Quantification of total apolipoprotein E and its specific isoforms in cerebrospinal fluid and blood in Alzheimer’s disease and other neurodegenerative diseases

EuPA Open Proteomics 8 (2015) 137–143

Contents lists available at ScienceDirect

EuPA Open Proteomics journal homepage: www.elsevier.com/locate/euprot

Quantification of total apolipoprotein E and its specific isoforms in cerebrospinal fluid and blood in Alzheimer’s disease and other neurodegenerative diseases Melinda Rezelia , Henrik Zetterbergb,c, Kaj Blennowb , Ann Brinkmalmb , Thomas Laurella , Oskar Hanssond , György Marko-Vargaa,* a

Division of Clinical Protein Science & Imaging, Department of Biomedical Engineering, Lund University, Lund, Sweden Clinical Neurochemistry Laboratory, Institute of Neuroscience and Physiology, The Sahlgrenska Academy at University of Gothenburg, Mölndal, Sweden c UCL Institute of Neurology, Queen Square, London WC1N 3BG, United Kingdom d Clinical Memory Research Unit, Department of Clinical Sciences, Lund University, Malmö, Sweden b

A R T I C L E I N F O

A B S T R A C T

Article history: Received 27 February 2015 Received in revised form 22 May 2015 Accepted 6 July 2015 Available online 7 August 2015

A targeted mass spectrometric assay was developed for identification and quantification of apoE isoforms (apoE2, E3 and E4), and it was utilized for screening of samples from AD patients (n = 39) and patients with other neurodegenerative disorders (n = 38). The assay showed good linearity with LOQ corresponds to total apoE concentration of 0.8 and 40 ng/mL in CSF and plasma/serum, respectively. We identified apoE phenotypes with 100% accuracy in clinical samples. We found strong association between genotypes of the individuals and their apoE levels in blood; e4 allele carriers had significantly lower apoE levels in blood than non-carriers. ã 2015 Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

Keywords: CSF biomarkers Alzheimer’s disease SRM Protein isoforms apoE

1. Introduction Alzheimer’s disease (AD) accounts for 60–70% of all cases of dementia [1]. It is a chronic neurodegenerative disorder characterized by neuronal degeneration, plaques composed of aggregated b-amyloid (Ab) and tangles composed of hyperphosphorylated tau protein [1].[21_TD$IF] Clinically, AD is characterized by an insidious onset and slowly progressive symptoms in the form of memory disturbances together with instrumental symptoms such as aphasia, agnosia and apraxia [1]. The cause of AD is unknown but most data suggest that aggregation of a 42 amino acid long amyloid b protein (Ab42) into senile plaques in the brain extracellular matrix occurs very early (10–20 years) before the first symptoms appear and may be pathogenic [2]. Genetic factors account for approximately 70% of the risk [3] with the apolipoprotein E gene ([2_TD$IF]APOE) being the major susceptibility gene for the sporadic form of the disease [4]. There are three common [23_TD$IF]APOE alleles (e2, e3 and e4) that encode three

* Corresponding author at: Division of Clinical Protein Science & Imaging, Department of Biomedical Engineering, Lund University, Lund, Sweden. Skåne University Hospital, Malmö, Sweden. E-mail address: [email protected] (G. Marko-Varga).

isoforms of the protein (E[24_TD$IF]2, E3 and E4). ApoE isoforms differ from each other by only one or two amino acids at positions 130 and 176 (112 and 158 in the mature protein) [5]. The E4 isoform appears to facilitate Ab42 aggregation into senile plaques by reducing its clearance [6], which is thought to be the underlying mechanism for the strong association of apoE e4 with AD. Plaque pathology, neurodegeneration and neurofibrillary tangle pathology can be monitored using three cerebrospinal fluid (CSF) markers, namely CSF Ab42, total tau (T-tau) and phospho-tau (Ptau), respectively [7]. Together, these markers are 85–95% sensitive and specific for AD in both pre-dementia and dementia stages of the disease. However, to date there are only a few studies investigation the association of CSF and plasma levels of apoE isoforms in the context of AD [8–10]. Identification and quantification of apoE4 in other matrices, such as astrocytes and brain tissue was reported previously [11,12]. The current study describes the development of an apoE isoform-specific mass spectrometry-based quantitative analysis. Selective reaction monitoring (SRM) assay was established and applied to CSF and blood samples from AD patients and patients with other neurodegenerative disorders in order to identify apoE phenotypes and quantify apoE levels including its specific isoforms.

