, De Geitere

, Fruchart

, Santens

, Mulleners

, Engelborghs

, De Deyn

, Vandekerckhove

, Rosseneu

, Labeur

(2000) Characterization and functional studies of lipoproteins, lipid transfer proteins, and lecithin:cholesterol acyltransferase in CSF of normal individuals andatients with Alzheimer’s disease. J Lipid Res 41, 963–974.

26.

Davidsson

, Westman-brinkmalm

CAA

, Nilsson

, Lindbjer

, Paulson

, Andreasen

, Sj

, Blennow

(2002) Proteome analysis of cerebrospinal fluid proteins in Alzheimer patients. Neuroreport 13, 611–615.

27.

Puchades

, Hansson

, Nilsson

, Andreasen

, Blennow

, Davidsson

(2003) Proteomic studies of potential cerebrospinal fluid protein markers for Alzheimer’s disease. Mol Brain Res 118, 140–146.

28.

Castaño

, Roher

, Esh

, Kokjohn

, Beach

(2006) Comparative proteomics of cerebrospinal fluid in neuropathologically-confirmed Alzheimer’s disease and non-demented elderly subjects. Neurol Res 28, 155–163.

29.

Roher

, Maarouf

, Sue

, Hu

, Wilson

, Thomas

(2009) Proteomics-derived cerebrospinal fluid markers of autopsy- confirmed Alzheimer’s disease. Biomarkers 14, 493–501.

30.

Bateman

, Xiong

, Benzinger

TLS

, Fagan

, Goate

, Fox

, Marcus

, Cairns

, Xie

, Blazey

, Holtzman

, Santacruz

, Buckles

, Oliver

, Moulder

, Aisen

, Ghetti

, Klunk

, McDade

, Martins

, Masters

, Mayeux

, Ringman

, Rossor

, Schofield

, Sperling

, Salloway

, Morris

(2012) Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N Engl J Med 367, 795–804.

31.

Jack

, Knopman

, Jagust

, Petersen

, Weiner

, Aisen

, Shaw

, Vemuri

, Wiste

, Weigand

, Lesnick

, Pankratz

, Donohue

, Trojanowski

(2013) Tracking pathophysiological processes in Alzheimer’s disease: An updated hypothetical model of dynamic biomarkers. Lancet Neurol 12, 207–216.

32.

Petersen

, Smith

, Waring

, Ivnik

, Tangalos

, Kokmen

(1999) Mild Cognitive impairment, clinical characterization and outcome. Arch Neurol 56, 303–309.

33.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

34.

Sperling

, Aisen

, Beckett

, Bennett

, Craft

, Fagan

, Iwatsubo

, Jack

, Kaye

, Montine

, Park

, Reiman

, Rowe

, Siemers

, Stern

, Yaffe

, Carrillo

, Thies

, Morrison-Bogorad

, Wagster

M V

, Phelps

(2011) Toward defining the preclinical stages of Alzheimer’s disease: Recommendations from the National Institute on Aging and the Alzheimer’s Association workgroup. Alzheimers Dement 7, 280–292.

35.

Jessen

, Amariglio

, van Boxtel

, Breteler

, Ceccaldi

, Chételat

, Dubois

, Dufouil

, Ellis

, van der Flier

, Glodzik

, van Harten

, de Leon

, McHugh

, Mielke

, Molinuevo

, Mosconi

, Osorio

, Perrotin

, Petersen

, Rabin

, Rami

, Reisberg

, Rentz

, Sachdev

, de la Sayette

, Saykin

, Scheltens

, Shulman

, Slavin

, Sperling

, Stewart

, Uspenskaya

, Vellas

, Visser

, Wagner

(2014) A conceptual framework for research on subjective cognitive decline in preclinical Alzheimer’s disease. Alzheimers Dement 10, 844–852.

36.

Van der Flier

, Pijnenburg

YAL

, Prins

, Lemstra

, Bouwman

, Teunissen

, van Berckel

BNM

, Stam

, Barkhof

, Visser

, van Egmond

, Scheltens

(2014) Optimizing patient care and research: The Amsterdam Dementia Cohort. J Alzheimers Dis 41, 313–327.

