[0001] The present invention pertains to the brain neuroscience field and the clinical medicine field, and relates to a multiplex surrogate biomarker for evaluating a cerebral amyloid β peptide (Aβ) accumulation state, and a method for analysis thereof. More specifically, the present invention relates to a multiplex surrogate biomarker for evaluating a cerebral Aβ accumulation state using, as an index, a level of Aβ and Aβlike peptides generated by cleavage of amyloid precursor protein (APP) in a living bodyderived sample, and a method for analysis thereof. The biomarker of the present invention is a marker to be used for, for example, presymptomatic diagnosis, screening for subjects of developing preventive intervention (preemptive therapeutic drug administration etc.) and evaluation of drug efficacy of therapeutic drugs and prophylactic drugs regarding Alzheimer's disease.
[0002] Alzheimer's disease (AD) is a principal cause of dementia, and occupies 50 to 60% of the entire dementia. The number of patients suffering from dementia was more than or equal to 24 million in the world in 2001, and is estimated to reach 81 million in 2040 (NonPatent Document 1). It is considered that amyloid β (Aβ) is deeply involved in development of Alzheimer's disease. Aβ is produced as a result of proteolysis of amyloid precursor protein (APP) which is a singlepass transmembrane protein by βsecretase and γsecretase. Appearance of senile plaques in brain due to aggregation of Aβ accompanying fibrosis triggers aggregation and accumulation of tau protein inside neurocytes to cause nerve malfunction and neuronal cell death. It is considered that this results in progressive deterioration of the cognitive ability. It has long been known that Aβ mainly consists of 40 residues (Aβ140) and 42 residues (Aβ142) and migrates into cerebrospinal fluid (CSF) and blood. Further, in recent years, existence of Aβlike peptides having lengths different from those of Aβ140 and Aβ142 in CSF or plasma has been reported (NonPatent Documents 2, 3).
[0003] Amyloid accumulation is considered as the earliest event among pathophysiological changes occurring in brain in the case of AD, and recent studies have revealed that amyloid accumulation in brain starts 10 years or more before onset of clinical symptoms. Therefore, it is important to exactly detect the amyloid accumulation in brain for enabling early diagnosis of AD. At present, amyloid PET and CSF Aβ examination are known as a method for detecting amyloid accumulation. The amyloid PET is a method of visualizing Aβ deposits by using a ligand molecule that specifically binds with Aβ, and an example of the amyloid PET includes PiBPET using Pittsburgh compoundB (PiB) . However, PET examination requires massive equipment, and thus an examination fee to perform one examination is high. Also, PET examination is invasive due to radiation exposure, and is not suited for a screening method of AD. On the other hand, a decrease in concentration of Aβ142 in CSF or a decrease in concentration ratio of Aβ142/Aβ140, and an increase in total tau value or phosphorylation tau value are reported to be a useful marker (Patent Document 1:
JPA201019864, NonPatent Document 4). However, collection of CSF is also highly invasive, and is not suited as a screening method of AD. Therefore, a blood examination that has low invasiveness and is low in cost is desired for the screening.
[0004] Under these circumstances, the potentiality of concentration of Aβ142 existing in blood as an Alzheimer's disease diagnostic marker is expected, and many researchers have reported the relationship between blood Aβ142 concentration and Alzheimer's disease development; however, consistent results have not been obtained (NonPatent Document 4) .
[0005] However, in recent years, a ratio of APP669711/Aβ142 was reported as a promising blood marker that reflects a cerebral amyloid accumulation state (NonPatent Document 5). NonPatent Document 5 indicates that the ratio of APP669711/Aβ142 has a strong correlation with a PiB accumulation degree obtained by PiBPET. Further, the results of ROC analysis between a PiB positive group and a PiB negative group indicate that the ratio of APP669711/Aβ142 is a marker capable of accurately distinguishing between a PiB positive person and a PiB negative person.
[0006] Also, Patent Document 2:
JPA201363976 discloses a monoclonal antibody that does not recognize a soluble Aβ monomer, but specifically binds only to a soluble Aβ oligomer, and also discloses a diagnostic method of Alzheimer's disease using the antibody. Paragraph [0104] of the publication discloses a method in which when the ratio of Aβ oligomer to Aβ monomer in a sample of a subject is higher than that of a normal healthy person, the subject is determined as being a candidate for Alzheimer's disease.
[0007] Patent Document 3:
JPT2014520529 discloses a method for diagnosing Alzheimer's disease by combining a plurality of miRNA levels in a sample obtained from an object.
[0009]
NonPatent Document 1: Blennow K, de Leon MJ, Zetterberg H.: Alzheimer's disease. Lancet. 2006 Jul 29; 368(9533): 387403
NonPatent Document 2: Portelius E, WestmanBrinkmalm A, Zetterberg H, Blennow K. : Determination of betaamyloid peptide signatures in cerebrospinal fluid using immunoprecipitationmass spectrometry. J Proteome Res. 2006 Apr; 5(4): 10106
NonPatent Document 3: Kaneko N, Yamamoto R, Sato TA, Tanaka K.: Identification and quantification of amyloid betarelated peptides in human plasma using matrixassisted laser desorption/ionization timeofflight mass spectrometry. Proc Jpn Acad Ser B Phys Biol Sci. 2014; 90(3):10417.
NonPatent Document 4: Hampel H, Shen Y, Walsh DM, Aisen P, Shaw LM, Zetterberg H, Trojanowski JQ, Blennow K. : Biological markers of amyloid betarelated mechanisms in Alzheimer's disease. Exp Neurol. 2010 Jun; 223(2): 33446
NonPatent Document 5: Kaneko N, Nakamura A, Washimi Y, Kato T, Sakurai T, Arahata Y, Bundo M, Takeda A, Niida S, Ito K, Toba K, Tanaka K, Yanagisawa K.: Novel plasma biomarker surrogating cerebral amyloid deposition. Proc Jpn Acad Ser B Phys Biol Sci. 2014; 90(9) : 35364.
[0010] It has been found that a large quantity of Aβ has been deposited before exteriorization of the cognitive function decline in an Alzheimer's disease (AD) patient. Although AmyloidPET is effective for detecting Aβ accumulation, it requires high examination cost and long time for executing the examination, and thus is not a diagnostic method that allows for a majority of elderly people to easily undergo the examination. Therefore, a simplified analytical method capable of detecting increase in Aβ accumulation before exteriorization of clinical symptoms has been demanded.
[0011] As described above, generally, an examination method using a biomarker existing in blood or cerebrospinal fluid (CSF) as an index is an effective method capable of conveniently detecting the development and progression of a disease on the molecular level. Patent Document 1 and NonPatent Document 4 described above have reported that in Alzheimer's disease, a decrease in concentration of Aβ142 in CSF is a useful diagnostic marker. On the other hand, however, NonPatent Document 4 has also reported that the relationship between blood Aβ142 concentration and AD development is low unlike the case of CSF Aβ142.
