WO2019195892A1

WO2019195892A1 - Dementia risk analysis

Info

Publication number: WO2019195892A1
Application number: PCT/AU2019/050327
Authority: WO
Inventors: Wei Ling Florence LIM; Ralph Nigel Walter Martins; Peter John Meikle; Kevin HUYNH
Original assignee: Baker Heart and Diabetes Institute; Edith Cowan University
Priority date: 2018-04-12
Filing date: 2019-04-12
Publication date: 2019-10-17

Abstract

A method of determining Alzheimer's Disease in sample by (i) detecting the level of a plurality of lipid species selected from one or two or more lipid classes from Hex3Cer, DE, and TG(O). A method of determining the severity of AD or Aβ burden in a subject, comprising (ii) detecting the level of a plurality of lipid species comprising one or two or more lipid species from one or two lipid classes or subclasses of ether lipids including alkylphosphatidylcholine, alkenylphosphatidylcholine, lysoalkylphosphatidylcholine, alkylphosphatidylethanolamine alkenylphosphatidylethanolamine, lysophosphatidylethanolamine, lysoalkenylphosphatidylethanolamine and alkyldiacylglycerol. A panel or kit or composition for use in determining a subject's risk of developing AD or for use in assessing or monitoring Aβ levels in a subject. A method of prophylaxis comprising determining a subject's risk of developing AD as per (i) and providing therapeutic or behavioral intervention. A method of AD treatment comprising assessing AD as per (ii) and providing AD treatment. A method or process of determining cognition in a subject, by detecting a level of a plurality of lipid species selected from DE, Hex3Cer, and TG(O).

Description

DEMENTIA RISK ANALYSIS

FIELD

The present disclosure relates generally to the field of circulating lipid biomarkers and their identity and profiling for populations who are vulnerable to dementia. The results of testing will inform analyses and should contribute to better diagnostic, prognostic and treatment outcomes. In one aspect the specification relates to methods and kits for assessing risk including a change in risk for developing dementia. In another aspect, the specification relates to methods and kits for assessing or monitoring dementia status, cognitive ability, and/or amyloid b (Ab) burden in a subject. In another aspect, methods of prophylaxis and treatment are also contemplated.

BACKGROUND

Bibliographic details of references referred to by author name in the subject specification are listed at the end of the specification.

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

Alzheimer’s disease (AD) is a neurodegenerative disorder characterised by the progressive loss of cognitive function resulting in dementia. Mild cognitive impairment (MCI) characterises an intermediate stage between the expected cognitive decline of normal ageing and the more serious decline of dementia. People with MCI are at increased risk of developing AD. There is currently no cure for AD and treatment options to halt or slow progression are limited. A key factor in the development and implementation of such therapies will be the early identification and characterisation of affected individuals and of individuals likely to become affected.

Alzheimer’s disease (AD) causes the predominant form of dementia in the older people and is reaching epidemic proportions. In the sporadic form of AD, symptoms begin to manifest after the age of 65, and with the ageing global population, the number of people with AD worldwide has been estimated to reach 81 million by 2040. This increased global incidence in AD means an increased demand for dementia care and an associated increased economic burden. Current and proposed treatment regimens focus on early intervention to prevent widespread neuronal cell death.

Lipids are among the least studied biomarkers and lipid profiling on the whole is in its infancy compared to the study of larger biomarkers. Levels of certain plasma lipid species have been associated with the presence of dementia but limited prospective studies and relatively small datasets have hampered appropriate data collection, management and interpretation.

Accordingly, there is an urgent need to identify individuals who are at greater risk for the onset of dementia to facilitate early interventions. This ideally requires a prognosis of AD in both healthy subjects who lack MCI and in subjects with MCI. There is also a need for non-imaging based methods for diagnosing or assessing AD or features of AD such as cognition in a subject and for identifying amyloid b (Ab) burden inter alia to facilitate assessment or diagnosis of Alzheimer's disease and disease pathogenesis.

SUMMARY

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 group of elements or integers but not the exclusion of any other element or integer or group of elements or integers. By "consisting of" is meant including, and limited to, whatever follows the phrase "consisting of". Thus, the phrase "consisting of" indicates that the listed elements are required or mandatory, and that no other elements may be present. By "consisting essentially of" is meant including any elements listed after the phrase, and limited to other elements that do not interfere with or contribute to the activity or action specified in the disclosure for the listed elements.

As used herein the singular forms "a", "an" and "the" include plural aspects unless the context clearly dictates otherwise. Thus, for example, reference to "a lipid species" includes a single lipid species, as well as two or more lipid species; reference to "an agent" includes one agent, as well as two or more agents; reference to "the disclosure" includes single and multiple aspects of the disclosure and so forth. Reference to "at least X or Y" means "at least X or at least X."

In one aspect, the disclosure enables a method of determining the presence, or risk of developing Alzheimer's Disease (AD) in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species wherein at least one or at least two lipid species are selected from one or two or more lipid classes or subclasses from GM3 (GM3 ganglioside), Hex3Cer (trihexosylcer amide), DE (dehydrocholesteryl ester), and TG(O) (alkyl -diacylglycerol);

(ii) comparing the levels or ratios of levels of the lipid species detected from (i) to reference levels or ratios of the lipid species; and

(iii) determining whether the subject has, or is likely or not to develop AD on the basis of the comparison.

In one embodiment, the plurality of lipid species comprises at least one or two or more lipid species set out in Table (a) or at least one or two lipid species set out in Table B (i) or (ii).

In another embodiment, the plurality of lipid species further comprises one or more lipid species selected from Tables C and/or Table K. In one embodiment, lipid species are selected from the most significant (top) species shown in Figure 1 (relevant for assessment of AD) and/or Figure 2 (relevant for assessing risk of AD onset). As determined herein, lipid species which are informative regarding prevalent AD may also be informative for determining an individual's risk of developing AD, that is some lipid species fall into both diagnostic groups. Flowever, as also determined herein, lipid species are also specific for the presence of dementia or specific for the risk of developing AD. The present specification discloses which lipid species are useful for which test/assay and permits the selection of lipid species and panels of lipid markers that can discriminate between healthy, prevalent and incident AD in a sample from a subject: whether a subject has dementia and/or their risk of developing dementia.

In one embodiment, the present disclosure employs a combinatorial approach to provide enhanced prognostic, diagnostic or monitoring capabilities. In this regard, the number of lipid species detected or compared may conveniently be between 2 and 5 or between 5 and 10 or 5 and 15 or 5 and 20. A large number of different lipid species were newly identified as associated with AD (prevalent) or associated with the risk of developing AD (incident) (see for example Table 21) and therefore many different sets of between 2 and 5 or between 5 and 10 or 5 and 15 or 5 and 20 lipid species are clearly encompassed. Greater lipid numbers were not found to provide further prognostic value for determining risk of developing AD but could be included as positive or negative controls. The number of lipid classes or subclasses may conveniently be between 1 and 5 or between 1 and 10 lipid classes or subclasses. In one embodiment, the number of lipid classes or subclasses is 2, 3, or 4. For example, in one embodiment the lipid class or subclass is selected from Hex3Cer., DE and TG(O), or DE and TG(O), Hex3Cer., or DE or TG(O). Multiple species within these classes have herein been shown to be positively associated with prevalent AD or risk of AD. As determined herein, additional specific lipid species include without limitation a sphingolipid species containing nervonic acid (24:1) and/or PE(P-l8:0/22:5), or a ceramide selected from Table 9a), Table H or Table A. For example, GM3 gangliosides are identified as plasma biomarkers of AD and may be used as described herein in combination with lipid species selected from further lipid classes or subclasses such as one or more of those listed in Table T, to provide prognostic value. In addition, newly identified lipid species may be employed in the subject methods and kits.

Reference to "incident AD lipid biomarkers" or "lipid species associated with risk of AD onset" or similar means lipid species identified herein as significantly associated, either alone or in combination, with subsequent onset of AD in a test population. "Significance" will be understood to those of skill in the art, and a base line of significance is set herein, as p< 0.05 (with or without correcting of multiple comparisons or covariates) using well established statistical methods. Reference to "prevalent AD lipid biomarkers" or "lipid species whose levels are associated with presence of AD" or similar means lipid species identified herein as significantly associated, either alone or in combination, with the presence of AD in a test population.

In one embodiment, a method is enabled for determining the risk of developing AD in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species wherein at least one or two lipid species are selected from one or two or more lipid classes or subclasses from GM3, DE and TG(O)

(iii) determining whether the subject is likely or not to develop AD on the basis of the comparison.

In another embodiment, a method is enabled for determining the presence of AD in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species wherein at least one or two lipid species are selected from one or more lipid classes or subclasses from GM3, DE, Hex3Cer, and TG(O)

(iii) determining whether the subject has AD on the basis of the comparison.

This method may be used for monitoring or assessing AD in a subject.

In another aspect is provided a method of determining the severity of Alzheimer's Disease (AD) or ab burden in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species comprising one or two or more lipid species from one or two lipid classes or subclasses of ether lipids including alkylphosphatidylcholine, alkenylphosphatidylcholine, lysoalkylphosphatidylcholine, alkylphosphatidylethanolamine alkenylphosphatidylethanolamine, lysophosphatidylethanolamine, lysoalkenylphosphatidylethanolamine and alkyldiacylglycerol;

(iii) determining the severity of AD or ab burden on the basis of the comparison.

In one embodiment, lipid species for use in the above method are also present in Table (a) or Table B (i) or Table B (ii) or are lipid species previously undisclosed as identified in Table 21. As determined herein, lipid species shown in Tables 19 (comparing Healthy controls, MCI subjects and AD subjects) and Table 20 (comparing Healthy control and MCI subjects) are significantly associated with cognitive ability (select lipid species with p< 0.05 with no adjustment for covariants or p< 0.05 with adjustment for covariants) with cognitive ability. Thus either lipid based tests disclosed herein or cognitive ability tests, such as the California Verbal Learning Test employed, provide a direct measure in a reciprocal fashion independent of MCI or AD status.

In another embodiment therefore, a method is enabled for determining a lipid based measure of cognitive ability, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species wherein at least one or two lipid species are selected from one or more lipid classes or subclasses from GM3, DE, Hex3Cer, and TG(O);

(iii) determining a lipid based measure of cognition or whether the subject has increased or decreased cognition relative to a control on the basis of the comparison.

In one embodiment, the plurality of lipid species are selected from lipid species showing a significant association with cognition set out in Table 19 or 20.

In one embodiment the method for determining a lipid based measure of cognitive ability (cognition) or a change in cognition in a subject, comprises:

(i) detecting in a biological sample from the subject a level of a plurality of lipid species selected from lipid species showing a significant association with cognition set out in Table 19 or 20.

In one embodiment, lipid species for use in the above method are also present in Table (a) or Table B (i) or Table B (ii) or are lipid species previously undisclosed as identified in Table 21.

