CA1302250C - Diagnostic for alzheimer's disease - Google Patents

Abstract

DIAGNOSTIC FOR ALZHEIMER'S DISEASE

ABSTRACT OF THE DISCLOSURE

It has been discovered that paired helical filament antigens, that are cross-reactive with the paired helical filament isolated from brains of persons who suffered from Alzheimer's Disease, are found at elevated levels in cerebral spinal fluid of persons suffering from said disease. Thus, a method is provided for indicating whether a person suffers from Alzheimer's Disease by measuring the level of paired helical filament antigens in a specimen of the person's cerebral spinal fluid. It has been discovered, further, that a major component of paired helical filament in brains of Alzheimer's Disease patients and of paired helical filament antigens in cerebral spinal fluid of persons suffer from the disease is abnormally phosphorylated microtubule-associated proteins known as tau. Thus, the invention also provides a method for indicating whether a person suffers from Alzheimer's Disease by testing for the presence of abnormally phosphorylated tau proteins in a specimen of the persons' cerebral spinal fluid.

Description

~3~)2~5~3 DIAGNOSTIC FOR ALZHEIM~R'S DISEASE

This invention was made with support under grants from the National Institute of Health. The United States Government has certain rights in this invention.
The present invention is directed to laboratory tests indicative of ~lzheimer's Disease and more particularly to assays for the levels of Alzheimer's Disease-associated antigenic materials in cerebral spinal fluid.

BACKGROUND OF THE INVENTION
A neurological degenerative disorder receiving increasing public attention is Alzheimer's Disease (AD) which is the cause of "senility" in a large number of older persons. This disease is characterized by alterations in the neurofibrils. In addition, argentophilic deposits, known as senile plaques, are observed.
Degeneration of the neurofibrils is evidenced by neurofibrillary tanglesl which are made of paired helical filaments, H. M. Wisniewski, et al. in: Marotta CA, ed. Neurofilaments, Minneapolis: University of Minnesota Press 1983: 196-221. The neurofibrillary tangles are found in neuronal perikaryon and in dystrophic neurites of the neuritic (senile) plaques in the cerebrum of a person suffering from AD.
The biochemical origin and composition of paired helical filaments have remained obscure.
A heterogeneous group of closely related microtubule-associated proteins, called "tau," with molecular weights on SDS-polyocrylamide gels mostly between about 55,000 and 62,000 daltons, has long been known from studies of various tissues, including :

~302;;:5~

mammalian brain. See, e.g., Weingarten et al., Proc.
Nat'l. Acad. Sci. (U.S.A.) 72, 1858-1862 (1975);
Cleveland et al., J. Mol. Biol. 116, 227-247 (1977); and Lindwall and Cole, J. Biol. Chem. 259, 12241-12245 (1984). In vitro, tau is known to promote microtubule assembly and to interconnect actin filaments, and both activities are known to be influenced by the degree of phosphorylation of tau. Selden and Pollard, J. Biol.
Chem. 258, 7064-7071 (1983); and Lindwall and Cole/ J.
Biol. Chem. 259, 5301-5305 (1984). The role of tau in vivo remains unclear.
The cross-reactivity of anti-paired helical filament antibodies with normal brain microtubules assembled in vitro has been reported.
Grundke Iqbal et al., Ann. Neurol. 6, 532-537 (1979);
and Grundke-Iqbal et al., Lancet I, 578-580 (1979~.
Antisera to brain microtubules are known to label isolated PHF which are washed repeatedly with SDS and neurofibrillary tangles and plaque neurites, but not amyloid, in tissue sections of Alzheimer's brain (i.e., brain from Alzheimer's Disease patients).
Wisniewski et alu, J. Neuropathol. Exp. Neurol. 43, 543-565 (1984). High molecular weight microtubule-associated proteins and various neurofilament peptides from normal brain are also known to be immuno-cross-reactive with isolated paired helical filament. Ishii et al., Acta. Neuropathol. 48, 105-112 (1979); Anderton et al., Nature 298, 84-86 (1982);
Dahl et al., J. Neurosci. (Baltimore) 2, 113-119 (1982);
Gambetti et al., J. Neuropathol. Exp. Neurol. 42, 69-79 (1983); Perry et al., Proc. Nat'l. Acadu Sci. (U.S.A.) 82, 3916-3920 (1985); and Kosik et al., Proc. Nat'l.
Acad. Sci. (U.S.A.) 81, 7941-7945 (1984).
A monoclonal antibody specifically reactive with paired helical filaments has been reported by G. P.
Wang, et al., Acta Neuropathol 62: 268-75 (1984), and the distribution of reactivity of the antibody in brain ~:311)22S~

