Intact Protein Analysis of Ubiquitin in Cerebrospinal Fluid by Multiple

Aug 5, 2014 - In conclusion, our LC–MS/MS method may facilitate ubiquitin determination to a broader community and might help to discriminate AD, CJ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/jpr

Intact Protein Analysis of Ubiquitin in Cerebrospinal Fluid by Multiple Reaction Monitoring Reveals Differences in Alzheimer’s Disease and Frontotemporal Lobar Degeneration Patrick Oeckl, Petra Steinacker, Christine A. F. von Arnim, Sarah Straub, Magdalena Nagl, Emily Feneberg, Jochen H. Weishaupt, Albert C. Ludolph, and Markus Otto* Department of Neurology, Ulm University Hospital, Oberer Eselsberg 45, 89081 Ulm, Germany ABSTRACT: The impairment of the ubiquitin-proteasome system (UPS) is thought to be an early event in neurodegeneration, and monitoring UPS alterations might serve as a disease biomarker. Our aim was to establish an alternate method to antibody-based assays for the selective measurement of free monoubiquitin in cerebrospinal fluid (CSF). Free monoubiquitin was measured with liquid chromatography−multiple reaction monitoring mass spectrometry (LC−MS/ MS) in CSF of patients with Alzheimer’s disease (AD), amyotrophic lateral sclerosis (ALS), behavioral variant of frontotemporal dementia (bvFTD), Creutzfeldt−Jakob disease (CJD), Parkinson’s disease (PD), primary progressive aphasia (PPA), and progressive supranuclear palsy (PSP). The LC−MS/MS method showed excellent intra- and interassay precision (4.4−7.4% and 4.9− 10.3%) and accuracy (100−107% and 100−106%). CSF ubiquitin concentration was increased compared with that of controls (33.0 ± 9.7 ng/mL) in AD (47.5 ± 13.1 ng/mL, p < 0.05) and CJD patients (171.5 ± 103.5 ng/mL, p < 0.001) but not in other neurodegenerative diseases. Receiver operating characteristic curve (ROC) analysis of AD vs control patients revealed an area under the curve (AUC) of 0.832, and the specificity and sensitivity were 75 and 75%, respectively. ROC analysis of AD and FTLD patients yielded an AUC of 0.776, and the specificity and sensitivity were 53 and 100%, respectively. In conclusion, our LC−MS/MS method may facilitate ubiquitin determination to a broader community and might help to discriminate AD, CJD, and FTLD patients. KEYWORDS: Ubiquitin, cerebrospinal fluid, MRM, mass spectrometry, biomarker, Alzheimer’s disease, frontotemporal lobar degeneration, amyotrophic lateral sclerosis, Creutzfeldt−Jakob disease, Parkinson’s disease



INTRODUCTION

promising strategy in biomarker identification. Cerebrospinal fluid (CSF) is the preferred biological fluid for the measurement of biomarkers in neurological disorders because it is in direct contact with the brain and is routinely collected during the diagnosis of neurological syndromes. Several studies investigated ubiquitin concentration in CSF using different antibody-based assays or proteomic approaches, and increased ubiquitin concentration was reported for AD,4,5 vascular dementia,6 Creutzfeldt−Jakob disease (CJD),7 and progressive supranuclear palsy (PSP)8 patients. Furthermore, subgroups of AD patients have been defined using a combination of CSF biomarkers including ubiquitin,9 which might be helpful in individualized therapy. However, ubiquitin concentrations vary considerably among control patients between studies,6,10−12 and there is uncertainty about what type of ubiquitin (e.g., mono- or polyubiquitin) is detected with the different immunoassays, highlighting the

Ubiquitin is a protein of 8.6 kDa, and its amino acid sequence is highly conserved in eukaryotes. It is attached to lysine residues of proteins as a mono- or polyubiquitin chain by a series of consecutively acting enzymes, E1−E3. Post-translational modification of proteins with ubiquitin is involved in different mechanisms including endocytosis, protein degradation, subcellular localization, and signal transduction. 1,2 One important function of ubiquitin is to mark proteins for degradation by the proteasome (ubiquitin-proteasome system, UPS) which is essential in protein quality control and regulation of protein function.3 Ubiquitin-positive, proteinaceous aggregates are a common histopathological feature of neurodegenerative diseases such as Alzheimer’s disease (AD) and Parkinson’s disease (PD). Protein aggregation is a key pathologic event in these disorders, and an impairment in protein clearance by the UPS is thought to contribute to neurodegeneration.3 To date, diagnosis of neurodegenerative diseases in an early disease state is not possible. Since the impairment of the UPS might be an early event in the pathogenesis of neurodegenerative diseases, monitoring UPS function could be a © XXXX American Chemical Society

Special Issue: Proteomics of Human Diseases: Pathogenesis, Diagnosis, Prognosis, and Treatment Received: June 16, 2014

A

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX

Journal of Proteome Research

Article

Table 1. Characteristics of Patientsa diagnosis Control (n = 28) ALS (n = 20) PD (n = 25) PSP (n = 17) bvFTD (n = 15) PPA (n = 6) Total FTLD (n = 38) AD (n = 20) CJD (n = 19)

age (years)b 63.5 57.5 70.0 64.0 60.0 72.0 64.0

(56.3−72.3) (51.0−67.8) (63.0−74.0) (57.0−70.0) (55.0−70.0) (66.0−74.3) (56.0−71.3)

