| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
REVIEW |
a I.R.C.C.S San Giovanni di Dio, Alzheimer's Disease Unit Sacred Heart Hospital-FBF, Brescia
b Department of Experimental Medicine and Biochemical Sciences, University of Roma Tor Vergata
c Internal Medicine Department, Geriatric Clinic, University of Pavia, Italy
d Laboratory of Adaptive Systems, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892, USA
e Institute for Cognitive and Computational Sciences, Georgetown University Medical Center, Washington D.C. 20007, USA
f Cornell University Medical College at Burke Medical Research Institute, New York 10605, USA
g Institute of Pharmacological Sciences, University of Milano
h University of Pavia, Italy
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
Key Words: peripheral cells • CSF • plasma • diagnosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
Much work has been done since then to identify the etiology and pathophysiological mechanisms that lead to the premature dysfunction and death of the neurons that characterize Alzheimer's disease. A molecular characterization of structures that constitute the neuropathological lesions typical of AD, the neuritic plaques and neurofibrillary tangles, are one approach to that question. Extracellular senile plaques in their classical form represent a compacted amyloid deposit consisting mainly of ß-amyloid (or Aß), a fragment of a larger protein defined as amyloid precursor protein (APP) (2, 3) that is expressed in all mammalian tissues except in nonnucleated red blood cells (4). Neurofibrillary tangles (NFT) represent intraneuronal bundles of paired helical filaments (PHF), which consist mainly of the microtubule-associated tau protein in an abnormally phosphorylated form and ubiquitin (5, 6). The reason why this abnormal accumulation of amyloid and PHF occurs in the AD brain is not yet clear.
Pathophysiological studies and diagnosis of the disease rely in part on the use of brain tissue from diseased patients. In particular, confirmation of AD diagnosis is only possible postmortem by neuropathological analysis, despite the lack of complete specificity of central markers such as amyloid plaques and NFT. Senile plaques and NFT are not unequivocal markers of Alzheimer's disease. In fact, amyloid plaques can be observed in the brain of normal aged donors (7). NFT can also be detected in the brain of other neurodegenerative disorders such as Pick's disease, progressive supranuclear palsy, and corticobasal degeneration (8).
Accordingly, there is an active search to identify accessible tissues or body fluids that are suitable for exploring pathophysiological hypotheses and that provide a biological marker with which to confirm diagnoses. Although this latter goal still eludes the efforts of the investigators, some relevant data have been obtained that contribute to the characterization of working pathophysiological hypotheses, such as in the case of amyloid precursor metabolism or oxidative metabolism dysfunction.
|
THE USE OF PERIPHERAL TISSUES: RATIONALE AND METHODOLOGICAL CONSTRAINTS |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
Peripheral tissues suitable for exploring pathophysiological hypotheses and possibly for providing a useful biological marker for diagnosis of AD include skin fibroblasts, platelets, lymphocytes, as well as body fluids such as plasma or cerebrospinal fluid (CSF). Among these tissues, cultured skin fibroblasts have been used successfully to elucidate the molecular and biochemical bases of many inborn errors of metabolism that cause neurological disease, i.e., Refsum's disease (9) and Lesch-Nyhan disease (10). Moreover, fibroblasts are an appropriate model for studies of those genetic diseases of the nervous system with late clinical onset, including familial Alzheimer' s disease, because they can be easily cultured and stored and contain the complete genomic information of the organism.
As always when studying the effect of a pathology on a given tissue, the choice of appropriate controls and the problem of distinguishing those changes that are due to the disease from effects related to other aspects of the patients, such as nutritional status and drug treatments, are key issues. This problem is most important for cells such as platelets and lymphocytes, which are influenced more by the patients' condition, but nutritional and drug effects are enormously diluted in cultured fibroblasts. When using cultured skin fibroblasts, some technical variables should also be considered. Growth properties and in vitro aging of AD and control fibroblast cultures do not differ under standard conditions (11–14), although this topic is still controversial. Since growth properties and biological age in culture can have profound effects on the properties of cells cultured from skin, including the expression of genes (which could be related to AD) (15), reproducible and interpretable results with the AD fibroblasts model require attention to the replicability of culture conditions including, but not limited to, matching AD and control cells for age, sex of the donors, and biological age in culture, i.e., cumulative population doubling level and percentage of life span completed (16). Different growth conditions, aging of cultures in vitro, and state of confluency of the cells at the time of the experiment may account for discrepancies between data from different laboratories. Only abnormalities replicated in larger series across different laboratories or characterizing subgroups of AD patients are likely to be of relevance (17) in diagnosing Alzheimer's disease. This diagnosis indeed appears to be the major challenge posed by AD in its sporadic late-onset form, which still represents the vast majority of all cases.