tissue from deceased Alzheimer's Disease patients has confirmed the localized high concentrations of paired helical filaments in neuronal perikaryon and dystrophic neurites of neuritic plaques in the cerebrum.
Alzheimer's Disease in living persons is now diagnosed primarily according to behavioral symptoms.
With present methods for diagnosing AD, it is known that a large fraction of patients are misdiagnosed as having the disease. There is clearly a need for improvement in methods for diagnosing the disease.
There is no known laboratory test of a body fluid or tissue which gives quantitative results that, alone or in conjunction with results of other tests, such as the tests based on behavioral symptoms now being used, are indicative of AD in a living patient.
Although paired helical filaments are known to occur in high concentration in certain locations in AD patients' brains, levels of paired helical filaments in brain tissue cannot be measured in living patients.
Thus, to improve the clinical usefulness and reliability of AD diagnoses, it would be highly desirable to have laboratory tests, to be conducted with body fluids or tissue of the living, which give quantitative results indicative of presence or absence of AD.

SUMMARY OF THE INVENTION
It has now been discovered that cerebral spinal fluid (CSF) contains trace levels of paired helical filament antigens ~referred to herein as PHFA), which are antigens that are cross-reactive with antibodies known to be reactive with isolated paired helical filaments (referred to herein as "PHF"), and that there is an indicative correlation between Alzheimer's Disease and high levels of PHFA in CSF. Thus, assays for P~IFA
in the CSF are provided, based upon binding of anti-PHF
antibody to PHFA in CSF specimens. The assays according :' 22~;~

to the invention are applicable to living patients and provide objective, quantitative results that are indicative of whether a patient suffers from AD. High levels of PHFA in CSF have been found to be indicative of, and can also be employed to confirm diagnoses of, AD. A preferred assay is a two-step, competitive enzyme-linked immunosorbent assay (ELISA), in which PHFA
in a spinal fluid specimen compete with surface-bound PHF for PHF-reactive antibody.
Further, it has been discovered that a major component of P~F in neurofibrillary tangles and of PHFA
in CSF of persons suffering from Alzheimer's Disease is abnormally phosphorylated tau protein. Thus, the present invention also provides assays, preferably immunoassays, for the Alæheimer's Disease-associated abnormally phosphorylated tau in CSF. The presence of said abnormally phosphorylated protein in the CSF of a person indicates that the person suffers from AD. A
finding of such presence, employing an assay of the invention for abnormally phosphorylated tau protein, can also be used to confirm a diagnosis of AD based on other techniques.

DETAILED DESCRIPTION OF THE INVENTION
The invention is based on the discovery that PHFA is found in the CSF of humans and that an elevated level of PHFA in the CSF is indicative of Alzheimer's Disease. Because CSF specimens can be relatively easily obtained, the present invention provides methods for indicating whether a living person suffers from AD by assaying for PHFA in a specimen of CSF from the person.
Prior to the discovery which underlies the present invention~ it was not known to be possible to test for AD-indicative PHFA in living persons, because it had only been established that elevated levels of PHF were found in localized areas of brain tissue which could only be examined after a patient had died.

o2~25~

One class of AD-indicative assays in accordance with the invention are immunoassays for PHFA in CSF.
In another aspect, it has been discovered that a major constituent of PHF in brains of persons who have died of Alzheimer's Disease and of the PHFA in the CSF
of persons suffering from the disease is abnormally phosphorylated tau proteins. Thus, the present invention also entails methods for indicating whether a person is suffering from Alzheimer's Disease by assaying a specimen of the person's CSF for these abnormally phosphorylated tau proteins.
The immunoassays according to the invention include solution-phase and solid-phase, and competitive as well as non competitive. They include enzyme-linked immunosorbent assays (ELISAIs) and radioimmunoassays (RIA's). The immunoassays are carried out by methods known in the immunoassay art for determining the concentration of an antigen (e.g., PHFA or abnormally phosphorylated tau protein) in a specimen using antibody (polyclonal or monoclonal) known to be reactive with the antigen.
One particular type of AD-indicative immunoassay according to the invention is a two-step, competitive solid-phase assay wherein a quantity of prepared PHF and an unknown quantity of PHFA in a CSF
specimen compete for binding by a PHF-specific antibody. Either the prepared PHF or the PHFA in the CSF is bound to a solid support (e.g., the inside wall of a microtiter plate well) while the other is held in solution. In the assay, an excess of the anti-PHF
antibody is first incubated with and complexed with the entire amount of either the support-bound antigen or the antigen in solution and the antibody remaining unbound is then incubated with the other of the antigens. The complexing of antibody with either the PHF or the PHFA
in the first step results in "inhibiting" antibody binding with the other of the antigens in the second "~%~s~

