Journal of Alzheimer’s Disease 24 (2011) 785–797DOI 10.3233/JAD-2011-100525IOS Press
785
Lysosomal �-Galactosidase and�-Hexosaminidase Activities Correlate withClinical Stages of Dementia Associatedwith Alzheimer’s Disease and Type 2Diabetes Mellitus
Roberto Tiribuzia, Antonio Orlacchiob,c, Lucia Crispoltonia, Mariangela Maiottid, Mauro Zampolinid,Massimiliano De Angelise, Patrizia Mecoccif , Roberta Cecchettif , Giorgio Bernardib,c,Alessandro Dattia, Sabata Martinoa and Aldo Orlacchioa,∗aDipartimento di Medicina Sperimentale e Scienze Biochimiche, Sezione di Biochimica e BiologiaMolecolare, University of Perugia, Perugia, ItalybLaboratorio di Neurogenetica, CERC-IRCCS Santa Lucia, Rome, ItalycDipartimento di Neuroscienze, University “Tor Vergata”, Rome, ItalydNeuropsicologia Clinica, Ospedale S. Giovanni Battista, Foligno, ItalyeU.O. S.S. Ambulatoriale di Diabetologia Di.M.I., University of Perugia, Perugia, Italyf Sezione di Gerontologia e Geriatria, Dipartimento di Medicina Clinica e Sperimentale, Universityof Perugia, Perugia, Italy
Handling Associate Editor: Annamaria Confaloni
Accepted 16 January 2011
Abstract. Multiple epidemiological studies have shown that individuals affected by type-2 diabetes mellitus (T2DM) carry a2-to-5-fold higher risk of developing Alzheimer’s disease (AD) when compared to non-diabetic subjects. Thus, biochemicalparameters that can be easily and routinely assessed for high-confidence evaluation of diabetic conditions leading to AD (AD-T2DM) are regarded as efficient tools aimed at early diagnosis and, in turn, timely AD treatment. In this regard, the activity oflysosomal glycohydrolases may of use, in light of the implication of these enzymes in early events that underlie AD pathology andan overt correlation, in diabetes, between altered metabolic homeostasis, abnormal glycohydrolase secretion in body fluids, andoccurrence of diabetic complications. Based on marked up-regulation previously shown in a peripheral, cell-based model of AD,we selected �-Galactosidase, �-Hexosaminidase, and �-Mannosidase to discriminate T2DM from AD-T2DM subjects. A screenof 109, 114, and 116 patients with T2DM, AD and AD-T2DM, respectively, was performed by testing enzyme activities in bothblood plasma and peripheral blood mononuclear cells. Compared to age-matched, healthy controls (n = 122), �-Galactosidaseand �-Hexosaminidase activities markedly diverged across the three groups, whereas virtually unchanged values were observedfor �-Mannosidase. In particular, plasma �-Galactosidase and �-Hexosaminidase levels were higher in patients with AD-T2DM
∗Correspondence to: Prof. Aldo Orlacchio, Ph.D., Dipartimentodi Medicina Sperimentale e Scienze Biochimiche, Sezione diBiochimica e Biologia Molecolare, University of Perugia, via delGiochetto, 06126 Perugia, Italy. Tel./Fax: +39 0755852187; E-mail:[email protected].
786 R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD
compared to those with T2DM, suggesting different mechanisms leading to enzyme secretion. Statistical analyses based on ROCcurves showed that both �-Galactosidase and �-Hexosaminidase activities, either intracellular or plasma-secreted, may be usedto discriminate AD patients from controls and AD-T2DM from T2DM patients.
Keywords: Alzheimer’s disease, diagnostic correlation, lysosomal glycohydrolases, type 2 diabetes
INTRODUCTION
Alzheimer’s disease (AD) is the most common formof dementia. The incidence of the disease is below 1%in individuals aged 60–64, but shows an exponentialincrease with age [1]. Thus, due to worldwide agingpopulation issues, AD can be considered a pandemic ofthe 21st century [2]. Only 5% of AD cases are describedas an inheritable disease, while the majority arises spo-radically. However, the disease may be caused by agenetic predisposition, as shown by the identificationof specific DNA mutations in a large number of fami-lies [3–7]. Currently, a definite diagnosis of AD is onlypossible by postmortem examination of the brain tissue[8, 9].
Recent evidence indicated that the formation ofamyloid-� (A�) plaques, a typical AD hallmark, maybe detected several years prior to the onset of clinicalsymptoms [10]. This is indicative of early occurrenceof biochemical players associated with the develop-ment and progression of pathological processes. Thus,the identification of predictive indicators in peripheralsystems would likely raise the prospect for early ther-apeutic interventions. For example, the dysregulationof signaling protein in blood plasma points to systemicdysregulation of haematopoiesis, immune responses,apoptosis, and neuronal support in presymptomaticAD [11], suggesting that blood plasma is a promisingsource of AD biomarkers [12].
Recently, clinical studies have revealed that patientswith type 2 Diabetes Mellitus (T2DM), a heteroge-neous metabolic disorder of glucose metabolism [13],are especially susceptible to AD [14–19], therebypointing to T2DM as a likely risk factor [20–26]. Inthis regard, it was reported that approximately 80% ofAD patients had T2DM or showed abnormal blood glu-cose levels [27]. T2DM and AD share several commonfeatures [13, 25, 28–29], which include a markedly dif-ferent modulation of lysosomal enzyme activities. Inparticular, several reports have shown that alterationsof the endosomal-lysosomal system emerge early inthe AD brain, prior to the development of plaques andtangles typically associated with AD neuropathology[14, 30–38]. For example, secreted Cathepsin D and�-Hexosaminidase A colocalize with A� plaques in
the cerebellum and striatum of AD individuals [39],while upregulated levels of �-Galactosidase, �-Hex-osaminidase, and �-Mannosidase and decreased activ-ity of Cathepsin D were observed in the fibroblastsof patients affected by either the familial or sporadicform of the disease [40–41]. On the other hand,T2DM patients show lower �-Galactosidase activityin the pancreas [42] and, in correlation with glycemiclevels, increased levels of �-Hexosaminidase and �-Mannosidase in plasma and serum [43–46].
Here, we asked whether the activities of �-Galac-tosidase (�-Gal, E.C.3.2.1.23), �-Hexosaminidase(�-Hex, E.C.3.2.1.52) and �-Mannosidase (�-Man,E.C.3.2.1.24) measured in peripheral blood mononu-clear cells (PMNCs) and plasma had the ability todiscriminate patients affected by AD from controls andAD-T2DM patients from T2DM patients.
Receiver operating characteristic (ROC) curvesrevealed that �-Gal and �-Hex activities, both intra-cellular and plasma-secreted, varied in correlation withaltered physiological conditions. Further, we observedthat these differences in enzyme levels were evidentduring the early stages of dementia in both AD and AD-T2DM patients, thus suggesting a diagnostic potentialfor early AD detection.