On the other hand, data obtained using fibroblasts from individuals with known gene defects, although representing only a small proportion of all AD cases, could be very informative about the cellular pathophysiology of Alzheimer's disease. Recent advances in understanding the genetics of AD allow identification of families bearing mutations in APP, presenilin 1 (PS1), or presenilin 2 (PS2) genes coded on chromosomes 21, 14 and 1, respectively (18). Fibroblast lines from familial AD patients can now be classified according to the specific gene defect to see whether a particular genetic abnormality alters cellular function in a unique manner (
Table 1).
|
Peripheral cells are therefore useful in identifying and testing hypotheses on the primary pathophysiological mechanisms leading to Alzheimer's disease and for avoiding variables derived from a postmortem state. On the other hand, peripheral cells cannot be used to answer other clinically relevant questions that require an intact organism. Low or absent expression of neuronal proteins by peripheral cells cultured under standard conditions is an important limiting factor. In the study of AD and other neurological diseases, peripheral cells are indeed an adjunct for studies of the brain and other clinically affected tissues, providing the tools to study in vitro the dynamic alteration of metabolic processes that neuropathological examination indicates might be targets of the disease. An example is given by the studies of protein kinase C (PKC) (detailed below), whose altered levels and activity in AD brain have also been consistently found in peripheral tissues. In turn, the use of fibroblasts has uncovered the correlation between altered PKC and APP metabolism. The same arguments can be applied to studies of cAMP production, which was found to be altered both in the brain and in fibroblasts of AD patients, and where the use of the peripheral tissue allowed a more detailed pharmacological dissection of the system and yielded important indications on the putative dysfunctional sites in AD.
|
PERIPHERAL CELLS |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
Fig. 1,
and Fig. 2,
step B) represent an example of the use of peripheral cells to examine biological abnormalities in AD. Decreased levels (19), activity, and translocation of PKC (20) have been demonstrated in the AD brain. Some authors have shown that levels of cytosolic PKC activity are lower in platelets from both AD and vascular dementia patients, whereas only AD patients have lower cytoskeletal PKC activity (21). On the other hand, a large percentage of patients diagnosed with vascular dementia also have AD, as suggested by autopsy series. In contrast, other authors observed that the PKC level is increased in AD platelets (22). A lower PKC activity has been demonstrated in cultured skin fibroblasts (23–25) of sporadic AD patients with respect to the cells of control subjects. An increase in the Kd for phorbol esters binding to PKC was also shown in the cytosolic fraction of AD fibroblasts, indicating a change in the characteristics of the enzyme in AD cells (25) perhaps due to a selective loss of PKC 
isoform (PKC-) (26), one of the several (27) PKC isoforms present in fibroblasts. Recent reports on fibroblasts from the Swedish kindred show that PKC is not altered in these cells (28), suggesting that the role of PKC in the pathogenesis of AD could be different in familial AD cases.
|
|
Similar to the results obtained in skin fibroblasts (25), a mild reduction of PKC activity was observed in natural killer (NK) cells from AD patients (29). However, the main differences were in PKC functional regulation instead of PKC basal activity. In fact, the observed reduction in NK-PKC activity in healthy elders after long-term exposure to stimulating cytokines (i.e., interleukin 2, IL-2; interferon ß, IFN-ß) might be ascribed to a process of down-regulation of the PKC tending to reduce cytotoxicity after previous activation. In AD patients, a lack of PKC down-regulation during IL-2 and IFN-ß exposure suggests a potentiated excitatory response of NK cells in these patients. An enhanced cytotoxic response to IL-2 and IFN-ß was observed in AD patients, whereas healthy elders had cytokine-induced NK cytotoxicity similar to that of healthy young subjects. This alteration seems to be specific for AD, since previous data (30, 31) reported normal NK activity and regulation in patients suffering from vascular dementia (clinically diagnosed).
Other transduction systems have also been investigated. A reduction in ß-adrenergic-stimulated cAMP increase (32) has been demonstrated. Diminished adenylate cyclase activity did not underlie these abnormalities, since direct stimulation of adenylate cyclase by forskolin elevated cAMP production equally in Alzheimer and control fibroblasts. Extensive studies using pharmacological probes indicated an abnormality in the coupling of G-proteins to the ß-adrenergic receptor in the cultured AD cells. Furthermore, the phosphoinositide cascade was investigated in cultured skin fibroblasts from AD and control subjects. Although basal inositol-(1,4,5)-trisphosphate (IP3) levels are similar in AD and control cells, the bradykin (BK) -stimulated increase in IP3 levels was much greater in AD fibroblasts than in controls. Elevated IP3 production in response to BK in AD fibroblasts is positively correlated with an increase in receptor number (33). The changes in bradykinin receptors and subsequent increased sensitivity to bradykinin in AD fibroblasts are unrelated to PKC (34). In fact, whereas acute activation of PKC with phorbol esters rapidly reduces the number of bradykinin receptors in normal cells, the prolonged down-regulation of PKC is not paralleled by a compensatory increase in BK receptors.
When considered together, studies of the transduction systems (summarized in 
Fig. 1) suggest that any pharmacological intervention with direct or in~direct receptor agonists aimed at ameliorating the impaired cerebral functions of AD may produce only modest results due to ineffective transduction of the agonist-promoted message, since some of these systems are impaired in AD patients.
