FILING DATA

[0001] This application is associated with and claims priority from Australian Provisional Patent Application No. 2014901970, filed on 26 May 2014, entitled "Biomarkers", the entire contents of which, are incorporated herein by reference.

BACKGROUND

FIELD

[0002] The present disclosure relates to nucleic acid biomarkers, the level of expression of which, is instructive as to Alzheimer's disease (AD) including identifying onset and various stages of AD, its various clinical manifestations and amyloidosis-related conditions.

DESCRIPTION OF RELATED ART

[0003] Bibliographic details of the publications referred to by author in this specification are collected alphabetically at the end of the description.

[0004] Reference to any prior art in this specification is not, and should not be taken as, an acknowledgment or any form of suggestion that this prior art forms part of the common general knowledge in any country.

[0005] Alzheimer's disease (AD) is a progressive degenerative disorder that causes loss of memory, cognitive disturbances and behavioral changes. It is a debilitating condition leading to significant rates of morbidity and mortality. It is the fourth largest cause of death in the United States and affects five percent of people over age 65 and 20 percent of people over age 80. To date, there has been no established treatment which will prevent the onset or delay the progression of AD although there is substantial research ongoing worldwide towards identifying potential treatments, including a number of drug trials.

[0006] This form of neurodegeneration is characterized by the formation of intracellular neurofibrillary tangles, neuronal and synaptic loss and the accumulation of beta amyloid (Αβ) into amyloid plaque material (Breteler et al. (1992) Epidemiol Rev 14: 59-82). Αβ is proteolytically cleaved from the amyloid precursor protein (APP) [Cole and Vassar (2008) J Biol Chem 253:29621-29625]. The most characteristic neuropathological feature of AD is the deposition of Αβ into plaques in the brain parenchyma and cerebral blood vessels leading to neuronal loss and cerebral atrophy (Terry et al. (1981) Ann. Neurol 70: 184-192).

[0007] Positron emission tomography (PET) [Cole and Vassar (2008) supra] scanning using a carbon- 11 -labeled Pittscurgh compound B ( ^u C-PiB) to observe Αβ burden of the brain has shown that Αβ deposition is a slow pathological process which occurs during a period of 17-30 years (Villemagne et al. (2013) The Lancet Neurology 72:357-367). Thus, the asymptomatic or presymptomatic nature of pre-clinical dementia presents a challenge for early diagnosis and stratification of AD patients. Diagnosis of AD has made some progress in the recent decade by recruiting large flagship studies which has produced promising diagnostic methods such as PET imaging (Villemagne et al. (2013) supra) and CSF biomarkers (Hansson et al. (2006) Lancet Neurology 5:228-234) to detect Αβ burden. Although these methods have shown high diagnostic accuracy for AD, the high costs and invasive nature of PET imaging and CSF collection are not ideal for routine clinical testing.

[0008] MicroRNAs (miRNAs) are non-coding RNA species of approximately 22 nucleotides in length which are transcribed in all tissues and cells (Krol et al. (2010) Nat Rev Genet 77:597-610) and which bind to complementary sites at the 3' untranslated region (3'UTR) of their corresponding mRNA targets resulting in down-regulation of gene expression (Hannon (2004) Nat Rev Genet 5:522-531). miRNAs can be released into the extracellular environment by binding to RNA-binding proteins or through secretion in cell- derived plasma microvesicles such as exosomes (Valadi et al. (2007) Nat Cell Biol 9:654- 659; Mitchell et al. (2008) Proceedings of the National Academy of Sciences 705: 10513- 10518; Bellingham et al. (2012) Front Physiol 3: 124). The presence of miRNA can, therefore, reflect the physiological state of the biological system. Recent developments in high throughout next-generation sequencing (NGS) have allowed the ability to profile miRNA in biological fluids (Cheng et al. (2013) Front Genet 4: 150). Profiles of deregulated miRNA isolated from plasma and serum (Mitchell et al. (2008) supra; Skog et al. (2008) Nat Cell Biol 70: 1470-1476) suggest they have diagnostic potential for human disease.

[0009] A non-invasive and high throughput biological fluid-based test is required for improved population-based screening and patient care in order to refer patients for further examination. Furthermore, with numerous drug trials aiming to treat AD, a fluid-based test is needed to enrich cohorts of AD cases followed by monitoring the potential benefits and side-effects of therapeutic drugs.

SUMMARY

[0010] Nucleotide sequences are referred to by a sequence identifier number (SEQ ID NO). The SEQ ID NOs correspond numerically to the sequence identifiers <400>1 (SEQ ID NO: l), <400>2 (SEQ ID NO: 2), etc. A summary of the sequence identifiers is provided in Table 1. A sequence listing is provided after the claims.

[0011] The microRNAs (miRNAs) defined by miR nomenclature are listed in Table 1 with reference to a sequence identifier number.

[0012] Abbreviations used herein are defined in Table 2

[0013] Enabled herein is a method for the stratification of human subjects with respect to AD and its various clinical manifestations and stages including the early diagnosis of asymptomatic or presymptomatic individuals. The method disclosed herein is also applicable to other amyloidosis-related conditions. The stratification is based on an association between deregulated miRNA expression and AD. Deregulated miRNAs can be detected in a range of biological samples such as samples comprising or enriched for membranous microvesicles, blood, serum, plasma, urine, lymph, cerebrospinal fluid, ascites, saliva, mucus, stool, biopsy specimens, breast milk, gastric juice, pleural fluid, semen, sweat, tears, hair, vaginal secretion and interstitial fluid. An example of a membranous microvesicle is an exosome. Such structures may first require chemical or mechanical disruption to release the miRNAs or the miRNAs may be free circulating miRNA. The association between miRNA expression and AD is validated through unbiased next generation sequencing (NGS) followed by qRT-PCR. An miRNA signature is selected from an NGS training set in the form of a panel which is used to predict future or potential cognitive decline, early onset AD, the various stages of AD as well as other amyloidosis-related conditions.

[0014] The miRNA expression profile is the level of expression of one or more miRNAs or 2 or more miRNA comprising a nucleotide sequence selected from the listing consisting of SEQ ID NOs: 1 through 16 or a precursor form of each of SEQ ID NOs: l through 16 compared to a first knowledge base generated from training data. The training data represent the correlation of expression levels of the miRNAs with subjects of known status with respect to AD. The first knowledge base, for example, may be miRNA expression levels in healthy individuals (healthy or normal controls). Alternatively, the first knowledge base may comprise levels of expression of the miRNAs in unhealthy controls. Hence, elevation or reduction in expression is dependent on the type of control. In an embodiment, the first knowledge base enables determination of ACt levels of a particular miRNAs obtained using qPCR and then ACt cut-off values established which correlate with the AD condition. Expression fold changes and levels of expression can also be measured of individual miRNAs and/or ratios determined of ACt levels, fold levels of expression between any 2 or more of SEQ ID NOs: l through 17 The assay may or may not require control sample to be run side-by-side.

[0015] Hence, enabled herein is a diagnostic rule based on the application of a comparison of levels of miRNA expression in a control sample or based on predetermined ACt cut-off values. In another embodiment, the diagnostic rule is based on application of statistical and machine learning algorithms. Such an algorithm uses the relationships between miRNA expression and disease status observed in training data (with known disease status) to infer relationships which are then used to predict the status of patients with unknown status (test data). Practitioners skilled in the art of data analysis recognize that many different forms of inferring relationships in the training data may be used without materially changing the scope of the present invention.

[0016] Enabled herein is an assay to stratify a human subject with respect to Alzheimer's disease (AD) or a symptom associated with AD or other amyloidosis-related condition, the method comprising screening a biological sample from the subject for an miRNA comprising a nucleotide sequence selected from the list consisting of SEQ ID NOs: l through 17 or a precursor form of each of SEQ ID NOs: l through 17 or their corresponding cDNA forms or their corresponding cDNA forms wherein the level of expression of a selected miRNA compared to a control is indicative of a stage in the development of AD. By "control" includes a ACt cut-off value based on known ACt levels of particular miRNAs obtained using qPCR. Ratios of expression or fold changes or ACt levels between 2 or more of SEQ ID NOs: 1 through 17 may also be determined.

[0017] In an embodiment, SEQ ID NO: 17 is excluded. Accordingly, the level of expression in the first knowledge base is based on levels in healthy individuals. In this aspect, the expression of one or more miRNAs comprising nucleotide sequences selected from SEQ ID NOs: l through 13 is elevated relative to the healthy control and the expression of one or more miRNAs comprising nucleotide sequences selected from SEQ ID NOs: 14 through 16 is reduced relative to a healthy control in a subject stratified as having a stage of AD or is at risk of developing AD. Reference to any of SEQ ID NOs: l through 16 as well as SEQ ID NO: 17 includes precursor forms thereof. In an embodiment, the biological sample comprises membranous microvesicles or a sample enriched for membranous microvesicles, blood, serum, plasma, urine, lymph, cerebrospinal fluid, ascites, saliva, mucus, stool, biopsy specimens, breast milk, gastric juice, pleural fluid, semen, sweat, tears, hair, vaginal secretion or interstitial fluid. Free circulating miRNAs may also be measured.