70.0 (61.3−72.0) 67.0 (59.0−74.0)

gender (m/f)

ubiquitin (ng/mL)b

10/18 14/6 17/8 11/6 8/7 3/3 22/16

32.2 31.3 32.3 30.6 33.6 28.4 31.1

9/11 8/11

43.6 (36.3−55.2)c,d 130 (84.3−262)e

(26.7−37.5) (24.3−36.1) (25.1−43.4) (27.2−42.7) (24.7−45.0) (22.9−37.2) (25.3−42.6)

T-Tau (pg/mL)b

Aβ42 (pg/mL)b

P-Tau (pg/mL)b

226 (178−339)

983 (835−1048)

46.0 (30.3−55.8)

225 319 223 250

983 990 981 990

39.0 54.0 37.5 40.5

(145−348) (241−468) (165−266) (169−393)

784 (445−1136)e,f 6140 (3490−11 730)e

(675−1160) (751−1123) (875−1120) (731−1133)

486 (419−527)e,f

(27.0−57.0) (39.0−77.0) (31.3−40.8) (30.8−58.0)

47.5 (36.0−57.8)

a

Groups were compared by Kruskal−Wallis test and Dunn’s posthoc test. ALS, amyotrophic lateral sclerosis; PD, Parkinson’s disease; PSP, progressive supranuclear palsy; bvFTD, behavioral variant of frontotemporal dementia; PPA, primary progressive aphasia; FTLD, frontotemporal lobar degeneration (includes PSP, bvFTD, and PPA); AD, Alzheimer’s disease; CJD, Creutzfeld−Jakob disease; m, male; f, female; T-Tau, total Tau; P-Tau, phosphorylated Tau. bMedian and interquartile range. cp < 0.05 compared with control. dp < 0.01 compared with FTLD. ep < 0.001 compared with control. fp < 0.001 compared with FTLD.

which includes PPA, PSP, and bvFTD) were diagnosed as described by Rascovsky et al.16 and Gorno-Tempini et al.17 AD patients fulfilled the NINCDS−ADRDA (National Institute of Neurological and Communicative Disorders and Stroke− Alzheimer’s Disease and Related Disorders Association) criteria, and CJD was confirmed neuropathologically according to the WHO consensus criteria.18 The Ethics Committees of the Ulm University Hospital and University of Göttingen approved the collection and analysis of the CSF samples. All patients or their relatives provided written consent to be included in this study.

need for more reliable and selective methods. Moreover, other neurodegenerative diseases have not been investigated so far, which is important in terms of the selectivity of this biomarker. The aim of the present study was to overcome the selectivity and reliability problems of the described immunoassays by establishing a method to determine free monoubiquitin with liquid chromatography−multiple reaction monitoring mass spectrometry (LC−MS/MS) in CSF and to measure CSF of patients with AD, CJD, amyotrophic lateral sclerosis (ALS), PD, PSP, behavioral variant of frontotemporal dementia (bvFTD), and primary progressive aphasia (PPA).



Stock Solutions, Calibration Standards, and Quality Control (QC) Samples

EXPERIMENTAL SECTION

Chemicals and Reagents

Ubiquitin and 13C-ubiquitin were corrected for purity and dissolved in water at 500 μg/mL. Stock solutions were stored at −80 °C. Calibration standards and QC samples were prepared by diluting the ubiquitin stock solution in aCSF containing 250 μg/mL bovine serum albumin (BSA). Eight different calibration standards were prepared at concentrations of 2, 5, 10, 20, 50, 100, 150, and 200 ng/mL. QC samples (low, medium, and high) had concentrations of 6, 90, and 170 ng/ mL, respectively. Additionally, a QCLLOQ-sample (2 ng/mL) was prepared for the validation experiments.

LC−MS grade methanol (MeOH), acetonitrile (ACN), and water was purchased from Carl Roth GmbH + Co. KG (Karlsruhe, Germany). LC−MS grade DMSO and formic acid was from Thermo Fisher Scientific GmbH (Dreieich, Germany). Ubiquitin from bovine erythrocytes was purchased from Sigma-Aldrich Chemie GmbH (Munich, Germany), and stable-labeled internal standard (IS) 13C-ubiquitin was purchased from Silantes GmbH (Munich, Germany). Artificial CSF (aCSF) was purchased from EcoCyte BioScience (Castrop-Rauxel, Germany).

Sample Preparation

Patients and Collection of CSF Samples

Fifty microliters of CSF sample, calibration standard, QC sample, or blank (aCSF + 250 μg/mL BSA) was mixed in a 96deep-well plate (Sarstedt AG & Co., Nümbrecht, Germany) with 50 μL of IS working solution (50 ng/mL in 90% ACN/2% formic acid). After that, ACN (100 μL) was added for protein precipitation, and after mixing for 2 min at 1400 rpm, the samples were centrifuged at 4000g for 30 min at room temperature. Seventy microliters of the supernatant was mixed with 400 μL of water in a new 96-deep-well plate, mixed for 2 min at 1400 rpm, and centrifuged again at 4000g for 10 min to remove air bubbles at the bottom of the wells and to pellet residual insoluble material. Afterward, the plate was put into the autosampler and stored at 4 °C until analysis.