Alterations in amyloid precursor protein metabolism
APP has the structure of an integral membrane protein and is composed of a large extracellular domain at the amino terminal, followed by a transmembrane domain and a 47 amino acid residue intracellular domain at the carboxyl terminus. The amyloidogenic sequence (Aß) begins near the end of the extracellular domain and extends approximately halfway through the transmembrane portion (35). APP follows the normal constitutive secretory pathway and is cleaved by a so-called secretase that releases the ectodomain (soluble APP, or sAPP) into the extracellular space (see Fig. 3)
(35). This proteolytic cleavage occurs within the Aß sequence, thereby preventing the formation of amyloidogenic fragments (36). On the other hand, there is evidence that Aß can be formed and secreted as a physiological product of cell metabolism (37). The routing of APP into these two pathways may be directed by protein phosphorylation. Several reports in the literature demonstrate that PKC activation with phorbol esters not only increases sAPP secretion via the nonamyloidogenic pathway (38–41), but can also decrease ß-amyloid secretion (42–44). In particular, experiments on transfected cells suggest that PKC- was involved in APP secretion (45).
|
By using skin fibroblasts from sporadic AD patients and control subjects, researchers demonstrated that the defective PKC activity (see above) is followed by a reduced basal sAPP secretion from AD cells. This indicates a constitutive deficiency in APP processing despite a lack of differences in APP expression in AD and non-AD cells in sporadic cases (46; S. Govoni, unpublished data). Furthermore, whereas high concentrations of phorbol ester (>75 nM) produced maximal, but equivalent, release of APP from both groups of cells, reduced sAPP release from AD cells was observed at low phorbol ester concentrations (<18 nM), with an EC50 in AD fibroblasts twofold higher than in control fibroblasts. The ratio between the EC50 values of AD and non-AD cells mirrors the ratio found for the Kd of labeled phorbol ester binding in the same cell lines (25, 26). Moreover, this defective regulation of amyloid precursor protein secretion might be correlated with a specific defect in PKC- (26). The data described were the first report of an alteration in APP secretion in fibroblasts derived from sporadic Alzheimer's disease patients, and led to the hypothesis that the altered APP metabolism that may underlie the pathology can be observed in peripheral cells as well as in the brain of sporadic AD patients.
This concept was even more strongly supported by data obtained using cells derived from patients with familial Alzheimer's disease (FAD). Recent findings demonstrated an alteration in APP metabolism, with a threefold increase in Aß secretion from skin fibroblasts derived from familial AD patients of the Swedish kindred (47) ( Table 1; Fig. 2, step C). Elevated ß-amyloid levels were found in conditioned media of fibroblasts both from patients with clinical AD and normal subjects presumed to have presymptomatic AD. Thus, ß-amyloid overproduction in this FAD pedigree may have a causal role in the development of the disease. Increased ß-amyloid secretion can begin many years before the onset of symptoms, even in peripheral tissues, indicating that it does not require preexisting neural abnormalities. An increased Aß secretion was also demonstrated in fibroblasts with FAD-linked PS1 and PS2 mutations (48). An abnormal processing of ß-amyloid precursor protein has also been observed in familial Alzheimer's disease lymphoblastoid cells (49, 50): APP processing via the nonamyloidogenic pathway appeared to be deficient in FAD cells, leading to increased production of a carboxyl-terminal, potentially amyloidogenic fragment.
Even if these data support the constitutive role of ß-amyloid in Alzheimer's disease, a word of caution should be sounded. No direct studies have been performed to determine whether there is any systematic biological difference between the symptomatic and presymptomatic patients apart from the presence of the signs and symptoms of illness. A comparison of data obtained using tissues derived from presymptomatic patients before the age of onset of the disease for the investigated family and after that age on both symptomatic cases and escapees is needed before making firm conclusions.
The view that fibroblasts can be used to effectively monitor pathology-induced defects of APP metabolism also emerges from experiments on Down's patients. Patients with trisomy 21 (Down's syndrome, DS) virtually always develop dementia of the Alzheimer's type after the fourth decade of life and show a histopathological phenotype indistinguishable from AD, presumably on the basis of increased APP gene dosage and transcription. Querfurth et al. (46) demonstrated an increased APP mRNA level in DS fibroblasts similar to that observed in FAD. They also demonstrated an increase in Aß secretion in fibroblasts from DS and FAD-affected subjects as compared to control cells. However, caution must be observed when using information obtained from Down's patients to model biochemical defects in Alzheimer's disease. In fact, data obtained from skin fibroblasts of DS subjects (51) show a completely different pattern of sAPP secretion with respect to sporadic AD cells under basal conditions and after PKC activation with phorbol esters. The sAPP basal release from DS fibroblasts is twofold that of age-matched control cells. Moreover, stimulated APP release is already maximal at low phorbol ester concentrations (9 nM), and the peak response is lower on a percentage basis than in control fibroblasts because of higher basal release. Thus, in DS fibroblasts, PKC-sensitive secretion can be stimulated only modestly, indicating a saturation of this pathway due to the higher APP content of DS cells when compared to control and AD fibroblasts. These results also indicate that the PKC-sensitive mechanisms that control APP release might be saturated and might depend on the level of APP expression. These observation stress the importance of evaluating APP expression after pharmacological treatments with a potential effect on APP secretion, since changes in this parameter may reset the system and affect the pool of the precursor available for secretion under basal or stimulated conditions.