[0018] An example of a membranous microvesicle is an exosome. Microvesicles and exosomes may first require chemical or mechanical disruption to release the miRNAs. The miRNAs may also be captured on a solid support such as a bead or the side of a reaction vessel. Hence, the sample may comprise or be enriched for membrane microvesicles or miRNAs released from exosomes or membrane vesicles. A non-microvesicle sample of free miRNA is also contemplated herein.

[0019] The ability to recognise asymptomatic or presymptomatic AD individuals enables early medical and behavioral intervention thereby improving quality of life outcomes.

[0020] Hence, taught herein is a cost effective, low risk, minimally invasive test to detect onset of and to monitor stages of AD as well as other amyloidosis-related conditions. The assay can also be used to monitor treatment protocols. Importantly, the miRNA profile enables diagnosis of AD in subjects who present as healthy individuals yet have amyloid burden and/or other risk factors.

Table 1

Summary of sequence identifiers

[0021] The present invention extends to the detection of precursor forms of the miRNAs which comprise the sequences of any one of SEQ ID NOs: 1 through 17. Table 2

Abbreviations

Ahhi eMiilioii Defin ition

Delta Ct

AD Alzheimer's disease

ΑροΕε4 Apolipoprotein ε4

AUC Area under receiver operating curve

CDR Clinical dementia ratings

FDR False discovery rate

HC Healthy control

MCI Mild cognitive impairment

miRNA Micro-RNA

MMSE Mini-mental state examination

NGS Next generation sequences

PAM Partitioning around meloids

PET Positron emission tomography

ROC Receiver operating characteristic

SUVR Standard uptake value ratio

BRIEF DESCRIPTION OF THE FIGURES

[0022] Some figures contain color representations or entities. Color photographs are available from the Patentee upon request or from an appropriate Patent Office. A fee may be imposed if obtained from a Patent Office.

[0023] Figure 1 is a diagrammatic representation showing hierarchical clustering of differentially expressed exosomal miRNA biomarkers obtained from healthy controls (HC) and patients with Alzheimer's disease (AD). Using the deep sequencing data obtained from the discovery set, hierarchical clustering is performed using Partek Genomics Suite on significantly differentially expressed miRNA using Euclidean average H05, p (AD, MCI and HC) < 0 05 and ± 1 -2 fold change); There are two major nodes of the dendrogram. Node 1 contains 15 miRNA which are found to be up-regulated (hsa-miR-361-5p, hsa- miR-30e-5p, hsa-miR-93-5p, hsa-miR-15a-5p, hsa-miR-143-3p, hsa-miR-335-5p, hsa- miR-106b-5p, hsa-miR-101-3p, hsa-miR-424-5p, hsa-miR-106a-5p, hsa-miR-18b-5p, hsa- miR-3065-5p, hsa-miR-20a-5p, hsa-miR-3065-5p and hsa-miR-582-5p). Node 2 contains 3 miRNA which are found to be down-regulated (hsa-miR-1306-5p, hsa-miR-342-3p, and 15b-3p, Table 3). Patient samples are arranged by attributes. HC, healthy control; MCI, mild cognitive impairment; and AD, Alzheimer's disease. The miRNAs are defined in Table 1 with reference to the sequence listing.

[0024] Figures 2A through D are graphical representations of clustering and random forest testing of correlated miRNA identified in healthy controls and patients with Alzheimer's disease within the discovery set. A) PAM clustering of the 16 miRNA used for validation. Two main clusters are found. One cluster included miRNAs up-regulated in AD, and one for those down-regulated in AD. These 2 main clusters are broken into 2 further cluster search generating 4 clustered miRNA groups. B) miRNA in clusters 1 and 3 showed increased mean expression across ΑροΕε4 allele and clinical classification. Cluster 2 and 4 showing decrease mean expression across ΑΡοΕε4 allele and clinical classification. C) Box plots of miRNA clusters displaying associations with ΑροΕε4 allele status and clinical classification. D) Variable selection via Random Forest analyses ordered by importance of contribution towards clinical classification. The miRNAs found in each cluster are listed in Table 4.

[0025] Figure 3 is a graphical representation of box plots showing validated miRNA differentially expressed in healthy controls (HC) and patients with Alzheimer's disease (AD). Mean centred and scaled data are plotted between HC and AD patients. HC patients are represented by blue circular dots while AD patients are represented by red triangular dots.

DETAILED DESCRIPTION

[0026] Throughout this specification, unless the context requires otherwise, the word "comprise" or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated element or integer or method step or group of elements or integers or method steps but not the exclusion of any element or integer or method step or group of elements or integers or method steps.

[0027] As used in the subject specification, the singular forms "a", "an" and "the" include plural aspects unless the context clearly dictates otherwise. Thus, for example, reference to "a biomarker" includes a single biomarker, as well as two or more biomarkers; reference to "an miRNA" includes a single miRNA, as well as two or more miRNAs; reference to "the disclosure" includes single and multiple aspects taught by the disclosure; and so forth. Aspects taught and enabled herein are encompassed by the term "invention". All such aspects are enabled within the width of the present invention. All variants referred to herein are included by the term "forms" of the invention.

[0028] The use of numerical values in this specification, unless expressly indicated otherwise, are stated as approximations as though the values are preceded by the word "about". Furthermore, the manner of defining a control, and the selection of miRNA may result in different outcomes. Notwithstanding, the skilled person would compensate for these parameters without departing from the essence or scope of the present invention. Also, the disclosure of any ranges is intended as a continuous range including every value between minimum and maximum values. In addition, the subject protocol extends to ratios of expression levels of two or more miRNA biomarkers providing a numerical value associated with a level of likelihood of an individual having AD or is at risk of developing AD or its symptoms such as cognitive impairment.

[0029] The present disclosure teaches the identification of microRNAs (miRNAs) or precursor forms thereof, the expression of which, is statistically associated with AD, a stage of AD including asymptomatic or presymptomatic AD and/or a risk of developing AD. AD manifests itself as a spectrum of symptoms from asymptomatic features to mild cognitive impairment to severe cognitive decline and other symptoms or outcomes of major neurodegeneration. The ability to identify asymptomatic or presymptomatic or early onset AD subjects enables early clinical and behavioral intervention leading to significant improvement in quality of life for the individual and surrounding social and family networks. The method described herein can also be used to detect other amyloidosis- related conditions.

[0030] Hence, enabled herein is a diagnostic rule based on the application of a comparison of levels of miRNA expression in a control sample or based on a ACt cut-off value selected from a set of predetermined ACt levels obtained using qPCR and which are associated with a healthy individual with or at risk of developing AD or other amyloidosis-related condition. In another embodiment, the diagnostic rule is based on application of statistical and machine learning algorithms. Such an algorithm uses the relationships between miRNA expression and disease status observed in training data (with known disease status) to infer relationships which are then used to predict the status of patients with unknown status (test data). Practitioners skilled in the art of data analysis recognize that many different forms of inferring relationships in the training data may be used without materially changing the scope or essence of the present invention.

[0031] The instant disclosure teaches an assay to stratify a human subject with respect to AD or a symptom associated with AD or other amyloidosis-related condition, the method comprising screening a biological sample from the subject for one or more miRNA comprising a nucleotide sequence selected from the list consisting of SEQ ID NO: l through 16 or precursor forms thereof or their corresponding cDNA forms wherein the level of expression of the one or more miRNAs is indicative of a stage of AD or of the amyloidosis-related condition. For the sake of brevity, reference to AD also includes other amyloidosis-related conditions. Ratios of expression levels or expression folds or ACt levels between 2 or more of SEQ ID NOs: 1 through 17 may also be determined.

[0032] The sample may comprise a sample comprising or enriched for membranous microvesicles (e.g. exosomes), blood, serum, plasma, urine, lymph, cerebrospinal fluid, ascites, saliva, mucus, stool, biopsy specimens, breast milk, gastric juice, pleural fluid, semen, sweat, tears, hair, vaginal secretion or interstitial fluid. Free circulating miRNA is hence encompassed as a sample. If membranous microvesicles such as exosomes are sampled, they may first require chemical or mechanical disruption to release the miRNA.

[0033] In an embodiment, enabled herein is an assay to stratify a human subject with respect to AD or a symptom associated with AD, the method comprising obtaining a sample comprising or enriched for membranous microvesicles and subjecting the microvesicles to mechanical or chemical disruption to release miRNA and screening for one or more miRNA comprising a nucleotide sequence selected from the list consisting of SEQ ID NO: l through 16 or precursor forms thereof wherein the level of expression of the one or more miRNAs is indicative of a stage of AD.

