M. Racchi, L. Gasparini, M. Trabucchi and S. Govoni IRCCS "Centro San Giovanni di Dio - Fatebenefratelli, Brescia and Institute of Pharmacology, University of Pavia, ITALY

Beta-amyloid, the key constituent of senile plaques in AD, derives from a large precursor protein, defined Amyloid Precursor Protein (APP), existing in several isoforms. The precursor protein can be processed by various protease activities leading to a large secreted non amyloidogenic fragment (sAPP) or alternatively releasing the amyloidogenic 39-42 aa peptide beta-amyloid. Both amyloidogenic and non-amiloidogenic metabolism appear to occurr normally as regulated processes

The significance of characterizing pharmacological modulation of APP metabolism relies on the fact that, according to one of the current hypotheses on AD pathogenesis, aberrant regulation of APP metabolism may be a causative factor in the overproduction of amyloidogenic fragments leading to the progressive deposition of extracellular amyloid and the development of Alzheimer's disease (AD) pathology. An action at this level might result in slowing the process of the disease and may be additive to other properties of any drug considered for the treatment of this devastating disease.

Cultured cells offer a convenient model to study the physiological control of the amyloid precursor protein as well as its pharmacological modulation. Fibroblasts, in particular, represent a testable model in which to study the effect of drugs potentially affecting APP synthesis and release (Govoni et al.Life Sci. 59:461-468, 1996). Moreover the fibroblasts derived from affected patients allow the testing of the drug directly on a tissue epressing pathology-related defects in the metabolism of the precursor. Accordingly, we studied the pharmacological modulation of sAPP release by testing the effect of various drugs and conditions potentially affecting sAPP metabolism using fibroblasts from AD as well as from control patients as detailed in methods.

Patients.
All the patients, except Down's syndrome patients (see Govoni et al., Neurology 47:1069-1075, 1996) were evaluated in the Alzheimer's disease Unit of "Sacro Cuore-Fatebenefratelli" Hospital, following approval by the local ethical committee. All patients met NINCDS-ADRDA criteria for probable AD. They did not have familiar history of dementia, and were classified as "sporadic AD" cases. The characteristics of all the patients used in the various sections of the study are summarized in table 1. In each set of experiments controls and AD were matched for age and sex.

Cell collection , colture and biochemistry.
The methods for fibroblast collection and culture, PKC activity and immunoreactivity measurement and sAPP detection in the conditioned medium have been published (see Govoni et al. Neurology 43:2581-2586,1993; Bergamaschi et al., Neurosci Letters 201:1-4,1995). In Brief, as far as sAPP detection and quantization is concerned, samples of conditioned media standardized to lysate protein concentration were run on 7,5% SDS-PAGE and then electrophoretically transferred onto Polyscreen PVDF membrane. For the detection of secreted APP, the monoclonal antibody 1G5 (kindly supplied by Athena Neurosciences) was used. Detection was carried out by incubation with alkaline phosphatase conjugated goat anti mouse IgG. The blots were then washed and sAPP visualized using a chemiluminescence method. sAPP is secreted into the medium of fibroblasts as a protein with an apparent molecular weight of 105 kDa. For Northern blot analysis total RNA was extracted, RNA samples (15 micrograms) were separated in 1% agarose-formaldehyde gel and then transferred overnight on a nylon membrane by capillarity. Nylon membrane were then hybridated using a fluorescein labeled probe for APP (273-2077 human cDNA fragment). RNA filter was washed and then incubated with alkaline phosphatase conjugated antifluorescein antibody. RNA membrane was then washed and the bands were detected by incubation with a chemiluminescence substrate, followed by exposure of an X-ray film.

Densitometric and statistical analysis.
The relative density of immunoreactive bands on Western and Northern blots was calculated from the area of the peak corresponding to the selected band following acquisition of the blot image through a Nikon CCD video camera module and analysis by means of the Image 1,47 program (Wayne Rasband, NIH, Research Service Branch, NIMH, Bethesda, MD). Normalization between different western blot was achieved by dividing each individual peak area by the average peak area of control samples within the respective blot. Normalization of Northern blot were achieved by dividing each individual peak area by the peak area of beta-tubulin (bTub., fig. 7) band of the same sample. Statistical analysis of the data was performed using a two tailed ANOVA followed when significant by a standard post hoc test (Dunnet t test).