Altered calcium homeostasis in AD peripheral cells
Several findings (for review, see ref 52) support the hypothesis that altered processing of the ß-amyloid precursor protein contributes to loss of neuronal calcium homeostasis and thus to neurofibrillary degeneration. These data suggest that a shift in APP processing, in favor of increased liberation of ß-amyloid and reduced release of sAPP, should destabilize intracellular calcium concentration ([Ca2+]i) and endanger neurons in two ways: 1) increased levels of ß-amyloid would make neurons more vulnerable to excitotoxicity (53); on the other hand, 2) reduced levels of sAPP would deprive neurons of a neuroprotective substance that can stabilize [Ca2+]i (54–57).
Alterations in calcium homeostasis have also been found in peripheral cells ( Fig. 2, step A). The first study utilizing fibroblasts from AD patients revealed that calcium uptake declines (58). Total membrane-bound calcium increases in fibroblasts in normal aging and is elevated even further in those with Alzheimer's disease (59). Other reports suggest that cytosolic free calcium decreases; however, this has not been replicated (60, 61). Cytosolic free calcium concentration is increased in platelets (62) and lymphocytes derived from AD patients (63), but not in lymphoblasts (64). In addition, freshly prepared human lymphocytes showed elevated mitogen-induced calcium responses after exposure to ß-amyloid (65–67). A diminished response of cytosolic free calcium to drugs such as bradykinin has been found in AD fibroblasts by some (61, 68, 69) but not all groups (70–73). The discrepancy may be due to the concentration of bradykinin used; Huang et al. (70) and Ito et al. (73) have observed an exaggerated response to low (nanomolar concentrations) BK concentrations by using AD fibroblasts.
Altered internal calcium stores have been demonstrated by different treatments in AD fibroblasts. Bombesin elevates cytosolic free calcium more in cells from AD patients than in controls (73). The calcium pool released by A23187 after treatment with high bradykinin concentrations is higher in AD than in control fibroblasts (74). Another study that demonstrated altered internal calcium pools in Alzheimer fibroblasts revealed that mitochondria from AD fibroblasts had decreased calcium uptake (75).
A greatly diminished response of cytosolic free calcium to high concentrations of the potassium channel blocker tetraethylammomium (TEA) also suggests that regulation of calcium entry is abnormal in AD fibroblasts. Patch clamp studies indicate that this abnormal calcium response reflects the functional absence of a particular K+ channel (76).
The apparent lack of replicability with fibroblasts has been particularly striking for the study of calcium homeostasis. Many laboratories are equipped to measure calcium, but methodological variations are great. In addition, controlling all the variables of tissue culture (presence or absence of antibiotics, synchrony of the cell cycle, cumulative population doubling level at the time of the experiments, interval between seeding and when the measurement is made, seeding density) is extremely difficult. In fact, the reduced resting calcium and the diminished response to bradykinin previously observed in AD fibroblasts (60, 68) have been difficult to replicate (70, 73). The inability to repeat these findings may have been partially related to the cell phase in which the various studies were performed. Recent results replicated the decreased response to bradykinin, vasopressin, or serum in familial AD cells if the cells were studied in S phase (69). No AD/control difference occurs after bradykinin stimulation when the cells are in G2 or M phase. Similarly, if cells are studied 7 days after plating, 40–60% decreases are apparent in the AD fibroblasts (69). Another measure of calcium homeostasis where the method of measurement accounts for the apparent discrepancies is where the potassium channel blocker TEA elevates cytosolic free calcium in controls, but not AD fibroblasts. In the original report, addition of TEA distinguished AD cell lines from controls (76). However, a recent report did not find this measure to discriminate between AD and controls (77). Methodological differences may account for the discrepancies. The original study used microscopic imaging to examine individual cells and calculate the percent of responding cells (76). However, experiments from Matsuyama et al. (77) averaged microscopic fields that contained many cells, which severely limits resolution and sensitivity.
Experiments on the bradykinin-induced IP3 response provide an example in which treatment of the cells accounts for the lack of replicability between studies. The initial study found no differences in IP3 formation in control and AD cells when IP3 release was measured after 3H-inositide in serum free medium for 4 days (70). Subsequent studies in the same laboratory revealed enhanced IP3 formation in AD cells compared to controls in cells that had not been serum deprived (33). Subsequent experiments showed that IP3 formation was enhanced in AD cells regardless of the method of measurement, as long as the treatment did not include prolonged serum deprivation (33). These discrepancies (also reviewed in ref 78) should not be viewed as a reason to dismiss this approach. Research on fibroblasts has progressed to a stage that elucidating the underlying mechanistic basis of any of these discrepancies is possible if adequate attention is paid to the details of experimental techniques used.
Alterations in oxidative metabolism in AD peripheral cells
Cellular calcium and oxidative metabolism are closely linked in fibroblasts, as in other tissues, and various parameters of altered energy metabolism have been reported in peripheral cells (particularly in fibroblasts) from Alzheimer patients.