MATERIALS AND METHODS
Patients
Patients and donor subjects were recruited fromthe following medical centers: I.R.C.C.S. Santa Luciaand Dipartimento di Neuroscienze, Policlinico “TorVergata”, Rome, Italy; Neuropsicologia Clinica,Ospedale S. Giovanni Battista, Foligno, Italy; andDIMISEM and Gerontologia e Geriatria, Universityof Perugia, Italy.
All AD patients fulfilled the NINCDS-ADRDA cri-teria for probable AD [47].
The severity of dementia was assessed by the Mini-Mental Status Examination (MMSE) [48] and throughthe Clinical Dementia Rating scale (CDR) [49, 50].Subjects with T2DM were diagnosed according to theAmerican Diabetes Association (ADA) criteria [51].Subjects with T2DM plus AD (AD-T2DM) were diag-
R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD 787
nosed as described for AD and T2DM. In all ADgroups, a computed tomography or magnetic reso-nance imaging (MRI) was also performed to validatethe diagnosis of AD. A control group composed ofelderly, healthy individuals was recruited from thecatchment’s area of each center. Exclusion criteriafor AD, T2DM, AD-T2DM patients and control sub-jects were based on history of depression or psychosis,alcohol or substance abuse, and use of psychoactivemedications. Informed consent was obtained from eachindividual, or his/her substitute decision maker, whoparticipated in this study.
Biological samples
Blood samples were collected, processed andanalyzed based on a standard operating procedureimplemented by the three medical centers involvedin the study (i.e., Policlinico “Tor Vergata”, Rome;Neuropsicologia Clinica, Ospedale S. Giovanni Bat-tista, Foligno; and DIMISEM Gerontologia e Geriatria,Perugia).
Peripheral blood (20 ml) was routinely obtainedfrom patients and healthy donors in the fasted state(generally in the morning, between 9:00–10:00 am)and processed within two hours to obtain plasma andPMNCs. Plasma was separated by a 20 min centrifu-gation (1500x g) performed at 4◦C [52].
To isolate PMNCs, peripheral blood was dilutedwith phosphate buffered saline (PBS), without Ca2+/Mg2+, containing 2 mM EDTA and subjected todensity gradient centrifugation using Lympholyte®
medium (Cedarlane Laboratories Limited) accordingto the manufacturer’s protocol. Cells were washedin PBS, counted, resuspended in 1 mL/106 cells ofice-cold 10 mM sodium phosphate buffer, pH 6.0, con-taining 0.1% (v/v) Nonidet NP40 detergent, and furtherdisrupted by sonication (three rounds of 30 s in ice-cold tubes) using ultrasonic baths generating operatingfrequencies of 25–35 KHz. Following the preparativesteps, both plasma and cell extracts were immediatelyfrozen in dry ice and stored at −80◦C in multiple,single-use aliquots. Enzyme activity determinationswere carried out within 1 month.
Enzyme assays
Enzyme activities were measured using commer-cially available, fluorometric substrates.
The �-Galactosidase assay was performed using1.5 mM 4-methylumbelliferyl-�-D-galactoside sub-
The �-Mannosidase assay was performed using3 mM 4-methylumbelliferyl-�-D-mannoside (MU-MAN, Sigma) dissolved in 0.2 M sodium acetate, pH4.0 [55]. Hexosaminidase activity was determinedwith either 3 mM 4-methylumbelliferyl-N-acetyl-�-d-glucosaminide (MUG, Sigma) or 3 mM4-methylumbelliferyl-6-sulfo-2-acetamido-2-deoxy-�-d-glucopyranoside (MUGS, Toronto ResearchChemicals) substrates dissolved in 0.1 M-citrate/0.2 Mdisodium phosphate buffer, pH 4.5 [56]. All reactionswere performed using 50 �l of test sample mixed with100 �l of substrate prior to incubation for 30 min at37◦C. Ice-cold 0.2 M Glycine/NaOH, pH 10.6, wasused to stop the assays.
Liberated 4-methylumbelliferone was measuredfluorometrically (λex 360 nm; λem 446 nm). One mUis defined as the amount of enzyme that hydrolyzes1 nmol/min of substrate.
To calculate specific activities, total protein contentwas measured via the Bradford method using serumbovine albumin as the standard [57].
Statistical analysis
Results are expressed as means ± SD. Based on non-Gaussian distributions of values; data analysis wascarried out using the one-way ANOVA and Dunn’sMultiple Comparison Test as post-hoc comparisontests (GraphPad 4.0 Software, San Diego, California,USA). p < 0.05 was considered significant.
ROC curves, used to evaluate the diagnostic sig-nificance of enzyme activities [58], were generatedby plotting sensitivity (true positives) versus 100-specificity (false positives) using the MedCalcTM
software.
RESULTS
Patients
This study involved a total of 461 subjects dis-tributed in the following groups: AD (n = 114), T2DM(n = 109), AD-T2DM (n = 116), and age-matched con-trols (n = 122). Each AD-T2DM patient had developedT2DM prior to AD diagnosis. Demographic and clin-ical data are summarized in Table 1.
Glycated hemoglobin levels were found to be agood metabolic control for T2DM subjects (HbA1C= 7.5% ± 1.0) and sub-optimal metabolic controlfor AD-T2DM patients (HbA1C = 8.1% ± 1.9) [59].
CTRL: age-matched healthy donors; AD: patients with Alzheimer’s disease; T2DM: individuals affected by type2 diabetes mellitus; AD-T2DM: diabetic patients with Alzheimer’s disease; The mean age ± SD is reported forall groups. Education: years of school, mean ± SD. HbA1C: Glycated hemoglobin; MMSE: Mini Mental StateExamination; CDR: Clinical Dementia Rating Scale; FAD: Family history of AD; ChEI: Cholinesterase inhibitor.
MMSE scores indicated moderate or severe cogni-tive impairment in AD (14.9 ± 3.6) and AD-T2DMpatients (10.54 ± 2.88), and no cognitive impairmentin healthy elderly controls and T2DM patients. AD andAD-T2DM patients were further sub-grouped basedon the severity of dementia using the Clinical Demen-tia Rating Scale (CDR) [49, 50], where CDR = 0,CDR = 1, CDR = 2, and CDR = 3 denote no cognitiveimpairment, mild dementia, moderate dementia, andsevere dementia, respectively. In particular, the ADgroup consisted of 12.3% patients with a CDR = 1,48.2% with a CDR = 2, and 39.5% with a CDR = 3.The AD-T2DM group was composed of 6.9% patientswith mild dementia, 33.6% with moderate dementia,and 59.8% with severe dementia. All AD patients wereunder cholinesterase inhibitor treatment.