[0012] Conventionally, in previous reports regarding Aβ in blood, the correlativity with AD has been examined only for concentrations of two kinds of Aβ140 and Aβ142 in blood. In NonPatent Document 5 described above, the ratio of APP669711/Aβ142 was reported as a promising blood marker that reflects a cerebral amyloid accumulation state. While cerebral amyloid accumulation can be determined with high sensitivity by the ratio of APP669711/Aβ142, a method that enables more accurate discrimination is demanded.
[0013] An object of the present invention is to provide a biomarker for evaluating a cerebral Aβ accumulation state using amyloid precursor protein (APP) derived Aβ and Aβlike peptides in a living bodyderived sample as an index, and a method for analysis thereof. In particular, an object of the present invention is to provide a biomarker for evaluating a cerebral Aβ accumulation state using amyloid precursor protein (APP)derived Aβ and Aβlike peptides in a blood sample as an index, and a method for analysis thereof. More specifically, an object of the present invention is to provide a marker to be used for, for example, presymptomatic diagnosis, screening for subjects of developing preventive intervention (preemptive therapeutic drug administration etc.) and evaluation of drug efficacy of therapeutic drugs and prophylactic drugs regarding Alzheimer's disease, and a method for analysis thereof.
[0014] As a result of diligent studies, the present inventors have completed the present invention by calculating a numerical value by a combination of two or more ratios selected from the group consisting of three ratios, Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142, regarding Aβ and Aβlike peptides derived from APP in a living body sample, through a mathematical technique.
[0015] In the present description, "Aβ" is used as an abbreviation of an amyloid β peptide. That is, "Aβ" includes Aβ140 and Aβ142. A peptide other than the Aβ generated by cleavage of amyloid precursor protein (APP) may be referred to as an Aβlike peptide. Aβ and an Aβlike peptides that are generated by cleavage of amyloid precursor protein (APP) may be referred to as "APPderived peptide".
[0016] The present invention relates to the following items:
 1. Use of a combination of:
a ratio of Aβ139 (SEQ ID NO.: 1) level to Aβ142 (SEQ ID NO.: 3) level:
Aβ139/Aβ142; and
a ratio of APP669711 (SEQ ID NO. : 4) level to Aβ142 (SEQ ID NO.: 3) level: APP669711/Aβ142,
in a living bodyderived sample,
as a marker for determining a cerebral Aβ accumulation state.
 2. An analytical method for determining a cerebral Aβ accumulation state, the method comprising:
a measurement step of subjecting a living bodyderived sample derived from a test subject to detection of a marker containing:
Aβ142 (SEQ ID NO.: 3);
Aβ139 (SEQ ID NO.: 1); and
APP669711 (SEQ ID NO.: 4),
to obtain measurement levels of:
Aβ142;
Aβ139; and
APP669711,
in the living bodyderived sample;
a calculation step of calculating:
a ratio of Aβ139 level to Aβ142 level: Aβ139/Aβ142; and
a ratio of APP669711 level to Aβ142 level:
APP669711/Aβ142;
a derivation step of deriving a composite variable by a combination of each of the ratios calculated, through a mathematical technique; and
an evaluation step of determining that an amount of cerebral Aβ accumulation of the test subject is larger than an amount of cerebral Aβ accumulation of a person who is negative for cerebral Aβ accumulation, when the composite variable of the test subject is higher than a standard level which is the composite variable of the person who is negative for cerebral Aβ accumulation.
 3. An analytical method for determining efficacy of a medical intervention regarding a cerebral Aβ accumulation state, the method comprising:
conducting examination, each of before and after a medical intervention performed for a test subject, the examination including:
a measurement step of subjecting a living bodyderived sample derived from the test subject to detection of a marker containing:
Aβ142 (SEQ ID NO.: 3);
Aβ139 (SEQ ID NO.: 1); and
APP669711 (SEQ ID NO.: 4),
to obtain measurement levels of:
Aβ142;
Aβ139; and
APP669711,
in the living bodyderived sample;
a calculation step of calculating:
a ratio of Aβ139 level to Aβ142 level: Aβ139/Aβ142; and
a ratio of APP669711 level to Aβ142 level: APP669711/Aβ142; and
a derivation step of deriving a composite variable by a combination of each of the ratios calculated, through a mathematical technique; and
comparing the composite variable of the test subject before the medical intervention and the composite variable of the test subject after the medical
intervention to determine efficacy of the medical intervention regarding a cerebral Aβ accumulation state.
 4. An analytical method for predicting progression of symptoms in future or predicting a risk of development of dementia regarding a cerebral A□ accumulation state, the method comprising:
conducting examination once or a plurality of times over time for a test subject, the examination including:
a measurement step of subjecting a living bodyderived sample derived from the test subject to detection of a marker containing:
Aβ142 (SEQ ID NO.: 3);
Aβ139 (SEQ ID NO.: 1); and
APP669711 (SEQ ID NO.: 4),
to obtain measurement levels of:
Aβ142;
Aβ139; and
APP669711, in the living bodyderived sample;
a calculation step of calculating:
a ratio of Aβ139 level to Aβ142 level: Aβ139/Aβ142;
and
a ratio of APP669711 level to Aβ142 level:
APP669711/Aβ142; and
a derivation step of deriving a composite variable by a combination of each of the ratios calculated, through a mathematical technique; and
predicting progression of symptoms in future or
predicting a risk of development of dementia regarding a cerebral Aβ accumulation state of the test subject based on a value(s) of the composite variable of the test subject in once or a plurality of times conducted over time.
 5. The analytical method according to any one of items 2 to 4, wherein the mathematical technique is a method using discriminant analysis, multiple regression analysis, principal components regression analysis, partial least squares regression, or logistic regression.
 6. The analytical method according to any one of items 2 to 4, wherein the mathematical technique is a method of normalizing each of the aforementioned ratios, and then deriving a mean value or a total value of the at least two ratios normalized.
 7. The analytical method according to any one of items 2 to 6, wherein the living bodyderived sample is selected from the group consisting of blood, cerebrospinal fluid, urine, feces, and body secreting fluid.
[0017] In the present invention, the term "level of marker" basically means a concentration, but may be other units applied correspondingly to concentration by a person skilled in the art. The term "test subject" includes human, and mammals other than human (rat, dog, cat etc.) . In the present invention, the living bodyderived sample is disposed of rather than being returned to the test subject (for example, subject) from which the biological sample is derived. The medical intervention includes administration of a therapeutic drug or a prophylactic drug, dietetic therapy, exercise therapy, learning therapy, surgical operation and the like.
[0018] Described herein is a marker for determining a cerebral Aβ accumulation state, including a combination of at least two ratios selected from the group consisting of:
a ratio of Aβ139 level to Aβ142 level: Aβ139/Aβ142;
a ratio of Aβ140 level to Aβ142 level: Aβ140/Aβ142; and
a ratio of APP669711 level to Aβ142 level: APP669711/Aβ142, in a living bodyderived sample. Also described herein is a method for analysis of the marker.