http://dx.doi.org/10.1016/j.euprot.2015.07.012 1876-3820/ ã 2015 Published by Elsevier GmbH. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

138

M. Rezeli et al. / EuPA Open Proteomics 8 (2015) 137–143

2. Materials and methods

2.4. nanoLC-SRM/MS analysis

2.1. Signal peptide selection

The SRM assay was performed on a TSQ Vantage mass spectrometer equipped with an Easy n-LC II pump (Thermo Scientific, Waltham, MA). Five microliter of plasma/serum digest (4.5 nL plasma/serum equivalent) and 10 mL of CSF digest (0.45 mL original CSF equivalent) were injected onto an Easy C18-A1 precolumn (Thermo Scientific, Waltham, MA), and following on-line desalting and concentration the tryptic peptides were separated on a 75 mm  150 mm fused silica column packed in-house with ReproSil C18 (3 mm, 120 Å from Dr. Maisch GmbH, Germany). Separations were performed in a 45-min linear gradient from 10 to 40% acetonitrile containing 0.1% formic acid; at the flow rate of 300 nL/min. The MS analysis was conducted with 1750 V spray voltage and 0 V declustering potential. The transfer capillary temperature was set to 270  C and S-lens RF amplitude was set to 143 V. At least 3 transitions per target peptide were monitored in scheduled SRM mode with 3-min detection windows. SRM transitions were acquired in Q1 and Q3 operated at unit resolution (0.7 FWHM), the collision gas pressure in [25_TD$IF]q2 was set to 1.2 mTorr. The cycle time was 1.5 s. Identical SRM acquisition parameters were used for the heavy and light forms of each peptide, while taking into account the Q1/ Q3 mass differences due to the stable isotope labeling. The three transitions that produce the highest signals while still maintaining the same relative intensities in both buffer and biological samples were selected as the final ion pairs for use for quantification.

Tryptic peptides unique for isoforms E2 (CLAVYQAGAR) and E4 (LGADMEDVR), and peptides common in all the three isoforms, in total 9 sequences (Table 1), were selected as signal peptides and used for the MRM assay development. Heavy peptides, isotope labeled with 15N and 13C in lysine (Dmass = +8) and arginine (Dmass = +10) of AQUA QuantPro quality (peptide purity higher than 97%, concentration precision equal or better than 25%), were purchased from Thermo Fischer Scientific (Ulm, Germany). These heavy isotope labeled peptides were spiked into the biological samples at known concentrations and the ratio between endogenous (light) and internal standard peptide was used to calculate the concentration of apoE in the samples. 2.2. Clinical materials Cohort A included plasma and cerebrospinal fluid (CSF) samples obtained at the Memory Clinic at Skåne University Hospital in Malmö (Sweden) from subjects with either Alzheimer’s disease (n = 13) or individuals with subjective cognitive decline but no dementia (SCD) diagnoses at sampling (n = 12). Cohort B included serum samples from AD patients (n = 26) and patients with other neurodegenerative diseases (n = 25; 13 with mixed AD/vascular dementia, 3 with vascular dementia, 3 with subjective memory complaints, 2 with frontotemporal dementia, 2 with Lewy body dementia, 1 with Parkinson’s disease and 1 with unspecified dementia). See Table 4 for patient characteristics. The clinical studies were approved by the Universities of Lund (Cohort A) and Umeå (Cohort B). All study enrollments followed the recommendations of the Declaration of Helsinki. 2.3. Sample preparation 100 mL of CSF or 10 mL of 10 times diluted plasma/serum were diluted with 50 mM AMBIC containing sodium deoxycholate (SDC) in a final concentration of 2 m/v%; and reduced with 10 mM dithiothreitol (60 min, 37  C) then alkylated with 50 mM iodoacetamide (30 min, at room temperature in dark). 50 mM AMBIC was added to the samples to reduce the SDC concentration to 0.5 m/v% before trypsin addition. The samples were digested with sequence grade trypsin (ca. 1:50 enzyme: total protein ratio) 18 h at 37  C and then the digestion was stopped by the addition of 10 v/v% formic acid. The acidified digests were centrifuged at room temperature for 5 min at 15,000  g to effectively pellet SDC and remove it from the samples. The digested samples were spiked with a mixture of heavy isotope-labeled peptide standards and finally analyzed by nanoLCSRM/MS.