Assay validation
To evaluate the robustness of the enzyme assaysimplemented by each of the medical centers involved inthis study, we first investigated whether the biologicalsamples were handled and stored in a manner to per-mit within-batch and between-batch comparisons. Inthis regard, approximately 10% of the biological sam-ples (i.e., plasma and PMNCs) from each of patientand control groups were tested prior to freezing, andafter 1 and 4 weeks of −80◦C storage. Determinationswere performed in triplicate using different, single-usealiquots. Measurements of �-Gal, �-Hex (MUG andMUGS substrates) and �-Man activities were virtually
unchanged with time of storage (i.e., >99% activity),thereby indicating that samples stored for up to 1 monthcould be accurately compared (not shown).
These results were also used to assess data repro-ducibility. We observed that both plasma and PMNCsamples displayed low intra- and inter-assay coeffi-cients of variation, which were in the 2% and 3.9%range, respectively, for all of the enzymes tested(Table 2).
These findings, taken together, revealed that thestandard operating procedure was suitable for use ina large-scale screen of samples that require collectionat disparate times.
Lysosomal β-Gal, β-Hex, and α-Man activitiesin plasma
Enzyme assays performed in the plasma of AD,T2DM and AD-T2DM patients revealed that �-Galactivity was markedly higher in AD (6.75 ± 1.47 mU/ml) than in AD-T2DM patients (4.7 ± 1.1 mU/ml),and that the values observed in the AD-T2DM groupwere comparable to controls (4.25 ± 1.15 mU/ml).The T2DM population, instead, showed the lowestlevels of activity (3.2 ± 0.98 mU/ml), suggesting adistinct mechanism(s) affecting �-Gal secretion rates(Fig. 1A).
By contrast, values of �-Man activity were virtuallyunchanged in all instances, and no statistically signifi-cant differences could be observed within any groups(Fig. 1B).
Samples were obtained from AD (n = 13) and AD-T2DM (n = 11) patients, T2DM (n = 12) subjects and age-matchedcontrols (n = 15) at different times. Samples were immediately processed and tested for enzyme activity prior to freezing,and after 1 and 4 week storage at −80◦C to determine inter-assay coefficients of variation (CVs). Intra-assay CVs werecalculated based on measurements performed in triplicate. Descriptions of plasma and PMNCs preparations and enzymeassay details are reported in the Materials and Methods section. Intra: intra-assay CVs (%). Inter: inter-assay CVs (%).
The activity of �-Hex was evaluated using twodifferent substrates, MUG and MUGS. The first is com-mon to both Hexosaminidase isoenzymes, HexA andHexB, while the other is specifically hydrolyzed byHexA [56]. Total �-Hex activity was higher in patientswith AD-T2DM (1192 ± 154 mU/ml), while in theAD and T2DM groups the enzyme displayed similarvalues (i.e., AD = 769 ± 193 mU/ml; T2DM = 768 ±145 mU/ml) and controls showed markedly lower lev-els (475 ± 149 mU/ml) (Fig. 1C). HexA exhibiteda similar pattern, with levels that were the highestin AD-T2DM patients (127 ± 21 mU/ml), compa-rable between AD (102 ± 21.8 mU/ml) and T2DM(89.1 ± 21.6 mU/ml) groups, and consistently lower incontrols (53 ± 19.6 mU/ml) (Fig. 1D).
Notably, statistical evaluations of the data revealedthat �-Gal and �-Hex activities were not affected bygender in any of the four groups.
To investigate a possible correlation betweenenzyme activity levels and AD progression, theresults were grouped based on the cognitive statusof AD and AD-T2DM patients. This analysis in-dicated that the increase in �-Gal activity was moreevident in AD patients with CDR1 (7.22 ± 0.9 mU/ml)and CDR2 (7.75 ± 1.15 mU/ml) (p<0.01 versus con-trols) than in AD patients with CDR3 (5.56 ±1.25 mU/ml; p<0.05 versus controls) (Fig. 1a′ ). Sim-ilarly, we observed increased �-Gal activity valuesin AD-T2DM patients with mild and moderatedementia (CDR1 = 5.4 ± 0.87 mU/ml; CDR2 = 5.14 ±1.09 mU/ml; p < 0.05 versus T2DM group) (Fig. 1a′′ ).
As expected, we noted a lack of correlation betweenthe activity of secreted �-Man and the severity ofdementia in both AD (Fig. 1b′ ) and AD-T2DMpatients (Fig. 1b′′ ).
The activity of total �-Hex increased in the earlystage of dementia and remained unchanged during theprogression of the disease in both AD and AD-T2DMpatients. In particular, total �-Hex, evaluated inAD patients with CDR1, CDR2, and CDR3, was788 ± 140 mU/ml, 781 ± 193 mU/ml, and 770 ± 208mU/ml (p < 0.05 versus controls), respectively(Fig. 1c′ ). In AD-T2DM patients, �-Hex levels were1315.9 ± 107 mU/ml (CDR1), 1210 ± 176 mU/ml(CDR2), and 1162.1 ± 133.6 mU/ml (CDR3) (p < 0.01versus T2DM group) (Fig. 1c′′ ). �-HexA showeda significant increase in AD patients with mild (93 ±10.98 mU/ml), moderate (115 ± 19.2 mU/ml) andsevere dementia (95 ± 17.3 mU/ml) (p < 0.05 versuscontrols) (Fig. 1d′), and similar results were also foundin AD-T2DM patients (CDR1 = 138.2 ± 13.1 mU/ml; CDR2 = 128.8 ± 24.9 mU/ml; CDR3 = 125 ± 18.6mU/ml; p < 0.05 versus T2DM group) (Fig. 1d′′).
Taken together, these results suggest a correlationbetween increased levels of �-Gal and �-Hex in plasmaand early stages of dementia in both AD and AD-T2DM patients.
Lysosomal β-Gal, β-Hex and α-Man activitiesin PMNCs
The activity of �-Gal was significantly lower inPMNCs isolated from both AD (125 ± 32 mU/mg)and AD-T2DM patients (98.7 ± 20.77 mU/mg) thanin cells from the control group (247 ± 50 mU/mg).Only a slight decrease was instead observed inT2DM cells (191.8 ± 32.16 mU/mg) (Fig. 2A). Simi-larly, levels of �-Man activity were high in controls(278.3 ± 58.5 mU/mg), as compared to AD (137.2 ±
790 R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD
Fig. 1. Activity of lysosomal enzymes in blood plasma. Scatter plots of enzyme activity (mU/ml) measured in plasma samples. A) �-Galac-tosidase; B) �-Mannosidase; C) total �-Hexosaminidase content (i.e. isoenzymes A and B) (MUG substrate); and D) �-Hexosaminidase A(MUGS substrate), were tested in 122 age-matched controls (CTRL), 114 Alzheimer’s disease patients (AD), 109 individuals with type 2diabetes (T2DM), and 116 patients with type 2 diabetes affected by AD. Horizontal lines indicate mean values. ∗: p ≤ 0.05 versus CTRL.a′: Activity of �-Galactosidase measured in plasma of healthy controls (CTRL) and AD patients with mild (CDR1), moderate (CDR2), andsevere dementia (CDR3). Values are expressed as means ± SD. ∗: p < 0.05 versus healthy donors (CTRL). a′′: Activity of �-Galactosidase inplasma of T2DM patients and AD-T2DM patients with mild (CDR1), moderate (CDR2), severe dementia (CDR3); ∗: p < 0.05 versus T2DM.The same biological samples were also employed to determine �-Mannosidase (panels b′; b′′ ), total �-Hexosaminidase (i.e., isoenzymes A andB) (panels c′ and c′′ ) and �-Hexosaminidase isoenzyme A (panels d′; d′′ ) activities.
