Immunostaining with 22C11 of whole platelet homogenate from 1 control subject (lane 1), 2 patients with non–Alzheimer disease (AD) dementia (lane 2), and 2 patients with AD (lane 3). Arrows indicate positions of amyloid β precursor protein isoforms and their apparent molecular weight.
Density ratio between the optical density of the upper band (130 kd) and the lower bands (106-110 kd) of platelet amyloid β precursor protein isoforms in 15 control subjects, 16 patients with non–Alzheimer disease dementia (NADD), and 32 patients with Alzheimer disease (AD). Ratio of platelet amyloid β precursor protein isoforms in the NADD group was not statistically significantly different from the control group (P=.55).
Double immunostaining in whole platelet homogenates from control subjects (lane 1) and patients with mild (lane 2), moderate (lane 3), and severe (lane 4) Alzheimer disease with m22C11 (solid arrowheads) and monoclonal antibody raised against actin (open arrowhead). Migration of prestained molecular weight standards is shown on the left.
Density ratio of platelet amyloid β precursor protein isoforms of patients with Alzheimer disease (AD) grouped according to severity based on Clinical Dementia Rating score.
Reverse transcriptase polymerase chain reaction (RT-PCR) products from platelet messenger RNAs of control subjects and patients with Alzheimer disease (AD). Lanes 1, 3, and 5, RT-PCR products from platelets of 3 control subjects; lanes 2, 4, and 6, RT-PCR products from platelets of 3 patients with severe AD; lane 7, DNA molecular weight standards ranging from 8 to 587 base pairs (bp) (Marker V; Boehringer Mannheim Corp).
Di Luca M, Pastorino L, Bianchetti A, Perez J, Vignolo LA, Lenzi GL, Trabucchi M, Cattabeni F, Padovani A. Differential Level of Platelet Amyloid β Precursor Protein IsoformsAn Early Marker for Alzheimer Disease. Arch Neurol. 1998;55(9):1195-1200. doi:10-1001/pubs.Arch Neurol.-ISSN-0003-9942-55-9-noc7558
Copyright 1998 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1998
To determine whether a differential level of platelet amyloid β precursor protein (APP) isoforms is specifically related to Alzheimer disease (AD) and whether it shows a correlation with the progression of clinical symptoms.
After subjects were grouped according to diagnosis and severity of dementia, APP isoform levels in platelets were compared.
University medical centers.
Thirty-two patients who fulfilled diagnostic criteria for probable AD, 25 age-matched control subjects, and 16 patients with non-AD dementia.
Main Outcome Measure
The levels of APP isoforms were evaluated by means of Western blot analysis and immunostaining of whole platelets. Messenger RNAs for APP transcripts were also evaluated by means of reverse transcriptase polymerase chain reaction.
The ratio between the intensity of the 130-kd and 106- to 110-kd APP isoforms was significantly lower in the AD group (0.31±0.15, mean±SD) compared with both controls (0.84±0.2) and non-AD subjects (0.97±0.4). The ratio of platelet APP isoforms in patients with AD grouped by Clinical Diagnostic Rating score significantly correlated with the severity of the disease (Pearson correlation coefficient, followed by Bonferroni correction, P =.01). Reverse transcriptase polymerase chain reaction experiments showed that APP transcripts in all experimental groups were equally expressed.
The pattern of platelet APP isoforms is specifically altered in patients with AD. In addition, the alteration of platelet APP isoforms shows a positive correlation with the progression of clinical symptoms, supporting the possibility to consider this peripheral parameter as a marker of progression of the disease. These alterations are not related to abnormalities of APP isoforms messenger RNAs in platelets.