Increased lactate production and altered glucose utilization occur (59, 79, 80) in AD fibroblasts. The activities of key mitchondrial enzymes such as -ketoglutarate dehydrogenase are diminished either slightly (81) or by as much as 40% (82) in Alzheimer fibroblasts. This last result has been confirmed in an independent series of experiments showing inherent defect of the -ketoglutarate dehydrogenase complex (83). Glutamine oxidation also declines (59, 79, 80) in AD cells, although not always (84). Furthermore, reduced cytochrome c oxidase (CO) activity was demonstrated in platelets derived from AD patients (85) and also in AD fibroblasts from our bank (Govoni et al., unpublished results). Recently published data (86) describe genetic defects in the mitochondrial DNA coding for CO subunits CO1 and CO2. However, these results await independent confirmation in a different series of patients. These results are all consistent with the hypothesis that abnormal oxidation in AD brain is not just secondary to neurodegeneration, but the hypothesis needs to be investigated further. Finally, abnormalities in cerebral metabolism of oxygen and glucose can be expected to interact synergistically with cerebrovascular factors in causing brain damage in AD.
Use of peripheral cells as an aid in diagnosing disease
Peripheral cells—and in particular cultured skin fibroblasts from AD patients—might provide a useful aid in AD diagnosis. Today, the criteria needed for a conclusive diagnosis of AD include clinical criteria in association with histopathologic evidence obtained from a biopsy or autopsy. Thus, the use of peripheral cells as a diagnostic tool, either to predict or at least to confirm the diagnosis, may be of great importance. None of the alterations described in peripheral cells can now be used as a diagnostic tool if considered alone. One possible approach to achieve this goal is to combine more than one set of data and thereby provide a profile for AD cells distinguishable from that of controls.
The first attempt in this direction was made by Hirashima et al. (87). Previous findings by the same authors demonstrated the absence of a specific potassium channel and of TEA-induced calcium increase in AD fibroblasts (76). The stimulation of calcium release from intracellular stores in response to bombesin or low BK concentrations was greatly enhanced in AD fibroblasts compared with cells from control donors (73). On the basis of these results, Hirashima (87) introduced a scoring system that integrated altered and normal responses to low BK concentrations, bombesin, and TEA stimulation of intracellular calcium elevation in AD vs. control fibroblasts. This scoring system also takes into account the degree of responsiveness and/or unresponsiveness of each particular cell line. It generated an index distinguishing AD patients from controls with both high specificity and sensitivity (87). The authors have also examined many non-Alzheimer's dementia controls and still found complete specificity. These results indicate that by taking the overall profile of responses into account, it is possible to enhance the diagnostic value of these cellular alterations. Furthermore, the diagnostic measurements obtained in fibroblasts appear to be closely correlated for two populations of patients living in distant geographic areas, since cells used for the study described were obtained from Coriell Cell Repository (Camden, N.J.) and a fibroblast bank established in Brescia, Italy (Project Ministry of Health #500.7 AG/4.8455/Regione Lombardia). Moreover, specificity and sensitivity of the assay should be established by including patients bearing different kinds of neurological diseases in order to determine the ability of the marker to distinguish AD from normal aging and from other dementias. Severity and length of the illness are other important factors to consider to determine whether the parameters measured also allow identification of stages of the disease.
Studies of platelets also suggest that it may be possible to detect alterations of APP synthesis/metabolism of potential diagnostic significance in sporadic AD patients. In fact, the presence of APP in platelet granules has generated great interest in the study of platelet function and biochemistry in AD patients. Several laboratories have shown that human platelets contain full-length APP (88–93) and are suitable for the study of APP processing. It has also been proposed that platelets may serve as a possible source of the Aß sequence (93), providing one mechanism by which cerebral amyloid might be derived from the circulation and contribute to cerebral or cerebrovascular amyloidosis. According to Van Nostrand et al. (88–90), the protease inhibitor protease nexin-2, the secreted form of APP that contains the Kunitz protease inhibitor domain, may play a role in regulating blood coagulation by inhibiting blood coagulation factor IXa, and platelets may serve as a vehicle to deliver this protein to sites of vascular injury. A recent study (94) has demonstrated the existence of a consistent alteration in platelet APP isoforms in AD and Down's syndrome (i.e., a lower 130–106/110 kDa ratio) that is of potential value as a biochemical marker of the disease.
|
CSF AND PLASMA. APP, Aß, TAU PROTEIN, AND OTHER MARKER PROTEINS |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
-sAPP and poor performance on neuropsychological tests in individuals from a Swedish family with AD and a double mutation at codons 670/671 of APP gene. Seubert et al. (102) first isolated and quantified Aß from human biological fluids (i.e., CSF and plasma). Alterations of CSF Aß levels were reported in AD patients. Nakamura et al. (103) found a significant elevation in the level of Aß in the CSF of early-onset AD patients. On the other hand, Nitsch et al. (104) found that Aß levels in the CSF of AD patients were inversely correlated both to cognitive and functional measures of dementia severity. Lannfelt et al. (105) demonstrated a correlation between decreased soluble Aß levels and duration of illness, but found no differences in CSF levels of Aß between APP670/671 mutation carriers and noncarriers from the Swedish family they studied. Comparable amounts of Aß in AD and control CSF were found by other laboratories (106–108), indicating that the total CSF Aß level is not a useful marker for current diagnosis of AD.