On the other hand, the activity of total �-Hex showedno significant differences between AD (2535 ±729 mU/mg), T2DM (3059.6 ± 385 mU/mg), and con-trol groups (3148 ± 362 mU/mg), whereas valuesassociated with AD-T2DM patients (1909 ± 322mU/mg) were found to be sharply lower (Fig. 2C).The same trend was shown by HexA, with val-ues of 231 ± 65 mU/mg, 289 ± 46 mU/mg, and 291 ±65 mU/mg in AD, T2DM and control samples, respec-tively, and 178 ± 32 mU/mg measured in the cells
of AD-T2DM patients (Fig. 2D). As observed inplasma, no relationship was found between intracel-lular enzyme activities and gender in any of the groupsstudied.
When the results were sorted based on the cog-nitive status of AD and AD-T2DM patients, wenoted that the activity levels were consistently loweracross the three clinical stages of dementia in bothgroups (Fig. 2a′, 2a′′ ). In particular, �-Man activ-ity was unchanged in AD patients with CDR1(128 ± 36 mU/mg), CDR2 (140 ± 30 mU/mg), andCDR3 (136 ± 26 mU/mg) (p < 0.01 versus controls)
R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD 791
Fig. 2. Activity of lysosomal enzymes in peripheral blood mononuclear cells (PMNCs). Scatter plots populated with values of enzyme activitymeasured in PMNCs and reported as mU/mg. Horizontal line indicate mean values. ∗: p ≤ 0.05 versus healthy donors (CTRL). Abbreviationsare as shown in Fig. 1.
(Fig. 2b′ ). Similarly, levels of this enzyme werelower in AD-T2DM patients with CDR1 and CDR2(71.8 ± 26.84 mU/mg and 76.15 ± 12 mU/mg, respec-tively) (p < 0.05 versus T2DM group), but showedcomparable values in AD-T2DM patients with CDR3and T2DM subjects (Fig. 2b′′ ).
�-Hex activity was found to decrease sharply dur-ing the initial stages of AD (CDR1 = 2304 ± 399 mU/mg; CDR2 2274 ± 677 mU/mg; p < 0.05 versus con-trols), however CDR3 patients displayed values(2985 ± 395 mU/mg) that were within the range ofthe control group (Fig. 2c′ ). Similarly, �-HexA activ-ity was lower in AD patients with CDR1 and CDR2(208.1 ± 48 mU/mg and 193.7 ± 39.6 mU/mg, respec-tively) (p < 0.05 versus controls), while measurements
performed in CDR3 patients (290 ± 35 mU/mg) wereidentical to controls (Fig. 2d′ ). No substantial differ-ences were found in the AD-T2DM group, except formarkedly decreased �-Hex and �-HexA activitiesin the CDR3 population. In particular, total �-Hexactivity was 1901.5 ± 202 mU/mg (CDR1), 1917.9± 221 mU/mg (CDR2) and 1917.2 ± 375 mU/mg(CDR3) (p < 0.01 versus controls) (Fig. 2c′′ ), while�-HexA activity was 163.7 ± 33.7 mU/mg (CDR1),175 ± 30.5 mU/mg (CDR2), and 182 ± 32.4 mU/mg(CDR3) (p < 0.01 versus T2DM group) (Fig. 2d′′ ).
Similar to the results observed in plasma, thisoutcome is indicative of significant changes in intracel-lular �-Gal and �-Hex activities occurring early duringthe progression of dementia.
792 R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD
Use of β-Gal and β-Hex assays to discriminateAD-T2DM and T2DM populations
The statistically different values of intracellularand secreted �-Gal and �-Hex activities measuredin patients and age-matched controls prompted us toinvestigate in greater detail the diagnostic potential ofthese enzymes. To this end, we generated ROC curves(Figs 3, 4), namely plots of sensitivity (true positives)against 100-specificity (false positives) [58]. In partic-ular, we investigated the potential of enzyme assaysto discriminate i) AD-T2DM patients from T2DMsubjects; and ii) AD patients from controls. AD andAD-T2DM patients were grouped based on mild-to-moderate (CDR ≤2) or severe (CDR3) dementia.
Enzyme assays using plasma as the sampleβ-Gal activity. With AD-T2DM samples from
CDR ≤2 patients, the optimal cut-off point of secreted�-Gal activity to distinguish AD-T2DM patientsfrom T2DM subjects was 4.21 mU/ml (true posi-tives = 79.49%, false positives = 22.13%, Area Underthe Curve (AUC) = 0.881 [0.828–0.934], p < 0.001).For patients with CDR3, the optimal cut-off pointwas 3.99 mU/ml (true positives = 72.46%, false pos-itives = 29.51%, AUC = 0.778 [0.71–0.84], p < 0.001)(Fig. 3A, upper panel). Secreted �-Gal showed sim-ilar results when AD patients were compared tocontrols: for patients with CDR ≤2, the optimal cut-off point was 6.01 mU/ml, with true positives andfalse positives rates of 90.9% and 7.38%, respectively(AUC = 0.97 [0.958–0.994], p < 0.001); for CDR3patients, the optimal cut-off point was 5.2 mU/ml,with true and false positive rates of 80.2% and 31%,respectively (AUC = 0.824 [0.771–0.876], p < 0.001)(Fig. 3A, lower panel).
β-Hex activity. With samples from AD-T2DMpatients with CDR ≤2, ROC analyses revealed anoptimal cut-off point of 999.7 mU/ml, correspond-ing to 92.31% true positives and 5.45% falsepositives (AUC = 0.984 [0.967–1.00], p < 0.001); forpatients with CDR3, the optimal cut-off point was981.6 mU/ml, corresponding to 94.2% true positivesand 7.4% false positives (AUC = 0.977 [0.985–0.996],p < 0.001) (Fig. 3B, upper panel). Comparable per-formances were observed when AD patients andthe control group were compared: for patients withCDR ≤2, optimal cut-off, true and false positive rateswere 571.3 mU/ml, 83.64% and 18.1%, respectively(AUC = 0.91 [0.853–0.948], p < 0.001); for patientswith CDR3, the cut-off was 576.8 mU/ml (true pos-itives = 75%, false positives = 16.4%, AUC = 0.858
[0.795–0.922], p < 0.001) (Fig. 3B, lower panel). Iden-tical discriminatory performances were observed byusing the corresponding values of HexA activity (notshown).
α-Man activity. As expected, due to largely over-lapping values across groups, �-Man activity did notshow any diagnostic potentials (Fig. 3C).
