Alzheimer's disease (AD) is a debilitating illness that affects millions of Americans. New strategies are beginning to emerge for diagnosis of this condition, a condition that currently can only be diagnosed with certainty at autopsy. If diagnostic strategies can improve, treatment of the disease at an earlier stage, before symptoms emerge, may be possible. In fact, the only drug currently available to treat the symptoms of AD tends to be most effective if given in early stages of disease. Clearly, early and definitive diagnosis is essential.
AD is characterized clinically by cognitive decline and memory loss and neuropathologically by the presence of neurofibrillary tangles (NFTs), neuropil threads (NTs), senile plaques (SPs), and regionally specific neuronal loss (Selkoe, D. J. 1994 Annu. Rev. Cell Biol. 10:373-403; Trojanowski, J. Q. et al. 1996 In: Current Neurology, Vol. XVI, Boston: Houghton Mifflin). The gradual accumulation of paired helical filaments composed of abnormal tau in NFTs and NTs (Lee, VM-Y et al. 1991 Science 251:675-678) as well as beta-amyloid-containing fibrils within SPs (Selkoe, D. J. 1994 Annu. Rev. Cell Biol. 10:373-403) have been implicated in the pathogenesis of AD. Similar neuropathological findings also are observed in the brains of elderly Down's syndrome (DS) patients who survive beyond the fourth decade of life.
Although the molecular mechanisms responsible for the pathogenesis of NFTs, NTs, and SPs remain to be clarified, immunohistochemical and Western blot analyses of AD brains have allowed detailed characterization of the abnormal tau, beta-amyloid-containing proteins in these lesions (Arai, H. et al. 1990 Proc. Natl. Acad. Sci. USA 87:2249-2253; Arai, H. et al. 1991. Ann. Neurol. 30:686-693; Selkoe, D. J. et al. 1986. J. Neurochem. 46:1820-1834). Congo Red and analogs thereof have been used to characterize amyloid plaques in Alzheimer's brains since this fluorescent dye and analogs thereof bind to peptides that make up the fibrils of these plaques (Ashburn et al. 1996 Chemistry and Biology 3:351-358). Northern blot analyses of AD brains also have identified changes in a variety of mRNAs, including those encoding the amyloid precursor proteins (Kang, J. et al. 1987 Nature 325:733-736; Goldgaber, D. et al. 1987. Science 235:877-880; Robakis, N. K. et al. 1987. Proc. Natl. Acad. Sci. USA 84:4190-4194; Tanzi, R. E. et al. 1987. Science 235:880-884; Golde, T. E. et al. 1990. Neuron 4:253-267). Furthermore, in situ hybridization histochemistry has localized abnormal tau and amyloid precursor mRNAs to neurons and glia in the normal and AD brain (Tanzi, R. E. et al. 1987. Science 235:880-884; Golde, T. E. et al. 1990. Neuron 4:253-267; Kosik, K. S. et al. 1989. Ann. Neurol. 26:353-361; Bahmanyar, S. et al. 1987. Science 237:77-88; Schmechel, D. E. et al. 1988. Alzheim. Dis. Assoc. Disord. 2:96-111). Although other protein components in NFTs and SPs have been identified (Schmidt, M. L. et al. 1994. Exp. Neurol. 130:311-322; Strittmatter, W. J. and A. D. Roses. 1995. Proc. Natl. Acad. Sci. USA 92:4725-4727), no data are available which provide information on whether RNAs exist in NFTs and SPs themselves. In fact, little is known about the non-proteinaceous components of SPs and NFTs.
The present invention provides a method for detecting the presence of and identifying RNAs, specifically mRNAs, in NFTs, NTs, and SPs of AD brain tissue. Using this method, it has now been found that neuronal mRNAs predominate in SPs in Alzheimer's disease.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a method of identifying senile plaques, neurofibrillary tangles and neuropil threads in brain tissue which comprises contacting brain tissue with a fluorescent dye capable of intercalating selectively into nucleic acids and detecting any fluorescence in the brain tissue indicative of senile plaques, neurofibrillary tangles and neuropil threads in the brain tissue.
Another object of the present invention is to provide a method of identifying RNAs in senile plaques, neurofibrillary tangles, and neuropil threads of brain tissue which encode proteins involved in the pathogenesis of Alzheimer's disease which comprises isolating single senile plaques in brain tissue by immunocytochemical techniques; identifying the presence of RNA by contacting said senile plaque with a fluorescent dye capable of intercalating selectively into nucleic acids; amplifying the identified RNA; and determining whether the amplified RNA product hybridizes to any known cDNAs for proteins involved in the pathogenesis of Alzheimer's disease.