step. The preferred of these competitive solid-phase immunoassays are ELISA's. Thus, as understood in the immunoassay art, the amount of antibody not complexed in the first step, but complexed in the second, whether the amount is support-bound or in solution, is measured by 1) reacting it with an anti-antibody~ preferably a polyclonal, subclass non-specific antibody, that reacts with immunoglobulin of the same class (e.g., IgG) and from the same species (e.g., mouse) as the anti-PHF
antibody and that has been previously linked to or complexed with an enzyme, 2) separating the portion of immunoglobulin-enzyme complex that binds to the PHF-antibody or PHFA-antibody complex from that portion which does not bind, and 3) determining the amount of enzyme in either the bound or unbound portion by exposing the same to a reagent or reagents which undergo a chromogenic reaction catalyzed by the enzyme.
Heretofore, there have been no reports of the presence of PHFA in the CSF, and there have been no reports of substances specifically associated with Alzheimer's Disease in the CSF. The discoveries of the presence of PHFA in the spinal fluid and the correlation of elevated spinal fluid PHFA levels with Alzheimer's Disease are highly significant because CSF is accessible to assay for diagnostic purposes with living patients.
Indeed, it may be expected that an assay for PHFA in CSF, such as the competitive ELISA described herein or other immunoassays, will eventually be performed on any patient suspected of having or being in a high risk category for Alzheimer's Disease.
More specifically, the immunoassay that we have developed for PHFA-level measurement in CSF is a two-step, competitive enzyme-linked immunosorbent assay (ELISA). The assay is based upon a sandwich technique, E. Engvall, et al., Method Enzymol., 70: 419-439 (1980), such as that which has been used previously to determine titers and specificity of antisera as well as ' -` , -2~1) to measure antigens both qualitatively and quantitatively. In the assay that we have developed, PHFA in CSF specimens is measured quantitatively through an antibody inhibition step.
Isolated (i.e., purified) paired helical filaments are obtained according to the method of K.
Iqbal, et al., Acta Neuropathol, 62, 167-177 (1984).
Briefly, by a standard technique, microtiter plate wells are coated with a fixed quantity of sonicated PHF per well. As the inhibition step, in two separate sets of vials, aliquots of a solution containing PHF-reactive antibody (preferably monoclonal antibody) are incubated with various known concentrations of fragmented (sonicated) PHF (in the set of vials for establishing a standard curve), as well as with CSF specimens (for dilutions thereof) containing unknown quantities of PHFA, for a period of -time that is sufficient to permit substantially complete reaction of the anti-PHF antibody with the available PHF or PHFA.
Thereafter, an equal volume of the incubation mixture from each vial is added to the PHF-coated plate wells and incubated for a period of time sufficient for antibody, which was not prereacted (inhibited) with the PHF (or PHFA) in the solution in the vial, to react with the PHF that is coated on the plates. Then the plates are washed thoroughly to remove all antibody, PHF, and antibody-PHF or antibody-PHFA complex that is not bound to the plate.
For assays, such as competitive assays, which require that the PHF be dissolved in or stably suspended in aqueous medium, it is necessary to fragment the PHF
so that it may be carried by the aqueous medium, intact PHF being substantially insoluble in aqueous medium. In accordance with an important aspect of the invention, fragmented PHF is prepared for use in assays by suspension and ultrasonication of the suspended, purified P~F. Chemical fragmentation of PHF in aqueous .