CSF was collected by lumbar puncture at the Department of Neurology, Ulm University Hospital (control patients and patients with PD, PSP, bvFTD, PPA, AD, ALS, and CJD) and University of Göttingen (CJD patients) (Table 1). Control patients had no neurodegenerative disease, and CSF samples were normal with respect to cell count; cell differentiation; albumin ratio; IgG, IgA, and IgM ratio; and oligoclonal bands.13 In control patients, a lumbar puncture was mainly performed to exclude bleeding, acute or chronic inflammation of the brain. The diagnoses comprised tension headache (5), neuropathia vestibularis (2), neuropathia of the trochlear nerve (2), fatigue syndrome (2), degenerative upper cervical spine syndrome, intoxication, paresis glaucoma, vertigo, transient global amnesia, ophthalmic zoster, exclusion primary lateral sclerosis, arteritis of the temporal artery, and schwannoma. ALS patients were diagnosed according to the El Escorial criteria,14 and PD diagnosis was performed according to the consensus criteria for PD.15 Cases with frontotemporal lobar degeneration (FTLD,

LC−MS/MS of Ubiquitin

The LC−MS/MS analysis of ubiquitin was carried out using an Agilent 1260 binary pump (Agilent Technologies Deutschland GmbH, Bö blingen, Germany), an Eksigent MicroLC200 including an CTC PAL autosampler, and an AB Sciex B

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX

Journal of Proteome Research

Article

precursor 714.7 on the QTRAP6500 using identical chromatographic settings. The fragment spectrum at the retention time of 1.6 min was compared with a MS/MS spectrum of purified ubiquitin. Criteria for LLOQ determination were precision and deviation ≤20% and signal intensity in blank and blank with IS samples must be ≤20% of the LLOQ. Each analytical run contained two sets of QC samples (low, medium, and high) and two sets of calibration standards, one at the beginning and one at the end of the run, also including a blank with IS and one blank without IS. A run was accepted if the calculated concentrations of four out of six QC samples and at least 75% of the calibration standards were within ±15% of the nominal concentration. Intra-assay accuracy and precision was determined by analyzing six replicates of low, medium, and high QCs, QCLLOQ, and a CSF sample of a control patient. Interassay accuracy and precision was calculated using the described QCs from three independent runs.

QTRAP6500 (all AB Sciex Germany GmbH, Darmstadt, Germany). Ten microliters of sample was injected onto a C8 PepMap100 Trap column (0.3 × 5 mm, 5 μm, Thermo Fisher Scientific GmbH, Dreieich, Germany) with a flow rate of 200 μL/min and a mobile phase composition of 5% MeOH and 0.1% formic acid (Table 2). Ubiquitin was eluted from the Trap Table 2. Gradient Profile for Ubiquitin Determination by LC−MS/MS Eksigent MicroLC200b

Agilent 1260a (200 μL/min), TCc

(20 μL/min), ACc time (min)

% Bd

5 5

0 0.5

22 22

2

80

1.2

22

3.4 3.5 4.5

80 5 5

1.25 2.3 2.4 2.7 2.8 3 3.1 3.2 3.5 3.6 3.8 3.9 4 4.5

85 85 22 22 85 85 85 22 22 85 85 85 22 22

time (min)

%B

0 1.2

d

switching valve TC and AC connected TC and AC disconnected

Determination of Aβ42, Total Tau (T-Tau), and Phosphorylated-Tau (P-Tau) in CSF

Aβ42, T-Tau, and P-Tau were determined using commercially available ELISAs (Fujirebio Germany GmbH, Hannover, Germany). Statistics

All statistical analyses were carried out with GraphPad Prism 5.00. Groups were compared with a Kruskal−Wallis test and Dunn’s posthoc test. Correlation of parameters was calculated using Spearman’s rank-correlation coefficient.



RESULTS

Method Establishment

a

Agilent 1260: (A): 0.1% formic acid in water; (B): 0.1% formic acid in MeOH. bEksigent MicroLC200: (A): 0.1% formic acid, 4% DMSO in water; (B): 0.1% formic acid, 4% DMSO in ACN. Included a prerun of 0.2 min with the initial mobile phase composition. cAC: analytical column (HALO Fused-Core C18 column); TC: trap column (C8 PepMap100 column). dMobile phase composition.