[0019] By combining at least two ratios selected from the group consisting of three ratios in a living bodyderived sample of a test subject: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142, it is possible to estimate a cerebral Aβ accumulation state with higher accuracy, as compared with the case where each of the three ratios is used singly. The composite variable using a mathematical technique can be obtained by combining at least two ratios selected from the group consisting of Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 with a weighting estimated from a statistical view, or with an equivalent weighting, and a cerebral Aβ accumulation state can be estimated more accurately from each of the ratios.
[0020] The present invention is applicable to detection of not only the advanced stage of Alzheimer's disease in which cerebral Aβ is excessively accumulated and a cognitive functional disorder has appeared, but also a mild cognitive impairment (MCI) which is an early stage of the advanced stage of Alzheimer's disease, and further a preclinical stage of Alzheimer's disease in which cerebral Aβ is excessively accumulated but a cognitive functional disorder has not been appeared.
[0021] According to the present invention, as the living bodyderived sample, not only blood, but also cerebrospinal fluid (CSF), urine, faces, and body secreting fluid (e.g., saliva, tear, sweat, nasal mucosal exudate, and sputum) can be used. Therefore, in the stage where the preventive method and the preemptive therapeutic method for Alzheimer's disease have established, analysis of a cerebral Aβ accumulation state for a person having normal cognitive function in a general medical examination, a complete physical examination and the like is effective for presymptomatic diagnosis of Alzheimer's disease.
[0022] By applying the present invention before and after a medical intervention performed for the test subject, it is possible to evaluate the drug efficacy of a therapeutic drug or a prophylactic drug for Alzheimer's disease, or to evaluate the efficacy of other treatment. Also, the present invention is useful for followup of a patient suffering from Alzheimer's disease.
[0023] By applying the present invention to the test subject once or a plurality of times over time, it is possible to predict progression of symptoms of Alzheimer's disease in the future or to predict a risk of development of dementia.
[0024] The figures show:
[Fig. 1] Fig. 1 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for Aβ139/Aβ142 in Experimental Example 1.
[Fig. 2] Fig. 2 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for Aβ140/Aβ142 in Experimental Example 1.
[Fig. 3] Fig. 3 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for APP669711/Aβ142 in Experimental Example 1.
[Fig. 4] Fig. 4 is an ROC curve for Aβ139/Aβ142 in Experimental Example 1.
[Fig. 5] Fig. 5 is an ROC curve for Aβ140/Aβ142 in Experimental Example 1.
[Fig. 6] Fig. 6 is an ROC curve for APP669711/Aβ142 in Experimental Example 1.
[Fig. 7] Fig. 7 is a scatter diagram of PiB accumulation mean value (mcSUVR) and Aβ139/Aβ142 ratio in Experimental Example 1.
[Fig. 8] Fig. 8 is a scatter diagram of PiB accumulation mean value (mcSUVR) and Aβ140/Aβ142 ratio in Experimental Example 1.
[Fig. 9] Fig. 9 is a scatter diagram of PiB accumulation mean value (mcSUVR) and APP669711/Aβ142 ratio in Experimental Example 1.
[Fig. 10] Fig. 10 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 1.
[Fig. 11] Fig. 11 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 1.
[Fig. 12] Fig. 12 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 1.
[Fig. 13] Fig. 13 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 1.
[Fig. 14] Fig. 14 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 1.
[Fig. 15] Fig. 15 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 1.
[Fig. 16] Fig. 16 is an ROC curve for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 1.
[Fig. 17] Fig. 17 is an ROC curve for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 1.
[Fig. 18] Fig. 18 is a scatter diagram of PiB accumulation mean value (mcSUVR) and discriminant score of a combination of Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 1.
[Fig. 19] Fig. 19 is a scatter diagram of PiB accumulation mean value (mcSUVR) and discriminant score of a combination of Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 1.
[Fig. 20] Fig. 20 is a scatter diagram of PiB accumulation mean value (mcSUVR) and discriminant score of a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 1.
[Fig. 21] Fig. 21 is a scatter diagram of PiB accumulation mean value (mcSUVR) and discriminant score of a combination of Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 1.
[Fig. 22] Fig. 22 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 2.
[Fig. 23] Fig. 23 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 2.
[Fig. 24] Fig. 24 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 2.
[Fig. 25] Fig. 25 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 2.
[Fig. 26] Fig. 26 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 2.
[Fig. 27] Fig. 27 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 2.
[Fig. 28] Fig. 28 is an ROC curve for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 2.
[Fig. 29] Fig. 29 is an ROC curve for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 2.
[Fig. 30] Fig. 30 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 2.
[Fig. 31] Fig. 31 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 2.
[Fig. 32] Fig. 32 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 2.
[Fig. 33] Fig. 33 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 2.
[Fig. 34] Fig. 34 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 3.
[Fig. 35] Fig. 35 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 3.
[Fig. 36] Fig. 36 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 3.
[Fig. 37] Fig. 37 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 3.
[Fig. 38] Fig. 38 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 3.
[Fig. 39] Fig. 39 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 3.
[Fig. 40] Fig. 40 is an ROC curve for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 3.
[Fig. 41] Fig. 41 is an ROC curve for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 3.
[Fig. 42] Fig. 42 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of Aβ139/Aβ142 and Aβ140/Aβ142 in Experimental Example 3.
[Fig. 43] Fig. 43 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of Aβ139/Aβ142 and APP669711/Aβ142 in Experimental Example 3.
[Fig. 44] Fig. 44 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of two markers : Aβ140/Aβ142 and APP669711/Aβ142 in Experimental Example 3.
[Fig. 45] Fig. 45 is a scatter diagram of PiB accumulation mean value (mcSUVR) and zscore of a combination of Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 in Experimental Example 3.
[1. Test subject]
[0025] In the present invention, the test subject includes human, and mammals other than human (rat, dog, cat etc.). Hereinafter, the description will be made mainly for the case of human, but the same applies to mammals other than human.
[0026] In the method of the present invention, the subject may be any individuals including a person expected to be a normal healthy person regardless of past clinical history. For a person expected to be a normal healthy person, a cerebral Aβ accumulation state can be determined in a general medical examination, or a complete physical examination, preferably by a blood test, and the method is particularly effective for early detection/diagnosis of Alzheimer's disease. For a subject suspected to be a candidate for Alzheimer's disease as a result of ADAScog, MMSE, DemTect, SKT, or a test of cognitive function such as a clock drawing test for examining clinical symptom, and confirmation of image findings of magnetic resonance imaging diagnosis (MRI), positron emission tomography (PET) and the like, the method of the present invention can be used as a determination material for diagnosing Alzheimer's disease more accurately from the view point of a fundamental view such as the presence or absence of a cerebral amyloid lesion.
[2. Living bodyderived sample]
[0027] The marker of the present invention can be detected and analyzed in a living bodyderived sample of a subject. Therefore, in the method of the present invention, a level of a marker in a living bodyderived sample of a subject is analyzed.
[0028] The living bodyderived sample can be selected from blood, cerebrospinal fluid (CSF), urine, faces, body secreting fluid (e.g., saliva, tear, sweat, nasal mucosal exudate, and sputum) and the like. Among these, blood is preferred for diagnosis and presymptomatic diagnosis of Alzheimer's disease in a general medical examination, a complete physical examination or the like.