~2:~5~

medium is more cumbersome and much less desirable than ultrasonic fragmentation because chemical fragmentation requires use of protein denaturants, which must be completely removed from the fragmented PHF before the PHF is suitable for use in an assay but which is very difficult to remove without causing reaggregation of the PHF. PHF fragments in the presence of a denaturant will not bind efficiently to microtiter plate wells or anti-PHF antibody in an assay. Further, as denaturant is separated from a solution of such PHF ~ragments, the fragments tend to form large, insoluble aggregates that are unsuitable for assay purposes.
The conditions of ultrasonication of PHF are not considered particularly critical, as long as the combination of intensity and exposure time are sufficient to fragment substantially all of the particles to a size where they are not visible under a light microscope, e.g., so that the fragments have greatest dimensions less than about 50 nm. E'ragments produced by ultrasonication are known to be stable in solution for at least about 24 hours, and it is expected that they are stable for considerably longer periods at temperatures approaching the freezing point; however, PHF fragment solutions should not be frozen, as freezing causes undesirable aggregation.
As a means of measuring the amount of antibody that is bound to the plate, the plate is exposed to an excess of a second antibody that reacts with the anti-PHF antibody and which is appropriately labeled with an enzyme. The second antibody is an anti-immunoglobulin which is reactive with immunoglobulins of the species from which the anti-PHF
~ antibody is derived and is usually non-subclass or - non-class specific. For example, if the first antibody is mouse anti-PHF monoclonal IgG antibody, the enzyme label may be linked to anti-mouse IgG or anti-mouse immunoglobulin that is not specific as to class. A

?~

2~i[) g preferred means of labeliny the second antibody is by linkage to an enzyme, such as alkaline phosphatase, which cataly~es the conversion of p-nitrophenyl phosphate to p-nitrophenol, which reaction is chromogenic, being observable by a rise in the optical density of the solution at 405 nm. The rate of increase in optical density of the solution per unit time is related to the amount of enzyme linked to the second antibody in the system, and measurement of this rate (the increase in optical density after a speciied period of time) is the basis for measuring the amount of first antibody bound to PHF in the microtiter well.
Other enzymatic reporter systems, such as the peroxidase reporter system, described in the Journal of Biological Chemistry 257: 14173-1~180 (1982), may be substituted for the alkaline phosphatase reporter system as a means for visualizing PHF-bound antibodies.
In this particular assay system, the more PHFA
that is present in the CSF specimen, the less color change that develops per unit time. Although the rate of color development decreases as a function of PHFA
concentration in the CSF specimen, it is not a linear relationship. Accordingly, it is necessary to prepare a standard curve, using solutions of known PHF
concentrations, against which the color development from a specimen with an unknown concentration of PHFA may be compared. The PHF for the solutions used to establish the standard curve can be Eragmented, purified PHE that is used to coat microtiter plate wells as described above. The standard curve for this assay will be approximately linear over only a limited PHF
concentration range; and, in order that the level of PHF
in an unknown CSF specimen may be read with accuracy from a standard curve, it is helpful in some cases, particularly in the case of abnormally high PHFA levels, to perform the assay on serial dilutions of a CSF
specimen to ensure that the concentrations of PHFA in ~3Z2S~
,, one or more of the serial dilutions o~ the specimen falls within the linear range. However, assay of a single CSF concentration is generally adequate to determine an abnormally high PHFA level; and, because a very high PHFA concentration is indicative of AD, precision of numerical values of highly elevated PHFA
levels will not be important in many cases.
A variation on the ELISA assay described above is to bind anti-PHE` antibody, rather than PHF, to the solid support. Also required in this case is enzyme-labeled PHF (PHF that is linked either directly or through a linking moiety, such as any bifunctional linker known in the art to covalently join protein molecules stably, to a suitable enzyme)~ A known quantity of the enzyme-linked PHF is added to CSF
specimens of known volume, and aliquots of such mixtures are placed in the antibody-coated wells. PHFA in the CSF specimens compete for antibody binding sites with the enzyme-labeled PHFo After an appropriate amount of time for PHF enzyme-labeled and PHFA in CSF to bind to the anti-PHF antibody, the wells are washed and subsequently exposed to a reagent system that undergoes a chromogenic reaction catalyzed by the enzyme linlced to the PHF. In such an assay, greater rate of color development indicates less PHFA from CSF in the reaction mixture. This assay system has the advantage of being somewhat simpler to perform, requiring only antibody linked to a plate, an enzyme-linked PHF solution, and substrate solution for enzymatic catalyzed reaction.
This assay system may be more suitable than the two-step, competitive ELISA for providing in kit form to a medical laboratory.
Radioimmunoassays represent another suitable technique whereby PHFA in a CSF specimen may be quantified. For example~ employing the well known chloramine-T reaction with fragmented, purified PHF, the labeling of PHF can be with a radioisotope, 125I, rather than an enzyme, and the amount of labeled PHF
measured (via radioactive emission) as bound to the microtiter plate wells will decrease as the amount of PHFA in CSF being assayed increases. However, enzyme-linked or other non-radioactive chromogenic assay procedures are preferred to radioimmunoassay techniques from the standpoint of safety considerations and reagent stability.
It is noteworthy, however, that this invention entails assay of PHFA in CSF by any available technique in the art, immunoassay or other. With respect to immunoassay techniques, the invention is not limited to any particular form that an immunoassay may take.
The development of an assay sensitive to levels of PHFA in CSF was made possible, in part, through the development of monoclonal antibodies specifically reactive with P~F. The first hybridoma cell line, which produces PHF-specific monoclonal antibody, was developed by~Wang, et al., supra. This hybridoma cell line is deposited at the American Type Culture Collection, Rockville, Maryland, U.S.A. (ATCC), where it is assigned accession number HB 9039. The deposit with the ATCC has been made under the terms of the Budapest Treaty on the ~eposit of Microorganism for the Purposes of Patent Procedure and the Regulations promulgated thereunder and samples of the cell line are and will be available from the ATCC to industrial property offices and other persons legally entitled to receive them in accordance with said Treaty and Regulations and in accordance with the patent laws and regulations of every country and international organization in which an application corresponding to this application is filed or a patent based on any such application is granted.
Reference herein to the hybridoma line with ATCC accession number HB9039 also encompasses subcultures of that line, as well as lines of mutants of cells from that line, or subcultures thereof, provided ~31[)Z2~