Also provided are a use of one or more lipid species identified herein as significantly associated with the presence of AD or the risk of developing AD in the manufacture of a composition, of an assay or kit to assess a subject's risk of developing AD or to diagnose AD or monitor AD progression or cognition. Panels or kits or compositions comprising combinations of two or more lipid species and, in some embodiments, including newly identified lipid species shown herein to be useful for prognosis of AD are enabled.

Panels or kits or compositions comprising combinations of lipid species including newly identified lipid species useful for assessment or monitoring of AD in a biological sample from a subject diagnosed with AD are enabled.

Panels or kits or compositions comprising combinations of two or more lipid species including newly identified lipid species useful for assessing Ab burden and/or cognition in a biological sample such as a blood sample from a subject are enabled.

In one embodiment, lipid species are labelled or tagged to facilitate identification and/or quantification.

In one embodiment, the specification provides a data collection method comprising detecting the level of one or more lipid species or one or more lipid classes or subclasses identified herein as incident or prevalent AD associated lipid species or classes or subclasses.

In one embodiment, the specification enables a data collection method comprising detecting in a biological sample from a subject a level of at least two incident or prevalent AD lipids as identified herein.

In one embodiment, lipid species for use in the above methods are also present in Table (a) or Table B (i) or Table B (ii) or are lipid species that are previously undisclosed as identified in Table 21. In one embodiment the application provides a composition comprising one or more of the isolated lipid species set out in Table B(i).

In one embodiment, the specification enables a data collection method comprising detecting in a biological sample from a subject a level of at least two lipid species identified herein to be associated with Ab burden and/or cognition in a subject.

In one embodiment, the subject-identified lipid species may be employed in kits in the form of particular concentrations of lipid species for use as a control, such as a quantification control or a positive and/or negative control.

Methods of prophylaxis or treatment are further contemplated as are the use of the herein described combinations of lipids species in the manufacture of kits, panels, compositions and assays for data collection, prognosis or assessment of AD or cognition in a subject or for selecting appropriate treatments. The above summary is not and should not be seen in any way as an exhaustive recitation of all embodiments of the present invention.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1: Association of lipid species with prevalent AD. Left panel- Logistic regression of lipid species against Alzheimer’s disease was performed without adjustment for covariates. Grey - non-significant after multiple comparison correction, orange - significant after multiple comparison correction. Red - Top 9 species ranked by lowest p-value: species whose levels are elevated in AD from the following: GM3, PE, acylcarnitine 16:1; species whose levels are significantly reduced in AD selected from; SM, PC, PE(O), and PE(P). Right panel- Logistic regression of lipid species against Alzheimer’s disease was performed with adjustment for age, BMI, sex, site, total cholesterol, HDL, triglycerides and ApoE4 status. Grey - non- significant after multiple comparison correction, Green - significant after multiple comparison correction. Red - Top 8 species ranked by lowest p-value are from Cer, GM3, PE and PC.

Figure 2: Association of lipid species with Incident AD. Left panel- Logistic regression of lipid species against incident Alzheimer’s disease was performed without adjustment for covariates. Grey - non-significant after multiple comparison correction, purple - significant after multiple comparison correction. Red - Top 8 species ranked by lowest p-value selected from ceramides including 24:1, PC(O), PE(O) and acylcarnitine 16:1 species. Right panel- Logistic regression of lipid species against incident Alzheimer’s disease was performed with adjustment for age, BMI, sex, site, total cholesterol, HDL, triglycerides and ApoE4 status. Grey - non- significant after multiple comparison correction.

BRIEF DESCRIPTION OF THE TABLES

Table (a) lists lipid species significantly associated (p< 0.05) with incident or prevalent AD not previously associated with the presence of AD. See also Table 21.

Table A lists lipid species whose levels are significantly associated (p< 0.05) with incident AD selected from Table (a) as identified in Table 6 and Table 21.

Table B(i) and B(ii) - B(i) lists newly identified lipid species significantly associated (p<0.05) incident AD; B(ii) lists newly identified lipid species significantly associated (p<0.05) prevalent AD. See also Table 21.

Table C lists lipid species significantly associated with incident AD (p< 0.05).

Table D lists the lipid species identified as members of prognostic groups of lipid species. Models were created to predict incident AD using only lipid species performed very well with the optimism corrected model containing 10 lipid species showing a C-statistic of 0.78 (95% Cl 0.72- 0.82) (lipid species 1 to 10 in Table E). The NRI for the same model was 0.79 (95% Cl 0.56-0.99). Only marginal improvements were observed for the same model containing 15 lipid species (comprising additionally lipid species identified as 11 to 15 in Table E)). Lipid species in this model included phospholipids (diacyl, ether and lyso species), sphingolipids as well as DG(l4:0_l6:0) and dehydrocholesterol(l8:l) (Taken from Table 13). Table E: as for Table D however lipid species are taken from Table 16 where a predictive combinatorial model was developed to predict the risk of AD onset independent of age. Panels comprising lipids identified as numbers 1 to 10 were increasingly prognostic by a number of performance metrics. The model was slightly improved by including lipids 1 to 15 or 1 to 20.

Table F: lipid species in number order taken from Table 17.

Table G: lipid species in number order taken from Table 18.

Table H lipid species useful for assessing prevalent AD as shown in Table 4 and Table 21 (p< 0.05) not previously associated with AD.

Table I lipid species useful for assessing prevalent AD as shown in Table 4 and Table 21 (p< 0.05).

Table Q lists lipid species that are significantly associated in a population with a risk for developing AD independently of any risk due to age, sex, BMI, total cholesterol, HDL-C, triglycerides, and ApoE4 genotype. Species are listed in decreasing order of significance (p< 0.05). (Taken from Table 6).

Table R: lists lipid species that are significantly associated in an age-matched sub population with a risk for developing AD independently of any risk due to age, sex, BMI, total cholesterol, HDL-C, triglycerides, and ApoE4 genotype. Species are listed in decreasing order of significance. The association was significant when values were adjusted for multiple comparisons by the method of Benjamin-Hochberg for PE (P-l8;0/22:5) (n6). (Taken from Table 8).

Table S lists lipid species whose levels are positively associated with onset of AD after correction for multiple comparisons and which displayed strong prognostic value when coupled as a ratio with any of the lipid species listed in Table T. Taken from Table 12.

Table T lists lipid species whose levels are negatively associated with onset of AD after correction for multiple comparisons and which displayed strong prognostic value when coupled as a ratio with any of the lipid species listed in Table S. Taken from Table 12.

Table U lipid species useful for assessing prevalent AD as shown in Table 4.

Table V : lipid species identified by comparing lipid species with SUVr status for assessing Ab burden when adjusted for covariates although not significant when corrected for multiple comparisons. Taken from Table 10.

Table W: lipid species useful for assessing Ab burden. Taken from Table 11.

Table X: lipid species useful for assessing Ab significant based upon multiple comparisons. Taken from Table 11.

Table 1: Baseline characteristics of the AIBL cohort for which plasma samples were available - Table Index: ¹ Values expressed as mean (standard deviation); ² Values expressed as number (% count). Table 2: Conditions for tandem mass spectrometry analysis of lipid species. - Table Index: ¹ PI, product ion; NL, neutral loss.

Table 3: Characteristics of the prevalent Alzheimer’s disease groups in the AIBL study at the 18 month time point. - Table Index: ¹ p-values are based on t-test (continuous) or chi-square test (categorical); ² Values expressed as mean (standard deviation); ³ Values expressed as number (% count).

Table 4: Association of lipid species with prevalent Alzheimer's disease. - Table Index: ¹ Logistic regression was performed on the 18 month time point samples (n=693 non-AD, n= 191 AD); ² Adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site; ³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0004<p<0.0l; green indicated 0.0l<p<0.05; ⁴ The p-values were corrected for multiple comparisons by the method of Benjamini-Hochberg.

Table 5: Baseline characteristics of convertors and non-convertors to Alzheimer's disease in the AIBL study. - Table Index: ¹ p-values are based on t-test (continuous) or chi-square test (categorical); ² Values expressed as mean (standard deviation); ³ Values expressed as number (% count).

Table 6: Association of lipid species with incident Alzheimer's disease. - Table Index: ¹ Logistic regression was performed on the baseline samples with six years follow up (h=515 no conversion to AD, n= 68 conversion to AD); ² Adjusted for age, sex, BMI, total cholesterol, HDL- C, triglycerides, ApoE4 genotype and collection site; ³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0004<p<0.0l; green indicated 0.0l<p<0.05; ⁴ The p-values were corrected for multiple comparisons by the method of Benjamini-Hochberg.

Table 7: Baseline characteristics of age matched convertors and non-convertors to Alzheimer's disease in the AIBL study. - Table Index: ¹ p-values are based on t-test (continuous) or chi-square test (categorical); ² Values expressed as mean (standard deviation); ³ Values expressed as number (% count).

Table 8: Association of lipid species with aged matched incident Alzheimer's disease subgroups. - Table Index: ¹ Logistic regression was performed on the age matched baseline samples with six years follow up (n=l30 no conversion to AD, n= 66 conversion to AD); ² Adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site;³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0004<p<0.0l; green indicated 0.0l<p<0.05; ⁴ The p-values were corrected for multiple comparisons by the method of Benjamini-Hochberg. Table 9: Baseline characteristics of participants with SUVr measurements in the AIBL study. - Table Index: ¹ p-values are based on t-test (continuous) or chi-square test (categorical); ² Values expressed as mean (standard deviation); ³ Values expressed as number (% count)

Table 10: Association of lipid species with Standard Uptake Value ratio (SUVr) status. - Table Index: ¹ Logistic regression was performed on the baseline samples (n=H5 SUVr < 1.4, n= 127 SUVr >1.4); ² Adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype, AD status and collection site; ³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0004<p<0.0l; green indicated 0.0l<p<0.05; ⁴ The p-values were corrected for multiple comparisons by the method of Benjamini-Hochberg.

Table 11: Association of lipid species with Standard Uptake Value ratio (SUVr) value. - Table Index: ¹ Linear regression was performed on the baseline samples (n=242); ² Adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype, AD status and collection site; ³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0001 <p<0.0l; green indicated 0.0l<p<0.05; ⁴ The p-values were corrected for multiple comparisons by the method of B enj amini-Hochberg.

Table 12: Association of lipid ratios with incident Alzheimer's disease. - Table Index: ¹ Logistic regression was performed on the baseline samples with six years follow up (h=515 no conversion to AD, n= 68 conversion to AD); ² Adjusted for age, sex, BMI, total cholesterol, HDL- C, triglycerides, ApoE4 genotype and collection site; ³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0001<r<0.01; green indicated 0.0l<p<0.05; ⁴ The p-values were corrected for multiple comparisons by the method of Benjamini-Hochberg. Ratios were generated from the top 55 lipids (p < 0.05, uncorrected no adjustment analysis, Table 6).

Table 13: Performance measures of multivariate models using lipid species only to predict Alzheimer's disease.

Table 14: Performance measures of multivariate models using lipid species, age and ApoE4 to predict Alzheimer's disease.

Table 15: Performance measures of multivariate models using lipid species, upon a base model of age and ApoE4, to predict Alzheimer's disease.