ALZHEIMER DISEASE (AD) is a neurodegenerative disorder characterized by progressive intellectual decline associated with senile plaques, neurofibrillary tangles, and amyloid angiopathy as main pathological hallmarks.1- 3 The major proteinaceous component of senile plaques is a self-assembling peptide, amyloid β peptide (Aβ). The Aβ originates by proteolytic processing from a larger precursor, the amyloid β precursor protein (APP),4- 6 which is a transmembrane protein present as numerous alternatively spliced isoforms derived from a single gene localized on human chromosome 21.7- 9 The APP isoforms are expressed in human brain as well as in several nonneural tissues and cell lines.10,11 Numerous studies12- 18 support the hypothesis that the presence of APP in peripheral cells, ie, endothelial and blood cells, may contribute to Aβ deposition.19,20 If this is the case, Aβ of circulating origin might have important implications both in the diagnosis and in the pathogenesis of the disease. Human platelets contain large (>95%) amounts of the circulating APP.21- 23 Three major APP isoforms are present in membrane of resting platelets, and both platelets and megakaryocytes express 3 APP transcripts encoding for APP 695, APP 751, and APP 770.24,25 Activated platelets release membrane fragments containing the mature full-length APP 751/77026 and the soluble, carboxyl-truncated form of APP, stored in alpha granules (protein nexin II).27 On stimulation, platelets are capable of releasing a protein functionally identical to the platelet coagulation factor XIa inhibitor, which has been shown to be a truncated form of APP containing the Kunitz domain.28 There is evidence that modifications in the concentration or processing of APP in platelets are expressed in advanced stages of AD,29- 33 and it has been reported that platelets from individuals with AD show alterations in membrane fluidity.34
These observations define the framework of the present study, which addresses the questions whether (1) differential levels of platelet APP isoforms are specifically related to AD and (2) levels of platelet APP isoforms can be considered a marker for the progression of the disease.
The study was undertaken on 32 patients with AD grouped according to the severity of the disease, 25 age-matched control subjects, and 16 patients affected by non–Alzheimer disease dementia (NADD). The diagnosis of probable AD was made according to National Institute of Neurological Disorders and Stroke–Alzheimer Disease and Related Disorders Association criteria.35 The NADD group consisted of a sample of patients affected by frontal lobe dementia, senile dementia of Lewy body type, systemic amyloidosis, Huntington disease, slow progressive aphasia, progressive supranuclear palsy, and Parkinson disease. Control subjects were drawn from a series of healthy subjects and nondemented hospitalized neurological patients.
Exclusion criteria for all groups were the following: head trauma, metabolic dysfunction, hematologic diseases, alcohol abuse, drug abuse, delirium, mood disorders, and actual treatment with acetylcholinesterase inhibitors or with medications affecting platelet function, such as anticoagulants, antiplatelet drugs, serotoninergic agonists-antagonists, and corticosteroids. All subjects included had a standardized clinical workup based on neurological examinations, laboratory blood and urine analysis, neuropsychological assessment, and neuroimaging study (head computed tomography and/or magnetic resonance imaging). Patients were followed up for at least 1 year before being included in the study. All patients were subjected to neurological and neuropsychological evaluation in the same week as platelet analysis. Before enrollment in the study, each subject or his or her legal caregiver signed an informed consent, after the nature and possible consequence of the study were explained.
Platelet preparation and subsequent analysis were carried out by personnel who were unaware of the patient's diagnosis.
Blood (27 mL) was collected into 3 mL of 3.8% sodium citrate (in the presence of 136-mmol/L glucose), mixed gently, and centrifuged at 200g for 10 minutes.32,36 The maximum time interval between blood drawing and the first centrifugation was 20 to 25 minutes. Platelet-rich plasma was separated from the blood pellet by means of a plastic pipette, with aspiration of the buffy coat avoided. The platelet pellet was washed twice, resuspended at a concentration of 5 mg/mL in ice-cold Tris hydrochloride, 10 mmol/L, pH 7.4, containing ethyleneglycotetraacetic acid, 1 mmol/L; phenylmethylsulfonyl fluoride, 0.1 mmol/L; and a complete set of protease inhibitors (Complete; Boehringer Mannheim Corp, Indianapolis, Ind), including E64, which has been reported to better inhibit APP degradation,18 and subjected to 2 rounds of freeze-thawing and 20 seconds of sonication at 0°C.