that the subcultures and lines of such mutants produce the monoclonal antibody reactive with PHF that is provided by the line with ATCC accession number HB9039.
The invention encompasses hybridoma cell lines which produce monoclonal antibodies which have specific activity for PHF, and the anti-PHF monoclonal antibodies produced by such cell lines. In particular this invention encompasses the cell line HB 9039 and the monoclonal antibody produced thereby. With the monoclonal antibody produced by the HB 9039 cell line, selection for hybridoma cell lines which produce antibodies reactive with the same antigenic determinant of PHF is greatly facilitated, and the invention encompasses other hybridoma cell lines which produce antibodies that react with the same antigenic determinant on PHF as the antibody produced by hybridoma line HB9039. ~inding of a monoclonal antibody to the same antigenic determinant is established if prior binding to PHF of the monoclonal antibody produced by the HB 9039 cell line blocks subsequent binding of the other monoclonal antibody.
As indicated above, it has been discovered that a major component of PHF isolated from Alzheimer's brains and and PHFA in CSF of persons suffering from Alzheimer's Disease is abnormally phosphorylated tau protein.
Consequently, the presence of the abnormally phosphorylated tau protein in a specimen of CSE from a person indicates that the person is afflicted with Alzheimer's Disease. The presence of abnormally phosphorylated tau in CSF can be determined by any of the immunoassay techniques described above, in connection with PHFA, using polyclonal or monoclonal antibody reactive with the abnormally phosphorylated tau. As positive control or as standard to establish standard curves in the assays, and as competitor in competitive assays, abnormally phosphorylated tau ~a~so protein prepared by electrophoresis or by immunoaffinity chromatography as described below, is employed~
Alternatively, employing polyclonal or monoclonal antibody that differs in affinity for abnormally phosphorylated and normal tau protein, the concentration of tau proteins in a CSE specimen can be measured by immunoassay both before and after subjecting the specimen, or portions thereof, to treatment, such as treatment with alkaline phosphatase, that dephosphorylates any abnormally phosphorylated tau protein. A difference in the amount of tau detected before and after dephosphorylation indicates the presence of abnormally phosphorylated tau in the specimen.
Microtubule-associated proteins t known as tau proteins, are known, as described above.
Also known and widely distributed among practitioners in the art is a tau-specific monoclonal antibody known as tau-l. Binder et al., J. Cell Biol. 101, 1371-1378 ~1985); Grundke-Iqbal et al., J.
Biol. Chem. 261, 6084-6089 (1986); Wood et al.l Proc.
Nat'1. Acad. Sci. (U.S.A.) 83, 4040-4043 (1986). Tau-1 binds to tau proteins from calf, rat and normal human brain and the epitope it recoynizes seems to be confined virtually exclusively to tau proteins.
Abnormally phosphorylated tau protein was identified as a component of PHF both by immonocytochemical analysis of neuro~ibrillary tangles in tissue section of Alzheimer's brains both before and after dephosphorylation with alkaline phosphatase and by immunoblots of polypeptides from isolated PHF, prepared as described by Iqbal et al., Acta ~europathol. 62, ; 167-177 (1984), both before and after dephosphorylation with alkaline phosphatase.
~ 35 Dephosphorylation of PHF (or PHFA in CSF) is ; accomplished with calf alkaline phosphatase at 40-50 ug/ml, pH 7.5-8Ø