Yet another object of the present invention is a method of diagnosing Alzheimer's disease comprising identifying the presence of RNA encoding a protein known to be involved in the pathogenesis of Alzheimer's disease.
DETAILED DESCRIPTION OF THE INVENTION
Acridine orange (AO) is a fluorescent dye that intercalates selectively into nucleic acids (Schummelfeder, N. 1958. J. Histochem. Cytochem. 6:392-393; von Bertalanffy, L. and I. Bickis. 1956. J. Histochem. Cytochem. 1956. 4:481-493; Rigler, R. 1966. Acta Physiol. Scand. 67 (Suppl.):7-122). AO histofluorescence has been used to detect RNA and DNA in malignant tumor cells as well as in cell and tissue homogenates separated by gel electrophoresis (Dart, L. H. and T. R. Turner. 1959. Lab. Invest. 8:1513-1522; McMaster, G. K. and G. G. Carmichael. 1977. Proc. Natl. Acad. Sci. USA 74:4835-4838; Pinto, A. et al. 1990. Arch. Pathol. Lab. Med. 114 (6):585-588). Cytoplasmic and nuclear RNA species also have been visualized in tissue sections of the developing and adult brain by AO histochemistry (Schmued, L. C. et al. 1982. J. Histochem. Cytochem. 30:123-128; Topaloglu, H. and H. B. Sarnat. 1989. Anat. Rec. 224:88-93; Mai, J. K. et al. 1984. J. Histochem. Cytochem. 32:97-104). Although AO also binds to mucopolysaccharides, they are not visualized in brain by AO histochemistry due to their low abundance (Szabo, M. M. and E. Roboz-Einstein. 1962. Arch. Biochem. Biophys. 98:406-412).
The affinity of AO for nucleic acids is dependent upon the concentration of the dye in the staining buffer and the pH of the solution. A low dye concentration (approximately 10-100 μg/ml) and a pH of 4.0 allows for the optimal intercalation of AO into RNA and DNA, thereby allowing the in situ visualization of these macromolecules (von Bertalanffy, L. and I. Bickis. 1956. J. Histochem. Cytochem. 1956. 4:481-493; Dart, L. H. and T. R. Turner. 1959. Lab. Invest. 8:1513-1522; Mikel, U. V. and R. L. Becker. 1991. Analyt. Quant. Cytol. Histol. 13:253-260). Specifically, upon excitation with ultraviolet/blue spectra (approximately 470-490 nm wavelength), AO intercalated into RNA emits a bright orange-red fluorescence, whereas AO intercalated into DNA emits a yellowish-green fluorescence (von Bertalanffy, L. and I. Bickis. 1956. J. Histochem. Cytochem. 1956. 4:481-493; Rigler, R. 1966. Acta Physiol. Scand. 67 (Suppl.): 7-122). In tissue sections, AO-labeled RNA and DNA stand out against the pale green background of the surrounding neuropil and white matter tracts that lack abundant nucleic acids. AO histochemistry can be used on paraffin-embedded brain sections (Topaloglu, H. and H. B. Sarnat. 1989. Anat. Rec. 224:88-93; Mai, J. K. et al. 1984. J. Histochem. Cytochem. 32:97-104), and can be combined with other histochemical/immunohistochemical techniques.
Using AO histochemistry, it has now been determined that cytoplasmic RNA species, including either ribosomal, transfer, or messenger (mRNA), are present in NFTs, NTs, and SPs of brains from AD and DS patients. AO histofluorescence was used to screen the brains of patients with AD and DS, as well as normal controls and non-AD patients with other neurodegenerative disorders, to determine whether cytoplasmic RNA species are detectable within NFTs, NTs, and SPs. In these experiments, the hippocampal formation and the entorhinal cortex were selected for analysis because NFTs, NTs, and SPs are abundant in these regions in the AD and DS brain (Hyman, B. T. et al. 1984. Science 225:1168-1170; Hyman, B. T. et al. 1990. Neurology 40:1721-1730). In addition to AO histochemistry, sections were double-labeled with AO and thioflavine-S (TS); AO histochemistry also was combined with immunocytochemistry for astrocytic and microglial markers to characterize the cellular distribution of cytoplasmic RNAs in AD brain.
Within the hippocampal formation and entorhinal cortex of normal brains, somatodendritic AO labeling of neurons was observed. AO-positive nuclei also were observed within neurons using the FITC and double cube filters. Ethanol fixation provided the most abundant and intense AO labeling in comparison to NBF and Bouin's fixed sections. Thus, using this AO labeling method cytoplasmic RNA species were identified in pyramidal cells and stellate cells within the normal aged human brain for the first time.