The data here reported include the characteristics of all the fibroblasts lines present in our repository. In each set of experiments we selected from the cell repository a group of controls, including non-AD (patients affected by neurological diseases other than AD) or of healthy controls, carefully matched for age and sex to AD cases.

Most of the fibroblast cell lines used for the experiment were screened for ApoE genotype. While the general pattern of ApoE allele distribution (reported as % below the graph) reflects the several published studies on ApoE4 frequency in AD patients, including ours (Frisoni et al Dementia 5, 240-242, 1994; Frisoni et al. Ann.Neurol., 37, 110-118, 1995), so far no correlation was found among the measured parameters reported in the present poster and ApoE genotype. Hovewer, it should be recognized that in each separate set of experiments the number of cases used may be insufficient to study a correlation between the measured parameter (PKC, sAPP secretion basal or stimulated) and ApoE distribution.

*P<0.05 compared to non AD. The values were derived from Govoni et al, Neurology 43, 2581-2586, 1993 and Bergamaschi et al., Neurosci Letters 201:1-4,1995.

The figure shows a representative experiment. Similar results were obtained using fibroblasts from at least 5 different healthy controls.

Figure 3
Decreased responsiveness of sAPP secretion to PdBu stimulation in fibroblasts from AD patients

Comment to Table 2, Figures 2,3
The data reported in table 2 indicate that the basal secretion of sAPP from AD cells (11 cases compared to 16 controls) is reduced, suggesting a constitutive deficiency in APP processing in spite of an absence of differences in APP expression (not shown). Furthermore, we observed that in normal human fibroblasts the secretion of sAPP can be stimulated by PKC activation with nanomolar concentrations of phorbol esters (Figure 2) and that this response is reduced in AD (Figure 3). High concentrations of PdBu (75-150 nM) produced maximal, and equivalent, release of APP from both groups of cells. Furthermore, we observed that this defective regulation of amyloid precursor protein secretion might be correlated with the specific defect in PKCa. Our data are the first report of an alteration in APP secretion in fibroblasts derived from sporadic AD patients, and extend to sporadic AD the concept that altered APP metabolism may underlie the pathology and can be observed in peripheral cells as well as in the brain.

Figure 5.
Effect of Sodium Azide on sAPP secretion from control and AD fibroblasts

Comment to Figures 4,5
Fibroblasts possess bradykinin B2 receptors, the stimulation of which induced a dose-dependent increase of sAPP secretion (up to 3 fold the basal) with an EC50 derived from a concentration-response curve (not shown) in the low nanomolar range (2.8 nM). Notably the effect of BK was independent from PKC, since it was not blocked by staurosporine, and was identical in AD and in controls (Figure 4, the figure shows a representative experiments, the averaged data from 6 controls and 5 AD cases did not reveal any statistically significant difference). In contrast, fibroblasts from AD donors were greatly more sensitive to energy and oxidative metabolism inhibition. In fact, the addition to the incubation medium of sodium azide, a substance interfering with cytochrome C oxidase and already shown to inhibit sAPP release from COS cells (Gasparini et al.Neuroscience Letters, in press 1997), significantly inhibited sAPP secretion from AD fibroblasts at concentrations (in the millimolar range) barely affecting sAPP release from control donors (Figure 5, the figure shows a representative experiment, the averaged data from 8 AD cases gave mean inhibition of 51%, P<0.01 compared with controls). The data support the concept that interference or malfunctioning of oxidative metabolism in AD tissues, if occurring at brain level, may contribute to aberrant APP metabolism and eventually to beta-amyloid deposition.