Table 16: Performance measures of multivariate models using lipid species only to predict Alzheimer's disease in an age and sex matched subcohort.

Table 17: Performance measures of multivariate models using lipid species and ApoE4 to predict Alzheimer's disease in an age and sex matched sub cohort.

Table 18: Performance measures of multivariate models using lipid species, upon a base model of age and ApoE4, to predict Alzheimer's disease in an age and sex matched sub cohort. Table 19: Association of lipid species with California Verbal Learning Test score (SUM) in healthy control MCI and AD patients - Table Index: ¹ Linear regression was performed on the entire baseline samples (n=l035); ² Adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype collection site and education level ³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0004<p<0.0l; green indicated 0.0l<p<0.05; ⁴ The p- values were corrected for multiple comparisons by the method of Benjamini-Hochberg.

Table 20: Association of lipid species with California Verbal Learning Test score (SUM) in healthy control and MCI patients - Table Index: ¹ Linear regression was performed on the HC and MCI baseline samples (n=864); ² Adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype collection site and education level; ³ Colour coding is as follows: red indicates p<0.000l; yellow indicates 0.0004<p<0.0l; green indicated 0.0l<p<0.05; ⁴ The p- values were corrected for multiple comparisons by the method of Benjamini-Hochberg.

Table 21: Lipid species associated with prevalent or incident Alzheimer's disease (p<0.05 unadjusted) - Table Index: ¹ Associated with prevalent AD (p<0.05, no adjustments) as shown in Table 4; ² Associated with incident AD (p<0.05, no adjustments) as shown in Table 6; ³ Previously reported to be associated with AD either in plasma or tissue; ⁴ Lipid species listed in the LIPID MAPS database (www.lipidmaps.org/data/structure/indexphp). Colour coding: Red indicates lipid species associated with prevalent and/or incident AD and not previously reported as such; Green indicates lipid species associated with prevalent AD; Orange indicates lipid species associated with incident AD.

LIST OF ABBREVIATIONS

List of Lipid classes and subclasses

DETAILED DESCRIPTION OF CERTAIN EMBODIMENTS

The naming convention for lipids used here follows the guidelines established by the Lipid Maps Consortium and the shorthand notation of Liebisch et al. (Liebisch et al., Fahy et. al. (2009), Fahy et al. ( 2005)]. Phospholipids typically contain two fatty acid chains and in the absence of detailed characterisation are expressed as the sum composition of carbon atoms and double bonds (i.e. PC(38:6). Flowever, where an acyl chain composition has been determined the naming convention indicates this (i.e. PC(38:6) is changed to PC(l6:0_22:6)). This is also extended into other lipid classes or subclasses. Species separated chromatographically but incompletely characterised were labelled with an (a) or (b), for example PC(P-l7:0/20:4) (a) and (b) where (a) and (b) represent the elution order.

The subject disclosure includes newly identified lipid species esterified with methylhexadecanoic acid (MHDA) including individual species of PE, LPC and PC. The new lipid species identified herein are claimed and are identified in Table 21 herein.

Reference to "risk for developing AD" includes assessing a subject's risk of developing AD in a blood sample from the subject at one or several time points and includes assessing a change in the lipid profile and therefore risk over time or during a treatment. Assessing risk in a subject equates to stratifying a subject as having a circulating lipid profile that increases or decreases the likelihood of AD or correlates with or provides a likelihood or degree of susceptibility or non-susceptibility that a subject will develop (i.e. a prognosis of) AD.

The methods described are suitable for use in a wide context and include determining the likelihood that a dementia treatment will reduce the risk of AD in a subject or to monitor the effect of a dementia treatment, or to select a subject for a dementia treatment.

The term "level" is used herein to include a concentration, relative amount, or other signature indicative of an amount or relative amount of a lipid species present in a blood or plasma sample obtained from an individual at given time point. The level of a lipid species may be determined directly such as by liquid chromatography electrospray ionization and tandem mass spectrometry using internal standards or indirectly, such as by assessing the level of a reporter molecule that is bound to a lipid species.

Reference to a "biological sample" includes a biological material such as whole blood, serum, plasma or other biological fluid, or a solid or semi-solid sample. Samples may be diluted or concentrated, treated or untreated. For applications such as nuclear magnetic resonance (NMR) spectroscopy and mass spectroscopy (MS) together with liquid or gas chromatography, samples are treated for lipid extraction prior to analysis. Blood samples may be dried down onto solid supports in for example, kits or panels prior to extraction. A high performance or ultra performance liquid chromatographic method might use a Cl 8 column that will separate out polar components at the solvent front with more non-polar/hydrophobic components at the end of the run. Serial analysis by MS provides increasing more detail on the identity of spectral fragments that may be previously undescribed.

The term "reference ratio or level" includes "control ratio or level" and includes data or a control a skilled person would use to facilitate the accurate interpretation of technical data. In one embodiment for example reference ratios or levels are obtained from an individual at an earlier or later time point than for test ratios or levels. Thus reference levels may be pre-determined or pre- selected. Reference levels may be expressed as a median, mode or mean level or range from an individual or a cohort or population of subjects or a mean together with standard deviation to determine suitable threshold or cut off levels. Reference levels may be expressed in any form such as by concentration, molality or a value associated with diagnostic or prognostic cut off levels or ranges. Reference levels or ranges may be levels determined from individuals or populations of individuals who do not have AD, and/or they may be levels or ranges determined from subjects or populations of subjects with AD.

Reference to a "subject" or "individual" includes any human, primate, mammalian or other species of veterinary importance, or test organism known to the skilled person.

Depending upon the embodiment or analysis format elected, as few as one or two lipid species may be employed in a kit or its/their level determined for the method. Alternatively, as described herein a focussed selection of lipid species (i.e., a panel or a plurality of lipid species) may be analysed. In one embodiment, the method comprises determining or determining and comparing the level of at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 or more lipid species. In one embodiment, the method comprises determining or determining and comparing the level of less than 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or less than 20 or 20-30 lipid species. In one embodiment, the method comprises determining or determining and comparing the level of less than 2, 3, 4, 5, 6, 7, 8, 9, or 10 lipid classes or subclasses. A large number of highly significant lipid species have been identified as described in the examples and therefore multiple different tests employing a combination of different lipid species are enabled. For example the combinations set out in Tables D to G are illustrative and the skilled person could select other combinations of 5 to 10 significantly associated lipid species from the tables.

The present disclosure used a multivariate approach to identify that combinations of lipid species can provide useful prognostic information and that the levels of as few as about 9 or 10 lipid species in combination are useful for AD prognosis. Additionally it has been determined that ratios of lipid species, including ratios between species that are upregulated and those that are down regulated compared to controls are highly significantly informative.

In one embodiment, the number of ratios determined or determined and compared is between one and twenty (i.e., 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 or more lipid species ratios.

Slight variations above or below the upper numerical limits defined herein are contemplated where they achieve effectively the same results. While the description provides a focussed approach to lipidomic analyses and identifies the least number of lipids for discriminative analyses, the actual number of lipids assessed may vary depending upon the specific platform for the analysis. For example, a laboratory based LC-MS analysis may detect multiple lipid species simultaneously and may employ less than 100 lipid analytes, while an enzyme or label based analysis might employ less than 10 lipid analytes.

The present disclosure identified 744 ratios that were highly significant on their own to determine AD risk. It is contemplated that combinations of ratios could be employed, optionally together with traditional dementia risk tests to develop improved AD prognosis tests. In many cases, the lipids are informative independent of known risk factors such as age, sex, BMI, etc. Illustrative methods capable of analysing lipid species include classical lipid extraction methods, mass spectrometry together with electrospray ionization and matrix-assisted laser desorption ionisation, with mass analysis such as quadruple and/or TOF (e.g.,. Quadrapole/TOF) or orbitrap mass analysers. Chromatographic methods are used for the separation of lipid mixtures such as gas chromatography, high pressure liquid chromatography (F1PLC), ultra -high pressure liquid chromatography (UF1PLC), capillary electrophoresis (CE). These may be used with mass spectrometry based detection systems or other detectors including optical detectors. Clinical mass spectrometry systems are used by clinical laboratories to provide lipid profiles and ratios upon request. Another suitable technique for quantitative lipid analysis is one or two dimensional nuclear magnetic resonance (NMR). Two dimensional techniques such as heteronuclear single quantum coherence (F1SQC) are suitable for lipid profiling through the ability to elucidate C-Fl bonds within a structure. Any technique capable of identifying individual lipid species in the sample can be used for collecting information on the lipid species. Typically MS is used coupled to a separation method such as various forms of chromatography.

In one embodiment enzymatic methods known in the art may be used to identify lipid classes or subclasses and/or species.

Lipid level data may be processed to produce a report of levels and/or ratios. In one embodiment, lipid data are processed as described herein to identify and/or report the risk that a subject will develop dementia or to report the dementia/Ab burden in the subject.

The methods enabled herein permit integration into pathology architecture or platform systems.

For example, the method described herein allows a user or client to determine AD/A /cognition status or AD risk status of an individual, the method including: (a) receiving data in the form of lipid levels, relative lipid levels or signature profiles developed from an individual's plasma or blood sample from the user via a communications network; (b) processing the individuals data via an algorithm which provides one or more status/risk value/s by comparing levels and/or ratios of lipid levels to those from one or more reference levels or ratios.

In some embodiments, an indication of the AD/A /cognition status or AD risk status transmitted to the user is transferred via a communications network. 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. When a server is used, it is generally a client server or more particularly a simple object application protocol (SOAP).

In one embodiment, the method is suitable to be practised as a home test kit or point-of- care method typically employing a device suitable for home use or point of care.

The kits or panels can be used in a laboratory or in a home use test kit. Blood, for example may be dried down onto a support material suitable for analysis at home or sent to a laboratory for analysis.

Biosensor technologies that permit less expensive equipment or fewer trained personnel are available for developing devices for lipid species analysis that may be used at point of care. This is particularly useful when as here a small number of lipid species can provide prognostic or monitoring data. Biosensors which recognise a target molecule and produce a measurable or observable signal may be for example, optical, electrochemical or mechanical biosensors. Assays that use a label indirectly measure the binding of an analyte lipid to a target molecule using a reporter molecule as an indication of binding and amount. Label free assays measure signal changes directly associated with target binding or cellular processes. Examples of label free optical sensors include surface plasmon resonance sensing (SPR), Interferometry (such as backscattering inferometry (BSI), ellipsometry, and assays based on UV absorption of lipid-functionalized gold nanorods. In optical assays using labels, the target lipid molecule is immobilized on the surface of a biosensor and then probed with a binding agent, such as an antibody couples to a label (many labels are known to the skilled person such as a flurophore, quantum dot, radioisotope, enzyme). Reference may be made to Sakamuri et al. "Detection of stealthy small amphiphilic biomarkers Journal of Microbiological Methods 103 : 112-117, 2014. These authors have used a waveguide based biosensor measuring only surface attached fluorescence antibody signals to detect lipids and amphiphilic targets in biological samples. Electrochemical sensors use an electrode to directly detect a reaction, typically a current from electron transfer during binding of an analyte and a chemically functionalized surface. Potentiometric sensors usefully measure charge accumulation to detect lipid antigens such as amphiphilic cholesterol using lipid films. Mechanical sensors are ideal for clinical applications and include cantilever and quartz crystal microbalances (QCM). The later detects changes in resonance frequency on the sensor surface from increased mass due to analyte binding.