We tested different mobile phase compositions for the chromatographic separation of ubiquitin, i.e., formic acid concentration and type of organic solvent. Using 0.1% formic acid in combination with ACN resulted in the most intense peak intensity. Addition of 4% DMSO significantly increased peak intensity, as previously described.19 The optimized method had a run time of 4.5 min and a retention time for ubiquitin of 1.6 min (Figure 1). MS parameters were optimized by direct infusion of ubiquitin and 13C-ubiquitin into the QTRAP6500. The M + 12H+ ion was the most intense charge state, reflecting the 7 lysine and 4 arginine residues. The most prominent product ion was the y58 fragment (Table 3), and its identity was confirmed by m/z accuracy (deviation 10 ppm) using a high-resolution mass spectrometer (Q Exactive, Thermo Fisher Scientific GmbH, Dreieich, Germany). The selectivity of the method for ubiquitin in CSF was tested using two additional ubiquitin transitions of different charge states (659.9 → 726.6 and 779.7 → 817.3), and peaks were observed in CSF for all three transitions at the expected retention times. After filtration of the CSF through a 3000 Da molecular weight cutoff centrifugal filter (Merck Chemicals

column with a flow rate of 20 μL/min onto an Eksigent HALO Fused-Core C18 MicroLC column (0.5 × 50 mm, 2.7 μm, Eksigent) that was kept at 60 °C and quantified by multiple reaction monitoring (MRM) with the QTRAP6500. A detailed description of the gradient profile and mobile phase composition is given in Table 2. The measured transitions and MS parameters are listed in Table 3. The settings for the ESI ion source were as follows: curtain gas, 15; CAD gas; high; nebulizer gas (GS1), 12; heater gas (GS2), 50; ion spray voltage, 5.5 kV; and temperature, 150 °C. The peak area was used for quantification and the calibration curve was calculated with a weighting of 1/x2. Validation and Acceptance Criteria for the LC−MS/MS Measurement of Ubiquitin

To confirm ubiquitin identity in CSF samples, a CSF sample was analyzed with an enhanced product ion scan (EPI) of the

Table 3. Transitions and MS Parameters for Ubiquitin and 13C-Ubiquitina

a

analyte

theoretical mass (Da)

precursor ion (m/z)

product ion (m/z)

dwell time (ms)

DP

CE

CXP

ubiquitin 13 C-ubiquitin (IS)

8564.8 8940.0

714.7 (M + 12H+) 745.4 (M + 12H+)

726.6 (y58 fragment) 757.8 (y58 fragment)

40 40

120 120

24 23

22 40

CE, collision energy; CXP, collision cell exit potential; DP, declustering potential; IS, internal standard; m/z, mass-to-charge ratio. C

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX

Journal of Proteome Research

Article

Figure 1. Representative chromatograms of ubiquitin and the internal standard (IS) 13C-ubiquitin. Chromatograms show a window of 2 min around the peak for calibration standards and blank samples. The samples were analyzed with liquid chromatography−multiple reaction monitoring mass spectrometry using an AB Sciex QTRAP6500 mass spectrometer. The measured transitions are given in parentheses. LLOQ, lower limit of quantitation.

PSP, bvFTD, and AD patients, whereas the correlation analysis for age was performed with all patient groups. CSF concentration of Aβ42 and P-Tau were measured in CSF of control, AD, bvFTD, PSP, and PPA patients, and T-Tau was also measured in CJD patients (Figure 3B−D). Aβ42 was significantly reduced in AD and T-Tau was significantly increased in AD and CJD. Correlation analysis of ubiquitin and T-Tau, Aβ42, and PTau was performed using data from control, PPA, PSP, bvFTD, AD, and CJD patients. T-Tau and Aβ values were missing for one PSP patient, and the P-Tau value was missing for two PSP patients. There was a strong correlation between ubiquitin and T-Tau concentration (r = 0.89, p < 0.0001, n = 104) and a small correlation with Aβ42 (r = −0.37, p < 0.0001, n = 104). Ubiquitin did not correlate with P-Tau concentration in the whole cohort (r = 0.15, p = 0.13, n = 103), but there was a good correlation when only AD and control patients were analyzed (r = 0.53, p = 0.0001, n = 48). Receiver operating characteristic (ROC) curve analysis of ubiquitin in CSF of CJD and control patients revealed an area under the curve (AUC) of 0.996 and a specificity and sensitivity of 100 and 95%, respectively (cutoff level, 60.3 ng/mL) (Figure 4). For AD and control patients, the AUC was 0.832, and the specificity and sensitivity were 75 and 75%, respectively (Figure 4). Because ubiquitin concentration in AD is also significantly different from that of FTLD patients (Figure 3), we performed a ROC analysis for the discrimination between FTLD and AD. The AUC was 0.776, and the specificity and sensitivity were 53 and 100% using a cutoff level of 32.1 ng/mL ubiquitin (Figure 4).

GmbH, Schwalbach, Germany), no peak was detectable (Figure 2). We further confirmed the identity of the ubiquitin peak in CSF by comparison of the MS/MS spectrum at the expected retention time with a spectrum from purified ubiquitin, and we could detect all major fragments of the purified ubiquitin in the CSF sample as well. Because selectivity was confirmed for the observed ubiquitin peak and due to the small peak width (3 s at baseline), we decided to use only the 714.7 → 726.6 transition for quantitation to increase the number of data points for the peak to ensure high precision and accuracy. The method was validated in a calibration range of 2−200 ng/mL and showed excellent intra- and interassay accuracy and precision (CV) for the QC samples (intra-assay, 100−107% and 4.4−7.4%; interassay, 100−106% and 4.9−10.3%) and CSF sample (intra-assay precision, 5.9%; interassay precision, 7.4%). Ubiquitin was stable in CSF for at least 4 days under benchtop conditions (room temperature). Ubiquitin Determination in Neurodegenerative Diseases