Total platelet proteins (6 µg) were separated in 6% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and electroblotted to nitrocellulose membranes in a running buffer containing 0.025-mol/L Tris hydrochloride, 0.192-mol/L glycine, and 20% methanol, pH 8.3, at 230 mA for 2 hours 30 minutes. After blocking with 10% nonfat milk, immunostaining reactions were performed with the following: 22C11 (Boehringer Mannheim Corp; dilution, 1:1500), clone AC40 (Sigma Chemical Co, St Louis, Mo; dilution, 1:1500), domain-specific antibodies 369s (antiserum to amino acid 645-694 COOH-terminus; dilution, 1:5000), 3129 antiserum distinguishing between APP and amyloid precursor–like protein37,38 (dilution, 1:1500), 6E10 (Senetek, St Louis), raised against the aa sequence 1-17 of Aβ peptide (dilution, 1:500), and 4G8 (Senetek), raised against the sequence 17-24 of Aβ peptide (dilution, 1:500). After incubation with peroxidase-conjugated IgG secondary antibodies (Kirkegaard, Gaithersburg, Md; dilution, 1:10000), blots were developed with enhanced chemiluminescence (Renaissance; NEN DuPont, Boston, Mass) and exposed to films (NEN DuPont).
Western blot analysis of platelet APP isoforms was quantified by means of computer-assisted imaging (Image, developed by W. Rasband, PhD, National Institutes of Health, Bethesda, Md). Statistical evaluations were performed according to 1-way analysis of variance followed by Scheffé test as a post hoc comparison test. When appropriate, Pearson correlation test, followed by Bonferroni correction, was applied.
Total cellular RNA was extracted from platelets collected by centrifugation from 50 mL of blood by the acid-phenol method with the use of an RNA extraction kit (Bio/RNA-X Cell; Bio/Gene Limited, Kimbolton, England). Total RNA, 200 ng, was used to perform the first strand complementary DNA synthesis by means of either APP-1213 as a specific primer or oligo(dt) as indicated in the text, and murine Moloney leukemia virus reverse transcriptase (RT) (Perkin Elmer, Branchburg, NJ).21
Reverse transcriptase–polymerase chain reaction (RT-PCR) primers were designed according to Golde et al11 and PCR was performed with minor modifications. Briefly, the reaction mixture contained the enzyme buffer (Perkin Elmer), 50 U of murine Moloney leukemia virus–RT, deoxynucleoside triphosphate at a final concentration of 0.2 mmol/L, and 10 pmol of APP-1213 primer. The RT was performed at 37°C for 1 hour and stopped at 65°C for 10 minutes. First-strand complementary DNA, 70 ng, was subjected to PCR amplification; the samples were brought to a final concentration of 0.2-mmol/L deoxynucleoside triphosphate, 0.5-mmol/L APP+958 and APP-1213, and 2-mmol/L magnesium chloride in the appropriate buffer in the presence of 2.5 U of Taq polymerase (Perkin Elmer). The PCR was carried out in a programmable heating block (Gene Amp PCR System 2400, Perkin Elmer) with the following scheme cycles: denaturation, 94°C, 30 seconds; annealing, 55°C, 30 seconds; extension, 72°C, 1 minute, for 34 cycles. Amplification was linear between 28 and 36 cycles (data not shown). The PCR products were electrophoresed in 3% agarose gel; gels were stained with ethidium bromide and photographed under UV light.
Demographic characteristics of the 3 groups of subjects included in the study are reported in Table 1.
Whole platelet homogenates from each subject were processed for Western blot analysis by means of monoclonal 22C11 antibody raised against the N-terminal domain of APP and therefore recognizing all APP isoforms.