In one embodiment the method is an enzyme -linked immunosorbent (ELISA) -type, flow cytometry, bead array, lateral flow, cartridge, microfluidic or immunochromatographic or enzyme- substrate based method or the like. Typically, such methods employ binding agents such as an antibody or an antigen-binding fragment thereof. Other suitable binding agents are known in the art and include antigen binding constructs such as affimers, aptamers, or suitable ligands (receptors) or parts thereof.

Antibodies, such as monoclonal antibodies, or derivatives or analogs thereof, include without limitation: Fv fragments; single chain Fv (scFv) fragments; Fab' fragments; F(ab')2 fragments; humanized antibodies and antibody fragments; camelized antibodies and antibody fragments, and multivalent versions of the foregoing. Multivalent binding reagents also may be used, as appropriate, including without limitation: monospecific or bispecific antibodies; such as disulfide stabilized Fv fragments, scFv tandems (scFv) fragments, diabodies, tribodies or tetrabodies, which typically are covalently linked or otherwise stabilized (i.e. leucine zipper or helix stabilized) scFv fragments.

Methods of making antigen -specific binding agents, including antibodies and their derivatives and analogs and aptamers, are well-known in the art. Polyclonal antibodies can be generated by immunization of an animal. Monoclonal antibodies can be prepared according to standard (hybridoma) methodology. Antibody derivatives and analogs, including humanized antibodies can be prepared recombinantly by isolating a DNA fragment from DNA encoding a monoclonal antibody and subcloning the appropriate V regions into an appropriate expression vector according to standard methods. Phage display and aptamer technology is described in the literature and permit in vitro clonal amplification of antigen-specific binding reagents with very affinity low cross-reactivity. Phage display reagents and systems are available commercially, and include the Recombinant Phage Antibody System (RPAS), commercially available from Amersham Pharmacia Biotech, Inc. of Piscataway, New Jersey and the pSKAN Phagemid Display System, commercially available from MoBiTec, LLC of Marco Island, Florida. Aptamer technology is described for example and without limitation in US Patent Nos. 5,270,163; 5,475,096; 5,840,867 and 6,544,776.

Methods of determining the presence, or risk of developing Alzheimer's Disease (AD)

In one embodiment the present disclosure enables a method of determining the presence of AD or the a risk of developing Alzheimer's Disease (AD), the method comprising

(i) detecting in a biological sample from the subject a level of at least two lipid species selected from at least one lipid classes or subclasses or sub class from ceramide, monohexosylceramide, GM3 ganglioside, sphingomyelin, phosphatidylcholine, alkylphosphatidylcholine, alkenylphosphatidylcholine, lysophosphatidylcholine, phosphatidylethanolamine, alkylphosphatidylethanolamine, alkenylphosphatidylethanolamine, lysophosphatidylethanolamine, acylcarnitine and triacylglycerol, or a level of at least two lipid species with one selected from the lipid classes or subclasses; ceramide, monohexosylceramide, GM3 ganglioside, lysophosphatidylcholine, phosphatidylethanolamine, lysophosphatidylethanolamine, acylcarnitine, triacylglycerol, and the second lipid species selected from the lipid classes or subclasses; sphingomyelin, phosphatidylcholine, alkylphosphatidylcholine, alkylphosphatidylethanolamine and alkenylphosphatidylcholine; or

(i) detecting in a biological sample from the subject a level of at least two lipid species selected from at least one lipid class or subclass from ceramide, monohexosylceramide, GM3 ganglioside, sphingomyelin, phosphatidylcholine, alkylphosphatidylcholine, alkenylphosphatidylcholine, lysophosphatidylcholine, phosphatidylethanolamine, alkylphosphatidylethanolamine, alkenylphosphatidylethanolamine, lysophosphatidylethanolamine, phosphatidylinositol, lysophosphatidylinositol, phosphatidylglycerol, dehydrocholesterol ester acylcarnitine, triacylglycerol and alkyldiacylglycerol, or a level of at least two lipid species selected from at least one lipid class or subclass from a ceramide, monohexosylceramide, GM3 ganglioside, sphingomyelin, phosphatidylcholine, alkylphosphatidylcholine, alkenylphosphatidylcholine, lysophosphatidylcholine, phosphatidylethanolamine, alkylphosphatidylethanolamine, alkenylphosphatidylethanolamine, lysophosphatidylethanolamine, phosphatidylinositol, phosphatidylglycerol, dehydrocholesterol ester acylcarnitine and triacylglycerol; and

(ii) comparing the levels of the lipid species detected from (i) to reference levels of the lipid species; and

(iii) determining whether the subject has AD or is likely or not to develop AD on the basis of the comparison.

In one embodiment, the disclosure enables a method of determining the presence, or risk of developing Alzheimer's Disease (AD) in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species wherein at least one or two lipid species are selected from one or two or more lipid classes or subclasses from GM3 (GM3 ganglioside), Hex3Cer (trihexosylcer amide), DE (dehydrocholesteryl ester), and TG(O) (alkyl -diacylglycerol);

In one embodiment, the plurality of lipid species comprises at least one or two or more lipid species set out in Table (a) and/or at least one or two lipid species set out in Table B (i) and (ii). Table (a)

Table B(i)

Table B (ii)

In one embodiment, the plurality of lipid species further comprises one or more lipid species selected from Tables C and/or Table I.

Table C

6

Table I

9

0

1

As determined herein specific ratios of lipid species provided within a single lipid class or subclass or between classes/subclasses provide significant prognostic or diagnostic value. Whilst not being limited to specific mode of action, the results and ratios identify lipid metabolic pathways associated with risk of developing AD or prevalence/severity of AD and thus provide targets for the development of therapeutic agents.

There are several interpretations that can be made for lipid species that show opposing associations with a specific endpoint. One interpretation is that the lipid species represent substrate and product across a specific enzyme that may be either down regulated or upregulated. The substrate and product would then be expected to show opposing directions of change associated with the altered regulation of the enzyme. One such example in this data is the opposing associations of Cer(dl 8:0/20:0) and Cer(dl 8: 1/20:0) which are substrate and produce for the enzyme dihydroceramide desaturase and show opposite odds ratios for prevalent AD of 0.85 (95%CI 0.72-1.00) and 1.64 (95%CI 1.39-1.94) respectively (Table 4).

A second interpretation is that the species represent effects of two distinct pathways that are differentially regulated in the disease state (or in those at risk of disease). These pathways may converge into the same pathway at some point. For example PE(l6:0_20:4) has an OR of 1.51 (95%CI 1.27-1.78) and PE(P-l6:0/20:4) has an OR of 0.70 (95%CI 0.59-0.83), although these are similar structures and are both synthesised using some common enzymes, they are derived from different starting pathways with the plasmalogen precursor (alkyl -acylglycerol) being synthesised in the peroxisome while the diacylglycerol (precursor to the PE species is synthesised in the endoplasmic reticulum.

A third option is that the different lipid species simply represent two entirely different pathways that may not intersect at all as is the case for Cer(dl 8:0/22:0) (negatively associated with prevalent AD) and DG(l8:l_l8:l) (positively associated with prevalent AD (Table 4).

In one embodiment, the detecting step (i) for determining risk of AD onset or prevalent AD comprises detecting at least one lipid species whose level is positively associated with prevalent AD or risk of AD onset from Table 4 or 6, respectively, and at least one lipid species whose level is negatively associated with prevalent AD or AD onset from Table 4 or 6, respectively; and step (ii) comprises comparing one or more ratios of positively and negatively associated lipid species to one or more reference ratios.

In one embodiment, at least one lipid species detected is a sphingolipid species containing nervonic acid (24:1).

In one embodiment, wherein at least one lipid species detected is a ganglioside GM3.

In one embodiment, at least one lipid species is a ceramide selected form Table (a), Table J or Table A. Methods of assessing risk of developing AD (i.e., risk of incident AD)

In one embodiment, the present disclosure enables kits and methods for assessing a subject for risk of developing AD. Subjects may be stratified as likely or not likely to develop AD based upon their lipid profile.

In one embodiment, the present methods or kits differentiate determine risk of incident AD independent of MCI status. That is, the subject to be tested may have no symptoms of dementia or MCI and the method is able to determine risk of incident AD for either subject. That is, a subject with MCI may be determined to have a low risk of incident AD according to the present methods or a person with no signs of dementia may be determined to be at risk of incident AD.

In one particular example, levels of alkenyl phosphatidylethanolamine (plasmalogen) PE (P-l8:0/22:5) alone were able to detect incident (prognose onset of ) AD in the presence of multiple risk factors including age, sex, BMI, total cholesterol, HDL-C, triglycerides and ApoE4 genotype. This lipid species was informative even after correction for multiple comparisons using the Benjamini-Hockberg (BH) method as described in the examples.

In another particular example, the subject prognostic methods and kits comprise the use or detection of lipid species that are in combination able to determining a subject's risk of developing AD independent of traditional risk factors including age, gender, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype. Additionally results were independent of MCI status.

Thus in one embodiment, the present disclosure provides a combinatorial approach to determining the likelihood that an individual will develop AD by detecting specific circulating lipid species in a sample from a subject and comparing their levels or ratios of levels to reference levels or ratios or ranges to determine the individual's risk of AD on the basis of the comparison. Specifically, ratios of levels of pairs of specific lipids were highly informative in this regard. Furthermore, tests were developed comprising assessing levels of between two and about 20 specific lipid species and these were found to be highly informative for incident AD prognosis (i.e., risk of onset) either alone or together with traditional risk factors, where enhanced prognosis was made available as described in the examples.

As the skilled person will appreciate, enhanced performance of a diagnostic or prognostic model is tested using one or more performance metrics including for example, C-statistic, IDI, relative IDI, sensitivity, specificity, accuracy, continuous NRI (events and non-events).

A similar combinatorial approach may be taken for assays designed to assess AD or monitor AD in a subject already diagnosed with AD.

In one embodiment, the lipid classes or subclasses or species are selected from the following lipid classes or subclasses: acylcarnitine, alkenylphosphatidylcholine (PC(P)), alkenylphosphatidylethanolamine (PE(P)), alkyldiacylglycerol (TG(O)), alkylphosphatidylcholine (PC(O)), alkylphosphatidylethanolamine (PE(O)), ceramide (Cer), cholesteryl ester (CE), dehydrocholesterol ester (DE), diacylglycerol (DG), dihexosylceramide (Hex2Cer), dihydroceramide (dhCer), free cholesterol (COH), GM1 Ganglioside (GM1), GM3 Ganglioside (GM3), lysoalkenylphosphatidylcholine (LPC(P)), lysoalkenylphosphatidylethanolamine (LPE(P)), lysoalkylphosphatidylcholine (LPC(O)), lysophosphatidylcholine (LPC), lysophosphatidylethanolamine (LPE), lysophosphatidylinositol (LPI), (monohexosylcer amide (HexCer), phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidylinositol (PI), phosphatidylserine (PS), sphingomyelin (SM), sulfatide, trihexosylcermide (Hex3Cer), and ubiquinone.