Because ubiquitin is known to be markedly increased in CSF of CJD patients,7 we used samples of CJD patients to validate the reliability of the LC−MS/MS method. There was a more than 5-fold increase of ubiquitin in CJD patients (Figure 3A), verifying our LC−MS/MS method for measuring diseaserelated changes of ubiquitin in CSF. We measured CSF ubiquitin concentration in patients of different neurodegenerative diseases (Figure 3A). AD patients showed significantly elevated ubiquitin concentration in CSF, but there was no significant alteration in patients with PD, ALS, PPA, PSP, and bvFTD. Age did not differ significantly between groups (p = 0.05). Ubiquitin concentration showed no correlation with total protein concentration (r = 0.20, p = 0.07, n = 86) or albumin quotient (r = 0.18, p = 0.10, n = 86) and only small correlation with age (r = 0.27, p = 0.0008, n = 150). Protein concentration and the albumin quotient were available only for control, PPA,



DISCUSSION Here, we established an alternative method to antibody-based assays for the measurement of ubiquitin in CSF with high selectivity for free monoubiquitin and with high precision and accuracy. We confirmed increased ubiquitin concentration in D

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX

Journal of Proteome Research

Article

Figure 2. Representative chromatograms of ubiquitin and the internal standard (IS) 13C-ubiquitin in cerebrospinal fluid (CSF) samples. Chromatograms show a window of 2 min around the peak for differently treated CSF samples to test the selectivity of the method. CSF was analyzed with or without addition of the IS or was filtered through a 3000 Da molecular weight cutoff filter and analyzed directly or after addition of IS and 40 ng/mL ubiquitin. Three transitions of different charge states were measured for ubiquitin to demonstrate the selectivity of the method.

CSF of AD and CJD patients and present, for the first time, data from patients with ALS, bvFTD, and PPA. Analysis of proteins by MRM typically includes enzymatic digestion of the proteins and subsequent analysis of unique peptides for quantification. However, ubiquitin not only appears in the monomeric form (monoubiquitin) but also as polyubiquitin chains; both forms can be conjugated to proteins or can be free. Furthermore, ubiquitin is encoded by different genes, and it is released from the resulting protein precursor proteolytically.1 Using a typical enzymatic digestion approach for sample preparation, all of the information about the ubiquitin form would be lost and is not useful. However, the small size of ubiquitin (76 amino acids) enables intact protein analysis of monoubiquitin by LC−MS/MS, resulting in a clear separation from the other ubiquitin forms. This is a major advantage over previously published ELISA techniques for ubiquitin, which suffer from the risk of cross-reactivity of the antibodies with different ubiquitin forms. Data for free monoubiquitin in CSF of control patients range from 13 to 110 ng/mL between studies,6,10−12 indicating that antibody selectivity is a serious problem. However, to deduce diseaserelated mechanisms, it is of great importance to determine the identity of the specific ubiquitin form measured, highlighting the need for a selective analysis method, as described here. We could confirm previous observations of elevated CSF ubiquitin in CJD7 and AD6 but not in PSP.8 Furthermore, our

study extends to other neurodegenerative diseases, including ALS, bvFTD, and PPA, which have not been investigated before. However, ubiquitin concentration in CSF was not different in these patients compared with that in controls. There are different sources of ubiquitin in CSF. Intracellular ubiquitin can be released by damaged cells into the extracellular space, which is supported by the observation of increased CSF ubiquitin in traumatic brain injury.20 Additionally, ubiquitin has been shown to be actively secreted by different cells including leptomeningeal21 or neuroblastoma cells.22 Several functions for extracellular ubiquitin have been described, such as antimicrobial and anti-inflammatory activity,23 showing that extracellular ubiquitin is more than a byproduct of cell damage. CSF ubiquitin may also originate from blood. Eighty percent of CSF protein concentration is blood-derived through passive diffusion via the blood−CSF barrier, and the concentration ratio of a protein in CSF vs serum gives information about its origin. For instance, albumin concentration, which is exclusively blood-derived, is 100-fold higher in serum than in CSF, and the albumin ratio is the accepted marker to describe the blood− CSF flow. Blood-derived proteins in CSF correlate with the albumin ratio.24 Ubiquitin concentration in serum was reported to be about 3- to 5-fold higher than that in CSF,23 which might indicate an additional contribution to the intrathecal ubiquitin pool. E

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX

Journal of Proteome Research

Article

Figure 3. CSF concentration of ubiquitin, Aβ42, T-Tau, and P-Tau in different neurodegenerative diseases. (A) Ubiquitin was measured in cerebrospinal fluid (CSF) by liquid chromatography−multiple reaction monitoring mass spectrometry (LC−MS/MS), and (B−D) Aβ42, total Tau (T-Tau,) and phosphorylated Tau (P-Tau) were measured by ELISA in patients with different neurodegenerative diseases. CSF ubiquitin and T-Tau were increased in AD (Alzheimer’s disease) and CJD (Creutzfeldt−Jakob disease), and Aβ42 was decreased in AD. There was no difference in P-Tau among groups (p = 0.21). Statistical analysis was perfomed with Kruskal−Wallis test and Dunn’s posthoc test. Individual values are shown (×). Bars and whiskers are the median and interquartile range. PD, Parkinson’s disease; ALS, amyotrophic lateral sclerosis; PPA, primary progressive aphasia; PSP, progressive supranuclear palsy; bvFTD, behavioral variant of frontotemporal dementia; FTLD, frontotemporal lobar degeneration (includes PSP, bvFTD, and PPA).