Figure 1 shows a representative immunoblot of platelet APP isoforms in 1 control subject, 2 patients with NADD, and 2 patients with AD. Staining with 22C11 showed that patients with AD have markedly lower levels of the highest platelet APP isoform (apparent molecular weight, 130 kd) than both control and NADD subjects, whereas the immunostaining of the lower bands (106 and 110 kd) showed a similar intensity in all experimental groups. To investigate further the identity of the 130-kd band, altered in platelets of patients with AD, domain-specific antibodies were used. In particular, the 130-kd form was recognized by (1) 369s, raised against the C-terminal domain of the protein; (2) 3129, a serum distinguishing between APP and highly homologous proteins, such as amyloid precursor–like protein39; and (3) 6E10 and 4G8, monoclonal antibodies raised against the aa sequence 1-17 and 17-24 of Aβ peptide, respectively, ruling out the possibility that the 130-kd band could correspond to amyloid precursor–like protein. In addition, 369s, 6E10, and 4G8 recognized the 106-kd band, thus suggesting that these bands are representing membrane inserted forms, with an intact Aβ sequence inserted (data not shown). On the other hand, the 110-kd band was not recognized by 369s serum, and this band might correspond either to amyloid precursor–like protein or to a secreted form. Further experiments are needed to clarify this point.
After the intensity of the bands at 130 and 106 to 110 kd were measured by image analysis, a highly statistically significant difference was found in the ratio of theme isoforms between patients with AD and both control subjects and patients with NADD (Figure 2) (mean±SD; control subjects, 0.84±0.24; NADD, 0.97±0.40; AD, 0.31±0.15; AD vs control group and vs NADD, P <.001). Patients with NADD showed no statistically significant difference when compared with control subjects (P=.55). The ratio between 130 and 106- to 110-kd bands for each patient was determined from at least 3 replications. When the ratio between the immunoreactivity of the 130-kd band and the 106-kd band was measured by means of 369s, a similar difference between control and AD patients was found (data not shown).
To investigate whether the ratio of platelet APP isoforms could show a correlation with the severity of the disease, patients with AD were grouped according to Clinical Dementia Rating score,40 and platelet APP isoforms were measured as described above. Figure 3 shows the 22C11 immunostaining of platelet APP isoforms in 2 control subjects and 2 patients each with mild, moderate, and severe AD. The ratio of platelet APP isoforms varied according to Clinical Dementia Rating score. Levels of platelet APP isoforms were measured in 11 patients with mild, 11 with moderate, and 10 with severe AD. The mean ratio of platelet APP isoforms and associated variance (SD) for the group with mild AD (0.51±0.13) was significantly different than the respective values for moderate (0.35±0.15) and severe (0.25±0.08) AD (mild vs moderate, P =.04; mild vs severe, P =.001). The values between moderate and severe failed to reach statistical significance. In addition, the ratio of platelet APP isoforms for the group with mild AD significantly differed from those for both controls and patients with NADD (P<.01) (Figure 4). A statistically significant correlation was observed also when individual Mini–Mental State Examination scores41 and APP isoforms ratios were considered (Pearson correlation coefficient, followed by Bonferroni correction, P=.01), thus indicating that this value could represent a sensitive measure for the progression of the disease. The ratio in the AD group does not correlate with the age of patients (r=0.1). In addition, immunostaining for actin in each sample showed no difference in the concentration of this cytoskeletal element (P=.70 vs controls), thus ruling out that the decrease observed in APP immunostaining in patients with AD could result from a general degradation of samples.
The mechanism by which the ratio in levels of platelet APP isoforms is decreased in patients with AD is at present only a matter of speculation. One possible explanation could reside in a decreased expression of a specific APP isoform in platelets of patients with AD. To address this point, RT-PCR experiments have been performed by using primers that flank the alternative splice site.11Figure 5 shows RT-PCR products from platelet messenger RNA of 3 control subjects and 3 patients with moderate to severe AD demonstrated by agarose gel and ethidium bromide. The primers used for PCR reaction amplify 3 partial transcript segments encoding APP 695 (87 base pairs), APP 751 (255 base pairs), and APP 770 (312 base pairs). For each sample of total RNA, RT-PCR for glyceraldehyde phosphate dehydrogenase, as housekeeping gene, was performed; amplification of glyceraldehyde phosphate dehydrogenase was similar in all samples (data not shown). This experiment was performed on 5 control subjects, 2 patients with NADD, and 5 patients with AD; in these samples we have confirmed previous findings showing 3 alternative spliced APP messenger RNAs corresponding to APP 695, APP 751, and APP 770. The pattern of APP gene expression was similar among groups, thus ruling out that the differential level of APP isoforms observed in AD can be ascribed to a decreased or absent level of messenger RNA encoding either for APP 770 or for APP 751.