[0034] Further taught herein is an assay to stratify a human subject with respect to AD or a symptom associated with AD, the method comprising subjecting a sample comprising or enriched for membranous microvesicles to mechanical or chemical disruption to release miRNA and treating one or more miRNA comprising a nucleotide sequence selected from the list consisting of SEQ ID NO: l through 16 or a precursor form thereof or a corresponding cDNA form to chemical processing and then determining the level of expression of the miRNAs wherein the level of expression of the one or more miRNAs is indicative of a stage of AD. By "chemical processing" in this context includes reverse transcription to generate cDNA, bisulfite treatment or chemical coupling to an immobilized support or to a solid phase. Free miRNA in a biological sample may also be assayed.

[0035] Reference to "membranous microvesicles" includes exosomes. The sample may or may not be a purified sample of microvesicles. A non-microvesicle sample may also be employed. A "stage of AD" means from presymptomatic to early onset to from mild to severe forms of AD. It also encompasses individuals with a likelihood of developing AD. [0036] Reference to "presymptomatic AD" includes a subject having no to mild cognitive impairment but having a risk factor for development of AD such as one or more ΑροΕε4 alleles and/or a high amyloid burden. This may also be determined by PET imaging.

[0037] The miRNA expression profile is the level of expression of one or more miRNAs comprising a nucleotide sequence selected from the listing consisting of SEQ ID NOs: 1 through 16 or precursor forms thereof or corresponding cDNA compared to a first knowledge base generated from training data. Ratios of levels of expression of 2 or more of SEQ ID NOs: l through 17 can also form the training data. The training data represent the correlation of expression levels of the miRNAs with subjects of known status with respect to AD. The first knowledge base, for example, may be miRNA expression levels in healthy individuals (healthy or normal controls). Alternatively, the first knowledge base may comprise levels of expression of the miRNAs in unhealthy controls. Hence, elevation or reduction in expression is dependent on the type of control. Alternatively, ACt levels in healthy or AD patients are predetermined using qPCR and a ACt cut-off value selected which is associated with AD or with a healthy person.

[0038] In an embodiment, the level of expression in the first knowledge base is based on levels in healthy individuals. In this aspect, the expression of one or more miRNAs comprising nucleotide sequences selected from SEQ ID NOs: l through 13 is elevated relative to the healthy control and the expression of one or more miRNAs comprising nucleotide sequences selected from SEQ ID NOs: 14 through 16 is reduced relative to a healthy control in a subject stratified as having a stage of AD or is at risk of developing AD.

[0039] In an embodiment, the level of deregulation relative to a healthy control of individual miRNA expression correlates to a sensitivity of from about 40% to about 100% and a specificity of from about 30% to 100%. Collectively, however, determination of miRNAs defined by SEQ ID NOs: l through 16 results in a sensitivity of about 87% or a specificity of about 77%. In this regard, another miRNA is hsa-miR-3065-5p (SEQ ID NO: 17) which is not always detected by qRT-PCR. This miRNA is still regarded as part of the present invention.

[0040] The term "deregulation" includes up-regulation and down-regulation. Hence, in an embodiment, the miRNA panel can be divided into two clusters based on up or down- regulation.

[0041] The up-regulation cluster includes SEQ ID NOs: l through 13 and the down- regulation cluster includes SEQ ID NOs: 14 through 16. Each of these clusters can be further divided based on APOE ε4 allele genotyping and clinical classification. Hence, SEQ ID NOs: l through 8 and 9 through 13 are in clusters 1 and 3, respectively, which exhibit a mean increase in expression relative to a healthy control and SEQ ID NO: 14 and 15 and SEQ ID NO: 16 are in clusters 2 and 4, respectively, which exhibit showing a mean decrease in expression relative to a healthy control. In a particular embodiment, SEQ ID NO: 16 (hsa-miRNA 1306-5p) is assayed. In an alternative embodiment, the level of miRNA expression correlates with an unhealthy subject having AD or a certain likelihood of developing AD.

[0042] To determine whether a subject is considered a candidate for AD or to determine the stage of AD, the expression of one or more of SEQ ID NOs: l through 16 may be assayed. This includes 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or all 16 miRNAs. SEQ ID NO: 17 may also be assayed. Precursor forms of RNAs comprising a nucleotide sequence selected from SEQ ID NOs: l through 17 may also be screened. Corresponding cDNA forms may also be assayed.

[0043] Enabled herein is an assay to stratify a human subject with respect to AD or a symptom associated with AD, the method comprising screening a biological sample from the subject for miRNA comprising a nucleotide sequence selected from the list consisting of SEQ ID NOs: l through 16 or precursor forms thereof or their corresponding cDNA forms wherein up-regulated levels of SEQ ID NOs: l through 13 or precursor forms thereof and/or down-regulated levels of SEQ ID NOs: 14 through 16 or precursor forms thereof or their corresponding cDNA forms relative to a healthy control are indicative of the subject having AD or is at risk of developing symptoms of AD. The biological sample may comprise or be enriched for a membranous microvesicles such as exosomes or may be a non-membranous fluid sample. The latter may comprise free circulating miRNAs.

[0044] In an embodiment, the level of deregulation is measured relative to a control amount. The control may be a statistically validated level or range which defines up- regulation or down-regulation or normal regulation or the control may be determined on a subject of known AD status (e.g. an unhealthy or disease control). Alternatively, predetermined ACt levels are obtained by qPCR associated with a healthy or not healthy subject and a ACt cut-off value selected to diagnose AD. This avoids the need for a healthy control to be run each time in the assay.

[0045] Hence, enabled herein is a first knowledge base of levels of expression of miRNAs in subjects with known status with respect to AD. The first knowledge base is determined from training data following a trial of subjects. When the level of expression of one or more miRNAs is determined in a subject with unknown status, this creates a second knowledge base. By comparison between the first and second knowledge bases, the clinician can determine an index of probability that any one of tested subject has AD or is at risk of developing AD.

[0046] The control level may be determined using any suitable method, such as the analysis of test results relative to a standard result which reflects individual or collective results obtained from individuals with known AD status. This form of analysis enables the design of kits which require the collection and analysis of a single membranous microvesicles sample, being a test sample of interest, relative to the control. The control level may be determined from the subjects of a specific cohort and hence there may be different controls to test samples derived from different cohorts. Accordingly, there may be determined a number of control values or ranges which correspond to cohorts which differ in respect of characteristics such as age, gender, ethnicity or health status. The "control level" may be a discrete level or a range of levels. [0047] Hence, the first knowledge base may comprise data from healthy controls or from patients with known levels of AD development. Hence, "deregulated" expression needs to be defined relative to whether the control (i.e. first knowledge base) is a healthy control or a disease control (or unhealthy control). A "disease control" is data from individuals with a known stage of AD. As indicated above, the first knowledge base may also comprise ACt levels based on qPCR from which a ACt cut-off value is selected based on correlation to AD. Expression levels or expression fold changes or ACt levels of individual miRNAs or ratios thereof from 2 or more of SEQ ID NOs: 1 through 17 can also be used.

[0048] To the extent that the miRNA product is present in a biological sample, the sample may be directly tested or else all or some of the nucleic acid material present in the sample may be isolated or enriched prior to testing. To this end, when screening for changes to the level of expression of the miRNA biomarkers, one may screen for the RNA transcripts themselves or cDNA which has been transcribed therefrom. The assay may be conducted on the sample directly or molecules derived therefrom pretreated prior to testing, for example, inactivation of live virus or being run on a gel. The sample may be freshly harvested or it may have been stored (for example by freezing) prior to testing or otherwise treated prior to testing. In an embodiment, the sample is a sample comprising or enriched for membranous microvesicles (e.g. exosomes), blood, serum, plasma, urine, lymph, cerebrospinal fluid, ascites, saliva, mucus, stool, biopsy specimens, breast milk, gastric juice, pleural fluid, semen, sweat, tears, hair, vaginal secretion or interstitial fluid.

[0049] As indicated above, whilst the subject assay is directed to AD, symptoms of AD can arise in other conditions associated with aberrant, unwanted or otherwise inappropriate amyloidosis. Hence, the miRNA expression profile may be used to screen for AD as well as Down's syndrome, hereditary cerebral hemorrhage with amyloidosis, amyloidosis associated with chronic inflammation, various forms of malignancy, familial Mediterranean fever, amyloidosis associated with multiple myeloma and other B-cell dyscrasias, amyloidosis associated with type II diabetes, prion disease, long term hemodialysis, carpal tunnel syndrome, amyloidosis associated with senile cardiac amyloid, familial amyloidotic polyneuropathy and amyloidosis associated with endocrine tumors. [0050] In an embodiment, the condition is AD or is a stage of AD. As indicated above, a stage of AD includes asymptomatic or presymptomatic AD.

[0051] The subject being assayed may undergo additional testing after being stratified as having or being at risk of developing AD. The additional testing includes a range of behavioral or psychological evaluations such as level of cognitive ability as well as genetic testing for missense mutations in the chromosome 21 gene coding for β-amyloid precursor protein, a chromosome 14q gene coding for a 467 amino acid protein with 7 putative membrane-spanning domains and inheritance of the ΑΡοΕε4 allele of the gene encoding apolipoprotein E. The presence of one or 2 alleles of ΑΡοΕε4 together with the miRNA signature of the present invention enables the clinician to diagnose for the presence or absence of AD or to assess the likelihood of development of AD or to assess the stage of AD. Physiological testing such as PET imaging may also be conducted.