Aß1-42 CSF levels seem to be a more specific marker of AD pathology than total Aß. In fact, CSF Aß1-42 levels were significantly lower in AD patients relative to a non-AD group (109–111). However, some overlapping of the control Aß1-42 levels with AD values (109) was present. Aß1-42 has recently been shown to deposit in the brain tissue of patients with AD (112, 113), suggesting that diminished clearance may account for its reduction in AD CSF. The peptide may be elevated early in the disease and decrease during progression (Lannfelt, working group on biological markers of AD, Alzheimer's Association).
Aß concentration also seems to be altered in plasma from AD patients. Song et al. (114) observed that plasma Aß1-42 is increased in carriers of familial AD linked to chromosome 14. Plasma Aß is also increased in carriers of the 670-671 mutation from the Swedish kindred (115). No differences in Aß levels were seen between symptomatic and presymptomatic carriers, which shows that the increase of Aß levels might be a very early event in the disease, at least in the Swedish kindred. On the other hand, it cannot be excluded that the increase in Aß levels is unrelated to the pathogenesis of the disease, being indistinctly elevated both in patients and in asymptomatic individuals. Follow-up studies investigating whether the elevated Aß levels in subjects with no symptoms are predictive of a future development of the illness will help to clarify this point. This observation has been interpreted as indicating a potential importance of plasma Aß as an early marker for identifying individuals who will develop AD because of elevated Aß concentrations, at least in cases of FAD. Moreover Younkin et al. (116) observed that an elevated plasma concentration of Aß1-42 may also play a role in some patients with sporadic AD cases and that, in at least some of these patients, Aß1-42 may be elevated before the onset of symptoms. This observation has not been replicated by other authors (see below); additional results are clearly needed before drawing any conclusions about the risk for AD that is conferred by elevation of plasma Aß1-42. Scheuner et al. (48) demontrated an increased Aß1-42 level in plasma from subjects with FAD-linked PS1, PS2, APPK670N, M671L, and APPV717I mutations. Plasma Aß1-42 level is also elevated in presymptomatic carriers of PS1 and PS2 mutations. Kosaka et al. (117) found an increased percentage of plasma amyloid ending at Aß42(43) in patients with the ßAPP717 mutation. An increased Aß1-40 plasma level was also observed in carriers of the APPK670N, M671L mutation. In contrast, Aß1-42 was not statisically increased in subjects with late-onset sporadic AD compared with age-matched control subjects. This observation and the finding that Aß1-42 was significantly increased in all presymptomatic gene carriers indicate that the increased Aß1-42 observed in subjects with FAD-linked APP and PS1 or PS2 mutations occurs as a direct consequence of the mutations and not as an indirect manifestation of the AD state. The diagnostic significance of levels of Aß1-42 in these cases obviously is secondary to the finding of the pathogenetic mutation.
Tau protein, the main component of neurofibrillary tangles, can also be detected in CSF. Tau levels in the CSF of AD cases are significantly elevated compared to healthy control individuals, vascular dementia patients, and neurological control subjects (118–127). It is debated whether tau levels are also elevated in other kinds of neurological diseases where neuronal death or damage occurs. In particular, increased CSF levels of tau are found in a definite proportion (10–20%) of neurological controls (119, 122, 123). Furthermore, there is an overlap between AD and other forms of dementia, reducing the predictive value of elevated CSF tau. The simultaneous analysis of tau and Aß1-42 in the same CSF sample appears to be of particular interest. In fact, the presence of elevated tau levels and reduced Aß1-42 seems to be highly predictive of AD (specificity 96%). Conversely, high Aß1-42 and low tau were observed only in control patients (specificity 100%). However, the combined CSF tau and Aß1-42 measurements were not informative in those patients who fell into the low Aß1-42/low tau group. The diagnostic potential of these results needs to be evaluated further (109, 110).
Among other proteins that are unrelated to the major Alzheimer's markers, such as Aß or tau, the iron binding protein p97 has recently been proposed as a possible diagnostic marker for the disease (128). Its level has been found to be significantly increased in freshly prepared plasma samples from AD patients. The data are promising because of the virtual absence of overlaps with the levels measured in the plasma of control subjects, including individuals suffering from other neuropathologies. As for many other proposed diagnostic markers for AD, the plasma level of p97 awaits a complete validation on a large number of patients in carefully designed multicenter studies.