37.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

jr , Kawa

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carillo

, Thies

, Weintraub

SPC

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

38.

Mulder

, Verwey

, van der Flier

, Bouwman

, Kok

, van Elk

, Scheltens

, Blankenstein

(2010) Amyloid-beta(1-42), total tau, and phosphorylated tau as cerebrospinal fluid biomarkers for the diagnosis of Alzheimer disease. Clin Chem 56, 248–253.

39.

Friedewald

, Levy

, Fredrickson

(1972) Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 18, 499–502.

40.

Rey

(1964) L’examen clinique en psychologi (The clinical examination in psychology), 2nd ed., Presses Universitaries De France, Paris.

41.

Reitan

(1958) Validity of the Trail making test as indicator of organic brain damage. Percept Mot Skills 8, 271–276.

42.

Folstein

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

43.

Reiber

(2003) Proteins in cerebrospinal fluid and blood: Barriers, CSF flow rate and source-related dynamics. Restor Neurol Neurosci 21, 79–96.

44.

Hong

, Shi

, Chung

K a.

, Quinn

, Peskind

, Galasko

, Jankovic

, Zabetian

, Leverenz

, Baird

, Montine

, Hancock

, Hwang

, Pan

, Bradner

, Kang

, Jensen

, Zhang

(2010) DJ-1 and alpha-synuclein in human cerebrospinal fluid as biomarkers of Parkinson’s disease. Brain 133, 713–726.

45.

Huang

, Potter

, Sigurdson

, Kasten

, Connors

, Morris

, Benzinger

TLS

, Mintun

, Ashwood

, Ferm

, Budd

, Bateman

(2012) Amyloid-β dynamics in human plasma. Arch Neurol 127, 358–366.

46.

Qiang

, Wong

, Siderowf

, Hurtig

, Xie

, Lee

VMY

, Trojanowski

, Yearout

, Leverenz

, Montine

, Stern

, Mendick

, Jennings

, Zabetian

, Marek

, Chen-Plotkin

(2013) Plasma apolipoprotein A1 as a biomarker for Parkinson disease. Ann Neurol 74, 119–127.

47.

Attems

, Jellinger

(2014) The overlap between vascular disease and Alzheimer’s disease – lessons from pathology. BMC Med 12, 206.

48.

Vollbach

, Heun

, Morris

, Edwardson

, McKeith

, Jessen

, Schulz

, Maier

, Kölsch

(2005) APOA1 polymorphism influences risk for early-onset nonfamiliar AD. Ann Neurol 58, 436–441.

49.

Smach

, Edziri

, Charfeddine

, Ben Othman

, Lammouchi

, Ltaief

, Nafati

, Dridi

, Bennamou

, Limem

(2011) Polymorphism in apoA1 influences high-density lipoprotein cholesterol levels but is not a major risk factor of Alzheimer’s disease. Dement Geriatr Cogn Dis Extra 1, 249–257.

50.

Jack

, Wiste

, Weigand

, Rocca

, Knopman

, Mielke

, Lowe

, Senjem

, Gunter

, Preboske

, Pankratz

, Vemuri

, Petersen

(2014) Age-specific population frequencies of cerebral β-amyloidosis and neurodegeneration among people with normal cognitive function aged 50–89 years: A cross-sectional study. Lancet Neurol 13, 997–1005.

51.

Sunderland

, Mirza

, Putnam

, Linker

, Bhupali

, Durham

, Soares

, Kimmel

, Friedman

, Bergeson

, Csako

, Levy

, Bartko

, Cohen

(2004) Cerebrospinal fluid beta-amyloid1-42 and tau in control subjects at risk for Alzheimer’s disease: The effect of APOE epsilon4 allele. Biol Psychiatry 56, 670–676.

52.

Prince

, Zetterberg

, Andreasen

, Marcusson

, Blennow

(2004) APOE epsilon4 allele is associated with reduced cerebrospinal fluid levels of Abeta42. Neurology 62, 2116–2118.