In one embodiment, the specification provides a method for determining a subject's risk of developing Alzheimer's Disease (AD), the method comprising

(i) detecting in a biological sample from the subject a level of at least two lipid species selected from at least one or two lipid classes or subclasses from ceramide, phosphatidylcholine, phosphatidylethanolamine, lysophosphatidylethanolamine, and diacylglycerol and , or a level of at least two lipid species selected from at least two lipid classes or subclasses from a ceramide, phosphatidylcholine, phosphatidylethanolamine, lysophosphatidylethanolamine, and diacylglycerol, and a level of at least two lipid species selected from at least one or two lipid classes or subclasses from a sphingomyelin, alkylphosphatidylcholine, alkenylphosphatidylcholine and alkyl phosphatidylethanolamine; and

In one embodiment lipids species are additionally selected from lipid classes or subclasses of phosphatidylethanolamine, alkenylphosphatidylethanolamine, alkylphosphatidylethanolamine, cholesteryl ester, lysophosphatidylcholine, phosphatidylglycerol, and phosphatidylinositol.

In one embodiment the method comprises:

(i) detecting in a biological sample from the subject a level of at least two lipid species selected from a GM3 ganglioside and a lipid class or subclass selected from ceramide, phosphatidylcholine, phosphatidylethanolamine, lysophosphatidylethanolamine, and diacylglycerol, or a level of at least two lipid species selected from a GM3 gangliosides and a lipid class or subclass selected from ceramide, phosphatidylcholine, phosphatidylethanolamine, lysophosphatidylethanolamine, and diacylglycerol, and a level of at least two lipid species selected from at least two lipid classes or subclasses from a sphingomyelin, alkylphosphatidylcholine, alkenylphosphatidylcholine and alkylphosphatidylethanolamine; and (ii) comparing the levels of the lipid species detected from (i) to reference levels of the lipid species; and

In one embodiment, the method comprises detecting in a biological sample from the subject a level of at least 2, 4, 6, 8, 10 or 12 to 20 lipid species selected from one of the lipid models set out in Tables 13 to 16. Kits and panels comprising these lipid models are expressly contemplated. As determined herein up to about 10 lipid species from each panel were found to add significantly to power of the test compared to risk factors such as age and/or ApoE4 genotype. In one embodiment, lipid species selected for use in the above method are also present in Table (a) or Table B (i) or Table B (ii) or are lipid species previously undisclosed as identified in Table 21. Table A

In one embodiment, the description enables a method for determining the risk of developing AD in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species wherein at least one or two lipid species are selected from one or two or more lipid classes or subclasses from GM3, DE and TG(O) (ii) comparing the levels or ratios of levels of the lipid species detected from (i) to reference levels or ratios of the lipid species; and

In one embodiment, the plurality of lipid species comprisies at least 1 to 6 lipid species set out in Table A (above) or at least 1 to 6 lipid species set out in Table B(i) (above)

In a further embodiment, the plurality of lipid species further comprisies one or more lipid species selected from Table C (above).

In one embodiment, the ratios are selected from Table 12.

In one embodiment, the detecting step (i) for determinig risk of AD onset comprises detecting at least one lipid species whose level is positively associated with prevalent AD or risk of AD onset from from Table 4 or 6, respectively, and at least one lipid species whose level is negatively associated with prevalent AD or AD onset from Table 4 or 6, respectively; and step (ii) comprises comparing one or more ratios of positively and negatively associated lipid species to one or more reference ratios.

In one embodiment, at least one lipid species detected is a ganglioside GM3.

In one embodiment, at least one lipid species is a ceramide selected form Table (a), Table J or Table A.

In one embodiment, at least one lipid species detected is PE (P-l8:0/22:5) and/or one ratio is Cer(dl8:l/24:l) | PC(0-l6:0/20:4) and/or acylcarnitine(l6:l) | PC(0-38:5).

As determined herein, comparing the detected levels of the at least two lipid species to their respective reference level from the at least two lipid classes or subclasses can provide an enhanced measure of a subject's risk of developing AD compared to individual lipid species or class comparisons, or compared to traditional risk factors.

In one embodiment, lipid species are selected from lipid species displaying a significant or highly significant (p< 10³) association with AD-onset as shown in Table Q (Table 6) or Table R (Table 8).

In one embodiment, at least 2 to 10 lipid species (e.g., 2, 3, 4, 5, 6, 7, 8, 9, or 10 lipid species) or 2 to 20 lipid species (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 lipid species) are detected in step (i). Table Q ( see also Table 6)

Table R (Table 8)

Accordingly, in one example the specification discloses a method for determining a subject's risk of developing Alzheimer's Disease (AD), the method comprising

(i) detecting in a biological sample from a subject a level of two or more lipid species (such as 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 lipid species) selected from Table Q (Table 6) or Table R (Table 8);

(ii) comparing the level of the one or two or more lipid species or one or two or more ratios of the level of two lipid species detected from (i) to a reference level or ratio for the lipid species; and

In one embodiment, the results from assessing lipid species are combined with one or more traditional risk factors for developing AD. Traditional risk factors for AD include age, sex, BMI, total cholesterol, HDL-C, triglycerides, and ApoE4 genotype.

In one embodiment, models and methods designed to assess traditional risk factors and the herein described prognostic lipid species provide an improved measure of a subject's risk of developing AD compared to traditional risk factors alone.

In another embodiment, the description provides a prognostic method wherein the detecting step (i) comprises detecting at least one lipid species positively associated with AD onset from Table S and at least one lipid species negatively associated with AD onset from Table T; step (ii) comprises comparing one or more ratios of positively and negatively associated lipid species to one or more reference ratios and (iii) determining whether the subject is likely or not to develop AD on the basis of the comparison.

Table S

Table T

In one embodiment, at least one lipid species detected is a ganglioside GM3.

In one embodiment, at least one lipid species detected is PE (P-l8:0/22:5).

In one embodiment, the lipid species/ratio is Cer(dl8:l/24:l) | PC(0-l6:0/20:4) and/or acylcarnitine(l6:l) | PC(0-38:5).

In one embodiment, the pluralities of lipid species are selected from Tables D, E, F and G or combinations thereof.

In one embodiment, lipid species selected for use in the above method are also present in Table (a) or Table B (i) or Table B (ii) or are lipid species previously undisclosed as identified in Table 21.

Table D ( Table 13)

Table E ( Table 16)

Table F ( Table 17)

Table G ( Table 18)

In one embodiment of the method, at least one lipid species is esterified with methylhexadecanoic acid. In one embodiment the lipid species is a PC, LPC or PE -MHDA species. PE species include l5-MHDA_l8:l, _l 8:2, _20:4 and/or _22:6. PC species include 15 MHDA_l8:l, _l 8:2, _20:4 and/or _22:6 species. LPC species include 15-MHDA [snl] or [snl] [l04_snl] or [sn2] species. In one embodiment at least one lipid species is selected from Table B(i).

The present specification also provides a kit, panel or solid support comprising one or more of the prognostic lipid species described herein for use in determining a subject's risk of developing AD according to the herein described prognostic methods. Suitable supports include any material upon which a biological sample can be stored until assessment and include paper, nitrocellulose, chromatographic or absorbent or filtering material.

Kits may optionally further comprise one or more test, control, reagent, or solvent containing portions known in the art, together with instructions for use. In one example, the lipid species are labelled or otherwise tagged to facilitate the quantification of test lipid species in the methods and kits. Suitable labels will be known in the art and are not limited to binding molecules, radiolabels/isotopes, antibody labels or fluorescent labels.

In another aspect, the description provides a method of prophylaxis comprising determining a subject's risk of developing AD according to a prognostic method described herein and exposing the subject to therapeutic or behavioral intervention on the basis that the subject tests positive for risk of developing AD.

Methods of assessing prevalent AD

In another aspect the present specification provides a method for assessing or monitoring AD in a subject with symptoms of dementia, the method comprising

(i) detecting in a biological sample from the subject a level of two or more lipids species from one or two or more lipid classes or subclasses including GM3 Ganglioside, and one or more of phosphatidylethanolamine, ceramide, dihydrocer amide, sphingomyelin, other sphingolipids, and alkenylphosphatidylethanolamine ;

(ii) comparing the level of the two or more lipid species or one or two or more ratios of the level of two lipid species detected from (i) to a reference level or ratio for the lipid species; and (iii) assessing or monitoring dementia/ AD in the subject on the basis of the comparison.

Table H

In one embodiment, the specification provides a method for determining the presence of AD in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species wherein at least one or two lipid species are selected from one or more lipid classes or subclasses from GM3, DE. Hex3Cer, and TG(O)

(iii) determining whether the subject has AD on the basis of the comparison.

In one embodiment, the plurality of lipid species comprises at least 1 to 6 lipid species set out in Table H (above) or at least 1 to 6 lipid species set out in Table B(ii) (above).

In one embodiment, the plurality of lipid species further comprises one or more lipid species selected from Table I (see above).

In another embodiment, the method of assessing or monitoring AD in a subject with symptoms of AD comprises

(i) detecting in a biological sample from the subject a level of two or more lipid species from Table U (Table 4); (ii) comparing the level of the two or more lipid species or one or two or more ratios of the level of two lipid species detected from (i) to a reference level or ratio for the lipid species; and

(iii) assessing or monitoring AD in the subject on the basis of the comparison.

Table U selects lipid species from Table 4 whose levels were significantly associated (p<0.05) with prevalent AD when adjusted for covariates and when adjusted for multiple comparisons. Lipid species may be further selected from those also present in Table (a) and/or Table B(ii).

Table U (Table 4)

In one embodiment, at least one lipid species is esterified with methylhexadecanoic acid. In one embodiment of the method or kit, at least one lipid species is esterified with methylhexadecanoic acid. In one embodiment the lipid species is a PC, LPC or PE -MHDA species. PE species include l5-MHDA_l8: l, _l 8:2, _20:4 and/or _22:6. PC species include 15 MHDA_l8: l, _l 8:2, _20:4 and/or _22:6 species. LPC species include 15-MHDA [snl] or [snl] [l04_snl] or [sn2] species.

In one embodiment, the description enables a panel or kit comprising one or more lipid species identified herein as significantly associated with AD for use in assessing or monitoring AD in a subject with symptoms of AD. In one embodiment, the lipid species are labelled.

Methods of assessing L/l burden

In another aspect, the description enables a method of assessing or monitoring Ab levels in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of two or more lipid species from one or two lipid classes or subclasses of ether lipids including alkylphosphatidylethanolamine alkenylphosphatidylethanolamine, lysophosphatidylethanolamine, and acylcarnitines and dehydrocholesterol ;

(ii) comparing the level of the two or more lipid species or one or two or more ratios of the level of two lipid species detected from (i) to a reference level or ratio for the lipid species; and

(iii) assessing or monitoring Ab levels in the subject on the basis of the comparison.