Figure 6.
Effect of 17-b-estradiol and cholesterol on sAPP release from COS cells

Figure 7
Effect of estradiol treatment on APP mRNA in COS cells

Comment to Figures 6,7

17beta-estradiol
Recent epidemiological data suggested that estrogen levels decay in postmenopausal women contributes to exacerbate molecular events taking place with aging thus leading to the development of AD. Within this context it has been demonstrated that in a human breast carcinoma cell line (which express high levels of the estrogen receptor), chronic 17beta-estradiol treatment increases the secretion of the soluble fragment of APP without modifying its cellular content. In light of these data we evaluated the effect of 17beta-estradiol (9 nM, 48 hrs) treatment in fibroblasts (not shown) and in COS cells (Figure 6 panel A, 17beta-estradiol for 2h at 37¡C) representing a cellular model independent from aberrant genetic profiles expressed by tumoral cells or genetically engeneered cells transfected with the estrogen receptor. However we did not observe changes in sAPP secretion (fibroblasts and COS cells) or mRNA (in COS cells, Figure 7). On the other hand the time chosen may have been too short to observe an effect in cells not overexpressing the receptor, suggesting the need to extend the time frame of estrogen exposure.

Cholesterol
Other steroids may also affect sAPP secretion by altering membrane characteristics, in particular we thouroughly investigated the effect of membrane cholesterol loading on sAPP release. Using COS cell, we observed (Racchi et al Biochem. J 322:893-898, 1997) that following membrane cholesterol loading, either by using betaVLDL from hyperlipidemic rabbit or cyclodextrin dissolved cholesterol (Figure 6, panel B, 30 mg/ml) for two hrs, there was an up to 70% inhibition of sAPP basal release. Although it is still not known whether fibroblasts are also sensitive to cholesterol membrane loading and whether cells from AD donors present a differential sensitivity, these data open the possibility that alterations in cholesterol metabolism, including a different ApoE genotype, may influence APP metabolism in nervous tissue. Since lipid and cholesterol transport may have great importance in neuronal repair the data also suggest the need to investigate the concerted role of sAPP and cholesterol metabolism in these events.

STUDY OF SAPP RELEASE FROM CELLS SPONTANEOUSLY OVEREXPRESSING THE PROTEIN

Fig. 8
PdBu (150 nM, 2hrs) stimulated sAPP secretion from control and Down's fibroblasts

Comment to Figures 8
The study of sAPP release from cells spontaneously overexpressing the protein, i.e. fibroblasts from patients with Down's syndrome, DS, showed a different pattern of sAPP secretion respect to AD cells. The sAPP basal release was twofold that in age-matched control cells (Govoni et al Neurology 47:1069-1075, 1996), while the pharmacological response to phorbol esters was blunted (Fig. 8) in DS fibroblasts, indicating a saturation of this pathway due to the higher APP content and basal sAPP release. An increased sensitivity to PdBu stimulus was instead observed in a case of Chromosome 21 deletion, leading to a reduced APP expression (Racchi et al. Soc. Neurosci. Abstract 730.14, 1997). These observation, as well as data we obtained in primary neuronal cultures (Salvietti et al, Neurosci Lett 212:199-203,1996) suggest that the pharmacological modulation of sAPP secretion may differ in cells with an altered APP expression.

Conclusions

The studied responses underscore the possibility to affect APP metabolism through pharmacological modulation using human tissue directly derived from patients affected by Alzheimer's disease. The data on fibroblasts from AD patients indeed suggest that APP metabolism is affected in sporadic AD. The disease appears to compromise the basal secretion of the non-amyloidogenic derivative sAPP and the response to some, but not all, stimuli favoring non-amyloidogenic cleavage. Moreover AD fibroblasts seem to be more susceptible to stimuli compromising sAPP secretion such as oxidative stress. The reported data suggest also caution in the interpretation and extrapolation of the in vitro results, since the observed effects may strongly depend on several variables including presence of pathology, cell type, culture conditions, levels of APP expression. This last point, underscored by the results obtained using DS fibroblasts, may be of particular relevance when studying the effect of chronic pharmacological treatments potentially affecting APP expression and secretion.