Enzyme assays using PMNCs as the sampleβ-Gal activity
To distinguish AD-T2DM patients with CDR ≤2from T2DM subjects, ROC analysis indicated an opti-mal cut-off point of 126.1 mU/mg, corresponding toa true positive rate of 99.49% and a false positiverate of 5.0% (AUC = 0.993 [0.98–1.00], p < 0.001);for patients with CDR3, the optimal cut-off pointwas 158.1 mU/mg (true positives = 99%, false pos-itives = 18.5%, AUC = 0.98 [0.973–0.997], p < 0.001(Fig. 4A, upper panel). A similar performance of �-Galactivity was observed when AD patients with CDR ≤2were compared to controls: the optimal cut-off pointwas 165.2 mU/mg, with true positive and false posi-tive rates of 98.8% and 2%, respectively (AUC = 0.998[0.996-1.00], p < 0.001); for CDR3 patients, the opti-mal cut-off point was 180 mU/mg, true positives were93.3% and false positives 5% (AUC = 0.989 [0.978-1.00], p < 0.001) (Fig. 4A, lower panel).
β-Hex activity. The optimal cut-off point to dis-tinguish AD-T2DM patients with CDR ≤2 fromthe T2DM group was 2328 mU/mg, correspond-ing to 100% true positives and 0% false positives(AUC = 1.00, p < 0.001); for AD-T2DM patientswith a CDR3 score, the cut-off value was instead2580 mU/mg, to which were associated 95.65% truepositives and 6.5% false positives (AUC = 0.992[0.985–1.00], p < 0.001) (Fig. 4B, upper panel). Thus,these results point to �-Hex activity as a good candi-date for the diagnostic evaluation of AD developmentand progression in type 2 diabetic populations.
To separate AD and control groups, the optimalcut-off value was 2841 mU/mg for AD patients withmild-to-moderate dementia (CDR ≤2); in this regard,true and false positives rates were 70.9% and 23%,respectively (AUC = 0.847 [0.786–0.908], p < 0.001).However, the enzyme test showed a poor discrimi-natory performance on the evaluation of AD patientsaffected by severe dementia (CDR3), based upon a truepositive rate of 54%, which is indicative of largelyoverlapping values (Fig. 4B, lower panel). Identicaldata were obtained with HexA measurements (data notshown).
R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD 793
Fig.
3.R
ecei
ver
oper
atin
gch
arac
teri
stic
(RO
C)
curv
esof
plas
ma-
secr
eted
lyso
som
alen
zym
eac
tiviti
es.
Show
nar
ese
nsiti
vity
(tru
epo
sitiv
es)
vers
us10
0-sp
ecifi
city
(fal
sepo
sitiv
es)
curv
esof
�-G
alac
tosi
dase
(A),
tota
l�-H
exos
amin
idas
e*(B
);an
d�
-Man
nosi
dase
(C)a
ctiv
ities
mea
sure
din
the
plas
ma
ofi)
AD
-T2D
Man
dT
2DM
grou
ps(t
oppa
nels
);an
dii)
AD
and
age-
mat
ched
cont
rol
grou
ps(b
otto
mpa
nels
).*
Tota
l�-H
exos
amin
idas
e:is
oenz
yme
A+
isoe
nzym
eB
.
Fig.
4.R
ecei
ver
oper
atin
gch
arac
teri
stic
(RO
C)
curv
esof
lyso
som
alen
zym
eac
tiviti
esm
easu
red
inPM
NC
s.Sh
own
are
sens
itivi
ty(t
rue
posi
tives
)ve
rsus
100-
spec
ifici
ty(f
alse
posi
tives
)cu
rves
of�
-Gal
acto
sida
se(A
),�
-Hex
osam
inid
ase*
(B),
and
�-M
anno
sida
se(C
)ac
tiviti
esm
easu
red
inPM
NC
sof
i)A
D-T
2DM
and
T2D
Mgr
oups
;and
ii)A
Dan
dag
e-m
atch
edco
ntro
lgro
ups.
*Tot
al�
-Hex
osam
inid
ase:
isoe
nzym
eA
+is
oenz
yme
B.
794 R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD
α-Man activity. Assays performed in AD-T2DM andT2DM groups revealed, for patients with CDR ≤2,an optimal cut-off value of 84.22 mU/mg (true pos-itives = 74.3%, false positives = 27.8%, AUC = 0.81[0.746–0.908], p < 0.001) (Fig. 4C, upper panel); bycontrast, when data were obtained from patients witha CDR3 score, the test showed an unsatisfactory per-formance, based on a true positive rate of 45%.
�-Man activity proved to be much more informa-tive in distinguishing AD patients with CDR ≤2 fromcontrols: the optimal cut-off point was 185.8 mU/mg,with true and false positive rates of 94.55% and 5.5%,respectively (AUC = 0.995 [0.99–1.00], p < 0.001;for CDR3 patients, the optimal cut-off value was173.5 mU/mg and the true/false positive rates were91.11% and 5%, respectively (AUC = 0.996 [0.991–1.00], p < 0.001) (Fig. 4C, lower panel).
Thus, these results suggest that both secreted andintracellular �-Gal and �-Hex activities, as well asintracellular �-Man activity, may be invaluable tools todiscriminate AD patients from age-matched individu-als, and/or AD-T2DM patients from T2DM subjects.In general, data proved to be much more robust andreliable when obtained from samples of AD patientsin the early stages (CDR ≤2) of dementia progression.
DISCUSSION
The identification of new peripheral biomarkerscombined with available diagnostic criteria may behelpful for the early diagnosis of AD. Here, we havereported the results from a prospective study thatincluded four groups of subjects, namely AD, T2DM,and AD-T2DM patients and healthy, elderly volun-teers. Based on ROC curve analyses, we show thatplasma- and PMNC-associated activities of two lyso-somal enzymes, �-Gal and �-Hex, have diagnosticvalue for the detection of early stages of dementia inAD and AD-T2DM patients. The activity of anotherlysosomal enzyme, �-Man, demonstrated a compa-rable diagnostic performance, despite being limitedto measurements performed in PMNCs. The resultswere generated from a large screen of patients (morethan 100 per group) and, for comparison purposes,122 age-matched, healthy donors. The research studyinvolved three medical centers that carried out, inde-pendently, collection, processing and analysis of thesamples based on a common standard operating pro-cedure. Despite being produced at separate locationsand by different personnel, the results yielded very lowand reproducible intra- and inter-assay CVs. Excel-
lent assay reproducibility was achieved in light ofvery well-established, routine protocols for samplecollection, and the simple and homogeneous assayformat configured for the determination of enzymeactivities.
ROC curve analyses revealed that these assays maydiscriminate AD patients from age-matched controlsand AD-T2DM patients from T2DM subjects. In thisregard, excellent diagnostic performance of the testswas revealed by the rates of true positives, which con-sistently exceeded 95%, when PMNCs were used as thesource of activity, or ranged between 80% and 90% inthe cell-free, plasma-based assay version. However, itmust be noted that such discriminatory potential wasachieved when the results of enzyme assays specifi-cally included AD and AD-T2DM patients with eithermild or moderate dementia (CDR ≤2), and ruled outdata obtained from patients affected by severe cogni-tive impairments (CDR = 3).