enzymes (DUBs) are also downregulated in AD,27 complicating the interpretation of these observations regarding the CSF ubiquitin concentration. In addition to brain-related changes, an impairment of the blood−CSF barrier could be a reason for the increase of ubiquitin in CSF. However, this is not common in AD patients,13 and the albumin quotient as a measure of blood− CSF barrier integrity was not significantly different among controls, ALS, PD, PPA, PSP, bvFTD, and AD patients in our study (data not shown) and ubiquitin does not correlate with the albumin quotient. Kudo and colleagues5 described a positive correlation of CSF and brain ubiquitin concentration, further supporting the intrathecal origin of increased CSF ubiquitin, at least in AD. However, with the exception of CJD, which is characterized by the most rapid neurodegeneration25 (thus increased CSF ubiquitin most likely arise from neuronal damage), the difference between AD and the other neurodegenerative diseases in our study cannot be explained with the aforementioned sources of increased CSF ubiquitin (i.e.,

The variety of CSF ubiquitin sources is also reflected in the possible causes of increased ubiquitin concentrations in AD and CJD in our study. Since CSF ubiquitin originates at least in part from dying cells,23 it is conclusive that it is increased in CSF during neurodegeneration, and it has also been shown in other types of neuronal damage.20 The high correlation of CSF ubiquitin and T-Tau concentration in our study also supports this idea because T-Tau is thought to be a general marker of axonal degeneration as well.25 However, the observation, that ubiquitin is elevated only in CJD and AD in our study argues against ubiquitin as a general marker of neuronal damage and implicates at least some kind of specificity with regard to disease-related processes or tissues. Cellular stress is known to increase ubiquitin expression1 and might be reflected in CSF. Because we selectively measured free monoubiquitin, alterations in the equilibrium between free and bound ubiquitin could also explain the observed concentration changes. Indeed, there is evidence that the ubiquitin ligase Parkin is downregulated in AD brains,26 which could shift the equilibrium to the side of monoubiquitin. On the other side, deubiquitinating F

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX

Journal of Proteome Research

Article

preparation method for ubiquitin that does not need special equipment, can be easily performed in 96-well plates, and can be performed within 1 to 2 h. The small size of monoubiquitin bears the advantage that it does not precipitate under high concentrations of organic solvent, making protein precipitation with ACN ideally suited for sample purification regarding speed and simplicity. In conclusion, we established a reliable method to selectively measure free monoubiquitin in CSF that might be adopted by other laboratories for improved comparability of ubiquitin measurements between centers. This represents a prerequisite and a first step toward the implementation of free CSF ubiquitin measurements in routine diagnosis.



Figure 4. Diagnostic potency of ubiquitin in cerebrospinal fluid (CSF) for AD and CJD. Receiver operating characteristic curve (ROC) analysis for the differentiation of AD and control patients, AD and FTLD patients, and CJD and control patients. Values for the area under the curve (AUC) and best combination of sensitivity and specificity are given in the Results section. AD, Alzheimer’s disease; FTLD, frontotemporal lobar degeneration; CJD, Creutzfeldt−Jakob disease.

AUTHOR INFORMATION

Corresponding Author

*Tel.: +49-731-500-63010. Fax: +49-731-500-63012. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



neuronal damage, cellular stress, downregulation of ubiquitinrelated enzymes) because these alterations are common among most of the investigated neurodegenerative diseases. Other disease-related mechanisms must contribute to the increased CSF ubiquitin concentration in AD, which needs further examination. Iqbal and colleagues9 hypothesized that alterations in CSF ubiquitin reflect early stages of the pathology in AD and found evidence that CSF ubiquitin concentration is highest in AD patients with recent onset. Furthermore, they defined subgroups of AD by combining different CSF biomarkers including Aβ42, T-Tau, and ubiquitin. Subgrouping of AD patients might be an important step for individualized therapy and successful drug discovery in the future. Although Aβ42 and Tau in CSF are routinely measured in the clinic to date, ubiquitin determination in CSF is not common due to the absence of reliable assays, which hinders the widespread implementation of subgrouping. In addition to confirming increased CSF ubiquitin concentration in AD patients compared with that in controls, we showed that CSF ubiquitin in AD patients also differs from that of other neurodegenerative diseases, e.g., FTLD. This might be promising for the differential diagnosis of AD and FTLD patients, which is hampered by overlapping clinical symptoms, resulting in a false diagnosis of 10−30% of clinically diagnosed FTLD patients that actually suffer from AD, not FTLD.28 In the present study, ubiquitin determination could discriminate AD and FTLD patients with a specificity and sensitivity of 53 and 100%, respectively. However, in our preliminary study, we used only a small cohort of patients from different neurodegenerative diseases, but for the reliable calculation of specificity, sensitivity, and cutoff levels, further and larger studies are needed. We are aware that our P-Tau levels do not have a similar differential diagnostic value as that of our T-Tau levels in AD. This might reflect a selection bias or a biological phenomenon, which, in a larger population, can now be investigated. In terms of the implication of ubiquitin determination in clinical routine or as a biomarker in drug discovery, high throughput, speed, and automation of the analysis becomes more important. Here, we established a simple and fast sample