Table 1 ApoE peptides used for SRM assay development. Position

Sequence

94–108 122–132 122–130 176–185 177–185 199–207 210–224 270–278 261–269

SELEEQLTPVAEETR LGADMEDVCGR LGADMEDVR CLAVYQAGAR LAVYQAGAR LGPLVEQGR AATVGSLAGQPLQER LQAEAFQAR LEEQAQQIR

2.5. Data analysis All raw data generated on the triple quadrupole mass spectrometer were imported to Skyline v2.5 software (MacCoss Lab.) for data analysis. Quantification was based on the corresponding light and heavy peak area ratios. The peak integration was done automatically by the software, using Savitzky–Golay smoothing, and all the data were manually inspected to confirm the correct peak detection. Further statistical analysis was done using Microsoft Excel, R and Matlab v7.11 (Mathworks, Natick, MA). 3. Results and discussion We have developed an SRM assay for the quantification of apoE isoforms by using a stable isotope dilution strategy. Isoform specific sequences and sequences common in each isoform were selected and synthesized in heavy form with stable isotope labeled C-terminal Arg or Lys. These peptides were used to optimize the SRM assay parameters, such as, the chromatographic separation and the selection of best transitions, free from interferences in complex biological media. We investigated the assay performance and the sample preparation in two biological matrices. The developed assay has been utilized for screening of both blood and CSF samples from AD and non-AD subjects; and additionally statistical analysis were done to find correlations between the experimental and clinical data. 3.1. Reproducibility and precision of the assay

Common E2, E3 E4 E2 E3, E4 Common Common Common Common

Both isoform specific peptides and peptides common in all apoE isoforms (Table 1) were targeted in the assay in order to measure the level of different isoforms and total apoE in biological samples. For detailed information about the monitored transitions see Supplementary Table 1. The linearity of the SRM assay was determined by spiking a mixture of heavy labeled internal standard (IS) peptides into a pooled sample of 18 [26_TD$IF]CSF samples at 8 different levels. The analysis of each sample was repeated 5 times and then the area ratios of

M. Rezeli et al. / EuPA Open Proteomics 8 (2015) 137–143

heavy labeled standard and corresponding endogenous peaks were plotted against the spiked IS concentrations (Supplementary Fig. 1). Excellent linear regressions were achieved for each peptide within the investigated concentration range (0.001–50 fmol/mL) with LOQ varied in the attomolar range (4–80 attomoles on column) (Supplementary Fig. 1). The LOQ of the apoE peptides by SRM analysis was estimated as the lowest concentration measured with CV <20% in 5 replicates. The best peptide (LGPLVEQGR) was used to calculate LOQ of the assay that corresponds to an apoE concentration of 0.8 and 40 ng/mL in CSF and plasma/serum, respectively. In order to evaluate the analytical performance of the SRM assay, including tryptic digestion and sample spiking with IS peptides we have investigated some key parameters. The variation of the LC-MRM/MS analysis was determined in 6 randomly selected CSF and plasma samples analyzed in triplicates. The CVs of the triplicate MS analyses ranged between 0.16% and 7.04% and the average was 2.28% in the overall dataset, [27_TD$IF] i.e., in the 12 investigated samples. Table 2 shows the average CV values for each peptide in the two different biological matrices. The reproducibility was slightly better in CSF for almost all peptides, except CLAVYQAGAR. Considerably lower variations were detected in CSF than in plasma with LGADMEDVCGR, LGADMEDVR and LEEQAQQIR peptides that most probably due to the higher complexity of plasma compared to CSF. Digestion replicates were also generated by digesting three individual aliquots from 5 selected CSF and plasma samples that were spiked with a mixture of heavy IS peptides and analyzed by nanoLC-MRM/MS. The overall variation of the sample processing was less than 20% for all target peptides, and it was only a slight difference between the distinct peptides in CSF (Table 2). In contrast, two peptides (SELEEQLTPVAEETR and LEEQAQQIR) showed high variability in the plasma samples. These two peptides seem to be more sensitive to slight changes in the digestion conditions and because of the higher risk of imprecision we excluded them from the final quantification. This finding is supported by our previous experiment, where different enzyme: protein ratios were investigated. We examined four different trypsin: protein ratios (1:100, 1:50, 1:25 and 1:12.5) and we found continuous increase in the endogenous peptide signals by increasing the trypsin amount only in the case of two peptides (SELEEQLTPVAEETR and LEEQAQQIR), all the other peptides gave approximately the same signals at different conditions (Supplementary Fig. 2). 3.2. Quantification of apoE isoforms and total apoE The endogenous levels of apoE peptides were calculated by taking the ratio between the peak areas of the endogenous (light)