Changes in enzyme activities are most likely due to aprogressive alteration of the lysosomal system, whichwas widely described as one of the earliest intracellularevents occurring in AD [32]. Our results are consistentwith previous observations showing that almost 10%of lysosomal enzymes are secreted in the extracellu-lar milieu under physiological conditions, whereas thispercentage increases in certain pathological conditionssuch as inflammatory or neurodegenerative diseases[60]. In particular, PMNCs boost secretion rates oflysosomal enzymes under inflammatory or cytokinestimulation [60], two conditions already reported inAD and T2DM [11, 61, 62]. Interestingly, levels ofpro-inflammatory cytokines are more abundant in ADpatients with CDR1 and CDR2, as compared to CDR3patients suffering major cognitive declines [11, 61].This may therefore justify, at least in part, the con-sistency of the results generated from the screen ofpatients with a CDR ≤2, which featured low false pos-itive (5%) and false negative (2%) rates as opposedto ≈15% for both rates in the CDR3 tests. Likely, theheterogeneity of the CDR3 group is due to a severeclinical status affecting glycemic control, nutrition,and/or therapeutic treatment.
The intracellular activity of �-Man proved to be avery effective tool, with performance comparable tothat of �-Gal and �-Hex, in distinguishing AD patientsfrom age-matched controls, but not AD-T2DM andT2DM groups. Further, no differences in enzyme activ-ity were found when blood plasma was used as thesource of activity across the four groups investigated.These findings suggest that �-Man, like �-Gal and�-Hex, is subject to a significant alteration of its expres-
R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD 795
sion in AD peripheral cells. However, it is possiblethat metabolic compensation can explain the resultsin AD-T2DM and T2DM patients, in light of largechanges of �-Man activity in plasma and serum ofpatients affected by diabetes mellitus [43–44, 46, 63]as well as related complications, such as diabeticretinopathy [64], nephropathy [65], and macrovascu-lar complications in the elderly [66]. The modulationof intracellular levels of lysosomal hydrolases andtheir secretory pathways may be related to an alteredganglioside metabolism. For example, the differen-tial expression of �-Gal (which hydrolyzes GM1 intoGM2) and �-HexA (which converts GM2 into GM3)in AD and T2DM patients and, more evidently, in AD-T2DM group, is consistent with previously publisheddata showing a dysfunctional ganglioside network inAD and T2DM patients leading to the accumulationof GM2 and GM3 gangliosides in both pathologicalconditions [67–71].
In conclusion, we have shown that both intracellu-lar and secreted forms of the lysosomal enzymes �-Galand �-Hex may be considered peripheral biomarkersfor the early diagnosis of AD, particularly for at-riskindividuals who are affected by T2DM. Their discrim-inating potential, demonstrated by ROC curves, offersa starting point towards the definition of reliable diag-nostic tools for routine applications.
FINANCIAL SUPPORT
This study was supported by the Italian Minis-tero della Salute [grant no. RF-UMB-2006-339457 toAn.O. (REG.17O), F.S., P.M., and Al.O., in additionto nos. EBRI.10, PS05.11, and PS05.21 to An.O.], theComitato Telethon Fondazione Onlus, the Amminis-trazione Autonoma dei Monopoli di Stato (AAMS), andthe City of Gubbio [grant no. GGP06209 to A.O.].
ACKNOWLEDGMENTS
We are grateful to patients, healthy donors and theirfamily members for their participation in this study.We wish to thank the Genetic Bank of the Labor-atorio di Neurogenetica, Centro Europeo del Cervello-Istituto di Ricovero e Cura a Carattere Scientifico(CERC-IRCCS) Santa Lucia, Rome, Italy (http://www.hsantalucia.it/neurogen/index en.htm) for invaluableassistance with samples and the “University of Peru-gia Statistical Computation Center” for help with dataanalysis.
Authors’ disclosures available online (http://www.j-alz.com/disclosures/view.php?id=738).
REFERENCES
[1] Ferri CP, Prince M, Brayn C (2005) Global prevalence ofdementia: a Delphi Consensus Study. Lancet 366, 2112-2117.
[2] Jellinger KA (2006) Alzheimer 100 – highlights in the historyof Alzheimer research. J Neural Transm 113, 1603-1623.
[3] Patterson CJ, Feightner W, Garcia A, Hsiung GY, MacKnightC, Sadovnick AD (2008) Diagnosis and treatment of demen-tia: 1. Risk assessment and primary prevention of Alzheimerdisease. Can Med Assoc J 178, 548-556.
[4] Ansbury PT, Lashuel HA (2006) A century-old debate onprotein aggregation and neurodegeneration enters the clinic.Nature 443, 774-779.
[5] Turner RS (2006) Alzheimer’s disease. Semin in Neurol 26,499-506.
[6] Ohyagi Y, Tabira T (2006) Intracellular amyloid beta-proteinand its associated molecole in the pathogenesis of Alzheimer’sdisease. Mini Rev Med Chem 6, 1075-1080.
[7] Goldgaber D, Lerman MI, Mcbride OW (1987) Characteriza-tion and chromosomal localization of a cDNA encoding brainamyloid of Alzheimer’s disease. Science 235, 877-880.
[8] Smach MA, Charfeddine B, Lammouchi T, Harrabi I, OthmanLB, Dridi H, Bennamou S, Limen K (2008) CSF �-amyloid1-42 and tau in Tunisian patients with Alzheimer’s disease:the effect of APOE �4 allele. Neurosci Lett 440, 145-149.
[9] Lewczuk P, Wiltfang J (2008) Neurochemical dementia diag-nostics: state of the art and research perspectives. Proteomics8, 1292-1301.
[10] Rowe CC, Ackerman U, Browne W, Mulligan R, Pike KL,O’Keefe G, Tochon-Danguy H, Chan G, Berlangieri SU,Jones G, Dickinson-Rowe KL, Kung HP, Zhang W, KungMP, Skovronsky D, Dyrks T, Holl G, Krause S, FriebeM, Lehman L, Lindemann S, Dinkelborg LM, Masters CL,Villemagne VL (2008) Imaging of amyloid beta inAlzheimer’s disease with 18F-BAY94-9172, a novel PETtracer: proof of mechanism. Lancet Neurol 7, 129-135.
[11] Ray S, Britschgi M, Herbert C, Takeda-Uchimura Y, BoxerA, Blennow K, Friedman LF, Galasko DR, Jutel M, KarydasA, Kaye JA, Leszek J, Miller BL, Minthon L, Quinn JF,Rabinovici GD, Robinson WH, Sabbagh MN, So YT, SparksDL, Tabaton M, Tinklenberg J, Yesavage JA, Tibshirani R,Wyss-Coray T (2007) Classification and prediction of clini-cal Alzheimer’s diagnosis based on plasma signaling proteins.Nat Med 13, 1359-1362.