ACKNOWLEDGMENTS

This work was supported by BMBF (federal ministry of education and research): Competence net neurodegenerative dementias (project: FTLDc), competence net multiple sclerosis (KKMS), MND-Net, the JPND networks for standardization of biomarkers (BiomarkAPD, SOPHIA), the JPND network STRENGTH, the EU (NADINE), foundation of the state Baden-Württemberg, and foundation Thierry Latran. We especially thank our patients for participating in these studies. We would like to thank Sandra Hübsch and Stephen Meier for their excellent technical assistance.



ABBREVIATIONS AC, analytical column; aCSF, artificial cerebrospinal fluid; AD, Alzheimer’s disease; ACN, acetonitrile; ALS, amyotrophic lateral sclerosis; AUC, area under the curve; BSA, bovine serum albumin; bvFTD, behavioral variant of frontotemporal dementia; CE, collision energy; CJD, Creutzfeldt−Jakob disease; CSF, cerebrospinal fluid; CXP, collision cell exit potential; IS, internal standard; CV, coefficient of variation; DP, declustering potential; DUB, deubiquitinating enzyme; f, female; FTLD, frontotemporal lobar degeneration; LC−MS/ MS, liquid chromatography multiple reaction monitoring mass spectrometry; LLOQ, lower limit of quantitation; m, male; MeOH, methanol; MRM, multiple reaction monitoring; NINCDS−ADRDA, National Institute of Neurological and Communicative Disorders and Stroke−Alzheimer’s Disease and Related Disorders Association; P-Tau, phosphorylated-Tau; PD, Parkinson’s disease; PPA, primary progressive aphasia; PSP, progressive supranuclear palsy; QC, quality control; ROC, receiver operating characteristic curve; T-Tau, total Tau; TC, trap column; UPS, ubiquitin-proteasome system



REFERENCES

(1) Kimura, Y.; Tanaka, K. Regulatory mechanisms involved in the control of ubiquitin homeostasis. J. Biochem. 2010, 147, 793−798. (2) Mukhopadhyay, D.; Riezman, H. Proteasome-independent functions of ubiquitin in endocytosis and signaling. Science 2007, 315, 201−205.