2S~

It was found that tau-l bound only weakly, if at all, to the neurofibrillary tangles, and the isolated PHF in immunoblots, without dephosphorylation, although it does bind significantly with tau proteins from normal brain without dephosphorylation. However, after dephosphorylation, the tau-l bound to the PHF, in tangles and in immunoblots, nearly as if the PHF were composed of normal tau protein. (In the immunoblots, the tau proteins from PHF migrate to essentially the same positions as tau proteins isolated from normal brain.) The tau-l monoclonal antibody, or any monoclonal antibody or polyclonal anti-serum which binds to abnormally phosphorylated tau and normal tau with different affinities, can be employed to assay CSF
specimens for abnormally phosphorylated tau, characteristic of PHF and Alzheimer's disease, by simply assaying one portion of the CSF specimen without dephosphorylation and another portion with dephosphorylation. If abnormally phosphorylated tau is - present in the specimen, the two assays will yield different concentrations for tau. Otherwise/ the assays will yield the same concentration for tau~
Abnormally phosphorylated tau protein can be isolated from normal tau protein in Alzheimer brain preparations by standard electrophoretic techniques, such as agarose gel electrophoresis, on the basis of the additional negative charges, due to the additional phosphate groups, on the abnormally phosphorylated protein.
Alternatively, the abnormally phosphorylated protein can be isolated by standard immunoaffinity chromatography techniques from isolated PHF or from preparations of Alzheimer brain tissue. Firs~, antibody which recognizes tau, both phosphorylated and not, is used to isolate tau proteins. Many such antibody preparations are known and are available in the art.

:~3~D2~Si~
- 15 ~
See Grundke-Iqbal et al., J. Biol. Chem. 261, 6084-6089 (1986); Drubin et al., J. Cell Biol. 98, 1090-1097 (1984); and Kosik et al., Proc. Nat'l. Acad. Sci.
(U.S.A.) 83, 4044-4048 (1986). Then either an antibody specific for normal tau protein, such as tau-l, or an antibody specific for abnormally phosphorylated tau, is employed to isolate the abnormally phosphorylated from the normal tau.
With purified abnormally phosphorylated tau available, polyclonal antibody preparations and monoclonal antibodies specific for the abnormally phosphorylated protein can be prepared by standard, well known techniques. In the preparation of monoclonal antibodies, hybridoma cultures are screened for antibody that recognizes abnormally phosphorylated tau but does not recognize normal tau (as derived, e.g., from brain tissue of persons younger than about 50 years old who have died but did not suffer from Alzheimer's Disease).

EXAMPLE I
CSF specimens were collected from nine patients with Alzheimer's Disease as determined by the criteria of J. Eisdorfer, et al. J. Fam. Pract. 11: 553-57 (1980), (mean age 64+10 SD years) and from nine patients (57+19 SD years) with non-AD neurological conditions, including stroke, seizures, multiple sclerosis, and other neurological conditions. The specimens were coded before analysis for PHF antigen.
Paired helical filaments were isolated at necropsy from the cerebral corte~ from patients with Alzheimer's Disease and senile dementia of the Alzheimer type according to the method of Iqbal, et al., supra.
Isolated paired helical filaments were fragmented to particles of sizes not visible by light microscopy by suspension in 0.32 M sucrose followed by ultrasonication of the suspension using a Branson sonicator at 10% pulse, 10 watts output and 4 second on-off cycle for .