[0052] A miRNA is typically approximately 22 nucleotides in length although the present assay extends to miRNAs of from 15 to 30 nucleotides in length or to any fragment or part of an miRNA. A precursor form may also be screened. A precursor may comprise 70-100 nucleotides and comprise a nucleotide sequence selected from SEQ ID NOs: l through 17. Corresponding cDNA forms of SEQ ID NOs: 1 through 17 may also be assayed.

[0053] Typically, miRNAs are endogenously transcribed from DNA, but not translated into protein. The DNA sequence that codes for an miRNA generally includes the miRNA sequence and an approximate reverse complement. When this DNA sequence is transcribed into a single-stranded RNA molecule, the miRNA sequence and its reverse- complement base hybridize to form a double stranded RNA hairpin loop, this forming the primary miRNA structure (pri-miRNA). A nuclear enzyme cleaves the base of the hairpin to form pre-miRNA. The pre-miRNA molecule is then actively transported out of the nucleus into the cytoplasm where the Dicer enzyme cuts 20-25 nucleotides from the base of the hairpin to release the mature miRNA. The miRNA can be released into the extracellular environment when the cell membrane is compromised during apoptotic or necrotic death, secreted when bound to lipoproteins or secreted in cell-derived plasma microvesicles such as exosomes.

[0054] The standard nomenclature system is used herein to identify miRNA. For example, a list of miRNAs (miRs) from a homosapien (hsa) having particular sequences is provided in Table 1.

[0055] In terms of screening for the "level of expression" or changes in fold of expression of miRNAs, this may be achieved in a variety of ways including screening directly for any of the forms of miRNA or cDNA generated therefrom. Accordingly, either RNA or DNA based screening may be employed. Changes to the absolute levels of any of these markers is indicative of changes to the expression of the subject miRNA. Furthermore, ratios of expression changes are also useful in determining a diagnosis. Still further, the miRNA which is identified and measured may be a whole molecule or a fragment thereof. For example, one may identify only fragments of miRNA from a membranous microvesicle sample, depending on how it has been processed. Provided that the fragment comprises sufficient sequence to indicate its origin with a particular miRNA, fragmented miRNAs are useful in the context of the method of the subject assay.

[0056] Reference to "nucleic acid molecule" should be understood as a reference to both ribonucleic acid molecules and deoxyribonucleic acid molecules and fragments thereof. The subject assay extends, therefore, to both directly screening for miRNA levels in a membranous microvesicle sample or screening for the complementary cDNA which has been reverse-transcribed from an miRNA population of interest. It is well within the skill of the person of skill in the art to design methodology directed to screening for either RNA or DNA.

[0057] The term "fragment" means a portion of the subject nucleic acid molecule. This is relevant with respect to screening for modulated miRNA levels in a fluid sample or membranous microvesicle samples which may have been chemically (e.g. enzymatically) or mechanically treated since the subject miRNA may have been degraded or otherwise fragmented. One may, therefore, actually be detecting fragments of the subject RNA molecule, which fragments are identified by virtue of the use of a suitably specific probe. Alternatively, one or more precursors of SEQ ID NOs: 1 through 17 may be detected.

[0058] Reference to a "membranous microvesicle" means any particle which is comprised of a cellular plasma membrane component. Such membranous microvesicles may adopt a structure which takes the form of a lumen surrounded by plasma membrane. Examples of membranous microvesicles include, but are not limited to, microparticles, exosomes, apoptotic blebs, apoptotic bodies, cellular blebs, extracellular vesicles and the like. In an embodiment, the membranous microvesicle is an exosome. Free circulating miRNAs may also be measured.

[0059] Reference to an "exosome" includes vesicles which are secreted by a wide variety of cell types.

[0060] The exosomes tested in the subject assay are generally enriched from a biological sample. By "biological sample" is meant any biological material derived from an individual. The miRNA may be located in fluid medium or in a medium which comprises exosomes. Hence, samples include, but are not limited to, a sample comprising or enriched from membranous microvesicles (e.g. exosomes), blood, serum, plasma, urine, lymph, cerebrospinal fluid, ascites, saliva, mucus, stool, biopsy specimens, breast milk, gastric juice, pleural fluid, semen, sweat, tears, hair, vaginal secretion, interstitial fluid and fluid which has been introduced into the body of an individual and subsequently removed such as, for example, the saline solution extracted from the lung following lung lavage or the solution retrieved from an enema wash. The biological sample may be tested directly or undergo some form of pre-treatment prior to testing. For example, the sample may require the addition of a reagent, such as a buffer, to immobilise the miRNA in the sample. It should be further understood that the sample which is the subject of testing may be freshly isolated or it may have been isolated at an earlier point in time and subsequently stored or otherwise treated prior to testing. For example, the sample may have been collected at an earlier point in time and frozen or otherwise preserved in order to facilitate its transportation to the site of testing. In yet another example, the sample may be treated to neutralize any possible pathogenic infection, thereby reducing the risk of transmission of the infection to the technician.

[0061] The choice of sample may be dependent on the nature of the subject including age and overall health status.

[0062] In an embodiment, the biological sample comprises or is enriched for membranous microvesicles or free miRNAs. In an embodiment, the sample is whole blood, plasma serum, cerebral fluid, spinal fluid, interstitial fluid or a fraction thereof.

[0063] To the extent that the subject biological sample is harvested from an individual, the term "individual" should be understood to include a human of any age or at an age regarded as being at risk. Such age selection includes subjects from 40 to 49, 50 to 59, 60 to 69, 70 to 79, 80 to 89 and 90 to 100 years old. Whilst the subject assay is directed to humans, the assay may be trialled in an animal model such as a non-human primate or a rodent (e.g. mouse) model.

[0064] The miRNA markers are defined by SEQ ID NOs: l through 17 although the preferred panel is SQ ID NOs: l through 16. Such markers include large precursor forms which comprise a sequence selected from SEQ ID NOs: l through 17 or their corresponding cDNA forms. In an embodiment, up-regulated levels of one or more of SEQ ID NOs: l through 13 and/or down-regulated levels of one or more of SEQ ID NOs: 14 through 16 relative to a healthy control is regarded as a signature of AD or a risk of developing AD. The miRNA display a fold expression difference of more than 1.2 and a significance of p<0.05. As indicated above, if a disease or unhealthy control is employed, the up- or down-regulation may be reversed. Also contemplated herein is a measurement of ratios of ACt values, fold changes and levels of expression between any 2 of SEQ ID NOs: 1 through 17 including any 2 of SEQ ID NOs: 1 through 16.

[0065] Whilst a signature profile of two or more miRNAs is likely to be more convenient, measurement of a single miRNA is also contemplated.

[0066] Hence, in another embodiment, there is provided a method of screening for the onset, predisposition to the onset or monitoring the progress of AD in an individual, the method comprising measuring the level of expression of SEQ ID NO: 16 (hsa-miRNA- 1306-5p) in a biological sample from individual wherein a higher level of expression of the miRNA relative to a healthy control level is indicative of the onset or predisposition of the onset of AD. A "biological sample" in this context includes a sample comprising or enriched from membranous microvesicles (e.g. exosomes) as well as whole blood, serum, plasma and the like.

[0067] Reference to the "onset" of AD should be understood as a reference to the commencement of changes to the brain which are characteristic of AD, such as Αβ plaque deposition or formation of neurofibrillary tangles. In this regard, these changes may be well advanced in that cognitive impairment has become evident. Alternatively, the physical changes to the brain may be at a very early stage such that symptoms of cognitive impairment are not yet evident. Assessment of an individual's predisposition to the development of AD is contemplated herein. Changed levels of the subject miRNA markers are proposed to be indicative of that individual's predisposition to developing AD.

[0068] Detection of converse changes in the levels of the miRNA marker may be desired under circumstances other diagnosis, for example, to monitor the effectiveness of therapeutic or prophylactic treatment directed to modulating AD onset or progression. For example, where elevated levels of the markers indicates that an individual has developed AD screening for a decrease in the levels of this markers subsequently to the onset of a therapeutic regime may be utilized to indicate reversal or other form of improvement of the subject individual's condition.

[0069] Contemplated herein is a method of monitoring treatment of a subject with or having a risk of developing AD or as a monitor of the effectiveness of therapeutic or prophylactic treatment regimes directed to inhibiting or otherwise slowing AD development. In these situations, mapping the modulation of miRNA marker expression levels in a biological sample is a valuable indicator of the status of an individual or the effectiveness of a therapeutic or prophylactic treatment regime which is currently in use. Accordingly, a method contemplated herein extends to monitoring for changes in miRNA marker expression levels in an individual relative to their levels prior to treatment or during treatment, or relative to one or more earlier marker expression levels determined from a biological sample of the individual. Furthermore, taught herein is a medical protocol to treat a subject, the protocol comprising assessing or stratifying the subject with respect to AD by the assay herein described and then providing therapeutic, psychological and/or behavioral intervention to ameliorate any symptoms of AD.