The main finding of the present study is that patients with AD show a specific alteration in levels of platelet APP isoforms. In particular, a decrease in the ratio between the 130-kd APP isoform and the lower (106-110 kd) forms was found in the AD patient group when compared with values of both control subjects and patients affected by other neurodegenerative disorders associated with dementia.
The alteration of platelet APP isoform ratio was different among subsets of patients with AD grouped according to Clinical Dementia Rating scores, and there was a positive correlation with individual Mini–Mental State Examination scores, thus suggesting that this value varies according to the progression of clinical symptoms. Indeed, the ratio of platelet APP isoforms in patients with mild AD showed a significant difference from the control value, indicating that this peripheral biochemical measure is modified early in the course of the disease. This difference was observed despite a minimal overlap between patients with mild AD and the control-NADD group. In a previous study,30 we reported that patients with AD show an altered pattern of APP isoforms in platelets. The results obtained in the present study confirm and extend this previous observation in a new larger population of patients. It is particularly interesting to note that the quantitative evaluation of levels of platelet APP isoforms in patients with moderate to severe AD is strikingly reproducible between this and the previous study, performed in a different population of patients (mean±SD ratio of APP isoforms in the first study: controls [n=10], 0.73±0.15; patients with AD [n=10], 0.16±0.06; second study: controls [n=25], 0.84±0.24; patients with AD [n=32], 0.31±0.15). Similar results have been reported by Rosenberg and colleagues,33 who also showed that the mean ratio of 120- to 130-kd isoform to the 110-kd APP isoform in patients with AD was significantly lower than that of the control subjects (P<.005). Although the altered ratio is consistent in our study and that of Rosenberg et al,33 the relative abundance of APP isoforms is different. Indeed, different groups described alteration in platelet APP in AD with the use of different procedures.18,29- 31 In addition, previous studies,21 although failing to show a clear difference between APP isoforms in platelet from patients with AD compared with control subjects, pointed out the possibility that the detection of APP isoform ratio in a large population of subjects may correlate with the disease. Differences occurring in platelet isolation and processing might, however, be responsible for these discrepancies. Recently, it has been described that the use of different anticoagulants associated with prostaglandin E1 might be responsible for differences in APP secretion in plasma during platelet preparation.42 These observations underlying the reproducibility of this finding, together with our data showing a positive correlation with the progression of clinical symptoms, argue against the hypothesis that a differential level of platelet APP isoforms can be a casual epiphenomenon unrelated to the disease. The biochemical cascade responsible for such an alteration of the ratio of platelet APP isoforms in AD is at present unknown.
However, RT-PCR experiments demonstrated that patients with AD showed the same levels of messenger RNAs encoding for the 3 major transcripts (APP 770, APP 751, APP 695) in platelets when compared with both control and NADD groups. These findings suggest that the observed reduction in the ratio of platelet APP isoforms cannot be ascribed to a marked alteration in the expression of 1 of the 3 transcripts, since we were never able to observe any difference in the levels of these transcripts, even in the most severe cases of AD, although we were using a semiquantitative procedure. Conversely, the observed decrease in platelet APP isoforms could be caused by alteration in the processing of mature APP in AD. It has been reported that platelet APP trafficking in AD is different from that in control subjects: both in basal condition and on activation, platelets in patients with AD retain more APP at the membrane surface, and an altered shuttling of the protein from beta granule to cell membranes occurs.30 In addition, Davies et al29- 31 showed a decrease in the secretion of the soluble nonamyloidogenic APPs. The authors concluded that this may account for a greater deposition of Aβ in the vasculature or, in any event, in the bloodstream. Indeed, our data showing a decrease in immunoreactivity of the 130-kd band, probably caused by increased processing, are in line with this hypothesis. Moreover, the action of calcium-dependent neutral cysteine proteases could also be excluded,18 since these enzymes exert proteolytic activity only on activation in platelet preparations, not in resting platelets, as does the preparation used in our study.32,36 Concomitant alterations of all these processes might result in the altered ratio in platelet APP isoforms observed in patients with AD.