G

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX

Journal of Proteome Research

Article

(3) Hegde, A. N.; Upadhya, S. C. Role of ubiquitin-proteasomemediated proteolysis in nervous system disease. Biochim. Biophys. Acta 2011, 1809, 128−140. (4) Wang, G. P.; Iqbal, K.; Bucht, G.; Winblad, B.; Wisniewski, H. M.; Grundke-Iqbal, I. Alzheimer’s disease: paired helical filament immunoreactivity in cerebrospinal fluid. Acta Neuropathol. 1991, 82, 6−12. (5) Kudo, T.; Iqbal, K.; Ravid, R.; Swaab, D. F.; Grundke-Iqbal, I. Alzheimer disease: correlation of cerebro-spinal fluid and brain ubiquitin levels. Brain Res. 1994, 639, 1−7. (6) Blennow, K.; Davidsson, P.; Wallin, A.; Gottfries, C. G.; Svennerholm, L. Ubiquitin in cerebrospinal fluid in Alzheimer’s disease and vascular dementia. Int. Psychogeriatr. 1994, 6, 13−22. (7) Steinacker, P.; Rist, W.; Swiatek-de-Lange, M.; Lehnert, S.; Jesse, S.; Pabst, A.; Tumani, H.; von Arnim, C. A. F.; Mitrova, E.; Kretzschmar, H. A.; et al. Ubiquitin as potential cerebrospinal fluid marker of Creutzfeldt−Jakob disease. Proteomics 2010, 10, 81−89. (8) Constantinescu, R.; Andreasson, U.; Li, S.; Podust, V. N.; Mattsson, N.; Anckarsäter, R.; Anckarsäter, H.; Rosengren, L.; Holmberg, B.; Blennow, K.; et al. Proteomic profiling of cerebrospinal fluid in parkinsonian disorders. Parkinsonism Relat. Disord. 2010, 16, 545−549. (9) Iqbal, K.; Flory, M.; Khatoon, S.; Soininen, H.; Pirttila, T.; Lehtovirta, M.; Alafuzoff, I.; Blennow, K.; Andreasen, N.; Vanmechelen, E.; et al. Subgroups of Alzheimer’s disease based on cerebrospinal fluid molecular markers. Ann. Neurol. 2005, 58, 748− 757. (10) Kandimalla, R. J.; S, P.; Bk, B.; Wani, W. Y.; Sharma, D. R.; Grover, V. K.; Bhardwaj, N.; Jain, K.; Gill, K. D. Cerebrospinal fluid profile of amyloid β42 (Aβ42), hTau and ubiquitin in North Indian Alzheimer’s disease patients. Neurosci. Lett. 2011, 487, 134−138. (11) Manaka, H.; Kato, T.; Kurita, K.; Katagiri, T.; Shikama, Y.; Kujirai, K.; Kawanami, T.; Suzuki, Y.; Nihei, K.; Sasaki, H. Marked increase in cerebrospinal fluid ubiquitin in Creutzfeldt−Jakob disease. Neurosci. Lett. 1992, 139, 47−49. (12) Skoog, I.; Vanmechelen, E.; Andreasson, L. A.; Palmertz, B.; Davidsson, P.; Hesse, C.; Blennow, K. A population-based study of tau protein and ubiquitin in cerebrospinal fluid in 85-year-olds: relation to severity of dementia and cerebral atrophy, but not to the apolipoprotein E4 allele. Neurodegeneration 1995, 4, 433−442. (13) Jesse, S.; Brettschneider, J.; Süssmuth, S. D.; Landwehrmeyer, B. G.; von Arnim, C. A. F.; Ludolph, A. C.; Tumani, H.; Otto, M. Summary of cerebrospinal fluid routine parameters in neurodegenerative diseases. J. Neurol. 2011, 258, 1034−1041. (14) Pradat, P.-F.; Bruneteau, G. Clinical characteristics of amyotrophic lateral sclerosis subsets. Rev. Neurol. (Paris). 2006, 162 Spec N, 4S29−4S33. (15) McKeith, I. G. Consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the Consortium on DLB International Workshop. J. Alzheimer’s Dis. 2006, 9, 417−423. (16) Rascovsky, K.; Hodges, J. R.; Knopman, D.; Mendez, M. F.; Kramer, J. H.; Neuhaus, J.; van Swieten, J. C.; Seelaar, H.; Dopper, E. G. P.; Onyike, C. U.; et al. Sensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia. Brain 2011, 134, 2456−2477. (17) Gorno-Tempini, M. L.; Hillis, A. E.; Weintraub, S.; Kertesz, A.; Mendez, M.; Cappa, S. F.; Ogar, J. M.; Rohrer, J. D.; Black, S.; Boeve, B. F.; et al. Classification of primary progressive aphasia and its variants. Neurology 2011, 76, 1006−1014. (18) Consensus on Criteria for Sporadic CJD; World Health Organization: Geneva, Switzerland, 1998. (19) Hahne, H.; Pachl, F.; Ruprecht, B.; Maier, S. K.; Klaeger, S.; Helm, D.; Médard, G.; Wilm, M.; Lemeer, S.; Kuster, B. DMSO enhances electrospray response, boosting sensitivity of proteomic experiments. Nat. Methods 2013, 10, 989−991. (20) Majetschak, M.; King, D. R.; Krehmeier, U.; Busby, L. T.; Thome, C.; Vajkoczy, S.; Proctor, K. G. Ubiquitin immunoreactivity in

cerebrospinal fluid after traumatic brain injury: clinical and experimental findings. Crit. Care Med. 2005, 33, 1589−1594. (21) Ohe, Y.; Ishikawa, K.; Itoh, Z.; Tatemoto, K. Cultured leptomeningeal cells secrete cerebrospinal fluid proteins. J. Neurochem. 1996, 67, 964−971. (22) Sandoval, J. A.; Hoelz, D. J.; Woodruff, H. A.; Powell, R. L.; Jay, C. L.; Grosfeld, J. L.; Hickeyd, R. J.; Malkas, L. H. Novel peptides secreted from human neuroblastoma: useful clinical tools? J. Pediatr. Surg. 2006, 41, 245−251. (23) Majetschak, M. Extracellular ubiquitin: immune modulator and endogenous opponent of damage-associated molecular pattern molecules. J. Leukocyte Biol. 2011, 89, 205−219. (24) Reiber, H.; Peter, J. B. Cerebrospinal fluid analysis: diseaserelated data patterns and evaluation programs. J. Neurol. Sci. 2001, 184, 101−122. (25) Blennow, K.; Hampel, H.; Weiner, M.; Zetterberg, H. Cerebrospinal fluid and plasma biomarkers in Alzheimer disease. Nat. Rev. Neurol. 2010, 6, 131−144. (26) Rosen, K. M.; Moussa, C. E.-H.; Lee, H.-K.; Kumar, P.; Kitada, T.; Qin, G.; Fu, Q.; Querfurth, H. W. Parkin reverses intracellular betaamyloid accumulation and its negative effects on proteasome function. J. Neurosci. Res. 2010, 88, 167−178. (27) Choi, J.; Levey, A. I.; Weintraub, S. T.; Rees, H. D.; Gearing, M.; Chin, L.-S.; Li, L. Oxidative modifications and down-regulation of ubiquitin carboxyl-terminal hydrolase L1 associated with idiopathic Parkinson’s and Alzheimer’s diseases. J. Biol. Chem. 2004, 279, 13256− 13264. (28) Rabinovici, G. D.; Miller, B. L. Frontotemporal lobar degeneration: epidemiology, pathophysiology, diagnosis and management. CNS Drugs 2010, 24, 375−398.

H

dx.doi.org/10.1021/pr5006058 | J. Proteome Res. XXXX, XXX, XXX−XXX