In one specific embodiment, the specification describes and enables a method of determining the severity of Alzheimer's Disease (AD) or ab burden in a subject, the method comprising

(i) detecting in a biological sample from the subject a level of a plurality of lipid species comprising one or two or more lipid species from one or two lipid classes or subclasses of ether lipids including alkylphosphatidylcholine, alkenylphosphatidylcholine, ly soalkylphosphatidylcholine , alkylphosphatidylethanolamine alkenylphosphatidylethanolamine , lysophosphatidylethanolamine, lysoalkenylphosphatidylethanolamine and alkyldiacylglycerol;

(iii) determining the severity of AD on the basis of the comparison.

In one embodiment, the plurality of lipid species comprises at least 1 to 6 lipid species set out in out in Table 10 or 11.

In one embodiment, at least 2 to 10 lipid species (e.g., 2, 3, 4, 5, 6, 7, 8, 9, or about 10 lipid species) or 2 to 20 lipid species (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or about 20 lipid species) are detected in step (i).

In one embodiment, the results from detecting and comparing lipid species are combined with one or more traditional risk factors for developing AD.

In a related embodiment, the description enables a method for assessing or monitoring Ab levels in a subject, the method comprising (i) detecting in a biological sample from the subject a level of two or more lipid species from one or two lipid classes or subclasses from Table V (Table 10), Table W (Table 11) or Table X (Table 11);

Table V (Table 10)

Table W (Table 11)

Table X (Table 11)

In one example, at least one lipid species is esterified with methylhexadecanoic acid. In one embodiment of the method or kit, at least one lipid species is esterified with methylhexadecanoic acid. In one embodiment the lipid species is a PC, LPC or PE -MHDA species. PE species include l5-MHDA_l8:l, _l 8:2, _20:4 and/or _22:6. PC species include 15 MHDA_l8:l, _l 8:2, _20:4 and/or _22:6 species. LPC species include 15-MHDA [snl] or [snl] [l04_snl] or [sn2] species. In one embodiment, lipid species for use in the above method assessing or monitoring Ab levels, or measuring cognitive ability, are also present in Table (a) or Table B (i) or Table B (ii) or are lipid species previously undisclosed as identified in Table 21. In one embodiment, there is provided a panel or kit comprising one or more lipid species identified or described herein a useful for assessing or monitoring Ab levels in a subject.

In one embodiment, the lipid species are labelled or tagged.

In one embodiment is provided a method of AD treatment comprising assessing or monitoring AD according to the methods disclosed herein or Ab levels in a subject according to a method disclosed herein, and administering AD treatment based upon the results of the method.

In one aspect, the application enables a data collection process or process of quantifying lipid levels in a subject, the method comprising detecting one or more of the specific combinations of lipid species identified herein in a lipid extracted plasma or serum or blood sample. Typically the sample comprises plasma mixed with a lipid extraction agent. Typically the sample further comprises a mixture of internal standards including lipid species representative of each lipid class or sub class to be to be analysed as illustrated in Table 2. As illustrated herein, in one embodiment, the process employs mass spectrometry such as tandem mass spectrometry and liquid chromatography. In some embodiment, the process as useful in assisting the presence or risk of AD or determining cognition levels in a subject. In one embodiment, the process is implemented using a system comprising at least one end station and a base station as described in International publication no. WO/02/090579. The present application contemplates a process including:

(a) receiving data in the form of one or more of the combinations of lipid species as described or defined herein, including previously unknown lipid species or lipid species not previously associated with AD or risk of developing AD, from the user via a communications network. In one embodiment, the process further comprises; (b) processing the subject data via multivariate analysis using 10 to 15 lipid species to provide an AD or cognition index value. In one embodiment, the process further comprises; (c) determining the status of the subject in accordance with the received results of the AD or cognition index value in comparison with predetermined values. In one embodiment, the process further comprises; (d) transferring an indication of the status of the subject to the user via. a communications network. In one embodiment, the process steps are (a) and (b). In one embodiment, the process steps are (c) and (d).

Each embodiment described here is to be applied mutatis mutandis to each and every other embodiment unless specifically stated otherwise.

The following methods and procedures are illustrative of the methods described herein. Lipid extraction

Lipid extraction was performed as described [Alshehry et al.]. In brief, lOpL of plasma was mixed with lOOpL of butanol: methanol (1 : 1) with lOmM ammonium formate which contained a mixture of internal standards (Table 2). Samples were vortexed thoroughly and set in a sonicator bath for 1 hour maintained at room temperature. Samples were then centrifuged (l4,000xg, 10 min, 20°C) before transferring the into sample vials with glass inserts for analysis.

Liquid chromatography mass spectrometry

Analysis of plasma extracts was performed on an Agilent 6490 QQQ mass spectrometer with an Agilent 1290 series HPLC system and a ZORBAX eclipse plus C18 column (2.lxl00mm 1 .8 pm. Agilent) with the thermostat set at 60°C. Mass spectrometry analysis was performed in positive ion mode with dynamic scheduled multiple reaction monitoring (MRM). Mass spectrometry settings and MRM transitions for each lipid class, subclass and individual species are shown in Table 2. The solvent system consisted of solvent A) 50% H20 / 30% acetonitrile / 20% isopropanol (v/v/v) containing lOmM ammonium formate and solvent B) 1% H20 / 9% acetonitrile / 90% isopropanol (v/v/v) containing lOmM ammonium formate. A stepped linear gradient was employed with a 15 minute cycle time and a lpL sample injection per sample. The following mass spectrometer conditions were used; gas temperature, l50°C, gas flow rate l7L/min, nebulizer 20psi, Sheath gas temperature 200°C, capillary voltage 3500V and sheath gas flow lOL/min.

Plasma quality control (PQC) samples consisting of a pooled set of 6 healthy individuals were incorporated into the analysis at 1 PQC per 18 plasma samples. Technical quality control samples (TQC) consisted of PQC extracts which were pooled and split into individual vials to provide a measure of technical variation from the mass spectrometer only. These were included at a ratio of 1 TQC per 18 plasma samples. TQCs were monitored for changes in peak area, width and retention time to determine the performance of the LC -MS/MS analysis and were subsequently used to align for differential responses across the analytical batches.

Relative quantification of lipid species was determined by comparison to the relevant internal standard (Table 2). As described by (Weir et al., 2013) response factors were generated for each cholesteryl ester species to better approximate their true concentrations.

Cross-sectional analyses: Using the lipidomic and clinical data and from the 18 month time point logistic regression was performed of lipid species against AD status with either no adjustment, or adjusting for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site.

Logistic regression was performed on baseline samples for all individuals for which Ab burden (SUVr score based on ⁿC-PiB PET data) was available (n=242). SUVr status (> 1.4 vs. <1.4) was used in the analysis. The analyses were adjusted for AD status alone or for AD status and age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site.

Linear regression was also performed on baseline samples for all individuals for which Ab burden (SUVr score based on ⁿC-PiB PET data) was available (n=242). Lipid species were analysed against SUVr score. The analyses were adjusted for AD status alone or for AD status and age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site.

Linear regression was also performed on baseline samples for all individuals for which California Verbal Learning Test (CVLT) scores (SUM) were available (n=l035). The CVLT is one of several tests that are commonly used to provide a measure of cognitive ability. These analyses were performed on Healthy control and MCI participants only (n=864) or on all participants (Healthy control MCI and AD, n=l035). The analyses were adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype, collection site and education level.

Analyses were corrected for multiple comparisons using the Benjamini-Hochberg method.

Longitudinal analyses: After six years of follow up of the AIBL inception cohort, 68 subjects had progressed to AD from the healthy control and MCI groups, while 515 individuals were followed for six years with no conversion. Using the lipidomic and clinical data from the baseline we performed logistic regression of lipid species or lipid ratios (created in an unbiased manner by calculating all possible ratios of individual lipid species) against incident AD status with either no adjustment, or adjusting for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site.

In order to allow for the large difference in age between the incident AD group (Progressors) and the non-progressor group (77yr vs. 68yr) these analyses were also performed on a subgroup consisting of 66 incident AD and 130 non-progressors, group matched for age and gender.

Creation of multivariable models to predict incident AD: Leature selection for multivariate models was performed on the entire dataset (or on an age and sex matched subset of data) by forward stepwise selection in a logistic regression model based on minimizing the Akaike information criterion (AIC). To correct for overfitting of the models, optimism correction was performed as described by Harrell et al (8). Bootstrapped samples (with replacement) were used to generate a multivariate model (logistic regression model based on minimizing the AIC). Each model which was then tested against the entire sample set, and the difference in performance metrics (c-statistic, IDI, relative IDI, sensitivity, specificity, accuracy, continuous NRI (events and non-events)) between the bootstrap sample and the whole dataset calculated. The averaged difference, in each performance measure, from the 1,000 bootstrap calculations, was subtracted from the original model to give the optimism corrected values and 95% confidence intervals.

In some analyses a base model consisting of age and ApoE4 status was used and the lipid species added to this base model to assess improvement of lipid species above these traditional risk factors. The following non-limiting examples are provided:

EXAMPLE 1

The AIBL study aimed to recruit 1000 individuals aged over 60 to assist with prospective research into AD (Alzheimer's disease). Initial recruitment took place from 2006-2008, and the inception cohort comprised 1112 participants, including 747 HC (Healthy Control), 126 MCI (Mild Cognitive Impairment) cases and 202 AD (Alzheimer's disease) cases for which plasma samples were available (Table 1). A clinical history was collected and anthropometric measurements were performed. Serum, plasma, platelets, red blood cell and white blood cells were collected and stored at -80°C. A full blood examination, urea and electrolytes, creatinine, androgen levels, globulin levels, sex hormone binding globulin, glomerular filtration rate, calcium, liver function tests, serum lipids, homocysteine, serum and red cell folate, B12, glucose, insulin, ceruloplasmin, ferritin/transferrin/iron, estradiol, luteinising hormone, thyroid function, and prostate specific antigen (males only) were measured. APOE genotyping was performed. All participants were assessed at baseline with a comprehensive battery of cognitive tests (see below) [Ellis et al.]. Participants were recalled every 18 months (we describe four recalls at 18, 36, 54 and 72 months) and all tests and collection of samples were repeated. 250 participants were assessed for Ab burden by ⁿC-PiB PET and cerebral atrophy by MRI; of these, 200 participants (145 HC, 36, MCI and 19 AD) were assessed at least four times at 18 month intervals [Villemagne et al.] Of the 200 participants 72 (28 HC, 28 MCI and 18 AD) had high Ab burden (standardised uptake value ratio (SUVR) scores, derived from ⁿC-PiB PET imaging, of 1.5 or higher). The main cognitive measures in the AIBL cohort include:

• Mini-Mental State Examination (MMSE): The MMSE is a 30-item test that is used extensively in clinical and research settings to screen for overall cognitive impairment.