139

and IS peptides (heavy) and correlate to the known concentration of the heavy peptides that were spiked into the samples. The calculations were made for each peptide individually. Total apoE amounts were obtained in two different ways, (1) by calculating directly from a peptide represents all the three isoforms (LGPLVEQGR, AATVGSLAGQPLQER and LQAEAFQAR) or (2) by adding the concentrations of each isoform (LGADMEDVCGR (E2 and 3) + LGADMEDVR (E4) and LAVYQAGAR (E3 and 4) + CLAVYQAGAR (E2)). The linear regression coefficients between total apoE concentrations calculated by different peptides were excellent with R2-values ranging between 0.75–0.97, however the absolute amounts were slightly different. Fig. 1 shows some examples of the correlations between different peptide signals in CSF (A and B) in plasma (C and D) and in serum (E and F). For all possible combinations see Supplementary Fig. 3. The differences between peptides could be attributed to alterations in digestion efficiency of different regions of apoE, and to variances of IS peptide concentrations; according to the manufacturer the concentration precision is equal or better than 25%. The total apoE concentration achieved from the sum of peptides LGADMEDVCGR and LGADMEDVR was lower than the sum of the other two specific peptides LAVYQAGAR and CLAVYQAGAR in all the three sample types, however they showed similar correlation with other peptide signals (Supplementary Fig. 3). This data suggest that using the signal of peptide LGADMEDVR will lead to an underestimation of the amount of apoE4 isoform. To address this problem we calculated a correction factor for LGADMEDVR by comparing the signals of LGADMEDVR (E4) and LAVYQAGAR (E3 and 4) in E4/E4 homozygous individuals. A specific correction factor was assigned to each sample type and used for calculation of apoE4 concentration. The concentration of apoE2 was calculated from the signal of peptide CLAVYQAGAR. Because there is no isoform specific peptide available for apoE3, its amount was calculated by subtracting the concentration of peptide LGADMEDVR (apoE4) from peptide LAVYQAGAR (apoE3 and 4) in samples with apoE3/E4 phenotypes, otherwise (for phenotypes apoE3/E3 and E2/E3) LAVYQAGAR was used. Peptide LGADMEDVCGR was used only for determination of the phenotypes. See Supplementary Fig. 2 for the summary of apoE quantification in the two cohorts. We further investigated the heterozygous individuals and found very similar isoform distribution in heterozygous individuals to what it was reported formerly using a similar mass spectrometry-based assay [13]. The distribution was 52/48% between E3 and E4 isoforms in CSF in [28_TD$IF]APOE e3/4 individuals, while in plasma and serum it was 66/34% and 63/37%, respectively. ApoE2 and ApoE3 isoforms have almost equal contribution to the total apoE level in plasma and serum in the heterozygous

Table 2 Analytical performance of the SRM assay in CSF and plasma.

SELEEQLTPVAEETRc LGADMEDVCGR LGADMEDVR CLAVYQAGAR LAVYQAGAR LGPLVEQGR AATVGSLAGQPLQER LQAEAFQAR LEEQAQQIRc a b c

Reproducibility of the MS analysis (average CV[15_TD$IF]%a )

Reproducibility of the sample processing (average CV[16_TD$IF]%b)

CSF

Plasma

CSF

Plasma

1.30 1.97 0.43 2.69 1.67 1.92 1.71 1.24 2.51

1.98 5.63 3.03 1.12 2.14 2.07 1.89 2.41 4.18

6.41 5.78 4.00 4.47 2.36 2.48 3.03 2.69 3.37

19.95 4.13 7.65 4.40 2.86 3.59 4.06 4.09 10.44

The CV value was calculated for each peptide based on triplicate MS analysis of 6 individual CSF and plasma samples. The CV was calculated for each peptide based on triplicate digestion of 5 individual CSF and plasma samples. Peptides excluded from the final SRM assay.