[0029] The blood sample is a sample that is directly subjected to a measurement step of expression level of a marker, and includes whole blood, plasma, serum and the like. The blood sample can be prepared by appropriately treating whole blood collected from a test subject. The treatment performed in the case of preparing a blood sample from collected whole blood is not particularly limited, and any treatment that is clinically acceptable, such as centrifugal separation may be performed. The blood sample subjected to the measurement step may be appropriately stored at a low temperature such as freezing in the intermediate stage of the preparation step or in the post stage of the preparation step. In the present invention, the living bodyderived sample such as a blood sample is disposed of rather than being returned to the subject from which it is derived.
[3. Marker]
[0030] The marker described herein comprises a composite variable using a mathematical technique from at least two ratios selected from the group consisting of:
a ratio of Aβ139 level to Aβ142 level: Aβ139/Aβ142;
a ratio of Aβ140 level to Aβ142 level: Aβ140/Aβ142; and
a ratio of APP669711 level to Aβ142 level: APP669711/Aβ142, in a living bodyderived sample. For the marker including a composite variable using a mathematical technique from these at least two ratios, a significant difference has been observed between the composite variable level in the plasma sample from a person having normal cognitive function who is negative for cerebral Aβ accumulation and the composite variable level in the plasma sample from a subject having excessively accumulated cerebral Aβ.
[0031]
APP672710 (Aβ139) (SEQ ID NO.: 1):
DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGV
APP672711 (Aβ140) (SEQ ID NO.: 2):
DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVV
APP672713 (Aβ142) (SEQ ID NO.: 3):
DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVVIA
APP669711 (SEQ ID NO.: 4):
VKMDAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVV
[0032] Amyloid precursor protein (APP) is a singlepass transmembrane protein and is composed of 770 amino acid residues . Amyloid precursor protein (APP) is proteolyzed by β secretase and γ secretase, and an amyloid β peptide (Aβ) is produced by the proteolysis. APP672713 and Aβ142 indicate the same peptide (SEQ ID NO. : 3). APP672711 and Aβ140 indicate the same peptide (SEQ ID NO.: 2).
[4. Analysis of marker]
[0033] A method for analysis of the marker described herein includes:
a measurement step of subjecting a living bodyderived sample derived from a test subject to detection of a marker containing:
Aβ142 (SEQ ID NO.: 3); and
at least two selected from the group consisting of Aβ139 (SEQ ID NO.: 1), Aβ140 (SEQ ID NO.: 2), and APP669711 (SEQ ID NO.: 4),
to obtain measurement levels of:
Aβ142; and
the at least two selected from the group consisting of Aβ139, Aβ140, and APP669711 in the living bodyderived sample;
a calculation step of calculating at least two ratios selected from the group consisting of:
a ratio of Aβ139 level to Aβ142 level: Aβ139/Aβ142;
a ratio of Aβ140 level to Aβ142 level: Aβ140/Aβ142; and
a ratio of APP669711 level to Aβ142 level: APP669711/Aβ142;
a derivation step of deriving a composite variable by a combination of each of the ratios calculated, through a mathematical technique; and
an evaluation step of determining that an amount of cerebral Aβ accumulation of the test subject is larger than an amount of cerebral Aβ accumulation of a person who is negative for cerebral Aβ accumulation, when the composite variable of the test subject is higher than a standard level which is the composite variable of the person who is negative for cerebral Aβ accumulation. This makes it possible to determine a cerebral Aβ accumulation state, or to use the marker as a determination material.
[0034] The term "level of marker" basically means a concentration, but may be other units applied correspondingly to concentration by a person skilled in the art, for example, a detected ion intensity in the mass spectrometry. Herein, the marker in the living bodyderived sample is analyzed by comparing the composite variable derived from a measurement value (measurement composite variable) and the composite variable derived from a standard value (standard composite variable) . For more accurate analysis, it is preferred that the measurement value and the standard value to be compared are values based on the living bodyderived samples prepared in the same conditions (pretreatment condition, storage condition and the like). As the standard composite variable of the marker, a composite variable that is derived from a measurement value for a person determined as negative for cerebral Aβ accumulation by a PiBPET image can be used. Alternatively, as the standard composite variable of the marker, a standard composite variable established for a normal person who is negative for cerebral Aβ accumulation by a PiBPET image can be used.
[0035] A marker is measured, preferably, by a test based on biological molecule specific affinity. The test based on biological molecule specific affinity is a method well known to a person skilled in the art and is not particularly limited, but is preferably an immunoassay. Specific examples of the immunoassay include competitive and noncompetitive assays such as western blotting, radioimmunoassay, ELISA (EnzymeLinked ImmunoSorbent Assay) (sandwich immunoassay, competitive assay, and direct binding assay are included), immunoprecipitation, precipitation reaction, immunodiffusion, immunoagglutination measurement, complementbinding reaction analysis, immunoradiometric assay, fluorescence immunoassay, and protein A immunoassay. In the immunoassay, an antibody that binds to the marker in a living bodyderived sample is detected.
[0036] In the present invention, the measurement of the marker may be performed by using an immunoglobulin having an antigen binding site capable of recognizing an amyloid precursor protein (APP)derived peptide, or an antibodyimmobilizing carrier prepared by using an immunoglobulin fragment having an antigen binding site capable of recognizing an amyloid precursor protein (APP)derived peptide. By immunoprecipitation using the antibodyimmobilizing carrier, a peptide in the sample can be detected by a mass spectrometer (ImmunoprecipitationMass Spectrometry: IPMS).
[0037] In the present invention, consecutive immunoprecipitation (cIP) may be conducted, and then a peptide in the sample may be detected by a mass spectrometer (cIPMS) . By conducting affinity purification twice consecutively, impurities that cannot be excluded by one affinity purification can be further reduced by the second affinity purification. Therefore, it is possible to prevent the ionization suppression of polypeptide due to impurities, and it becomes possible to measure even a very small amount of polypeptide in a living body sample with high sensitivity by mass spectrometry.
[0038] By combining at least two ratios selected from the group consisting of three ratios in a living bodyderived sample of a test subject: Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142, it is possible to estimate a cerebral Aβ accumulation state with higher accuracy, as compared with the case where each of the three ratios is used singly. The composite variable using a mathematical technique can be obtained by combining at least two ratios selected from the group consisting of Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 with a weighting estimated from a statistical view, or with an equivalent weighting, and a cerebral Aβ accumulation state can be estimated more accurately from each of the ratios.
[0039] As the mathematical technique for combination with a weighting estimated from a statistical view, for example, using at least two ratios selected from the three ratios, a composite variable is calculated by discriminant analysis, multiple regression analysis, principal components regression analysis, partial least square, or logistic regression. This can be a combined composite variable.