[0070] Insofar as the sample comprises or in enriched for membranous microvesicles (e.g. exosomes), the membranous microvesicles may be derived from any suitable biological sample and may be either isolated from that sample or enriched therein. Methods for performing isolation or enrichment are known and it is within the skill of the person in the art to select and apply a method appropriate to the particular circumstances. For example, the microvesicles may be enriched for by subjecting the biological sample of which they are part to mechanical or chemical disrupture such that nucleic acid material is release. The nucleic acid material would include miRNA.

[0071] When using biological fluids, generally, cellular contamination is removed by low speed centrifugation.

[0072] To the extent that it is sought to isolate and analyse the miRNA within the membranous microvesicles, it is necessary to lyse the microvesicles in order to expose their nucleic acid content and to thereafter analyse the miRNA subpopulation of nucleic acid molecules. To this end, the analysis of exosome RNA is often based on isolation of total RNA followed by PCR amplification of specific transcripts of interest. Methods for isolating and analysing total RNA are well known. Once a cell-free biological fluid is obtained, a high speed centrifugation spins can be used to pellet the exosomes which are lysed. Thus, the nucleic acids are exposed and extracted. However, a kit can also be used which extracts the exosomal RNA without use of high-speed centrifugation.

[0073] There is a wide variety of methods which can be and have been used to isolate total RNA from samples.

[0074] miRNA amplification or probing steps require use of primers. Reference to a "primer" or an "oligonucleotide primer" should be understood as a reference to any molecule comprising a sequence of nucleotides, or functional derivatives or analogues thereof, the function of which includes hybridization to a region of a nucleic acid molecule of interest. It should be understood that the primer may comprise non-nucleic acid components. For example, the primer may also comprise a non-nucleic acid tag such as a fluorescent or enzymatic tag or some other non-nucleic acid component which facilitates the use of the molecule as a probe or which otherwise facilitates its detection or immobilization. The primer may also comprise additional nucleic acid components, such as the oligonucleotide tag. In another example, the primer may be a protein nucleic acid which comprises a peptide backbone exhibiting nucleic acid side chains. The design and synthesis of primers suitable for use in the subject assay would be well known to those of skill in the art.

[0075] Various techniques can be used to analyse an amplification product in order to determine relative miRNA expression levels. Their operational characteristics, such as ease of use or sensitivity, vary so that different techniques may be useful for different purposes. They include but are not limited to sequencing, pyrosequencing, enzyme digestion, microarray analysis, denaturing gradient gel electrophoresis, agarose gel based separation, melt curve analysis on real-time PCR cyclers, quantitative real-time PCR, denaturing high performance liquid chromatography, mass spectrometry, primer extension, oligonucleotide-ligation, mutation specific polymerase chain reaction, denaturing gradient, electrophoresis (DGGE), temperature gradient denaturing electrophoresis, constant denaturing electrophoresis, single strand conformational electrophoresis and denaturing high performance liquid chromatography (DUPLC). [0076] It is well within the skill of the person of skill in the art to select and apply an appropriate method of screening for the miRNA marker expression levels hereinbefore discussed.

[0077] Further enabled herein is an algorithm-based screening assay to screen biological samples from subjects for levels of selected miRNAs including a panel of miRNA. Generally, input data in the form of expression levels are collected based on miRNAs defined by SEQ ID NOs: l through 16 or precursor forms thereof or their corresponding cDNA forms and subjected to an algorithm to assess the statistical significance of any elevation or reduction in levels which information is then output data. Computer software and hardware for assessing input data are encompassed by the present invention. In an embodiment, the samples include a sample comprising or enriched for membranous microvesicles, blood, serum, plasma, urine, lymph, cerebrospinal fluid, ascites, saliva, mucus, stool, biopsy specimens, breast milk, gastric juice, pleural fluid, semen, sweat, tears, hair, vaginal secretion or interstitial fluid.

[0078] Another aspect of the present invention contemplates a method of stratifying a subject with respect to AD a female, the method comprising subjecting the subject to a diagnostic assay to determine the levels of one or more of miRNAs defined by SEQ ID NOs: l through 16 or precursor forms thereof or their corresponding cDNA forms to generate an index of probability of the subject having a stage of AD or is at risk of developing AD.

[0079] The present invention further provides the use the levels of one or more miRNAs selected from SEQ ID NOs: l through 16 or precursor forms thereof or their corresponding cDNA forms in the generation of an index of probability for use in a diagnostic assay to predict the presence of AD or a risk of developing same.

[0080] The assay of the present invention permits integration into existing or newly developed pathology architecture or platform systems. For example, the present invention contemplates a method of allowing a user to determine the status of a subject with respect to AD, the method including:

(a) receiving data in the form of levels of expression of one or more miRNAs defined by SEQ ID NOs: 1 through 16 and/or from the user via a communications network;

(b) processing the subject data via an algorithm which provides a likelihood index value;

(c) determining the status of the subject in accordance with the results of the likelihood index value in comparison with predetermined values; and

(d) transferring an indication of the status of the subject to the user via the communications network.

[0081] Conveniently, the method generally further includes:

(a) having the user determine the data using a remote end station; and

(b) transferring the data from the end station to the base station via the communications network.

[0082] The base station can include first and second processing systems, in which case the method can include:

(a) transferring the data to the first processing system;

(b) transferring the data to the second processing system; and

(c) causing the first processing system to perform the algorithmic function to generate the likelihood index value.

[0083] The method may also include:

(a) transferring the results of the algorithmic function to the first processing system; and

(b) causing the first processing system to determine the status of the subject.

[0084] Reference to an "algorithm" or "algorithmic functions" as outlined above includes the performance of a multivariate analysis function. A range of different architectures and platforms may be implemented in addition to those described above. It will be appreciated that any form of architecture suitable for implementing the present invention may be used. However, one beneficial technique is the use of distributed architectures. In particular, a number of end stations may be provided at respective geographical locations. This can increase the efficiency of the system by reducing data bandwidth costs and requirements, as well as ensuring that if one base station becomes congested or a fault occurs, other end stations could take over. This also allows load sharing or the like, to ensure access to the system is available at all times. This particular method is also amenable for high throughput screening of ACt levels based on qPCR wherein a ACt value is selected for cutoff in terms of a subject having or is at risk of developing AD or who is healthy.

[0085] In this case, it would be necessary to ensure that another base station contains the same information and signature such that different end stations can be used.

[0086] It will also be appreciated that in one example, the end stations can be hand-held devices, such as PDAs, mobile phones, or the like, which are capable of transferring the subject data to the base station via a communications network such as the Internet, and receiving the reports.

[0087] In the above aspects, the term "data" means the levels or concentrations of the biomarkers. The "communications network" includes the internet. When a server is used, it is generally a client server or more particularly a simple object application protocol (SOAP).

EXAMPLES

[0088] Aspects disclosed herein are further described by the following non-limiting Examples. The ensuing Examples employ the following Materials and Methods.

Participants

[0089] Patients are divided into three groups: healthy controls (HC), mild cognitive impairment (MCI) and Alzheimer's disease (AD) based on the establish criteria from the National Institute of Neurological and Communicative Diseases and Stroke-Alzheimer's Disease and Related Disorders Association (NINCDA-ADRDA) [McKhann et al. (1984) Neurology 3 :939-944], which has been described in detail elsewhere (Ellis et al. (2009) Int Psychogeriatr 27:672-687). Blood samples are collected from all patients (fasting) from whole blood venepuncture into Sarstedt s-monovette serum-gel 7-5 ml tubes 01 1602 001 (Sarstedt, Germany). Serum-gel tubes are processed within two hours of collection and serum is snap frozen in liquid nitrogen. In the discovery set, serum from 49 patients (HC, 23; MCI, 3; AD, 23) are collected for deep sequencing and serum from 60 patients (HC, 36; MCI, 8; AD, 16) are collected for validation. All individuals included are assessed for full blood pathology testing (Melbourne Health and PathWest Laboratory Medicine, Melbourne, Australia), apolipoprotein ε4 (ΑροΕε4) genotyping and assessment of cognitive functions (mini-mental state examination, MMSE) at the time of collection (Ellis et al. (2009) supra). Of these, 83 (discovery set, 23 and validation set, 60) patients hadAmyloid-PET neuroimaging data available for assessment of cerebral Αβ accumulation. Αβ burden is expressed as (Cheng et al. (2013) supra) C-PiB standardized uptake value ratio (SUVR) as previously described (Villenmagne et al. (2013) supra). Patient data from both baseline and 54-month time points are available for analyses.