140

[(Fig._1)TD$IG]

M. Rezeli et al. / EuPA Open Proteomics 8 (2015) 137–143

Fig. 1. Correlations between total apoE concentrations measured by different peptides in CSF (A and B), in plasma (C and D) and in serum (E and F). Total apoE concentrations were achieved from peptides, common in each apoE isoforms (LGPLVEQGR and LQAEAFQAR) or calculated by summing the signals of peptides representing different isoforms (LAVYQAGAR (apoE3 and 4) + CLAVYQAGAR (apoE2)).

individuals (54/46% and 52/48%), while in CSF the proportion of E3 isoform is more pronounced (61/39%). The differences in isoform distribution in between CSF and blood, additionally the lack of significant association between CSF and plasma total apoE levels support the concept [14] that apoE from blood does not exchange with apoE from brain because of the blood–brain barrier. 3.3. ApoE phenotypes and concentration in clinical samples In this study we analyzed blood samples collected from two cohorts of AD and non-AD subjects, additionally CSF samples from patients belonging to the first cohort were also analyzed. In order to identify different phenotypes we monitored four peptide signals (LGADMEDVCGR (apoE2 [29_TD$IF]and 3), LGADMEDVR (apoE4), LAVYQAGAR (apoE3 and 4), CLAVYQAGAR (apoE2)) each and every phenotype has a specific pattern (Supplementary Fig. 4) that allows phenotyping of biological samples. We managed to

identify five different phenotypes (E2/E3, E2/E4 and E3/ E4 heterozygotes; additionally E3/E3 and E4/E4 homozygotes) in our sample cohorts (four types in the first cohort and 5 types in the second cohort) (Table 3); our sample characterization matched well with genotyping data, 100% agreement was found between the two methods. Our data generated from plasma samples belonging to the first cohort indicated a strong relationship between [30_TD$IF]APOE genotypes of the subjects and their apoE levels in plasma. Patients carrying e4 allele had significantly lower apoE levels in blood than noncarriers (Fig. 2A and B). The relationship between genotypes and apoE levels was allele dose-dependent; more specifically the protein level of apoE in e4 non-carriers was the highest, in heterozygotes e4 carriers the expression was lower and in homozygotes e4 carriers the expression was the lowest (e4 / > e4 /+ > e4+/+). In contrast there was no clear genotypedependent relationship in CSF (Supplementary Fig. 5).

M. Rezeli et al. / EuPA Open Proteomics 8 (2015) 137–143 Table 3 ApoE phenotype frequencies in the two cohorts. Study

ID

E4 freq. (%)

E2/E3

E2/E4

Cohort A

AD n = 13 Non-AD n = 13 Total

62

2

–

3

4

4

38

3

–

5

4

1

50

5

–

8

8

5

65

3

1

6

7

9

40

5

1

10

5

4

53

8

2

16

12

13

Cohort B

AD n = 26 Non-AD n = 25 Total

E3/E3

E3/E4

E4/E4

141

In order to validate our observation on the apoE genotype dependent relationship in blood, we analyzed another cohort (cohort B) with a larger sample number. We detected similar tendency in the apoE levels in relation to e4 allele carriers and noncarriers in the second cohort (Fig. 2C and D). One outlier with e4/ e4 genotype, which gave extremely high total apoE concentration (174.5 mg/mL vs. 46.1 19.0 mg/mL), was removed from the final statistical analysis. This finding about the genotype dependence of apoE level in blood is in agreement with previous studies [15,16,8,9]. Inconsistent results were published previously in regard to apoE concentration differences in CSF between different apoE genotypes [17,18]. Our data showed no association between [31_TD$IF]APOE genotype

[(Fig._2)TD$IG]

Fig. 2. Total apoE concentration in plasma samples from cohort A (A and B) and in serum samples from cohort B (C and D). Samples are grouped by phenotypes on Fig 2A and C (E2/E3: heterozygous apoE2/3 individuals, E2/E4: heterozygous apoE2/4 individuals, E3/E3: homozygous apoE3/3 individuals, E3/E4: heterozygous apoE3/4 individuals and E4/E4: homozygous apoE4/4 individuals) and by e4 status on B and D (E4 / : e4 non-carriers, E4 /+: heterozygous e4 carriers, E4+/+: homozygous e4 carriers). Total apoE concentration was calculated from the measured LGPLVEQGR peptide signal.