[0040] As the mathematical technique for combination with an equal weighting, for example, for at least two ratios selected from the three ratios, the ratios are normalized, and then a mean value or a total value of the at least two ratios normalized is derived, and the derived mean value or total value is given as a composite variable of the at least two ratios. More specifically, for at least two ratios selected from the three ratios, normalization based on all cases of the test subject is conducted, or normalization based on the control group (PiBgroup: a group determined as being negative for cerebral Aβ accumulation by a PiBPET image) is conducted, and then a mean value of the at least two ratios normalized (zscore) is calculated, and this may be a combined composite variable.
[0041] By using such a mathematical technique, even when one ratio of at least two ratios selected from the three ratios is too large or too small as compared with the other one or two ratios, it is possible to combine these ratios with an equal weighting, and by using the composite variable combined, it is possible to estimate the cerebral Aβ accumulation state more accurately from each of the aforementioned ratios.
EXAMPLES
[0042] Hereinafter, the present invention will be described specifically with reference to examples, but is not limited to these examples. In the following, the amount of a matter indicated by % is based on weight when the matter is solid, and based on volume when the matter is liquid unless otherwise indicated.
[Experimental Example 1]
[11. Plasma sample]
[0043] Plasma samples (76 specimens) of cases classified into groups of PiB and PiB+ were prepared at National Center for Geriatrics and Gerontology.
PiB (person determined as negative for cerebral Aβ accumulation by PiBPET image): 50 cases
PiB+ (person determined as positive for cerebral Aβ accumulation by PiBPET image): 26 cases
[0044] In order to determine the positivity or negativity of cerebral Aβ accumulation, PiBPET images of the brains of the subjects were acquired. When the PiB accumulation amount of the cerebral cortex is larger than or equivalent to the nonspecific PiB accumulation amount of the white matter, the subject was determined as positive. When only nonspecific accumulation to the white matter was observed, and little accumulation was observed in the cortex, the subject was determined as negative. The cognitive impairment was determined in conformity with the NIAAA criteria published in 2011.
[0045] Regarding PiB accumulation mean value (mcSUVR: mean cortical Standard Uptake Value Ratio), cortical PiB accumulation was quantified, and an accumulation ratio of cerebrum based on cerebellum was determined. However, in PiB, there were two cases of missing values.
[12. Preparation of antibodyimmobilizing beads]
[0046] Clone 6E10 (available from Covance) of an antiAβ antibody (IgG) recognizing 38 residues of amyloid β protein (Aβ) as an epitope was prepared.
[0047] For 100 µg of an antiAβ antibody (IgG), about 3.3 × 10
^{8} magnetic beads (Dynabeads (registered trademark) M270 Epoxy) were reacted in an immobilizing buffer (0.1 M phosphate buffer containing 1 M ammonium sulfate (pH 7.4)) at 37°C for 16 to 24 hours, to prepare antiAβ IgG immobilizing beads.
[13. Consecutive Immunoprecipitation (cIP)]
(First reaction step)
[0048] Into 250 µL of human plasma, 250 µL of a first IP reaction buffer (0.2% (w/v) DDM, 0.2% (w/v) NTM, 800 mM GlcNAc, 100 mM TrisHCl, 300 mM NaCl, pH 7.4) containing 10 pM stable isotope labeled Aβ138 (SILAβ138) was mixed, and then the mixture was allowed to stand for 5 to 30 minutes on ice. SILAβ138 in which carbon atoms in Phe and Ile are substituted by
^{13}C was used as an internal standard for standardization of signal intensity of a mass spectrum. The plasma was mixed with antiAβ IgG immobilizing beads, and shaken for 1 hour on ice.
(First washing step, First elution step)
[0049] Then, the antibody beads were washed three times with 100 µL of a first IP washing buffer (0.1% DDM, 0.1% NTM, 50 mM TrisHCl (pH 7.4), 150 mM NaCl), and washed twice with 50 µL of a 50 mM ammonium acetate buffer, and then Aβ and Aβlike peptides (namely, APPderived peptides) bound to the antibody beads were eluted with a first IP eluent (50 mM Glycine buffer containing 0.1% DDM (pH 2.8)).
(Neutralization step)
[0050] The obtained eluate was mixed with a second IP reaction buffer (0.2% (w/v) DDM, 800 mM GlcNAc, 300 mM TrisHCl, 300 mM NaCl, pH 7.4), to obtain a first purified solution.
(Second reaction step)
[0051] The obtained first purified solution was mixed with another antiAβ antibody immobilizing beads, and shaken for 1 hour on ice.
(Second washing step, Second elution step)
[0052] Then, the antiAβ antibody immobilizing beads were washed five times with 50 µL of a second washing buffer (0.1% DDM, 50 mM TrisHCl (pH 7.4), 150 mM NaCl), washed twice with 50 µL of a 50 mM ammonium acetate buffer, and washed once with 30 µL of H
_{2}O, and then Aβ and Aβlike peptides (APPderived peptides) bound to the antibody beads were eluted with 5 µL of a second IP eluent (70% (v/v) acetonitrile containing 5 mM hydrochloric acid) . In this manner, a second purified solution was obtained. The second purified solution was subjected to mass spectrometry.
[0053] nDodecylβDmaltoside (DDM) [critical micelle concentration (cmc): 0.009%]
nNοnylβDthiomaltoside (NTM) [cmc: 0.116%]
[14. Detection by MALDITOF MS]
[0054] As a matrix for Linear TOF, αcyano4hydroxycinnamic acid (CHCA) was used. A matrix solution was prepared by dissolving 1 mg of CHCA in 1 mL of 70% (v/v) acetonitrile. As a matrix additive, 0.4% (w/v) methanediphosphonic acid (MDPNA) was used. After mixing equivalent amounts of a 1 mg/mL CHCA solution and 0. 4% (w/v) MDPNA, 0.5 µL of the resultant mixture was dropped on a µFocus MALDI plate™ 900 µm (Hudson Surface Technology, Inc., Fort Lee, NJ) and dried and solidified.
[0055] One µL of the second purified solution obtained by the aforementioned immunoprecipitation was dropped into the matrix on the µFocus MALDI plate™ 900 µm.
[0056] Mass spectrum data was acquired by Linear TOF in a positive ion mode by using AXIMA Performance (Shimadzu/KRATOS, Manchester, UK). For 1 well, 400 spots, or 16,000 shots were integrated. A criterion of a detection limit of a peak was an S/N ratio of 3 or more. An m/z value of Linear TOF was indicted by an average mass of a peak. An m/z value was calibrated by using human angiotensin II and human ACTH fragment 1839, bovine insulin oxidized betachain, and bovine insulin as external standards.
[15. Normalization of peak intensities of Aβ and Aβlike peptides]
[0057] In each mass spectrum, by dividing a signal intensity of each of Aβ and Aβlike peptides (APPderivedpeptide) by a signal intensity of the internal standard peptide (SILAβ138), signal intensities of Aβ and Aβlike peptides were normalized. Thereafter, a mean value of normalized intensity of each APPderived peptide obtained from four mass spectra per one specimen was calculated, and used in a statistical analysis. Among the four normalized intensities used in averaging, a normalized intensity that is out of the range of 0.7 to 1.3 times of the median was regarded as an outlier, and removed in the data processing for averaging. When the number of data of normalized intensity to be used in averaging is less than 3 because the data does not reach the detection lower limit (S/N < 3), or an outlier occurs, the analysis result is "undetectable".
