Field of Invention

The invention relates to a diagnostic test for schizophrenia. Background

Schizophrenia is a common and devastating psychiatric disorder which affects between 0.5% and 1.5% of the population in most parts of the world. Perhaps ten times as many relatives, friends, colleagues, teachers and employers have had their lives adversely affected by this disease. The economic costs are high since most patients begin to develop symptoms between the ages of 15 and 35, yet live a relatively normal life span. Many homeless people are schizophrenic.

There is a strong genetic basis for the disease, with affected families commonly having schizophrenic members in successive generations. However, the mode of inheritance is not clear and simple genetic models do not explain the pattern of occurrence of the illness. In identical twins, the concordance rate for schizophrenia is only approximately 50%, even though both members of a discordant pair must carry the genes conferring susceptibility. Thus, environmental factors must play an important role in the phenotypic expression of this disorder.

The diagnosis of schizophrenia has not been clearly elucidated. At the time that manifestations are first apparent, schizophrenia, bipolar manic-depressive disorders, severe depression, drug-related disorders, stress-related disorders and other conditions may be difficult to distinguish from each other. However, optimum treatments for schizophrenia differ from optimum treatments for other disorders and the long term prognosis of the different disorders also differs extensively.

Therefore, an objective test that augments the validity of a diagnosis of schizophrenia would be of great value. It would enable a more accurate diagnosis to be made at a much earlier stage than is possible at present, allowing reliable early initiation of the best available treatment and also the development of a more accurate prognostic evaluation. It might even, in future, allow the prevention of schizophrenia. Many

families which carry schizophrenic genes are known. Children in those families could be identified prior to puberty as individuals potentially at risk. Appropriate low stress social, educational and employment environments could be provided and relevant treatment and nutritional regimes instituted, in particular diets high in unsaturated fats. There is evidence that diets low in saturated fats and high in unsaturated fats may lead to a more benign course of schizophrenia (Christensen O & Christensen E, Acta Psychiatr Scand 1988, 78:567-591; Horrobin DF, Glen AIM & Vaddadi KS, Schizophrenia Research 1994, 13: 195-207). The development of these techniques could lead to elimination of the phenotypic expression of the disease in those who are genotypically affected.

Over the past twenty years evidence has accumulated that there may be abnormalities in membrane phospholipids in schizophrenic patients. Much of that evidence is summarised in the paper by Horrobin, Glen and Vaddadi referenced above. One important finding is that in patients with schizophrenia, there may be reduced levels of the important fatty acids - arachidonic acid (AA) and docosahexaenoic acid (DHA) - in membrane phospholipids. The two highly unsaturated fatty acids are predominantly found in the sn2 position of membrane phospholipids and are removed from that position by a number of enzymes collectively known as phospholipases A2 (PLA2s). When considering the pathogenesis of schizophrenia, an abnormality in one or more of the PLA2s is therefore a reasonable possibility. Abnormal blood levels of one type of PLA2 in schizophrenic patients have already been reported (Gattaz WF et al, Biol Psychiatry 1990, 28:495-501). Present Work

There are several forms of PLA2. In particular, we have investigated the cytoplasmic form of PLA2 (cPLA2) which is activated by calcium and which catalyses the release of arachidonic acid (A A) from membrane phospholipids. The cPLA2 gene has been identified, as has a region adjacent to the cPLA2 gene on chromosome 1 which is believed to be the promoter region for the gene (Tay et al, Biochim Biophys Acta 1994; 1211: 345-347 and Tay et al, Genomics 1995; 26: 138-141). Promoter regions

are believed to regulate the expression of the gene and to determine the amount of active gene product produced.

In each normal individual there are two copies of each gene, one copy being derived from the mother and one from the father. Each gene may have several possible different variants known as alleles. Which alleles are present will determine what characteristics are expressed in an individual. If the two genes derived from the maternal and paternal chromosomes are the same alleles, the individual is said to be homozygous for that gene. If the alleles are different, the individual is said to be heterozygous for that gene.