142

M. Rezeli et al. / EuPA Open Proteomics 8 (2015) 137–143

Table 4 Patient characteristics in the two cohorts. Study cohort

ID

Age

Gender% (M/F)

MMSE score

QAlb

CSF T-tau (ng/L)

CSF P-tau (ng/L)

CSF Ab42 (ng/L)

Cohort A

AD n = 13 Non-AD n = 13

75.3 (64–83)

21/79

26.5 (25–29)

8.1 (6.1–16.2)

685 (354–900)[17_TD$IF]a

91 (46–119)b

469 (353–562)

79.2 (57–89)

62/38

25.1 (21–29)

12.2 (5.4–41.7)

408 (180–730)

[18_TD$IF]a

b

502 (170–690)

AD n = 26 Non-AD n = 25

75.3 (61–89)

54/46

25.0 (11–30)

6.1 (2.0–32.0)

591 (264–1110)

–

534 (289–1060)

74.6 (58–88)

44/56

23.9 (17–30)

7.0 (3.1–13.7)

651 (252–1420)

–

634 (272–1410)

Cohort B

[19_TD$IF]a b

significant difference at p < 0.001 level. significant difference [20_TD$IF]at p < 0.005 level.

and apoE level in CSF, which is in line with a previous study from Martinez-Morillo et al. [9]. 3.4. Comparison of apoE levels with clinical markers We analyzed the relationship between clinical markers (T-tau, P-tau and Ab42) of AD and the determined apoE levels in CSF and blood, additionally we investigated a possible link between clinical diagnoses and apoE level. However, we are aware that the size of our clinical cohorts may be too small to draw consistent conclusions. The demographic data of subjects involved in this study is presented in Table 4. The role of apoE in AD pathogenesis is unquestionable, however the results from studies considering apoE as potential biomarker are inconsistent and contradictory. ApoE levels have been measured both in CSF and in blood previously. Some studies reported decreased apoE levels in AD patients in plasma [16,19] and also in CSF [20], while elevated apoE levels were also measured in AD [32_TD$IF][21,22,16]. Additionally, some studies could not find differences between controls and AD patients [8,23,24]. The contradictions may arise from the use of different methods, variances in sample collection and handling or selection of the patient cohorts. Our results indicated no significant differences of the apoE level in between AD and non-AD groups neither in CSF nor in plasma (p = 0.361, p = 0.333). The most probable explanation for the inability to differentiate healthy from AD patients is the relatively

[(Fig._3)TD$IG]

56 (29–104)

A

small sample number and the large heterogeneity within groups. The unequal gender distribution between control and AD group may also be a possible cause, as higher blood apoE level was previously reported in females older than 50 [8,25]. Our results also showed slightly higher apoE concentrations in female subjects (35.7 vs. 33.24 mg/mL), although this was not significant. The analysis of the second cohort showed the same results, [27_TD$IF]i.e., no alterations in apoE level between the two patient groups (p = 0.244). These results confirm the previous reports that the determination of total apoE biofluid level itself is not supporting the Alzheimer’s diagnosis [8,24,26]. We found that total apoE concentrations in CSF were positively associated with CSF P-tau in e4 non-carriers (s = 0.881, p = 0.027) but not in e4 carriers (s = 0.104, p = 0.737). This finding is inconsistent with previous publications [9,27], where genotype independent correlations were reported between CSF apoE and tau levels (both Tand P-tau). Further, a weaker correlation between total CSF apoE and CSF Ab1-42 (s = 0.385, p = 0.105) was detected in e4 carriers contrary to e4 non-carriers (s = 0.310, p = 0.354), which is in accordance with a previous report from Martinez-Morillo et al. [9]. It was found that apoE involved in Ab oligomerization [17,27] where E4 isoform promotes Ab42 aggregation into senile plaques by reducing its clearance [6]. Our results, however the association was not significant, also suggest isoform-specific interaction between apoE and Ab. No associations were found between AD biomarkers and apoE concentration in plasma, although a slight decrease in plasma apoE

B

Fig. 3. Association between CSF Ab42 and apoE levels in plasma (A) and in CSF (B). Individuals were divided into two groups based on CSF Ab42 level (low and high), cut-off level was set to 500 pg/mL. Non-AD subjects are represented by red dots and AD patients are represented by blue dots.