Procedures

[0090] Serum exosomal RNA is isolated by using the Plasma/serum exosomal RNA isolation kit (Norgen Biotek, Canada) from 1 ml serum per participant whereby the manufactures' protocol is followed. The total exosomal RNA yield, composition and quality is analyzed by the Agilent 2100 Bioanalyser using the Small RNA kit (Agilent). Exosomal RNA is converted into cDNA libraries using the Ion Total RNA-Seq Kit V2 (Life Technologies, Australia) and prepared for sequencing as described previously (Cheng et al. (2013) Kidney International). Pooled libraries with unique barcodes are loaded on 318 sequencing chips and run on the Ion Torrent Personal Genome Machine (Life Technologies, Australia) [Cheng et al. (2013) supra). The sequences are then assessed for quality and primer-adapter sequences are trimmed by the Torrent Suite software (version 3 -4- 1), followed by alignment to the human reference genome (HG19). The trimmed and aligned data is transferred to Partek Genomics Suite (Partek, Singapore) and mapped to known miRNA using miRBase V-20.Bioinformatics analysis and differential expression is performed using Partek Genomics Suite. The panel of candidate miRNA found highly associated with AD (including positive and negative controls) are used for validation on a new set of serum samples collected from AIBL. Serum samples for the validation study are spiked with synthetic C. elegan miR-39 (Qiagen, Australia) during exosomal RNA extraction to monitor extraction efficiency and for normalization purposes. Upon processing of the serum samples as above, lng of exosomal miRNA is converted to cDNA (TaqMan MicroRNA Reverse Transcription Kit, Applied Biosystems, Australia) according to the manufacturers' protocol with a primer pool containing 23 miRNA assays (TaqMan microRNA assays, 5x, Applied Biosystems). cDNA samples are pre-amplified (TaqMan PreAmp Master Mix Kit, Applied Biosystems) and qRT-PCR (TaqMan Fast Advanced Master Mix, Applied Biosystems) is performed using individual miRNA assays (TaqMan microRNA assays, 20x, Applied Biosystems) and run on the ViiA (Trade Mark) 7 Real- Time PCR System (Life Technologies) across a 384-well format. Reverse transcription and pre-amplification no template controls using primer pools and individual assays are also prepared to ensure there is no background amplification of miRNA assays. Raw data are uploaded to DataAssist (Applied Biosystems) to calculate delta Ct (ACt) and normalization against controls further statistical analysis.

Statistical analysis

[0091] To analyze the deep sequencing data, the number of reads of each miRNA are normalized to reads per million (RPM) across all samples. Low abundant miRNAs with less than 50 read counts across all samples are removed thus high abundant miRNAs are analyzed. Initial statistical analysis of miRNA expression changes is performed using the Partek Genomics Suite. Selection of miRNAs is based upon ANOVA comparing healthy control and AD groups (clinical classification at the time of collection). Probability values are adjusted for multiple testing using the False Discovery Rate (FDR) method. Significant changes in miRNA expression are expressed in fold change [LOG ₂ ] and defined as p (AD Vs HC) < 0 05, p (AD, MCI and HC) < 0 05 and ± 1 2 fold change. Generalized Linear Modelling is performed to compare each miRNA between clinical classification chosen from the initial analyses adjusted for age, gender and ΑροΕε4 allele status. Further classification (discovery data set) and prediction (validation data set) analyses are performed using Random Forest analyses. From these models, sensitivity and specificity and related statistics are calculated. All statistical analyses are performed using the R statistical environment, version 3.02.

EXAMPLE 1

Exosomal miRNA discover set analyses

[0092] The characteristics of the participants included in this study are shown in Table 3. The prevalence of ΑροΕε4 is significantly higher in the AD group compared with health controls (p = 0.006). AD patients performed significantly poorer in the MMSE (p = 0.006) and obtained low clinical dementia ratings [CDR]( < 0.0001) and composite scores (composite score 1 : p< 0.0001, composite score 2: p< 0.0001). Deep sequencing of the exosomal RNA extracted from serum samples is performed and sequences are mapped to miRBase V-20. Approximately half of the reads obtained (43%) from small RNA sequencing are miRNA. Overall, the sample cohort mapped to 1419 known miRNA. Upon performing normalization of reads and performing ANOVA analyses, 17 miRNA are found to be significantly deregulated (p ANOVA (AD Vs HC) < 0 05, ANOVA (AD, MCI and HC) < 0 05 and ± 1 -2 fold change, Figure 1); 14 miRNA are found to be up- regulated (hsa-miR-361-5p, hsa-miR-30e-5p, hsa-miR-93-5p, hsa-miR-15a-5p, hsa-miR- 143-3p, hsa-miR-335-5p, hsa-miR-106b-5p, hsa-miR-101-3p, hsa-miR-424-5p, hsa-miR- 106a-5p, hsa-miR-18b-5p, hsa-miR-3065-5p, hsa-miR-20a-5p, and hsa-miR-582-5p) and 3 miRNA are found to be down-regulated (hsa-miR-1306-5p, hsa-miR-342-3p, and 15b-3p, Table 3). Upon post adjustment of age, ΑροΕε4 allele status and sex (Generalized Linear Modelling), the majority of miRNA remained statistically significant upon contrasting within all attribute groups (AD, MCI and HC) and between AD and HC (Table 4). The miRNAs listed above are defined by sequence identifier number (Table 1).

[0093] Upon performing Partitioning Around Medoids (PAM) clustering of the 17 miRNA, two main clusters are found, one for those miRNAs up-regulated in AD, and one for those down-regulated in AD. Further investigation of these revealed four clusters (Figure 2A), with stepwise increases in miRNA expression across ΑροΕε4 allele carriers and clinical classification (Figure 2B). Delineation into miRNA cluster averages, showed clear associations with ΑροΕε4 allele status (Figure 2C), confirming the stepwise increases shown by assessing miRNA by cluster. Variable selection via Random Forest analyses (Figure 2D) defined hsa-miR-1306-5p as contributing more towards clinical classification than age, indicating this miRNA is important in class separation.

[0094] Receiver Operating Characteristic (ROC) analyses statistics for each miRNA marker are shown in Table 5. The ROC analyses is performed using the raw sequence count data to estimate the performance of each miRNA to predict AD. Criterion cut points are chosen by taking the closest point on the curve to the top left corner of the ROC curve; approximating the most optimum performance for each marker. Independently, the miRNAs displayed between 35 and 100% sensitivity and specificity. Similar to the Random Forest model, hsa-miR-1306-5p showed the greatest combined, sensitivity and specificity (87.5% and 10%, p = 0.0008).

EXAMPLE 2

qRT-PCR validation set analyses

[0095] All patients in the validation set (n = 60, Table 3) went through full assessment including neuroimaging performed at baseline (prior to this study) and 54 months (time of collection). Samples are spiked with synthetic cel-miR-39 as an external control and three highly abundant miRNA are run as potential endogenous controls (hsa-miR-451, hsa-miR- 223-3p, and hsa-miR-339-5p,). For quality control assurance and normalization controls, cel-miR-39 is used as an external control and hsa-miR-451 is chosen as the endogenous control owing to its stable expression across all samples. Upon completion of the qRT- PCR assay, two patient samples did not display appropriate threshold levels of cel-miR-39 and hsa-miR-451, thus did not fulfil quality control measures and are removed. In addition, hsa-miR-3065-5p is found to be undetected (Ct >35) in 40 samples. As qRT- PCR could not detect hsa-miR-3065-5p in all patient samples by qRT-PCR, this miRNA is removed from the list of biomarker candidates.

[0096] Model validation (Random Forest, Figure 2D) using the 16 miRNA markers, age, gender and ΑροΕε4 allele status with 54 month clinical classification correctly diagnosed 13 out of 15 AD patients (sensitivity 87%, Table 6) and confirmed 27 out of 35 healthy control patients (specificity 77%, Table 6). Five HC patients (samples 7, 9, 16, 34 and 51) who are incorrectly classified as having AD are found to have high Αβ burden with an SUVR of greater than 1 -5 across four time points of the AIBL longitudinal study, suggestive of future risk of clinical manifestation of AD. Three HC patients (sample 8, 17 and 39) had low SUVR, but who are APOE ε4 carriers are also classified as having AD. Case by case information is presented in Table 6. Box plots of qRT-PCR data for each miRNA marker stratified by clinical classification are presented in Figure 3.

[0097] As the cohort size for MCI patients is small, the model could not be trained to predict progression from HC to MCI or MCI to AD. Instead HC and AD thresholds are applied to predict the MCI patients in the validation set. Of the eight MCI patients at 54 months, two are classified as having AD (samples 27 and 53), while the remaining six are classified as HC (Table 6). Further investigation using levels of neocortical Beta Amyloid burden show those MCI patients classified as having AD had high levels of SUVR (greater than 1.5), suggesting that these patients may be on an AD type dementia pathway. Furthermore, three MCI patients who are classified as HC all have low SUVR and are negative for APOE ε4, suggesting that they are not on an AD type dementia pathway. The remaining patients (samples 26, 28, 37, 49 and 29) incorrectly classified do not carry APOE ε4 and are classified as HC. Interestingly a clear outlier (sample 49), which is APOE ε4 negative, displayed the highest SUVR of 3 03 and a MMSE score of 10 at 54 months, was classified as healthy using miRNA biomarkers.