[16. Statistics]
[0058] For comparison between group the PiB and the group PiB+, evaluation using a ttest was conducted. For evaluating the performance of discriminating between PiB and PiB+, an area under the curve (AUC), a sensitivity (Sensitivity), a specificity (Specificity), and an accuracy (Accuracy) were determined using a Receiver Operatorating Characteristic (ROC) curve. Every test was conducted by a twosided test, and P < 0.05 was used as a significant level. Correlation analysis between each marker value and mcSUVR was evaluated with Pearson productmoment correlation coefficient. However, there were two cases of missing values in mcSUVR, so that analysis was conducted for 74 cases.
[17. Comparison between groups for each marker]
[0059] Using a ratio of a normalized intensity of Aβ139, Aβ140, or APP669711 to a normalized intensity of Aβ142 (i.e., Aβ139/Aβ142, Aβ140/Aβ142, APP669711/Aβ142) as a marker, comparison between the group PiB and the group PiB+ was conducted (Figs. 1, 2, 3). Any P value obtained in the ttest satisfied P < 0.001, revealing that the value increased statistically significantly in PiB+ as compared with PiB.
[0060] Fig. 1 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for Aβ139/Aβ142. Likewise, Fig. 2 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for Aβ140/Aβ142. Fig. 3 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for APP669711/Aβ142. These are results for a single marker.
[0061] In each boxandwhisker plot, the range indicated by the box in each group represents the intensity ratio contribution range (quartile range) of the samples whose intensity ratio is rated between 25 to 75% of all specimens, and the horizontal lines shown above and below the box respectively indicate the maximum value and the minimum value of the samples within the range from the upper end and the lower end of the box to 1.5 times the quartile range, and the horizontal bar in the box indicates the median of the intensity ratio. The same applies in each boxandwhisker plot below.
[18. ROC analysis for each marker]
[0062] For evaluating the determination performance of Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142, ROC analysis of the group PiB+ versus the group PiB was conducted with PiB+ as positive (Figs. 4, 5, 6) . As a result, Aβ139/Aβ142 showed the highest AUC=0.898, APP669711/Aβ142 showed the second highest AUC=0.894, and Aβ140/Aβ142 showed AUC=0.828. The AUC was 0.8 or more for any marker, and this reveals that the markers are capable of discrimination between the group PiB and the group PiB+ with high accuracy.
[0063] Fig. 4 is an ROC curve for Aβ139/Aβ142. Likewise, Fig. 5 is an ROC curve for Aβ140/Aβ142. Fig. 6 is an ROC curve for APP669711/Aβ142. These are results for a single marker.
[0064] In an ROC curve of each marker, the value showing the highest "sensitivity+specificity1" was set as a cutoff value. The set cutoff values, and Sensitivity, Specificity, and Accuracy at each cutoff value are shown in Table 1. Aβ139/Aβ142 showed the highest Accuracy=0.855. In Table 1, Numbers 1 to 3 show analysis of a single marker.
[Table 1]
Single Marker  AUC  Correlation coefficient (r)  Cutoff  Sensitivity  Specificity  Accuracy 
No.  Aβ139 /Aβ142  Aβ140 /Aβ142  APP669711 /Aβ142 
1 
X 


.898 
0.630 
1.311 
0.923 
0.820 
0.855 
2 

X 

0.828 
0.477 
17.587 
0.808 
0.760 
0.789 
3 


X 
0.894 
0.489 
0.684 
0.923 
0.740 
0.803 
[19. Correlation analysis between each marker value and mcSUVR]
[0065] In order to investigate whether Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 reflect a cerebral amyloid accumulation amount, correlation between each index and mcSUVR was analyzed (Figs. 7, 8, 9, Table 1). As a result, in all the three markers, the Pearson productmoment correlation coefficient (r) was 0.4 or more, and existence of correlation was proved, and in particular, the correlation coefficient (r) of Aβ139/Aβ142 showed the strongest correlation of 0.630.
[0066] This indicates that Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 have the potential to be an index that is useful for determination of a cerebral amyloid accumulation state. In the present analysis for the 76 cases, Aβ139/Aβ142 showed the most excellent determination performance.
[0067] Fig. 7 is a scatter diagram for Aβ139/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents Aβ139/Aβ142 ratio. Likewise, Fig. 8 is a scatter diagram for Aβ140/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents Aβ140/Aβ142 ratio. Fig. 9 is a scatter diagram for APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents APP669711/Aβ142 ratio. These are results for a single marker. In the diagram, "○" indicates the group PiB , and "●" indicates the group PiB+. The same applies to the following scatter diagrams.
[110. Method for combining markers using discriminant analysis]
[0068] Discriminant analysis was conducted by using combinations of two markers, Aβ139/Aβ142 and Aβ140/Aβ142, Aβ139/Aβ142 and APP669711/Aβ142, and Aβ140/Aβ142 and APP669711/Aβ142, and a combination of three markers Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142, and discriminant function z in each combination was obtained. A discriminant score was calculated from marker values to be combined by using discriminant function z.
[0069] A discriminant score was calculated from marker values to be combined by using discriminant function z, and then subsequent statistical analysis was conducted.
[111. Comparison between groups by discriminant score]
[0070] For discriminant scores obtained by combinations of markers, comparison between the groups PiB and PiB+ was conducted (Figs. 10, 11, 12, 13). Any P value obtained in the ttest satisfied P < 0.001, and it was confirmed that even in the case of combination, the value increased statistically significantly in the group PiB+ compared with the group PiB.
[0071] Fig. 10 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 11 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142. Fig. 12 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ140/ Aβ142 and APP669711/Aβ142. Fig. 13 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[112. ROC analysis by discriminant score]
[0072] For evaluating the determination performance of discriminant scores, ROC analysis of the group PiB+ versus the group PiB was conducted with PiB+ as positive (Figs. 14, 15, 16, 17) . As a result, combinations of two makers, Aβ139/Aβ142 and Aβ140/Aβ142 (Fig. 14), Aβ139/Aβ142 and APP669711/Aβ142 (Fig. 15), Aβ140/Aβ142 and APP669711/Aβ142 (Fig. 16), and a combination of three markers (Fig. 17) each showed an AUC of 0.9 or more, which is higher than that by a single marker (Figs. 14, 15, 16, 17, Table 2) . That is, by combining markers, the discrimination performance improved and discrimination between PiB and PiB+ with very high accuracy was enabled. In the present ROC analysis, the combination of Aβ139/Aβ142 and APP669711/Aβ142 showed the highest AUC.