|
NEUROENDOCRINE SYSTEM |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
Studies (136) in healthy elderly humans suggest that with aging, the human HPA system becomes more disinhibited, showing an age-related increase in HPA system activity. Thus, one would expect that, in patients with AD, the age-associated HPA system changes are aggravated by the disease process itself. However, studies of HPA system function in AD patients have so far yielded controversial findings. Hatzinger et al. (137) demonstrated profoundly altered regulation of the HPA system in patients with AD. In fact, they demonstrated that patients with AD had significantly higher basal cortisol concentrations, and a higher percentage of AD patients escaped the dexamethasone (DEX) suppression of cortisol release when compared to elderly controls. On the other hand, the DEX suppression test is typically abnormal in depression, so the observed changes cannot be considered specific for dementia. The alteration of the HPA axis in AD patients is further supported by a significantly lower adrenocorticotropin hormone (ACTH) and cortisol release after administration of additional corticotropin-releasing hormone (CRH), as compared with the control group. In addition, it was demonstrated that CRH and ACTH levels are significantly reduced in the CSF of AD patients (138–140) in the absence of differences in plasma levels of CRH, ACTH, and cortisol (141). However, none of the patients showed evidence of pituitary-adrenal dysfunction, suggesting an involvement of extrahypothalamic CRH (138). Several studies indicate a reduction in CRH-like immunoreactivity in the cerebral cortex of AD patients (142–144), particularly in temporal, frontal, and occipital areas. In the same cortical areas, an increased CRH receptor number was reported (142). Nevertheless, these findings are not specific to Alzheimer's dementia. In fact, reduced levels of CRH in CSF were also demonstrated in patients with vascular dementia (139–141), and reduced CRH-like immunoreactivity at the level of the cerebral cortex level was also found in Parkinson's disease and progressive supranuclear palsy (143, 145). Other alterations in the neuroendocrine system of AD patients have also been reported. An attenuated growth hormone-releasing hormone (GHRH) -induced growth hormone response specific to AD was demonstrated by Lesh et al. (146) and Ghiso et al. (147), but not by others (148, 149). Furthermore, a reduction in CSF levels of antidiuretic hormone was observed not only in AD patients, but also in patients with frontal lobe dementia (150). No alteration of the thyrotropin-releasing hormone/thyrotropin hormone (TRH/TSH) and prolactin (PRL) systems was reported in AD patients (151, 152).
A reduced somatostatin-like immunoreactivity (SLI) in some regions of the AD brain was reported by several groups (153–156). However, a reduction is also detectable in the brain of patients with Parkinson's disease (154, 155) or major depression (155). Also, one report (157) shows an increased SLI in both AD and vascular dementia. Several papers (150, 158–165) even demonstrated reduced SLI in the CSF of AD patients at early stages of the disease, whereas just one group found no differences between AD and control subjects (165). However, low SLI values in CSF are not specific to AD; in fact, reduced SLI was demonstrated in the CSF of patients affected not only by neurological disorders such as frontal lobe dementia (150), Parkinson disease (166), vascular dementia, normal pressure hydrocephalus (161, 167), senile dementia of the Lewy body type (168), dementia with frontotemporal degeneration (154), and dementia associated with alcoholism (164), but also in psychiatric diseases like major depression (158) and schizophrenia (156). Therefore, the measurement of SLI in CSF in demented patients may not be helpful in the diagnostic procedure. These observations do not support the existence of a specific neuroendocrine pattern for AD patients. However, the new findings on the relationship between AD and inflammation and estrogens and AD risk in postmenopausal women strongly suggest the need to explore further the potential neuroendocrine alterations in this disease (169–173).
|
APOLIPOPROTEIN E AND DEMENTIA |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
The ApoE gene is localized on chromosome 19 and presents three alleles—
2, 3, and 4—that determine ApoE polymorphism. Analysis of ApoE alleles in AD and control subjects demonstrated a highly significant association of ApoE type 4 allele (ApoE 4) and late-onset familial Alzheimer's disease (180). The allele frequency of ApoE 4 in 30 random affected patients, each from a different AD family, was 0.50 ± 0.06 vs. a frequency of 0.16 ± 0.03 in 91 age-matched nonrelated controls. The analysis was then extended to sporadic late-onset AD patients by Saunders et al. (181): the same kind of association with allele 4 was demonstrated in living and autopsy-documented AD patients vs. controls. The increased frequency of the 4 allele in late-onset sporadic and familial AD was demonstrated in U.S. (180–182) as well as Japanese, Finnish, and Italian populations (183–187). Furthermore, the frequency of the 4 allele also seems to be increased in early onset sporadic AD patients in a Japanese and a Dutch population (188, 189).
The inheritance of one or two ApoE 4 alleles seems to influence Alzheimer pathology in a gene dose-dependent manner: the amount of histologically identified Aß in vessels and plaques in the cerebral cortex of patients with sporadic late-onset AD is a direct function of their ApoE genotype. Increased amyloid deposits in vessels and increased density of strongly Aß immunoreactive plaques in patients homozygous for ApoE 4 compared with patients homozygous for ApoE 3 have been demonstrated, although the topic is controversial (185). ApoE 3/4 patients have an intermediate phenotype (190). A gene dose effect can also be detected in the age of onset of AD. Inheritance of the ApoE 4 allele is associated with an earlier onset of the disease: the median age at onset among AD patients decreased from 83 to 78 to 74 years as the number of ApoE 4 alleles increased from 0 to 1 to 2, respectively (191). These findings led to the hypothesis of ApoE 4 as a risk factor for Alzheimer's disease and suggested a potential use of ApoE genotype as a diagnostic test, although the issue is still controversial (192–194).