~3~2;~5~

30 min. The protein concentration was determined by amino acid analysis and adjusted to 75 ug/ml with phosphate-buffered saline (PBS), pH 7.2.
A two-step compe~itive ELISA was done in duplicate. Microtiter plate wells were coated with P~F
by incubation overnight at 4C in a humid chamber with 100 ul per well of a dilution of the sonicated P~F
suspension to 0.5 ug/ml PHF with 0.05 M Na carbonate-bicarbonate buffer, pH 9.6. After the incubation~ the wells were washed three times with PBS
* *
containing 0.05% "Tween-20" (PBS-Tween).
Mouse anti-PHF-specific monoclonal antibody, Wang, et al., supra., from hybridoma line HB 9039, at an initial concentration of 30 mg/ml protein, was diluted 1:20,000 with PBS-Tween. 100 ul aliquots of the diluted antibody were mixed 1:1 with 1) PBS-Tween as a negative control, 2) serial dilutions of sonicated PHF suspension in PBS-Tween for establishing a standard curve and 3) the CSF specimens. The mixtures were incubated for one hour at 22C. The monoclonal antibody solutions were then transferred to the wells of the microtiter plate with bound PHF. The plate was incuba~ed at 22C
for three hours. After washing three times with PBS-Tween, 100 ul of alkaline phosphatase conju~ated to goat anti-mouse IgG (0.25 ug/ml) was added to each of the wells and incubated at 22C for three hours.
After again washing three times with PBS-Tween, 100 ul of substrate (p-nitrophenyl phosphate, 1 ug/ml) was added and reacted for 90 and 120 minute periods at room temperature. Optical density was measured at both times at 405 nm in a standard ELISA reader. A standard curve was established, and PHFA levels in CSF specimens were determined with reference to the standard curveO
An arbitrary unit of PHF antigen was established as the inhibition of the rate of color development in the ELISA (from the rate with PBS-Tween negative control) obtained with 180 ng protein of PHF

* trade mark 13~2;~S~

suspension (i.e., the inhibition with the PHF serial dilution containing fragmented PHF at 1.8 ug/ml PHF
protein and treated as described above). The PHF
levels, in these arbitrary units, of the nine CSF
specimens obtained from patients with Alzheimer's Disease were as follows: .01, .22, .23, .26, .34, .47, .48, .55, and .64; the mean and standard deviation being 0.35 ~ 0.06. The PHF levels, in the arbitrary units, of the nine specimens obtained from patients without Alzheimer's Disease were as follows: .01, .08, .12~
.16, .18, .19, .28, .39, and .40, the mean and standard deviation being 0.21 ~ .05.
It can be seen that the Alzheimer's Disease patients, as a group, have substantially higher levels of PHF antigen in their CSF than that of other neurological patients (p less than 0.05, one-tailed t test). It is also appreciated that there is considerable overlap of the PHFA levels in the CSFs of these two groups. Thus, while an elevated level of PHFA
in a CSF specimen is indicative oE Alzheimer's Disease, it cannot be said to be conclusive. Nevertheless, the indicative nature of this test, along with other clinical symptoms, has important utility in establishing whether or not a patient has Alzheimer's Disease. For example, any PHFA level above about 0.4 unit in the CSF
of a patient is highly indicative of Alzheimer's Disease in the patient. A level of below about 0.2 unit is indicative of the patient's being healthy or having a neurological disorder other than Alzheimer's Disease.
As noted above, the numerical unit by which the two groups were compared is arbitrary and other units may be used. For example, if an antigen from a non-human animal is found to be cross-reactive with an antibody that reacts with human PHF, whether or not the antibody has similar degree of affinity for the antigens from the different species, a PHF unit may be established relative to the animal antigen; or ~3~

alternatively, tests using such a non-human antigen may be correlated with tests using the less readily available human PHF so that results may be expressed in terms of human PHF activity.
Although the assay is preferably performed with a PHF-specific monoclonal antibody, the test may be performed using polyclonal antiserum generated against PHF (Grundke-Iqbal, et al. Acta Neuropathol., 62:259~267, (1984); Grundke-Iqbal, et alO, Acta Neuropathol., 66:52-81 (1985). Monoclonal antibody has the advantage of uniformity, allowing better standardization of results~ a result that is highly desirable when kits for the assay are to be prepared for commercial distribution. It is inherently possible, however, to practice the invention using PHF from a variety of sources and with a variety of antibodies, setting up standard curves and establishing normal and abnormal values in each case.
The PHF levels in the above example are determined for a given volume of CSF. It will be appreciated that values may be otherwise calculated, e.g., relative to a measured level of total protein per volume in each CSF specimen.
It should be appreciated that the non-Alzheimer's Disease controls in the above Example were not normal, healthy persons, but that each of the controls had a neurological disorder. The comparison in the example was between Alzheimer's Disease and non-Alzheimer's Disease patients suffering from various neurological disorders because CSF specimens are seldom taken from patients without compelling medical reason.
Therefore~ normal, CSF specimens from healthy adult controls are not as readily available as are CSF
specimens from non-Alzheimer's Disease neurological patients from which CSF is frequently obtained for a variety of diagnostic purposes.