M. Rezeli et al. / EuPA Open Proteomics 8 (2015) 137–143

level was detected in the group of patients with low CSF Ab42 level in comparison with the high level group (Fig. 3). The same cut-off level of Ab42 (500 pg/mL) was used to distinguish the two groups as in previous studies [9,28]. In contrary, we found no difference in CSF apoE levels between the same groups of subjects, which is in contrast with former publication from Cruchaga et al. [18]. 4. Conclusions In the current study we presented a highly specific mass spectrometry-based assay for reliable quantification of apoE including its specific isoforms in biological fluids. The assay allows apoE phenotype determination with 100% accuracy. We showed that neither CSF nor blood level of total apoE is enable to differentiate AD patients from non-AD subjects. Additionally we demonstrated that [28_TD$IF]14APOE e4 carriers have lower levels of apoE in blood independent of clinical diagnosis, which is in accordance with previous studies [8,9]. We further showed correlation between CSF apoE and P-tau levels in e4 non-carriers, and a weak association between CSF apoE and CSF Ab1-42 concentrations in e4 carriers only. Further studies are needed to clarify contradictions in publications concerning apoE fluid levels in relation to AD, in addition to determine the mechanisms by which apoE isoforms are involved in AD development. Acknowledgements This work was supported by grants from the Swedish Research Council, grant # 14,002,Linnaeus Bagadilico, Swedish Academy of Pharmaceutical Sciences, Vinnova, Ingabritt & Arne Lundbergs forskningsstiftelse, the Crafoord Foundation, the Knut and Alice Wallenberg Foundation, the Alzheimer Foundation, the Torsten Söderbergs Stiftelse at the Royal Swedish Academy of Sciences and Hjärnfonden. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.euprot.2015.07.012. References [1] K. Blennow, M.J. de Leon, H. Zetterberg, Alzheimer’s disease, Lancet 368 (2006) 387–403. [2] J. Hardy, N. Bogdanovic, B. Winblad, E. Portelius, N. Andreasen, A. CedazoMinguez, et al., Pathways to Alzheimer’s disease, J. Intern. Med. 275 (2014) 296–303. [3] M. Gatz, C.A. Reynolds, L. Fratiglioni, B. Johansson, J.A. Mortimer, S. Berg, et al., Role of genes and environments for explaining Alzheimer disease, Arch. Gen. Psychiatry 63 (2006) 168–174. [4] V. Chouraki, S. Seshadri, Genetics of Alzheimer’s disease, Adv. Genet. 87 (2014) 245–294. [5] K.H. Weisgraber, S.C. Rall Jr., R.W. Mahley, Human E poprotein heterogeneity. Cysteine–arginine interchanges in the amino acid sequence of the apo-E isoforms, J. Biol. Chem. 256 (1981) 9077–9083. [6] D.M. Holtzman, J. Herz, G. Bu, Apolipoprotein E and apolipoprotein E receptors: normal biology and roles in Alzheimer disease, Cold Spring Harb. Perspect. Med. (2012) 2012.