[12] Lovestone S, Francis P, Kloszewska I, Mecocci P, SimmonsA, Soininen H, Spenger C, Tsolaki M, Vellas B, WahlundLO, Ward M (2009) AddNeuroMed Consortium AddNeu-roMed – the European collaboration for the discovery of novelbiomarkers for Alzheimer’s disease. Ann N Y Acad Sci 1180,36-46.
[13] Sims-Robinson C, Kim B, Rosko A, Feldman EL (2010) Howdoes diabetes accelerate Alzheimer disease pathology? NatRev Neurol 6, 551-559.
[14] Mattson MP (2004) Pathways towards and away fromAlzheimer’s disease. Nature 430, 631-639.
[15] Selkoe DJ (1991) The molecular pathology of Alzheimer’sdisease. Neuron 6, 487-498.
[16] Goedert M, Spillantini MG (2006) A century of Alzheimer’sdisease. Science 314, 777-781.
796 R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD
[17] Orlacchio A, Bernardi G, Orlacchio A, Martino S (2007) RNAinterference as a tool for Alzheimer’s disease therapy. MiniRev Med Chem 7, 1166-1176.
[18] Costanzi E, Martino S, Persichetti E, Tiribuzi R, Massini C,Bernardi G, Orlacchio AN, Orlacchio A (2008) Effects ofvtamin C on fibroblasts from sporadic Alzheimer’s diseasepatients. Neurochem Res 33, 2510-2515.
[19] Martino S, di Girolamo I, Orlacchio A, Datti A, OrlacchioA (2009) MicroRNA implications across neurodevelopmentand neuropathology. J Biomed Biotechnol 2009, 654346.
[20] Haan MN (2006) Therapy insight: type 2 diabetes mellitusand the risk of late-onset Alzheimer’s disease. Nat Clin PractNeurol 2, 159-166.
[21] Strachan MWJ, Deary IJ, Ewing FME, Frier BM (1997) Istype II diabetes associated with an increased risk of cognitivedysfunction? A critical review of published studies. DiabetesCare 20, 438-445.
[22] Neumann KF, Rojo L, Navarrete LP, Farıas G, Reyes P,Maccioni RB (2008) Insulin resistance and Alzheimer’s dis-ease: molecular link & clinical implication. Curr Alz Res 5,438-447.
[23] Takeda S, Sato N, Uchio-Yamada K, Sawada K, Kunieda T,Takeuchi D, Kurinami H, Shinohara M, Rakugi H, MorishitaR (2009) Elevation of plasma beta-amyloid level by glucoseloading in Alzheimer mouse models. Biochem Biophys ResCommun 385, 193-197.
[25] Li L, Holscher C (2007) Common pathological processes inAlzheimer disease and type 2 diabetes: a review. Brain ResRev 56, 384-402.
[26] Zhao WQ, Townsend M (2009) Insulin resistance and amy-loidogenesis as commo molecular foundation for type 2diabetes and Alzheimer’s disease. Biochim Biophys Acta1792, 482-496.
[27] Janson J, Laedtke T, Parisi JE, O’Brien P, Petersen RC, ButlerPC (2004) Increased risk of type 2 diabetes in Alzheimerdisease. Diabetes 53, 474-481.
[28] Bradley W, Doble, Woodgett JR (2003) GSK-3: tricks of thetrade for a multi-tasking kinase. J Cell Sci 116, 1175-1186.
[29] Qiu WQ, Folstein MF (2006) Insulin, insulin-degradingenzyme and amyloid-beta peptide in Alzheimer’s disease:review and hypothesis. Neurobiol Aging 2, 190-198.
[30] Nixon RA, Yang DS, Lee JH (2008) Neurodegenerative lyso-somal disorders: a continuum from development to late age.Autophagy 4, 590-599.
[31] Adamec E, Mohan PS, Cataldo AM, Vonsattel JP, Nixon RA(2000) Up-regulation of the lysosomal system in experimen-tal models of neuronal injury: implications for Alzheimer’sdisease. Neuroscience 100, 663-675.
[32] Nixon RA, Mathews PM, Cataldo AM (2001) The neu-ronal endosomal-lysosomal system in Alzheimer’s disease.J Alzheimers Dis 3, 97-107.
[33] Hardy J, Selkoe DJ (2002) The amyloid hypothesis ofAlzheimer’s disease: progress and problems on the road totherapeutics. Science 297, 353-356.
[34] Lee HG, Zhu X, Nunomura A, Perry G, Smith MA (2006)Amyloid beta: the alternate hypothesis. Curr Alz Res 3, 75-80.
[35] La Ferla FM, Green KN, Oddo S (2007) Intracellular amyloid-beta in Alzheimer’s disease. Nat Rev Neurosci 8, 499-509.
[36] Cataldo AM, Peterhoff CM, Troncoso JC, Gomez-Isla T,Hyman BT, Nixon RA (2000) Endocytic pathway abnormali-ties precede amyloid beta deposition in sporadic Alzheimer’sdisease and Down syndrome: differential effects of APOE
genotype and presenilin mutations. Am J Pathol 157, 277-286.
[37] Cataldo AM, Nixon RA (1990) Enzymatically active lyso-somal proteases are associated with amyloid deposits inAlzheimer brain. Proc Natl Acad Sci U S A 87, 3861-3865.
[38] Hook Vivian YH, Kindy M, and Gregory H (2008) Inhibitorsof cathepsin B improve memory and reduce A�42 in trans-genic Alzheimer’s disease mice expressing the wild-type, butnot the swedish mutant, �-secretase app site. J Biol Chem 283,7745-7753.
[39] Cataldo AM, Barnett JL, Mann DM, Nixon RA (1996) Colo-calization of lysosomal hydrolase and �-amyloid in diffuseplaques of the cerebellum and striatum in Alzheimer’s diseaseand Down’s syndrome. Neuropathol Exp Neurol 55, 704-715.
[41] Urbanelli L, Emiliani C, Massini C, Persichetti E, OrlacchioA, Pelicci G, Sorbi S, Hasilik A, Bernardi G, Orlacchio A(2008) Cathepsin D expression is decreased in Alzheimer’sdisease fibroblasts. Neurobiol Aging 29, 12-22.
[42] Maciejewski R, Burski K, Baj J, Madej B, Burdan F,Dabrowski A, Brakowiecki F (2001) Changes in pancreaticlysosomal enzymes activity as the potential factors leading todiabetic enteropathy. J Physiol Pharmacol 52, 823-834.
[43] Waters PJ, Flynn MD, Corrall RJ, Pennock CA (1992)Increases in plasma lysosomal enzymes in type 1 (insulin-dependent) diabetes mellitus: relationship to diabeticcomplications and glycaemic control. Diabetologia 35, 991-995.