. .

~3~2:~0 Because the non-Alzheimer's Disease controls exhibit neurological de~eneration for other reasons, it is expected that their PHFA levels may well prove to be elevated relative to a normal, healthy population, and that the difference in mean PHFA levels in Alzheimer's Disease and normal, healthy adults may prove to be even greater than the difference found in the Example.
Nevertheless, the comparison in the Example is meaningful because it is expected that the test will initially be used in helping to decide with what neurological disorder a patient is afflicted rather than to distinguish an Alzheimerls Disease patient from a normal, healthy person.
While the invention has been described with some specificity, modifications apparent to one with ordinary skill in the art may be made without departing from the spirit and scope of the invention.
Various features of the invention are set forth in the following claims.

;

.

Claims (20)

1. An in vitro assay for use in the diagnosis of Alzheimer's Disease in a living person, comprising detecting abnormal concentrations of paired helical filament antigens (PHFA) in a sample of cerebral spinal fluid (CSF), wherein the concentration of PHFA in said sample is greater than the concentration of PHFA in a sample of CSF from a person who does not have Alzheimer's Disease.

2. A method according to Claim 1 wherein the level of PHFA in said CSF specimen is determined by immunoassay.

3. A method according to Claim 2 wherein the level of PHFA in said specimen is determined by enzyme-linked immunosorbent assay.

4. A method according to Claim 2 wherein said immunoassay utilizes a monoclonal antibody reactive with human PHF.

5. A method according to Claim 3 wherein said ELISA
utilizes a monoclonal antibody reactive with human PHF.

6. A method according to Claim 5 wherein the ELISA
is a competitive assay in which PHFA from the CSF specimen competes with fragmented PHF for binding by said anti-human PHF monoclonal antibody.

7. A method according to Claim 4 wherein the monoclonal antibody is that secreted by hybridoma cell line HB 9039.

8. A method according to Claim 5 wherein the monoclonal antibody is that secreted by hybridoma cell line HB 9039.

9. A method according to Claim 6 wherein the monoclonal antibody is that secreted by hybridoma cell line HB 9039.

10. A method of preparing PHF fragments comprising providing substantially purified PHF, immersing said substantially purified PHF in an aqueous medium, and subjecting said aqueous medium with said immersed PHF
to ultrasound waves to fragment said substantially purified PHF to a size where the fragments are non-visible under a light microscope.

11. PHF fragments produced in accordance with the method of Claim 10.

12. A method for indicating Alzheimer's Disease in a human comprising detecting the presence of abnormally phosphorylated tau protein in a specimen of CSF from the human.

13. A method according to Claim 12 wherein the detection is by immunoassay with polyclonal or monoclonal antibody with higher affinity for abnormally phosphorylated tau protein than for normal tau protein,

14. A method according to Claim 13 wherein said antibody is a monoclonal antibody.

15. A method according to Claim 12 which comprises (i) measuring the level of tau protein in a portion of said CSF specimen, without treatment of said portion to dephosphorylate tau protein which might be present in said portion; and (ii) measuring the level of tau protein in a portion of said CSF specimen after treatment of said portion to dephosphorylate tau protein which might be present in said portion, provided that the measurement of the level of tau protein in both of said portions is by immunoassay with a monoclonal or polyclonal antibody which has a different affinity for abnormally phosphorylated tau protein than for normal tau protein.

16. A method according to Claim 15 wherein said antibody is monoclonal and binds with a higher affinity to an abnormally phosphorylated tau protein than to normal tau protein.

17. A method according to Claim 15 wherein said antibody is monoclonal and binds with a higher affinity to normal tau protein than to abnormally phosphorylated tau protein.

18. A method according to Claim 15 wherein the dephosphorylation comprises treatment with alkaline phosphatase.

19. A method according to Claim 16 wherein the dephosphorylation comprises treatment with alkaline phosphatase.

20. A method according to Claim 17 wherein the dephosphorylation comprises treatment with alkaline phosphatase.