143

[7] K. Blennow, H. Hampel, M. Weiner, H. Zetterberg, Cerebrospinal fluid and plasma biomarkers in Alzheimer disease, Nat. Rev. Neurol. 6 (2010) 131–144. [8] R. Simon, M. Girod, C. Fonbonne, A. Salvador, Y. Clemen, P. Lanteri, et al., Total ApoE and ApoE4 isoform assays in an Alzheimer’s disease case-control study by targeted mass spectrometry (n = 669): a pilot assay for methioninecontaining proteotypic peptides, Mol. Cell. Proteomics 11 (2012) 1389–1403. [9] E. Martinez-Morillo, O. Hansson, Y. Atagi, G. Bu, L. Minthon, E.P. Diamandis, et al., Total apolipoprotein E levels and specific isoform composition in cerebrospinal fluid and plasma from Alzheimer’s disease patients and controls, Acta Neuropathol. 127 (2014) 633–643. [10] B. Krastins, A. Prakash, D.A. Sarracino, D. Nedelkov, E.E. Niederkofler, U.A. Kiernan, et al., Rapid development of sensitive, high-throughput, quantitative and highly selective mass spectrometric targeted immunoassays for clinically important proteins in human plasma and serum, Clin. Biochem. 46 (2013) 399–410. [11] K.R. Wildsmith, B. Han, R.J. Bateman, Method for the simultaneous quantitation of apolipoprotein E isoforms using tandem mass spectrometry, Anal. Biochem. 395 (2009) 116–118. [12] M. Wang, J. Chen, I.V. Turko, N-15-labeled full-length apolipoprotein E4 as an internal standard for mass spectrometry quantification of apolipoprotein E isoforms, Anal. Chem. 84 (2012) 8340–8344. [13] E. Martinez-Morillo, H.M. Nielsen, I. Batruch, A.P. Drabovich, I. Begcevic, M.F. Lopez, et al., Assessment of peptide chemical modifications on the development of an accurate and precise multiplex selected reaction monitoring assay for apolipoprotein E isoforms, J. Proteome Res. 13 (2014) 1077–1087. [14] K.R. Wildsmith, J.M. Basak, B.W. Patterson, Y. Pyatkivskyy, J. Kim, K.E. Yarasheski, et al., In vivo human apolipoprotein E isoform fractional turnover rates in the CNS, PLoS One 7 (2012) . [15] M. Smit, P. Deknijff, M. Rosseneu, J. Bury, E. Klasen, R. Frants, et al., Apolipoprotein-e polymorphism in the Netherlands and its effect on plasma– lipid and apolipoprotein levels, Hum. Genet. 80 (1988) 287–292. [16] V.B. Gupta, S.M. Laws, V.L. Villemagne, D. Ames, A.I. Bush, K.A. Ellis, et al., Plasma apolipoprotein E and Alzheimer disease risk The AIBL study of aging, Neurology 76 (2011) 1091–1098. [17] T. Darreh-Shori, A. Forsberg, N. Modiri, N. Andreasen, K. Blennow, C. Kamil, et al., Differential levels of apolipoprotein E and butyrylcholinesterase show strong association with pathological signs of Alzheimer’s disease in the brain in vivo, Neurobiol. Aging 32 (2011) . [18] C. Cruchaga, J.S.K. Kauwe, P. Nowotny, K. Bales, E.H. Pickering, K. Mayo, et al., Cerebrospinal fluid APOE levels: an endophenotype for genetic studies for Alzheimers disease, Hum. Mol. Genet. 21 (2012) 4558–4571. [19] A.J.C. Slooter, P. de Knijff, A. Hofman, M. Cruts, M.M.B. Breteler, C. Van Broeckhoven, et al., Serum apolipoprotein E level is not increased in Alzheimer’s disease: the Rotterdam study, Neurosci. Lett. (1998) 21–24. [20] K. Blennow, C. Hesse, P. Fredman, Cerebrospinal-fluid apolipoprotein-e is reduced in Alzheimers-disease, Neuroreport 5 (1994) 2534–2536. [21] A. Merched, H. Blain, S. Visvikis, B. Herbeth, C. Jeandel, G. Siest, Cerebrospinal fluid apolipoprotein E level is increased in late-onset Alzheimer’s disease, J. Neurol. Sci. 145 (1997) 33–39. [22] H. Song, K. Saito, M. Seishima, A. Noma, K. Urakami, K. Nakashima, Cerebrospinal fluid apo E and apo A-I concentrations in early- and late-onset Alzheimer’s disease, Neurosci. Lett. 231 (1997) 175–178. [23] S. Hahne, C. Nordstedt, A. Ahlin, H. Nyback, Levels of cerebrospinal fluid apolipoprotein E in patients with Alzheimer’s disease and healthy controls, Neurosci. Lett. 224 (1997) 99–102. [24] R.J.L. Kandimalla, W.Y. Wani, R. Anand, A. Kaushal, S. Prabhakar, V.K. Grover, et al., Apolipoprotein E levels in the cerebrospinal fluid of North Indian patients with Alzheimer’s disease, Am. J. Alzheimers Dis. Other Demen. 28 (2013) 258–262. [25] F. Schiele, D. De Bacquer, M. Vincent-Viry, U. Beisiegel, C. Ehnholm, A. Evans, E. et al. Apolipoprotein, serum concentration and polymorphism in six European countries: the ApoEurope Project, Atherosclerosis 152 (2000) 475–488. [26] C. Schmidt, N. Gerlach, C. Peter, K. Gherib, K. Lange, T. Friede, et al., Cerebrospinal fluid apolipoprotein E concentration and progression of Alzheimer’s disease, J. Alzheimers Dis. 43 (2015) 1229–1236. [27] J.B. Toledo, X. Da, M.W. Weiner, D.A. Wolk, S.X. Xie, S.E. Arnold, et al., CSF Apo-E levels associate with cognitive decline and MRI changes, Acta Neuropathol. 127 (2014) 621–632. [28] A.M. Fagan, M.A. Mintun, R.H. Mach, S.Y. Lee, C.S. Dence, A.R. Shah, et al., Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid A beta(42) in humans, Ann. Neurol. 59 (2006) 512–519.