Cytoplasmic neuronal AO labeling also was identified within the hippocampal formation and entorhinal cortex of AD patients. However, in contrast to the control tissue intense labeling with the fluorescent dye was observed within NFTs and NTs. Specifically, AO-labeled puncta were arrayed in a filamentous pattern in NFTs and NTs throughout the hippocampal formation and entorhinal cortex. Moreover, SPs were consistently observed to contain AO labeling, both in the fibrillar corona and the core region. Upon further examination, it was found that AO-labeled NFTs and SPs were observed infrequently within the hippocampal and entorhinal cortex of the normal aged controls.
To examine the nucleic acid specificity of the AO-labeling, tissue sections were pretreated with RNase. This treatment abolished the AO labeling of NFTs, NTs, and SPs, whereas pretreatment with DNase had little or no effect upon cytoplasmic AO labeling of these lesions. Pretreatment with proteinase K also had no effect on the visualization of AO-labeled NFTs and SPs. Accordingly, it is RNA species within NFTs, NTs, and SPs which are AO-labeled.
NFTs, NTs, and SPs in DS patients also exhibited intense AO histofluoresence in the hippocampal and entorhinal regions. The distribution and density of NFTs and SPs was greater in the DS cases as compared to AD patients, however, the AO staining intensity was qualitatively similar. In contrast, AO labeling within the hippocampal formation and entorhinal cortex of the brains of patients with other neurodegenerative disease (e.g., ALS, DLDH and S-D) was similar to the pattern seen in the normal controls, with abundant cytoplasmic neuronal staining and infrequent labeling of NFTs and SPs.
To confirm that the AO-labeled profiles were indeed NFTs, NTs, and SPs, double-labeling with AO and TS was compared with adjacent single-labeled AO and TS stained sections. Qualitative observations showed that the majority of NFTs co-localized AO and TS, whereas a subpopulation of SPs contained only AO labeling. Quantitative analysis demonstrated that AO co-localized to approximately 80% of the NFTs within the stratum pyramidale of CA1 and layers II/III of entorhinal cortex. AO-stained SPs comprised approximately 55% of the entorhinal and CA1-subiculum plaques quantified.
The specificity of this AO labeling technique for NFTs, NTs and SPs thus provides a means for detecting these brain pathologies characteristic of AD in brain tissue. Further, as will be obvious to those of skill in the art upon this disclosure, while the experiments described herein are specific to acridine orange, other fluorescent dyes capable of intercalating selectively into nucleic acids could also be used. Examples of other nucleic acid binding dyes include, but are not limited to bis-benzimide, ethidium bromide and ethidium homodimer. It is believed that these fluorescent dyes, and in particular AO, labeled with a reporter molecule in accordance with well known techniques, could be used in vivo in the diagnosis of Alzheimer's disease in patients suspected of suffering from this diseases by detecting dye bound to NFTs, NTs and SPs in the brain of these patients.
The method of the present invention has been used to identify RNAs within NFTs, NTs, and SPs of AD and DS patients, which play a role in the pathogenesis of these hallmark AD lesions. An amplification RNA (aRNA) technique (Eberwine, J. et al. 1992. Proc. Natl. Acad. Sci. 89:3010-3014; Eberwine, J. et al. 1995. The Neuroscient.1:200-211; VanGelder, R. et al. 1990. Proc. Natl. Acad. Sci. USA 87:1663-1667) was used that allows identification and quantitation of multiple mRNAs of variable abundance in single immunocytochemically identified cells (Crino, P. B. and J. Eberwine. 1996. Neuron 17:1173-1187; Crino, P. B. et al. 1996. Proc. Natl. Acad. Sci. USA 93:14152-14157), as well as a method for the analysis of multiple DNA sequences from single cells in fixed tissue sections (Becker, I. et al. 1996. Lab. Invest. 75:801-807; D'Amore, F. et al. 1997. Lab. Invest. 76:219-224; Emmert-Buck, M. R. et al. 1996. Science 274:998-1001) and in situ hybridization. The presence of RNA in SPs was verified by AO histofluoresence as before, and mRNAs from individual, immunocytochemically identified SPs were amplified. The amplified products were then hybridized to known cDNAs on reverse Northern blots. By “known cDNAs” is meant cDNAs encoding proteins which have been implicated in the pathogenesis of plaques, tangles, or the degeneration of neurons in AD. Examples include, but are not limited to, cDNAs corresponding to the following classes of proteins: 1) plaque-associated proteins (amyloid-beta protein precursor, presenilin 1, heparin sulfate proteoglycan); 2) cytoskeletal proteins (tau, high, low and medium molecular weight neurofilament subunits, beta-actin, microtubule-associated proteins, nestin); 3) protein kinases and phosphatases (PP1-alpha, PP1-gamma, PP2-alphac, cyclin-dependent kinase, glycogen synthase kinase 3 beta); 4) neurotrophins and neurotrophin receptors; 5) glial enriched proteins (glial fibrillary acidic protein, apolipoprotein E, alpha 1-antichymotrypsin, interleukins); 6) transcriptional activators/cell death mediators (cyclic AMP response element binding protein, c-fos, c-jun, cyclin D1); 7) glutamate receptors and calcium channels; and 8) others (superoxide dismutase 1, dopamine-beta hydroxylase, glutamate decarboxylase, and glyceraldehyde-3′-phosphate dehydrogenase).