[0073] Fig. 14 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 15 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142. Fig. 16 is an ROC curve for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142. Fig. 17 is an ROC curve for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[0074] The discriminant score > 0 is discriminated as PiB+, and the discriminant score < 0 is discriminated as PiB. That is, Sensitivity, Specificity, and Accuracy in each combination when the cutoff value is set at 0 are shown in Table 2 (Numbers: 11, 12, 13, 14). All the combinations showed Accuracy that is higher than that by a single marker. That is, by combining markers, the probability of exact determination was improved. In the present analysis, the combination of three markers showed the highest Accuracy. In Table 2, Numbers 11 to 14 show the results obtained by conducting discriminant analysis by using each marker value, and using a discriminant score of combined markers for analysis.
[Table 2]
Combination of Markers  AUC  Correlation coefficient (r)  Cutoff  Sensitivity  Specificity  Accuracy 
No.  Aβ139 /Aβ142  Aβ140 /Aβ142  APP669711 /Aβ142 
11 
X 
X 

0.909 
0.634 
0 
0.769 
0.920 
0.868 
12 
X 

X 
0.966 
0.653 
0 
0.923 
0.880 
0.895 
13 

X 
X 
0.923 
0.581 
0 
0.846 
0.880 
0.868 
14 
X 
X 
X 
0.964 
0.656 
0 
0.885 
0.900 
0.895 
[113 . Correlation analysis with mcSUVR by discriminant score]
[0075] In order to investigate whether discriminant scores of combined markers reflect a cerebral amyloid accumulation amount, correlation between each discriminant score and mcSUVR was analyzed. Combinations of two markers, Aβ139/Aβ142 and Aβ140/Aβ142, Aβ139/Aβ142 and APP669711/Aβ142, and Aβ140/Aβ142 and APP669711/Aβ142, as well as a combination of three markers each showed improved Pearson productmoment correlation coefficient (r) as compared with that by a single marker, and reflected the PiB accumulation degree more favorably (Figs. 18, 19, 20, 21, Table 2). In the present correlation analysis, the combination of three markers showed the highest correlation coefficient.
[0076] Fig. 18 is a scatter diagram for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents discriminant score of a combination of Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 19 is a scatter diagram for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents discriminant score of a combination of Aβ139/Aβ142 and APP669711/Aβ142. Fig. 20 is a scatter diagram for a combination of two markers: Aβ140/ Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents discriminant score of a combination of two markers: Aβ140/ Aβ142 and APP669711/Aβ142. Fig. 21 is a scatter diagram for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents discriminant score of a combination of Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[Experimental Example 2]
[21. Method for combining markers using normalization based on distribution of all specimens]
[0077] In combining respective values of markers, the combined value is greatly influenced by a marker having a larger value if the values are directly averaged or summed. For combining the markers equally, first, normalization was conducted for each of Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142 based on distribution of all the 76 cases. Normalization was conducted by calculating mean value (X) and standard deviation (S) of marker values of all the 76 cases, and converting marker value (x
_{i}) of each sample into zscore (zi) according to the following formula.
[0078] After averaging zscores of markers to be combined, the following statistical analysis was conducted.
[0079] Combinations of two markers, Aβ139/Aβ142 and Aβ140/Aβ142, Aβ139/Aβ142 and APP669711/Aβ142, and Aβ140/Aβ142 and APP669711/Aβ142, as well as a combination of three markers, Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142 were implemented.
[22. Comparison between groups by combination of markers]
[0080] For zscore of each combination of markers, comparison between the groups PiB and PiB+ was conducted (Figs. 22, 23, 24, 25) . Any P value obtained in the ttest satisfied P < 0.001, and it was confirmed that even in the case of combination, the value increased statistically significantly in the group PiB+ compared with the group PiB.
[0081] Fig. 22 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 23 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142. Fig. 24 is a boxandwhisker plot showing comparison between the group PiBand the group PiB+ for a combination of two markers: Aβ140/ Aβ142 and APP669711/Aβ142. Fig. 25 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[23. ROC analysis by combination of markers]
[0082] For evaluating the determination performance of combined markers, ROC analysis of the group PiB+ versus the group PiBwas conducted with PiB+ as positive (Figs. 26, 27, 28, 29). As a result, combinations of two makers, Aβ139/Aβ142 and APP669711/Aβ142 (Fig. 27), Aβ140/Aβ142 and APP669711/Aβ142 (Fig. 28), and a combination of three markers (Fig. 29) showed an AUC of 0.9 or more, which is higher than that by a single marker (Figs. 27, 28, 29, Table 3). That is, by combining markers, the discrimination performance improved and discrimination between PiB and PiB+ with very high accuracy was enabled. In the present ROC analysis, the combination of Aβ139/Aβ142 and APP669711/Aβ142 showed the highest AUC.
[0083] Fig. 26 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 27 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142. Fig. 28 is an ROC curve for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142. Fig. 29 is an ROC curve for a combination of three markers : Aβ139/Aβ142 , Aβ140/Aβ142 and APP669711/Aβ142.
[0084] In an ROC curve of each combination, the value showing the highest "sensitivity+specificity1" was set as a cutoff value. The set cutoff values, and Sensitivity, Specificity, and Accuracy at each cutoff value are shown in Table 3 (Numbers : 21, 22, 23, 24). All the combinations showed Accuracy that is higher than that by a single marker. That is, by combining markers, the probability of exact determination was improved. In the present analysis, the combination of three markers showed the highest Accuracy. In Table 3, Numbers 21 to 24 show the results obtained by normalizing each marker value based on distribution of all the specimens, and then using an averaged value of markers to be combined for analysis.
[Table 3]
Combination of Markers  AUC  Correlation coefficient (r)  Cutoff  Sensitivity  Specificity  Accuracy 
No.  Aβ139 /Aβ142  Aβ140 /Aβ142  APP669711 /Aβ142 
21 
X 
X 

0.890 
0.607 
0.168 
0.808 
0.900 
0.868 
22 
X 

X 
0.965 
0.649 
0.223 
0.962 
0.880 
0.908 
23 

X 
X 
0.912 
0.585 
0.453 
0.846 
0.940 
0.908 
24 
X 
X 
X 
0.945 
0.650 
0.434 
0.808 
0.980 
0.921 
[24. Correlation analysis with mcSUVR by combination of markers]
[0085] In order to investigate whether zscores of combined markers reflect a cerebral amyloid accumulation amount, correlation between each zscore and mcSUVR was analyzed. Combinations of two markers, Aβ139/Aβ142 and APP669711/Aβ142, and Aβ140/Aβ142 and APP669711/Aβ142, as well as a combination of three markers each showed improved Pearson productmoment correlation coefficient (r) as compared with that by a single marker, and reflected the PiB accumulation degree more favorably (Figs. 30, 31, 32, 33, Table 3). In the present correlation analysis, the combination of three markers showed the highest correlation coefficient.