EXAMPLE 3

Development of assay

[0098] The use of exosomes compared to cell-free or whole blood has a number of disease specific advantages for diagnostic purposes. First, isolating the enriched miRNA in exosomes from biological samples of AD patients removes the saturation of insignificant miRNA expressed in both AD and HC patients thus, providing a better performing predictive AD diagnosis. Secondly, miRNA originating from the brain have been demonstrated to cross the endothelial cellular layers of the blood brain barrier by transcytosis of exosomes across the endothelial layer in order to communicate between the brain and distant organs via biological fluids (Haqqani et al. (2013) Fluids barriers CNS 10:4). Thirdly, evidence has revealed that exosomes serve as a RNase-protective vesicle which shield miRNA from RNase-rich environments such as the circulatory system (Huang et al. (2013) Bmc Genomics 14319). Furthermore, specific packaging of miRNA into exosomes (Gibbings et al. (2009) Nat Cell Biol 11.1143-1149) and increased secretion of microvesicles into peripheral blood of cancer patients compared to healthy patients (Mitchell et al. (2008) supra) suggests the importance of exosomes in the role of extracellular communication during disease. These factors allow the possibility of profiling disease specific miRNA that are found enriched from exosomes.

[0099] This study produced a large coverage of miRNA (1419 known miRNA) extracted from exosomal serum samples isolated for AD biomarker discovery by next-generation sequencing (NGS). The advantage of NGS is its ability to detect absolute counts of all miRNA present in samples. However, in order to successfully validate miRNA assays by qRT-PCR technology for diagnostic assays, miRNAs need to lie within the detectable range of qRT-PCR. Using the workflow carried out in this study, it is estimated that approximately 50 RPM must be obtained across the majority of biological samples for successful biomarker validation by qRT-PCR. Biomarkers studies attempting to validate a large panel consisting of low abundant miRNAs will result in unsuccessful validation of miRNAs and/or changes in direction of expression (Lei dinger et al. (2013) Genome Biol 7 :R78; Kumar et al. (2013) PLoS ONE 5:e69807). In this study, from the 1419 mapped miRNA in this cohort, 220 highly abundant miRNA are present in all samples across the study. Seventeen significantly differentially expressed exosomal miRNA are found in AD patients and 16 miRNA are successfully validated displaying consistent expression changes with AD and HC across both NGS and qRT-PCR data sets. The commonly used comparative delta Ct approach in qRT-PCR may not be applicable for diagnostic use as a healthy control group is required for every run. Thus, normalized delta Ct is applied into a Random Forest model to predict clinical classification. Such a model is able to predict clinical classification (at 54 months) with high accuracy (87% sensitivity and 77% specificity).

[0100] A possible limitation in the assay is an ability to diagnose those with late AD where the rate of pathogenesis, in particular Αβ deposition, slows in the later stages of AD when patients experience cerebral atrophy and severe cognitive impairment (Villenmagne et al. (2013) supra). However, in that case diagnosis is less critical and the assay may be used with other markers. Patients of the AIBL study are diagnosed according to NINCDS- ADRDA criteria, which is heavily reliant on cognitive and neuropsychological testing for a clinical diagnosis of probable AD. Thus, in this study, the diagnosis obtained by miRNA analysis is compared with the participant's classification by NINCDS-ADRDA criteria. The additional strength of this study comes from the extensive database of metadata collected at baseline through to 54 months. Each participant underwent a battery of assessment including APOE ε4 genotyping, cognitive examination and Amyloid-PET neuroimaging. Amyloid-PET and ΑροΕε4 genotyping information for those HC patients incorrectly classified with AD suggested progression towards AD. Taking into consideration that these HC patients could actually be prodromal AD, the predicted specificity would increase to 91 -4%. Ultimately, the objective and purpose of the exosomal miRNA biomarker test is to predict future cognitive decline in asymptomatic or presymptomatic individuals and during the progression of patients with early dementia. This study represents a significant step towards developing a cost-effective, non-invasive and low risk diagnostic test to detect the onset and monitor various stages of AD in order for physicians to provide optimal care for patients. [0101] Here, serum samples are analyzed for exosomal miRNA associated with AD together with metadata which included cognitive and amyloid-PET imaging. A16-miRNA expression profile is generated for the diagnosis of AD with sensitivity and specificity of 87% and 77%, respectively. Importantly, the exosomal miRNA AD signature is able to predict cognitively healthy patients presenting with high amyloid burden as probable AD.

[0102] Those skilled in the art will appreciate that the disclosure described herein is susceptible to variations and modifications other than those specifically described. It is to be understood that the disclosure contemplates all such variations and modifications. The disclosure also enables all of the steps, features, compositions and compounds referred to or indicated in this specification, individually or collectively, and any and all combinations of any two or more of the steps or features or compositions or compounds.

Table 3

Demographics and clinical make up of discovery and validation set

Discovery Set HC MCI AD p value*

N 23 3 23

Age (mean ±SD) 73 -07 (7-57) 76-91 (4-4) 78-89 (7-33) 0 034

Gender (F/M) 10/13 1/2 13/9 0-484

ApoE4 (-ve/+ve) 18/5 1/2 7/15 0 006

MMSE (mean iSD) 29 (1 -25) 23 (5 -66) 15 (8-66) O-0001

Composite score 1 (mean ±SD) 0-26 (0-73) -2 04 (0 -80) -2-41 (0-25) O-0001

Composite score 2 (mean ±SD) 0 10 (1 10) -1 -79 (1 18) -2- 10 (0-90) O-0001

Validation Set HC MCI AD p value*

N 36 8 16

Age (mean ±SD) 78-55 (6-52) 78-91 (6-3) 78-29 (7-55) 0-963

Gender (F/M) 21/14 4/4 10/5 0-737

ApoE4 (-ve/+ve) 27/8 6/2 2/13 0 0001

MMSE (mean iSD) 30 (1 08) 29 (2-03) 21 (8- 14) O-0001

Composite score 1 (mean ±SD) 0 (0-59) -1 16 (0-65) -2 02 (0-65) O-0001

Composite score 2 (mean ±SD) 0 (0-91) -0-63 (0-7) -1 -64 (0-4) 0 046

MMSE = mini-mental state examination. APOE = apolipoprotein E. ^Compared with healthy controls

Table 4

Number of reads obtained from deep sequencing data per miRNA across each clinical classification

Cluster miRNA HC MCI AD p Value p value (GLM)*

Fold All groups

Mean RPM change (Unadjuste All groups HC vs AD HC V! (SD) Mean RPM (SD) Mean RPM (SD) ANOVA d) (Adjusted) (Unadjusted) (Adju; hsa-miR- ■30e-5p 85-38 (18-81) 160-39 (34-93) 111-26 (35-59) 1 ^• 30 110E-04 002 003 001 001 hsa-miR- ■101-3p 7514 (20-29) 164-98 (118-45) 107-38 (63-26) 1 ^• 43 001 010 0-44 0-07 001 hsa-miR- ■15a-5p 2472-22 4168-34 (2182-56) 3254-91 1 ^• 32 002 005 013 004 001

(769-66) (1407-36)

hsa-miR- ■20a-5p 1774-54 3191-99 (1824-21) 2818-21 1 ^• 59 9-53E-03 002 010 002 002

(484-33) (1568-42)

hsa-miR- ■93 -5p 2577 (601-67) 394108 (1307-43) 338906 1 ^• 32 9-26E-03 003 018 003 002

(1236-83)

hsa-miR- ^■ 106b- 658-56 1244-95 (693-21) 900 -7 (462 -92) 1 ^• 37 002 009 0-40 0-07 001

5p (209-64)

hsa-miR- ■18b-5p 64-7 (20-93) 10509 (53-34) 96-48 (42-46) 1 ^• 49 6-85E-03 001 005 001 003 hsa-miR- ^■ 106a- 396-55 (13017) 450-94 (218-61) 1 41 003 004 0-23 005 001

5p

2 hhssaa--mmiiRR-- ■11330066-- 46-72 (22-26) 49-45 (14-6) 30-72 (1519) 0 ^• 66 001 001 005 001 004

55pp

3 hhssaa--mmiiRR-- ■33006655-- 34-45 (23-2) 5411 (7-36) 52-86 (27-71) 1 ^• 53 004 001 010 001 002

55 _PP ^##

3 hhssaa--mmiiRR-- ■558822--55pp 15-84 (10-82) 29-42 (16-55) 25-41 (12-43) 1 ^• 60 002 001 006 001 002

3 hhssaa--mmiiRR-- ■114433--33pp 11169 (53-6) 181-94 (51-29) 149-21 (52-8) 1 ^• 34 002 002 0-07 002 003