Recently it has been demonstrated that the frequency of ApoE 4 allele is increased not only in AD, but also in other types of dementia. Frisoni et al. (186) confirmed increased ApoE 4 allele frequency in 93 sporadic AD patients, but also demonstrated a similar increase in 4 frequency in 23 vascular dementia patients. Moreover, it has been demonstrated that 4 frequency is also increased in patients affected by Creutzfeld-Jacob disease (195) and by Lewy body dementia (196), but not in Down's demented subjects (197, 198).
Therefore, the association of the ApoE 4 allele is not unique to Alzheimer's disease. The use of ApoE genotype as a diagnostic test in symptom-free individuals has insufficient epidemiological support at present to be useful. In fact, it cannot predict when, or if, an individual with ApoE 4/4 will get Alzheimer's disease. On the other hand, the genotype may be used to confirm that patients with early dementia have Alzheimer's disease. ApoE genotyping may also be used to subdivide symptomatic patients by biological risk in order to foresee the response to drug treatment (199) and the clinical progression of the pathology (200), since the proportion of patients with slow progression seems to increase with increasing 4 gene dose, suggesting that disease duration might be longer in 4 carriers (201).
In addition to genotyping, several groups have measured ApoE levels in CSF alone or in conjunction with tau and Aß1-42. Various authors found reduced CSF ApoE concentrations in AD patients, irrespective of the ApoE genotype (202–204). However, contrasting results have been published; one recent report found no difference (205) in CSF ApoE concentrations, and another found an increase (206). The decrease of CSF ApoE, on the other hand, does not appear to be limited to AD and, rather, may represent an unspecific marker for neurodegenerative disorders (204). It is tempting to speculate that since lipoproteins participate in lipid transport during cell repair, the relative lack of ApoE may impair repair mechanisms, making the brain more susceptible to damage. In studies that measured CSF levels of Aß1-42 (reduced in AD) or tau (which increases in AD), the two parameters were found to be independent (102, 107) of ApoE genotype. However, in a recent study (207) of a limited number of AD patients followed for 14 months, the ApoE 4 carriers presented increasing tau CSF levels. While this observation awaits confirmation, it may suggest that there are ApoE isoform-specific differences in the regulation of the homeostasis of tau.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
Fig. 2being an example. According to data reviewed here, a defect in transduction systems might play a key role in the pathogenesis of Alzheimer's disease. For example, the reduced PKC activity observed in AD fibroblasts might influence APP metabolism and reduce APP processing via the nonamyloidogenic pathway with an increased formation of Aß. Many events could follow enhanced Aß production (Aß deposition, facilitation of neurotoxicity, disruption of calcium homeostasis, abnormal tau phosphorylation), resulting in the neurodegenerative pattern and loss of function typical of AD patients. However, this hypothesis needs to be investigated further, and peripheral tissues could be an important tool in the clarification of several other points. The search for a biological marker that could help in predicting or confirming AD diagnosis is still a rich area for research. The attempt to build a biological profile that is characteristic of AD cells, with the association of different variables, seems a more promising approach (87, 110), at least for sporadic cases. Furthermore, the initial data indicate a need to design studies that are carefully targetted at detecting the diagnostic potential of observed biological differences. A more clinically oriented experimental design, with double-blind and multicenter tests with preestablished inclusion criteria, should be adopted.
|
FOOTNOTES |
|---|
2 Abbreviations: Aß, ß-amyloid; ACTH, adrenocorticotropin hormone; AD, Alzheimer's disease; ApoE, apolipoprotein E; APP, amyloid precursor protein; BK, bradykinin; CO, cytochrome c oxidase; CRH, corticotropin-releasing hormone; CSF, cerebrospinal fluid; DEX, dexamethasone; DS, Down's syndrome; FAD, familial Alzheimer's disease; GHRH, growth hormone-releasing hormone; HPA, hypothalamic-pituitary-adrenal; IL-2, interleukin 2; IFN-ß, interferon ß; IP3, inositol-(1,4,5)-trisphosphate; NFT, neurofibrillary tangles; NK, natural killer; PHF, paired helical filaments; PKC, protein kinase C; PKC-, protein kinase C isoform; PRL, prolactin; PS1, presenilin 1; PS2, presenilin 2; sAPP, soluble APP; SLI, somatostatin-like immunoreactivity; TEA, tetraethylammonium; TRH/TSH, thyrotropin releasing hormone/thyrotropin hormone.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONTHE USE OF PERIPHERAL...PERIPHERAL CELLSCSF AND PLASMA. APP,...NEUROENDOCRINE SYSTEMAPOLIPOPROTEIN E AND DEMENTIACONCLUSIONREFERENCES |
|---|
secretases and then released into the extracellular milieu. The routing of APP into these two pathways is under the control of protein phosphorylation, in particular by PKC, a mechanism that is defective in sporadic AD fibroblasts (see text). The mechanism is presently unknown, but presenilins are also presumed to play a role in the regulation of these pathways. Finally, the endosomal/lysosomal pathway (7) is called into action for degradation of the carboxyl-terminal fragment left over by secretase processing (5) and for the degradation of full-length APP molecules recycling from the cell surface or directly from the biosynthetic pathway through the Golgi cisternae.