[0086] Fig. 30 is a scatter diagram for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 31 is a scatter diagram for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of Aβ139/Aβ142 and APP669711/Aβ142. Fig. 32 is a scatter diagram for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of Aβ140/Aβ142 and APP669711/Aβ142. Fig. 33 is a scatter diagram for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[Experimental Example 3]
[31. Method for combining markers using normalization based on distribution of control group]
[0087] In Experimental Example 2, for combining each marker value, normalization was conducted based on distribution of 76 cases including all of the group PiB and the group PiB+. In this Experimental Example 3, normalization based on distribution of 50 cases of the group PiB as a control was conducted, and evaluation of combination was conducted. Normalization was conducted by calculating mean value (X) and standard deviation (S) of marker values of 50 cases of the group PiB, and converting each marker value (x
_{i}) into zscore (z
_{i}) according to the following formula.
[0088] After averaging zscores of markers to be combined, the following statistical analysis was conducted.
[0089] As is the case with Experimental Example 2, combinations of two markers, Aβ139/Aβ142 and Aβ140/Aβ142, Aβ139/Aβ142 and APP669711/Aβ142, and Aβ140/ Aβ142 and APP669711/Aβ142, as well as a combination of three markers, Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142 were implemented.
[32. Comparison between groups by combination of markers]
[0090] For zscore of each combination of markers, comparison between the groups PiB and PiB+ was conducted (Figs. 34, 35, 36, 37) . Any P value obtained in the ttest satisfied P < 0.001. Statistically significant increase in PiB+ compared with PiBwas observed even when normalization based on the group PiBwas used.
[0091] Fig. 34 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 35 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142. Fig. 36 is a boxandwhisker plot showing comparison between the group PiBand the group PiB+ for a combination of two markers: Aβ140/ Aβ142 and APP669711/Aβ142. Fig. 37 is a boxandwhisker plot showing comparison between the group PiB and the group PiB+ for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[33. ROC analysis by combination of markers]
[0092] For evaluating the determination performance of combined markers, ROC analysis of the group PiB+ versus the group PiBwas conducted with PiB+ as positive (Figs. 38, 39, 40, 41). As a result, combinations of two makers, Aβ139/Aβ142 and APP669711/Aβ142 (Fig. 39), and Aβ140/Aβ142 and APP669711/Aβ142 (Fig. 40), as well as a combination of three markers (Fig. 41) showed an AUC of 0.9 or more, which is higher than that by a single marker (Figs. 39, 40, 41, Table 4) . That is, even when normalization based on the group PiB was used, the discrimination performance improved and discrimination between PiB and PiB+ with very high accuracy was enabled by combining markers. Also in the present ROC analysis, the combination of Aβ139/Aβ142 and APP669711/Aβ142 showed the highest AUC.
[0093] Fig. 38 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 39 is an ROC curve for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142. Fig. 40 is an ROC curve for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142. Fig. 41 is an ROC curve for a combination of three markers : Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[0094] In an ROC curve of each combination, the value showing the highest "sensitivity+specificity1" was set as a cutoff value. The set cutoff values, and Sensitivity, Specificity, and Accuracy at each cutoff value are shown in Table 4. All the combinations showed Accuracy that is higher than that by a single marker. That is, by combining markers, the probability of exact determination was improved. In the present ROC analysis, the combination of Aβ139/Aβ142 and APP669711/Aβ142 showed the highest Accuracy. In Table 4, Numbers 31 to 34 show the results obtained by normalizing each marker value based on distribution of 50 cases of the group PiB, and then using an averaged value of markers to be combined for analysis.
[Table 4]
Combination of Markers  AUC  Correlation coefficient (r)  Cutoff  Sensitivity  Specificity  Accuracy 
No.  Aβ139 /Aβ142  Aβ140 /Aβ142  APP669711 /Aβ142 
31 
X 
X 

0.898 
0.612 
0.781 
0.846 
0.880 
0.868 
32 
X 

X 
0.966 
0.655 
1.194 
0.885 
0.940 
0.921 
33 

X 
X 
0.912 
0.585 
1.154 
0.846 
0.940 
0.908 
34 
X 
X 
X 
0.947 
0.653 
0.748 
0.923 
0.880 
0.895 
[34. Correlation analysis with mcSUVR by combination of markers]
[0095] In order to investigate whether zscores of combined markers reflect a cerebral amyloid accumulation amount, correlation between each zscore and mcSUVR was analyzed. Combinations of two markers, Aβ139/Aβ142 and APP669711/Aβ142, and Aβ140/Aβ142 and APP669711/Aβ142, as well as a combination of three markers each showed improved Pearson productmoment correlation coefficient (r) as compared with that by a single marker (Figs. 42, 43, 44, 45, Table 4) . That is, even when normalization based on the group PiB was used, combinations resulted in reflection of PiB accumulation degree more favorably. In the present correlation analysis, the combination of Aβ139/Aβ142 and APP669711/Aβ142 showed the highest correlation coefficient.
[0096] Fig. 42 is a scatter diagram for a combination of two markers: Aβ139/Aβ142 and Aβ140/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of Aβ139/Aβ142 and Aβ140/Aβ142. Likewise, Fig. 43 is a scatter diagram for a combination of two markers: Aβ139/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of Aβ139/Aβ142 and APP669711/Aβ142. Fig. 44 is a scatter diagram for a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of two markers: Aβ140/Aβ142 and APP669711/Aβ142. Fig. 45 is a scatter diagram for a combination of three markers: Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142, in which the horizontal axis represents PiB accumulation mean value (mcSUVR) and the vertical axis represents zscore of a combination of Aβ139/Aβ142, Aβ140/Aβ142 and APP669711/Aβ142.
[0097] These analytical results revealed that the accuracy of discrimination between PiB and PiB+ is improved and correlation with PiB accumulation degree is enhanced by combining markers rather than using the markers singly. That is, it is shown that by combining blood markers, Aβ139/Aβ142, Aβ140/Aβ142, and APP669711/Aβ142, these markers have the effects of complementarily detecting cerebral amyloid accumulation with higher accuracy. Regarding the combining method, the effect by the combination of markers was observed both by the discriminant analysis using each marker and the method of normalizing each marker value and then combining the markers . Regarding the method for normalizing each marker value, it was shown that both normalization based on all the specimens and normalization based on the group PiB have the effect by combination of markers.
[0098] The results demonstrated above indicate that the combined marker of the present invention is useful as a blood marker for determining a cerebral Aβ accumulation state. This also has indicated the applicability to assist diagnosis of Alzheimer's disease and to presymptomatically diagnose Alzheimer's disease.
SEQUENCE LISTING
[0099]
<110> SHIMADZU CORPORATION
NATIONAL CENTER FOR GERIATRICS AND GERONTOLOGY
<120> Multiplex biomarker for evaluating accumulation state of amyloid
\203À in brain and Analytical method for same
<130> G116251WO
<150> JP 2015183372
<151> 20150916
<160> 4
<210> 1
<211> 39
<212> PRT
<213> Homo Sapiens
<400> 1
<210> 2
<211> 40
<212> PRT
<213> Homo Sapiens
<400> 2
<210> 3
<211> 42
<212> PRT
<213> Homo Sapiens
<400> 3
<210> 4
<211> 43
<212> PRT
<213> Homo Sapiens
<400> 4