The average number of total reads obtained per sample was 780, 371. Reads were mapped to 1419 known mature miRNA using miRBase V-20. miRNA with lower than 50 reads across all samples were removed leaving 225 abundant miRNA for analysis. Raw reads were normalized to read per million (RPM). RPM values for each sample. ANOVA analysis was performed by comparing cohort groups with healthy controls. *P values are the result of Generalized Linear Modeling pre and post adjustment with age, sex and APOE ε4 allele. miRBase v.20 accession numbers. ^# Hsa-miR-3065 -5p was not used in the validation set as it was found to be undetectable using qRT-PCR

Table 5

Receiver operating characteristics of miRNA exosomal Alzheimer's disease biomarkers miRNA AUC(95%CI) Rvalue Sensitivity Specificity Threshold hsa-miR-30e-5p 75 ^• 31 (57 ^• 92 -92 ^• 71) 0 ^• 01 56 ^• 25 9000 10807 hsa-miR-101-3p 60 ^• 31 (39 15 -81 ^• 48) 0 ^• 31 43 ^• 75 10000 11701 hsa-miR-15a-5p 70 00 (52 ^• 25 -87 ^• 75) 0 ^• 04 37 ^• 50 10000 3741-71 hsa-miR-20a-5p 64 ^• 69 (45 04 -84 ^• 34) 0 14 43 ^• 75 10000 2803-99 hsa-miR-93-5p 67 ^• 50 (48 ^• 33 -86 ^• 67) 0 08 37 ^• 50 10000 3869-47 hsa-miR-106b-5p 57 ^• 50 (36 ^• 66 -78 ^• 34) 0 ^• 46 43 ^• 75 9000 866-54 hsa-miR-18b-5p 71 ^• 25 (53 ^• 33 -89 ^• 17) 0 03 50 00 10000 101-53 hsa-miR-106a-5p 62 ^• 81 (41 ^• 53 -84 09) 0 ^• 20 50 00 9000 391-25 hsa-miR-1306-5p 81 ^• 88 (67 ^• 37 -96 ^• 39) 0 0008 87 ^• 50 7000 34-99 hsa-miR-582-5p 70 ^• 94 (53 ^• 59 -88 ^• 29) 0 03 43 ^• 75 9500 26-88 hsa-miR-143-3p 71 ^• 25 (53 ^• 66 -88 ^• 84) 0 03 87 ^• 50 5500 11139 hsa-miR-335-5p 66 ^• 88 (48 ^• 86 -84 ^• 89) 0 09 93 ^• 75 3500 119-78 hsa-miR-361-5p 68 13 (50 ^• 34 -85 ^■ 91) 0 ^• 07 93 ^• 75 5000 93-82 hsa-miR-424-5p 74· 06 (57 ^• 62 -90 ^• 50) 0 01 50 00 9000 208-91 hsa-miR-342-3p 69 06 (51 ^• 46 -86 ^• 66) 0 ^• 05 81 ^• 25 5500 1146-51 hsa-miR-15b-3p 66 ^• 25 (48 15 -84 ^• 35) 0 10 10000 3000 184-99

A UC=Area under receiver operating curve. CI = Confidence interval.

Table 6

Demographics, clinical make up and classification of patients of the validation set at baseline and 54 months of the AIBL flagship study

PiB PiB

Sample MMSE MMSE Class. Class. Validation

Age Gender ΑροΕε4 SUVR SUVR

Baseline 54 m Baseline 54 m Class.

Baseline 54 m

Patients diagnosed correctly according to clinical classification

2 67 F + 28 24 2-57 2-97 AD AD AD

46 73 F + 25 21 204 NI AD AD AD

50 59 F + 23 25 1-59 NI AD AD AD

40 69 F + 30 26 2-6 NI HC AD AD

48 78 M + 30 16 2-25 2-39 HC AD AD

43 68 M + 30 17 2-27 2-55 HC AD AD

5 75 F + 27 21 2-64 2-82 MCI AD AD

4 84 M + 26 23 2-44 2-48 MCI AD AD

58 77 F + 27 26 206 NI MCI AD AD

59 80 F + 26 26 2-2 NI MCI AD AD

55 76 F + 27 27 21 2-31 MCI AD AD

47 84 F + 22 2 2-55 NI MCI AD AD

1 69 M + 26 10 216 NI MCI AD AD

3 81 F - 29 29 106 113 HC HC HC

23 81 F - 28 30 116 107 HC HC HC

12 81 F - 30 30 1-26 1-24 HC HC HC

13 88 F - 30 30 218 2-3 HC HC HC

19 80 M - 30 29 2-03 2-41 HC HC HC

42 86 F - 30 30 1-36 NI HC HC HC

10 74 F - 30 30 114 104 HC HC HC

11 63 M - 29 28 1-36 1-31 HC HC HC

36 65 F - 30 30 118 1-23 HC HC HC

32 72 F - 26 29 1-35 161 HC HC HC

25 78 M - 28 29 201 2-27 HC HC HC

18 78 F - 29 28 1-87 NI HC HC HC

15 76 F - 28 26 1-79 2-03 HC HC HC

14 74 F - 28 28 107 112 HC HC HC

22 76 F - 28 30 113 NI HC HC HC

20 76 M - 26 29 1-27 1-25 HC HC HC

21 74 F - 27 28 102 102 HC HC HC PiB PiB

Sample . „ . MMSE MMSE Class. Class. Validation

Age Gender ΑροΕε4 _ SUVR SUVR

# Baseline 54 m Baseline 54 m Class.

Baseline 54 m

35 67 F 29 30 108 104 HC HC HC

33 63 M 30 29 106 105 HC HC HC

45 61 F 29 30 111 1-22 HC HC HC

38 72 M 30 30 121 121 HC HC HC

44 71 M 30 30 1-22 1-3 HC HC HC

52 74 M 30 27 115 109 HC HC HC

54 77 M 30 30 118 1-3 HC HC HC

60 61 M 29 30 107 103 HC HC HC

56 76 F 29 28 119 139 HC HC HC

24 72 M 29 30 117 1-2 HC HC HC

Patients misdiagnosed from clinical classification

9 73 F + 29 29 1-69 1-75 HC HC AD

51 82 M + 30 29 2-33 2-37 HC HC AD

7 79 F + 28 27 2-07 NI HC HC AD

16 74 M + 29 30 1-35 1-52 HC HC AD

34 72 M + 29 30 1-57 1-78 HC HC AD

17 76 F + 30 30 111 108 HC HC AD

8 69 F + 30 30 121 NI HC HC AD

39 68 F + 30 30 106 115 HC HC AD

27 70 M + 26 28 2-05 2-56 MCI MCI AD

53 76 M + 26 30 165 206 MCI MCI AD

57 82 F 29 29 11 1-22 MCI MCI HC

41 78 F 27 30 105 1-2 HC MCI HC

6 79 M 27 29 1 103 MCI MCI HC

26 68 F 30 29 1-4 NI MCI MCI HC

28 78 M 28 25 216 2-38 MCI MCI HC

37 64 F 30 25 1-62 1-82 MCI MCI HC

49 82 M 20 10 304 3-03 AD AD HC

29 65 F 24 6 1-92 212 MCI AD HC

Positive APOE ε4 is indicated by "+". Negative for APOE ε4 is indicated as "-".MMSE = mini-mental state examination. Classification obtained using miRNA exosomal biomarkers are indicated under validation classification. PiB = Carbon- 11 -labelled Pittsburgh compound B. SUVR = standard uptake value ratio. Patients who presented with a low Αβ burden in the brain have a SUVR <l-5. Patients who presented with high Αβ burden in the brain have a SUVR > 1-5. ² Class. = Classification. Clinical classification at baseline and 54 months is diagnosed by NINCDS-ADRDA criteria for AD. Classification obtained using miRNA exosomal biomarkers are indicated under validation classification.

BIBLIOGRAPHY

Bellingham et al. (2012) Front Physiol 3: 124) Breteler et al. (1992) Epidemiol Rev 14:59-82

Cheng et al. (2013) Kidney International

Cheng et al. (2013) Front Genet 4Λ50

Cole and Vassar (2008) J Biol Chem 253:29621-29625 Ellis et al. (2009) Int Psychogeriatr 27:672-687 Gibbings et al. (2009) Nat Cell Biol 77: 1143-1149 Hannon {2004) Nat Rev Genet 5:522-531

Hansson et al. (2006) Lancet neurology 5:228-234 Haqqani et al. (2013) Fluids barriers CNS 10:4 Huang et al. (2013) Bmc Genomics 14:319

Krol et al. (2010) Nat Rev Genet 77:597-610

Kumar et al. (2013) PLoS ONE 5:e69807

Lei dinger et al. (2013) Genome Biol 7 :R78;

McKhann et al. (1984) Neurology 34:939-944 Mitchell et al. (2008) Proceedings of the National Academy of Sciences 705:10513-10518

Skog et al. (2008) Nat Cell Biol 10: 10- 16

Terry etal. (\9%\)Ann. Neurol 70:184-192

Valadi etal. (2007) Nat Cell Biol 9:654-659

Villemagne etal. (2013) The Lancet Neurology 72:357-367