Due to its consistent high abundance in SPs, cyclic AMP response element binding protein mRNA (CREB mRNA) was used to normalize the hybridization signal to other mRNAs on each blot, enabling quantitative comparisons of the relative mRNA levels in individual SPs with the mRNA levels in the neurons and neuropil of CA1 in elderly control brains. Empty vector was used as a negative control. Following CREB mRNA normalization, SPs were shown to harbor two distinct populations of mRNAs, high abundance and low abundance mRNAs. There were 18 high abundance mRNAs identified including amyloid-beta protein precursor (APP), tau, bcl-2, bax, PP1-alpha, PP1-gamma, and six different AMPA/kainate glutamate receptor mRNAs. There were 33 low abundance mRNAs identified that included neurofilament subunit mRNAs and glial enriched mRNAs. Comparison of mRNAs amplified from SPs as well as CA1 neurons and neuropil revealed some distinct differences. For example, bax mRNA, which encodes a protein that promotes apoptotic cell death by interacting with the anti-apoptotic bcl-1 protein, was present in SPs at significantly higher levels than in control CA1 neurons and neuropil. It is known that bax expression is upregulated in the AD brain (MacGibbon, G. A. et al. 1997. Brain Res. 750:223-234; Su, J. H. et al. 1997. J. Neuropathol. Exp. neurol. 56:86-93). In contrast, the mRNA levels of high affinity nerve growth factors, trkB and trkC, were significantly lower in SPs compared to control CA1 neurons, a finding that is paralleled by evidence that neurotrophin signaling may be impaired in AD due to diminished expression of neurotrophin receptors (Salehi, A. et al. 1996. Neurosci. 75:373-387; Mufson, E. J. et al. 1997. Exp. Neurol. 146:91-103). Several mRNAs preferentially enriched in glia (alpha 1-act, APOE, GFAP, and IL-1) and a housekeeping gene (GAPDH) were also consistently less abundant in SPs relative to control neuropil.
In situ hybridization was then used to localize selected mRNAs detected in SPs by the aRNA method to SPs in tissue sections. The presence of CREB mRNA in SPs was confirmed using a digoxigenin labeled CREB probe; and CREB mRNA was shown to be distributed throughout the corona of hippocampal SPs double-labeled with TS. Digoxigenin labeled CREB also was observed in neurons, but not in vascular amyloid deposits. Because amyloid beta is the major component of amyloid precursor plaques, and APP mRNAs have been localized to TS stained SPs by in situ hybridization (Hyman, B. T. et al. 1993. Mol. Brain Res. 18:253-258), detection of APP mRNA in SPs, CA1 neurons, and neuropil was performed by PCR. The PCR primers used differentially recognize the most abundant isoforms of APP in brain (APP695, APP751, and APP770) . APP695, the splice variant enriched within neurons, was the predominant species of APP detected in single SPs, CA1 neurons, and neuropil. To confirm the identity of the 87 base pair PCR product obtained from single SPs and CA1 neurons, this PCR product was sequenced and shown to be 100% identical to the corresponding segment of wild type APP695. This confirms that the individual mRNAs amplified from single SPs, neurons and the neuropil correspond to the cDNAs used to detect and identify these same mRNAs. Accordingly, using the method specific characterization of mRNA species sequestration in AD SPs can be performed.
Further, multiple mRNA species have now been found in individual, extracellular SPs of the AD hippocampus with this method combined with aRNA expression profiling, PCR and in situ hybridization methods. The expression profile of mRNAs amplified from these SPs is predominantly neuronal. Thus, using this method of the present invention, studies can be performed to explore the role of these mRNAs in the pathogenesis of AD as well as to identify brains with AD specific changes. Further, expression profiling information derived from the method of the present invention can be used in the design of probes specific to identified mRNA encoding proteins involved in the pathogenesis of AD. Such probes can be used in vivo to detect abnormal levels of these mRNAs which may be an early indicator of AD. The existence of a variety of specific mRNAs in AD brain is also a potential tool for use in designing new therapeutics to be used in the treatment of AD.