• Composite episodic memory score: The average of z scores for the Rey Complex Figure Test (RCFT) long delay recall, the California Verbal Learning Test-Second edition (CVLT-II) long delay free recall (LDFR), and Logical Memory II delayed recall (Story A only) [17].

• Composite non-memory score: The average of z scores of the Boston Naming Test, letter fluency, category fluency, Digit Span (forward and backward), Digit Symbol-Coding (DSC) and RCFT copy.

Lipid associations with prevalent Alzheimer’s disease

The associations between lipid species and AD relative to HC and MCI combined at the 18 month time point were determined. The characteristics of these groups are shown in Table 3. There were 308 lipid species associated with prevalent AD in unadjusted models. When corrected for multiple comparisons, 275 remained significantly associated with prevalent AD. In models adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site, there were 150 lipid species associated with prevalent AD with 60 of these remaining significantly associated after correction for multiple comparisons (Table 4, Figure 1). These included species of ceramide dihydroceramide sphingomyelin and other sphingolipids as well as phosphatidylethanolamine and some alkenylphosphatidylethanolamine species.

EXAMPLE 2

Lipid associations with incident Alzheimer’s disease

The associations between lipid species and incident AD relative to HC and MCI individuals who did not convert to AD over the six year follow up period were determined. The characteristics of these groups at baseline are shown in Table 5. There were 164 lipid species associated with incident AD in unadjusted models with 55 species remaining significant after correction for multiple comparisons. In models adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site, there were 61 lipid species associated with incident AD with none of these significant after correction for multiple comparisons. These also included species of ceramide and sphingomyelin and a range of phospholipids including ether and plasmalogen species (Table 6, Figure 2).

In consideration of the large difference between ages of the incident AD groups and the non-convertors (average of 77 vs 68 years) we performed a sub -analyses where we selected 2:1 age matched sub-groups for incident AD (Table 7). There were 74 lipid species associated with incident AD in unadjusted models with none remaining significant after correction for multiple comparisons. In models adjusted for age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site, there were 74 lipid species associated with incident AD with one of these (RE(R-18:0/22:5) (n6)) significant after correction for multiple comparisons (Table 8).

EXAMPLE 3

Lipid associations with Afi burden

Logistic regression of lipid species against SUVr status (>1.4 vs. <1.4) (Table 9) with no adjustment for covariates showed 73 lipid species with p<0.05, however after correction for multiple comparisons no associations had a p<0.05 (Table 10). When the analyses were repeated adjusting for covariates (age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype, AD status and collection site), there were 42 lipid species with p<0.05, but again none were significant after correction for multiple comparisons (Table 10).

In contrast when we performed linear regression analysis of lipid species against the SUVr score with no adjustment for covariates, there were 118 lipid species with p<0.05 and following correction for multiple comparisons 28 remained significant (p<0.05). These lipid species were predominantly ether lipids (including plasmalogens. We also note dehydrocholesterol was significantly associated with SUVr (Table 11). When we repeated the analysis adjusting for covariates (age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype, AD status and collection site), there were 68 lipid species with p<0.05 but none remained significant after correction for multiple comparisons (Table 11).

EXAMPLE 4

Lipid ratios show stronger association than individual lipid species

We observed 31 lipid species that were positively associated and 24 lipid species that were negatively associated with incident AD after correction for multiple comparisons (Table 6). When we calculate the ratios of each positively associated species with each negatively associated species we observed stronger associations with incident AD with all 744 lipid ratios being significantly associated with incident AD and remaining significant after correction for multiple comparisons. When these analyses were adjusted for covariates (age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype and collection site), 624 lipid ratios were significantly associated with incident AD and 608 of these remained significant after correction for multiple comparisons (Table 12). The association with the lowest p-value was with the lipid ratio Cer(dl8: 1/24:1) | PC(0-l6:0/20:4) which had an odds ratio of 2.19 (95%CI, 1.72-2.79), p = 2.05E-10. The association with the highest odds ratio was with the lipid ratio AcylCarnitine(l6:l) | PC(0-38:5) which had an odds ratio of 4.63 (95%CI, 2.67-8.02), p = 4.80E-08.

EXAMPLE 5

Multivariable models to predict incident Alzheimer’s disease

Models created to predict incident AD using only lipid species performed very well with the optimism corrected model containing 10 lipid species showing a C-statistic of 0.78 (95% Cl 0.72-0.82). The NRI for the same model was 0.79 (95% Cl 0.56-0.99). Only marginal improvements were observed for the model containing 15 lipid species (Table 11). Lipid species in this model included phospholipids (diacyl, ether and lyso species), sphingolipids as well as DG(l4:0_l6:0) and dehydrocholesterol(l8:l) (Table 13). When models were made using lipid species and risk factors age and ApoE4 were selected as the first two features in the model (Table 14) and the addition of lipid species made little improvement in any performance measures. This was repeated with the base model fixed as age and ApoE4 status, and selected lipid species above the base model (Table 15).

The main discriminator of incident AD in these analyses was age (77yr vs. 68yr), which in the AIBL cohort was a greater discriminator that has been reported in the general population. To assess the predictive ability of lipid species above independently of age the same multivariable model was developed on a sub-cohort consisting of (66 incident AD and 130 age and sex matched non-progressors. Model development using lipid species only again performed well with the optimism corrected model containing 10 lipid species showing a C-statistic of 0.73 (95% Cl 0.67- 0.78). The NRI for the same model was 0.65 (95% Cl 0.41-0.86) (Table 16).

When models were made using lipid species and risk factors ApoE4 were selected as the first feature in the model (Table 17), however here the addition of lipid species made a significant improvement with the 10 feature model (Apoe4 status and nine lipid species) showing a C-statistic of 0.79 (95% Cl 0.74-0.84). The NRI for the same model was 0.76 (95% Cl 0.55-0.97) (Table 17).

When the base model was fixed as age and ApoE4 status, the improvement in the C- statistic with the addition of 10 lipid species was 0.13 (95% Cl 0.02-0.23). The NRI for the same model was 0.67 (95% Cl 0.47-0.89) (Table 18). This clearly demonstrates the ability of lipid species to improve on ApoE4 in the prediction of incident AD independently of age.

EXAMPLE 6

Lipids associated with cognition as measured by the California Verbal Learning Test (CVLT) scores (SUM):

Linear regression of lipid species against CVLT scores in the entire cohort with no adjustment for covariates showed 281 lipid species with p<0.05, after correction for multiple comparisons 239 associations had a p<0.05 (Table 19). When the analyses were repeated adjusting for covariates (age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype, collection site and education), there were 128 lipid species with p<0.05, and 35 remained significant after correction for multiple comparisons (Table 19).

When these analyses were repeated on only health control and MCI individuals; with no adjustment for covariates showed 241 lipid species with p<0.05, after correction for multiple comparisons 185 associations had a p<0.05 (Table 20). When the analyses were repeated adjusting for covariates (age, sex, BMI, total cholesterol, HDL-C, triglycerides, ApoE4 genotype, collection site and education), there were 46 lipid species with p<0.05, and none remained significant after correction for multiple comparisons (Table 20). EXAMPLE 7

Many of the lipid species described herein have not previously been reported in relation to AD. It is demonstrated here that many of the lipid species described here are associated with prevalent and incident AD. In many instances these are independent of known risk factors (age, sex, BMI, total cholesterol, HDL-C, triglycerides and ApoE4 genotype) demonstrating the potential of these lipid species to be useful biomarkers for the prediction of future AD.

Sphingolipids are associated with prevalent and incident Alzheimer’s disease

Associations of specific n-acylated ceramides with prevalent AD, positive associations were observed with species containing stearic acid (18:0), arachidic acid (20:0) and nervonic acid (24:1) while negative or neutral associations were observed for species containing behenic acid (22:0), lignoceric acid (24:0) and creotic acid (26:0), irrespective of the sphingoid base. Significant associations of sphingolipids with incident AD were only observed for species containing nervonic acid. Nervonic acid is particularly abundant in both the central and peripheral nervous system, where it is the major fatty acid present in myelin sphingolipids. Synthesis of nervonic acid is thought to be through the elongation of oleic acid (18:1).

While specific cer amide synthases are responsible for the n -acylation of longer species (ceramide synthase 2 and 3), different associations were observed here for the 24:0 and 24:1 species, suggesting that our observations of associations with incident AD relate to the synthesis of 24:1. An alternate explanation may be that the liberation of nervonic acid containing sphingolipid from myelin during AD pathology is reflected in the periphery.

Gangliosides are a group of sphingolipids with oligosaccharide groups linked to the sphingoid base. The available methodology can measure the most abundant circulating ganglioside class (G_VB) which exhibits a positive association, in particular when n-acylated with nervonic acid, GM3(dl 8: 1/24:1). Gangliosides, notably the GM1 and GM3 classes or subclasses, have been reported to accelerate Ab aggregation, leading to deposits in the brain [Yamamoto et al., Hoshino et al.,]. While the mechanism increasing the circulatory amount of gangliosides is currently not known, GM3 is not a commonly measured sphingolipid and so far no other reports have attempted to link associations between peripheral GM3 and AD.

Ether lipids are associated with prevalent and incident Alzheimer’s disease and with Afi burden

Ether lipids, in particular plasmalogens, have been demonstrated to be associated with AD in multiple studies [Goodenowe et al., Han et al., Grimm et al.] in both the periphery and brain. As determined herein specific alkyl phospholipids and plasmalogens associate negatively with prevalent AD, incident AD and SUVr in all our analysis (Table 4, 6, 8, 10, 11). Species containing both omega-3 and omega-6 polyunsaturated fatty acids (lineoic acid, 18:2, arachidonic acid, 20:4, eicosapentaenoic acid 20:5, docosahexaenoic acid 22:6) associate negatively with AD. Their downstream metabolites (LPC(O), LPC(P) and LPE(P)) remain largely unchanged with AD.

Lipid ratios show stronger association to Alzheimer’s disease than individual species

Some lipid species showed a positive association with AD risk (are increased in those who develop AD) while other lipid species showed a negative association with AD risk (are decreases in those who develop AD). This raises the possibility of using ratios of lipid species to improve prediction of AD. Indeed the ratios created from those positively associated lipid species with the negatively associated lipid species show a stronger association with incident AD than some of the individual lipid species.

Multivariable models can improve the prediction of incident AlzJtiemer’s disease above traditional risk factors

Multivariable models created with up to 20 lipid species can further improve the prediction of future AD (above traditional risk factors and above individual lipid species. In an age and sex matched subcohort we demonstrate that the addition of 10 lipid species to a base model of age and ApoE4 resulted in a increase in the C-Statistic from 0.73 (95%CI, 0.65-0.80) to 0.80 (95%CI, 0.75- 0.85) and an continuous net reclassification index of 0.67 (95%CI, 0.47-0.89).

These results clearly demonstrate the potential of a relatively small number of lipid species to improve the prediction of incident AD above traditional risk factors as described in the examples and tables.