[44] Burlina AB, Goi G, Fabi A, Lombardo A, Gaburro D,Tettamanti G (1987) Behaviour of some lysosomal enzymesin the plasma of insulin dependent diabetic patients duringartificial pancreas treatment. Clin Biochem 20, 423-427.
[45] Walkley SU (2009) Pathogenic cascades in lysosomal disease– Why so complex? J Inherit Metab Dis 32, 181-189.
[46] Poon PY, Davis TM, Dornan TL, Turner RC (1983) PlasmaN-acetyl-beta, D-glucosaminidase activities and glycaemiain diabetes mellitus. Diabetologia 24, 433-436.
[47] McKhann G, Drachman D, Folstein M, Katzman R, PriceD, Stadlan EM (1984) Clinical diagnosis of Alzheimer’s dis-ease: report of the NINCDS-ADRDA Work Group under theauspices of Department of Health and Human Services TaskForce on Alzheimer’s Disease. Neurology 34, 939-944.
[48] Folstein MF, Folstein SE, McHugh PR (1975) “Mini-MentalState”. A practical method for grading the cognitive state ofpatients for the clinicians. J Psychiatr Res 12, 189-198.
[49] Morris JC (1997) Clinical dementia rating: a reliable and validdiagnostic and staging measure for dementia of the Alzheimertype. Int Psychogeriatr 9, 173-176.
[50] Chaves ML, Camozzato AL, Godinho C, Kochhann R, SchuhA, de Almeida VL, Kaye J (2007) Validity of the clini-cal dementia rating scale for the detection and staging ofdementia in Brazilian patients. Alzheimer Dis Assoc Disord21, 210-217.
[51] American Diabetes Association (2004) Diagnosis and classi-fication of diabetes mellitus. Diabetes Care 27, S5-S10.
[52] Rai AJ, Gelfand CA, Haywood BC, Warunek DJ, Yi J,Schuchard MD, Mehigh RJ, Cockrill SL, Scott GB, TammenH, Schulz-Knappe P, Speicher DW, Vitzthum F, Haab BB,Siest G, Chan DW. (2005) HUPO Plasma Proteome Projectspecimen collection and handling: towards the standardiza-tion of parameters for plasma proteome samples. Proteomics5, 3262-3277.
R. Tiribuzi et al. / Lysosomal Enzymes and Clinical Stages of AD 797
[53] Martino S, Tiribuzi R, Tortori A, Conti D, Visigalli I, LattanziA, Biffi A, Gritti A, Orlacchio A (2009) Specific determina-tion of {beta}-Galactocerebrosidase activity via competitiveinhibition of {beta}-Galactosidase. Clin Chem 55, 541-548.
[54] Sopelsa AM, Severini MH, Da Silva CM, Tobo PR, GiuglianiR, Coelho JC (2000) Characterization of beta-galactosidasein leukocytes and fibroblasts of GM1 gangliosidosis het-erozygotes compared to normal subjects. Clin Biochem 33,125-129.
[55] Beccari T, Mancuso F, Costanzi E, Tassi C, Barone R,Fiumara A, Orlacchio A, Aisa MC, Orlacchio A (2000)�-Hexosaminidase, �-mannosidase and �-mannosidaseexpression in serum from patients with carbohydrate deficientglycoprotein syndrome type I. Clin Chim Acta 302, 125-132.
[56] Martino S, di Girolamo I, Cavazzin C, Tiribuzi R, Galli R,Rivaroli A, Valsecchi M, Sandhoff K, Sonnino S, VescoviA, Gritti A, Orlacchio A (2009) Neural precursor cell cul-tures from GM2 gangliosidosis animal models recapitulate thebiochemical and molecular hallmarks of the brain pathology.J Neurochem 109, 135-147.
[57] Bradford MM (1976) A rapid and sensitive method forthe quantitation of microgram quantities of protein utiliz-ing the principle of protein-dye binding. Anal Biochem 72,248-254.
[58] Zweig MH, Campbell G (1993) Receiver-operating charac-teristic (ROC) plots: a fundamental evaluation tool in clinicalmedicine. Clin Chem 39, 561-565.
[59] Schernthaner G (2010) Diabetes and cardiovascular disease:is intensive glucose control beneficial or deadly? Lessonsfrom ACCORD, ADVANCE, VADT, UKPDS, PROactive,and NICE-SUGAR. Wien Med Wochenschr 160, 8-19.
[61] Motta M, Imbesi R, Di Rosa M, Stivala F, Malaguarnera L(2007) Altered plasma cytokine levels in Alzheimer’s disease:correlation with the disease progression. Immunol Lett 30,46-51.
[62] Wu W, Wang M, Sun Z, Wang X, Miao J, Zheng Z (2010)The predictive value of TNF-alpha and IL-6 and the incidenceof macrovascular complications in patients with type 2 dia-betes. Acta Diabetol, Published online May 21. DOI 10.1007/s00592-010-0198-0
[63] Hultberg B, Isaksson A, Agardh E, Agardh CD (1995) Plasmabeta-hexosaminidase isoenzymes A and B exhibit differentrelations to blood glucose levels in a population of type 1diabetic patients. Scand J Clin Lab Invest 55, 723-728.
[64] Pitkanen E, Kyllastinen M, Koivula T, Hormila P (1980)beta-N-acetylglucosaminidase and beta-glucuronidase activ-ities in insulin-dependent diabetic subjects with retinopathy.Diabetologia 18, 275-278.
[65] Jung K, Pergande M, Schimke E, Ratzmann KP, Ilius A (1988)Urinary enzymes and low-molecular-mass proteins as indica-tors of diabetic nephropathy. Clin Chem 34, 544-547.
[66] Weitgasser R, Schnoell F, Gappmayer B, Kartnig I (1999)Prospective evaluation of urinary N-acetyl-beta-D-gluco-saminidase with respect to macrovascular disease in elderlytype 2 diabetic patients. Diabetes Care 22, 1882-1886.
[67] Wakabayashi M, Okada T, Kozutsumi Y, Matsuzaki K (2005)GM1 ganglioside-mediated accumulation of amyloid �-protein on cell membranes. Biochem Biophys Res Commun328, 1019–1023.
[68] Barrier L, Ingrand S, Damjanac M, Rioux Bilan A, HugonJ, Page G (2007) Genotype-related changes of gangliosidecomposition in brain regions of transgenic mouse models ofAlzheimer’s disease. Neurobiol Aging 28, 1863-1872.
[69] Yanagisawa K (2007) Role of gangliosides in Alzheimer’sdisease. Biochim Biophys Acta 1768, 1943-1951.
[70] Matsuzaki K, Kato K, Yanagisawa K (2010) Abeta poly-merization through interaction with membrane gangliosides.Biochim Biophys Acta 1801, 868-877.
[71] Yamamoto N, Taniura H, Suzuki K (2010) Insulin inhibitsAbeta fibrillogenesis through a decrease of the GM1 gangli-oside-rich microdomain in neuronal membranes. J Neuro-chem 113, 628-636.