The observation that patients with AD show a differential level of platelet APP isoforms has several implications. First, APP processing abnormalities, believed to be a very early change in AD in the neuronal compartment, do occur in extraneuronal tissues, such as platelets, thus suggesting that AD is a systemic disorder.
Second, our data suggest that APP 751 and APP770 may be oversecreted in platelets from patients with AD and that APP of circulating origin might reflect similar changes in the brain parenchyma. In fact, Aβ deposits in brains of patients with AD show the same anatomical localization of systemic amyloidosis known to be of circulating origin, and Aβ deposits are often found in specific structures that are not readily accessible to tissue-derived peptides.10 More recently,19,20 it has been inequivocably shown that Aβ of circulating origin may be carried into brain parenchyma, most likely through chaperone proteins, thus indicating the possibility of a fine interplay between Aβ from vascular compartment and from neuronal origin. This could have potentially important implications for the design of novel therapeutic approaches to prevent the progressive deposition of Aβ.
Finally, our data strongly indicate that a differential level of platelet APP isoforms can be considered a potential peripheral marker of AD, allowing for discrimination between AD and other types of dementia. Several other biochemical markers have been proposed as adjunctive diagnostic tools in AD. The large majority of these involve measurements of different proteins, such as τ, Aβ, sAPP (ie, the soluble, carboxyl-truncated form of APP), and acetylcholine esterase in cerebrospinal fluid.43- 45 Although all of these markers represent an important step forward for the in vivo diagnosis of AD, no correlation between cerebrospinal fluid levels of these proteins and the progression of the disease has been reported so far. In this respect, our data represent the first observation of a positive correlation between the modification of a peripheral biochemical marker and the progression of clinical symptoms. Furthermore, the observation that this biochemical measure changes early in the development of the disease lends support for the use of analysis of platelet APP isoforms as an early marker for AD.
Recent studies have demonstrated that the ϵ4 allele of apolipoprotein E is associated with increased risk to developing sporadic AD.46 The issue of whether the presence of the ϵ4 allele might further affect the ratio of platelet APP has been recently pointed out by Rosenberg et al.33 These authors reported, although in a limited number of subjects, that patients with AD show a lowered platelet APP ratio independently from the presence of apolipoprotein E ϵ4 allele, although the ϵ4 allele might influence this ratio.
We have demonstrated that patients with AD are characterized by a differential level of platelet APP isoforms and that this measure positively correlates with the progression of clinical symptoms. Whether this change will be suitable for an adjunctive diagnostic tool in AD in a large population of patients is under investigation.
Reprints: Monica Di Luca, PhD, Institute of Pharmacological Sciences, University of Milano, via Balzaretti, 9-20133 Milano, Italy (e-mail: firstname.lastname@example.org).
Accepted for publication February 10, 1998.
We thank Joseph D. Buxaum, PhD, for kindly providing domain-specific antibodies, and Giuliano Binetti, MD; Emiliana Corazzina, MD; and Giovanni Fabbrini, MD, for help in clinical examination of patients. We are particularly grateful to Francesca Colciaghi, PhD, for invaluable scientific help and discussion; to Mauro Cimino, PhD, and Sabrina Melino, PhD, for help in RT-PCR experiments. The authors dedicate this work to the memory of Prof Luigi Amaducci.