Lipid markers can predict Ab burden

It is further demonstrated herein that the lipid species are associated with Ab burden as determined from the SUVr score, with multiple lipid species showing significant associations with this score independent of AD status. Thus there is a clear potential to use such markers to assess individuals for their Ab burden without the need for expensive and invasive ⁿC-PiB PET imaging.

Lipid markers can predict cognition

As determined herein the lipid species are associated with cognition as determined by the CVLT score, with multiple lipid species showing significant associations with this score even when only examining the healthy control and MCI participants. This indicates that lipid species may also have application to assess individuals for their cognition levels. Such assessments are proposed to be useful in the diagnosis, monitoring or assessment of MCI and AD.

Novel lipids identified associated with incident and prevalent Alzheimer’s disease

The unprecedented power of the AIBL study cohort design combined with the extensive lipidomic profiling performed has enabled the identification of 308 lipid species that are associated with prevalent AD (p<0.05, unadjusted) and 164 lipid species associated with incident AD (p<0.05, unadjusted), to give a total of 345 lipid species associated with prevalent and/or incident AD. To the inventor's knowledge, at least 117 of these lipid species have not previously been associated with AD (Table 21). Indeed, 64 of these lipid species have not previously been reported and are not included in the LIPID MAPS database (lipidmaps.org/data/structure/index.php).

Further to this, in many previous reports of lipids that may be useful markers for AD, statistical significance was not achieved in the study and so it is only the present study that demonstrates lipid species showing a significant association with incident and/or prevalent AD in plasma. Thus for the majority of lipid species reported here this is the first statistically verified report for significant associations with prevalent and/or incident AD. Furthermore, lipid ratios containing at least one lipid species that has not previously been associated with prevalent or incident AD represent new lipid biomarkers.

It will be appreciated by persons skilled in the art that many modifications may be made to the above-described embodiments.

BIBLIOGRAPHY

Fahy E, Subramaniam S, Brown HA, Glass CK, Merrill AH, Jr., Murphy RC, Raetz CR, et al. A comprehensive classification system for lipids. J Lipid Res 2005;46:839-861.

Fahy E, Subramaniam S, Murphy RC, Nishijima M, Raetz CR, Shimizu T, Spener F, et al. Update of the LIPID MAPS comprehensive classification system for lipids. J Lipid Res 2009;50 Suppl:S9- 14.

Goodenowe DB, Cook LL, Liu J, Lu Y, Jayasinghe DA, Ahiahonu PW, Heath D, et al. Peripheral ethanolamine plasmalogen deficiency: a logical causative factor in Alzheimer's disease and dementia. J Lipid Res 2007;48:2485-2498.

Grimm MOW, Kuchenbecker J, Rothhaar TL, Groesgen S, Hundsdoerfer B, Burg VK, Friess P, et al. Plasmalogen synthesis is regulated via alkyl-dihydroxyacetonephosphate-synthase by amyloid precursor protein processing and is affected in Alzheimer's disease. Journal of Neurochemistry 2011;116:916-925.

Han X, Holtzman DM, McKeel DW. Plasmalogen deficiency in early Alzheimer's disease subjects and in animal models: molecular characterization using electrospray ionization mass spectrometry. Journal of neurochemistry 2001;77:1168-1180.

Hoshino T, Mahmood MI, Mori K, Matsuzaki K. Binding and aggregation mechanism of amyloid beta-peptides onto the GM1 ganglioside-containing lipid membrane. J Phys Chem B 2013;117:8085-8094.

Ifuku M, Katafuchi T, Mawatari S, Noda M, Miake K, Sugiyama M, Fujino T. Anti- inflammatory/anti-amyloidogenic effects of plasmalogens in lipopolysaccharide -induced neuroinflammation in adult mice. J Neuroinflammation 20l2;9:l97.

Liebisch G, Vizcaino JA, Kofeler H, Trotzmuller M, Griffiths WJ, Schmitz G, Spener F, et al. Shorthand notation for lipid structures derived from mass spectrometry. J Lipid Res 20l3;54:l523- 1530.

Mandel H, Sharf R, Berant M, Wanders RJA, Vreken P, Aviram M. Plasmalogen Phospholipids Are Involved in HDL-Mediated Cholesterol Efflux: Insights from Investigations with Plasmalogen-Deficient Cells. Biochemical and Biophysical Research Communications 1998;250:369-373.

Mankidy R, Ahiahonu PW, Ma H, Jayasinghe D, Ritchie SA, Khan MA, Su-Myat KK, et al. Membrane plasmalogen composition and cellular cholesterol regulation: a structure activity study. Lipids Health Dis 20l0;9:62. Pike LJ, Han XL, Chung KN, Gross RW. Lipid rafts are enriched in plasmalogens and arachidonate -containing phospholipids and the expression of caveolin does not alter the lipid composition of these domains. Faseb Journal 200l;l5:A20-A20.

Schooneman MG, Vaz FM, Houten SM, Soeters MR. Acylcarnitines: Reflecting or Inflicting Insulin Resistance? Diabetes 2013;62:1-8.

Vetrivel KS, Thinakaran G. Membrane rafts in Alzheimer's disease beta-amyloid production. Biochim Biophys Acta 2010;1801:860-867.

Wang X, Wang W, Li L, Perry G, Lee H-g, Zhu X. Oxidative stress and mitochondrial dysfunction in Alzheimer's disease. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 2014;1842:1240-1247.

Weir JM, Wong G, Barlow CK, Greeve MA, Kowalczyk A, Almasy L, Comuzzie AG, Mahaney MC, Jowett JB, Shaw J, Curran JE, Blangero J, Meikle PJ. Plasma lipid profiling in a large population-based cohort. J. Lipid Res. 2013;54:2898-2908.

Yamamoto N, Hirabayashi Y, Amari M, Yamaguchi H, Romanov G, Van Nostrand WE, Yanagisawa K. Assembly of hereditary amyloid b-protein variants in the presence of favorable gangliosides. FEBS Letters 2005;579:2185-2190.

Zoeller RA, Grazia TJ, LaCamera P, Park J, Gaposchkin DP, Farber HW. Increasing plasmalogen levels protects human endothelial cells during hypoxia. Am J Physiol Heart Circ Physiol 2002;283:H67l-679.

Claims

CLAIMS:

1. A method of determining the presence of, or risk of developing Alzheimer's Disease (AD) in a subject, the method comprising

2. The method of claim 1 wherein the plurality of lipid species comprises at least one or two ore more lipid species set out in Table (a) or at least one or two lipid species set out in Table B (i) and (ii).

3. The method of claim 1 or 2 wherein the plurality of lipid species further comprises one or more lipid species selected from Tables C and/or Table I.

4. The method of any one of claim 1 to 3 wherein the method determines the risk of developing AD in a subject, the method comprising

5. The method of claim 4 wherein the plurality of lipid species comprisies at least 1 to 6 lipid species set out in Table A or at least 1 to 6 lipid species set out in Table B(i).

6. The method of claim 5 wherein the plurality of lipid species further comprisies one or more lipid species selected from Table C.

7. The method of any one of claims 4 to 6 wherein the ratios are selected from Table 12.

8. The method of any one of claims 4 to 6 wherein the plurality of lipid species are selected from Tables D, E, F and G or combinations thereof.

9. The method of any one of clai s 1 to 8 wherein the detecting step (i) comprises detecting at least one lipid species whose level is positively associated with prevalent AD or risk of AD onset from from Table 4 or 6, respectively, and at least one lipid species whose level is negatively associated with prevalent AD or AD onset from Table 4 or 6, respectively; and step (ii) comprises comparing one or more ratios of positively and negatively associated lipid species to one or more reference ratios.

10. The method of any one of claims 1 to 9 wherein at least one lipid species detected is a sphingolipid species containing nervonic acid (24:1).

11. The method of any one of claim 1 to 10 wherein at least one lipid species detected is a ganglioside GM3.

12. The method of any one of claims 1 to 11 wherein at least one lipid species is a ceramide selected form Table (a), Table H or Table A.

13. The method of any one of claims 1 to 12 wherein at least one lipid species detected is PE (P-l8:0/22:5) and/or one ratio is Cer(dl8: 1/24:1) | PC(0-l6:0/20:4) and/or acylcarnitine(l6:l) | PC(0-38:5).

14. The method of any one of claim 1 to 3 wherein the method determines the presence of AD in a subject, the method comprising

(iii) determining whether the subject has AD on the basis of the comparison.

15. The method of claim 14 wherein the plurality of lipid species comprises at least 1 to 6 lipid species set out in Table H or at least 1 to 6 lipid species set out in Table B(ii).

16. The method of claim 15 wherein the plurality of lipid species further comprises one or more lipid species selected from Table I.

17. A method of determining the severity of Alzheimer's Disease (AD) or Ab burden in a subject, the method comprising

(iii) determining the severity of AD on the basis of the comparison.

18. The method of claim 17 wherein the plurality of lipid species comprisies at least 1 to 6 lipid species set out in out Table 10 or 11.

19. The method of any one of claims 1 to 18 wherein at least 2 to 10 lipid species (e.g., 2, 3, 4,

5, 6, 7, 8, 9, or about 10 lipid species) or 2 to 20 lipid species (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,

13, 14, 15, 16, 17, 18, 19, or about 20 lipid species) are detected in step (i).

20. The method of any one of claims 1 to 19 wherein the results from detecting and comparing lipid species are combined with one or more traditional risk factors for developing AD.

21. A panel or kit comprising one or more lipid species identified in any one of claims 1 to 13 for use in determining a subject's risk of developing AD.

22. A panel or kit comprising one or more lipid species identified in any one of claims 14 to 16 for use in assessing or monitoring AD in a subject with symptoms of AD.

23. A panel or kit comprising one or more lipid species identified in any one of claims 17 or 18 for use in assessing or monitoring Ab levels in a subject.

24. The panel or kit of any one of claims 21 to 23 wherein the lipid species are labelled.

25. A method of prophylaxis comprising determining a subject's risk of developing AD according to a method of any one of claims 1 to 13 and providing therapeutic or behavioural intervention on the basis that the subject tests positive for risk of developing AD.

26. A method of AD treatment comprising assessing or monitoring AD according to a method of any one of claims 14 to 16 or Ab levels in a subject according to a method of any one of claims 17 or 18 and providing AD treatment and or behavioural intervention based upon the results of the method.

27. A method of determining a lipid based measure of cognition or a change in cognition in a subject, the method comprising (i) detecting in a biological sample for the subject a level of a plurality of lipid species wherein at least one or two lipid species are selected from one or more lipid classes or subclasses from GM3, DE, Hex3Cer, and TG(O);

28. A method of determining a lipid based measure of cognitive ability or a change in cognitive ability in a subject, the method comprising;

(i) detecting in a biological sample for the subject a level of a plurality of lipid species selected from lipid species showing a significant association with cognition set out in Table 19 or 20.

29. The method of claim 27 or 28 wherein at least 2 to 10 lipid species (e.g., 2, 3, 4, 5, 6, 7, 8, 9, or about 10 lipid species) or 2 to 20 lipid species (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or about 20 lipid species) are detected in step (i).