As described in the papers referenced above, Tay et al have identified the promoter region for the cPLA2 gene. This is at the 5' end of the cDNA sequence for cPLA2. This region was rapidly amplified using a kit from Gibco, BRL, Gaithersburg, MD, USA. Within the promoter region, two simple sequence repeats (SSRs) were identified. One of those was a string of cytosine-adenine pairs and another a straight string of linked adenines, on which the present work has concentrated. Such SSRs in the promoter regions of genes may often show different lengths in different individuals, and the different SSR lengths may indicate different levels of gene expression. Each different length represents a different allele.

In Tay et al's work, in 50 unrelated healthy controls, ten alleles were identified in the polyadenine SSR, designated A1-A10. In the first six of these (A1-A6), each allele differed from the next by only one or two repeats; there were then five repeats between A6 and A7. In this group of individuals from the normal population, 92% of the total of 100 alleles of the 50 individuals were A1-A6 and only 8% A7-A10.

Using the techniques described by Tay et al, we have now studied 50 schizophrenic individuals in a similar way. They were diagnosed on the basis of the Structured Clinical Interview for DSM-IIIR (criteria set by the American Psychiatric Association). The last four alleles, A7-A10, accounted for 62% of all the alleles in this group as compared to 8% in the normal group.

We have also studied 40 further normal controls and have found three more alleles with sequence repeats shorter than those designated Al by Tay et al.

Niacin, given at a dose of over lOOmg by mouth, induces marked facial flushing in normal individuals. Some schizophrenic individuals flush normally, whereas others fail to flush in response to niacin. There is some evidence that schizophrenics who fail to flush in response to niacin represent a more tightly defined group. In a group of 10 schizophrenics who were defined as niacin non-flushers, this was reflected in the A7-A10 alleles present. Since niacin non-flushers appear to represent a core group of schizophrenics with a poor long term prognosis, this finding of a high proportion of A7- A10 alleles in this population may be important. The Invention

We propose that identification of the type of allele present in the promoter region of the cPLA2 gene is of value in predicting those persons potentially at risk for developing schizophrenia.

More particularly the invention provides a method of identifying individuals genotypically predisposed to schizophrenia characterised by establishing the presence of abnormal alleles in the promoter region of the cytoplasmic phospholipase A2 (CPLA2) gene of chromosome 1.

Suitably what is established is the presence in that region of alleles with polyadenine simple sequence repeats (SSRs) longer than those of alleles A1-A6.

In particular the method may be carried out by (i) taking a tissue sample and extracting genetic material from it; (ii) amplifying the poly-A sequence of the promoter region of the CPLA2 gene; (iii) determining whether A7-A10 or other longer alleles than A1-A6 are present, such presence indicating a positive diagnosis.

The invention also provides correspondingly a diagnostic test or diagnostic kit for such methods, as set out in the claims hereafter.

An alteration associated with the cPLA2 gene may give rise to disease, either causing the disease itself or acting as a modifier that allows for the expression of another mutant gene. It is possible that, alone, a mutation of the cPLA2 gene may be necessary

but not a sufficient pathologic process in the development of schizophrenia. Either way the presence of the abnormal alleles will identify individuals at risk of developing schizophrenia.

The invention may be further stated as:-

1. A diagnostic indicator of schizophrenia, based on identification of the alleles present at the promoter region of the cPLA2 locus on human chromosome 1, and thus a method of indicating which individuals may be schizophrenic, or at risk of developing schizophrenia, by identification of such alleles.

2. A diagnostic test or diagnostic kit for, and use of such a test or kit in, diagnosis of individuals at risk of schizophrenia or who are carrying genes which confer a risk of schizophrenia, which test or kit comprises means for determining the types of alleles present in the promoter region of the cPLA2 gene of human chromosome 1.

The methodology for identifying the alleles may be that described by Tay et al or that modified by us as described in detail later or may be any type of methodology known or yet to be discovered which allows accurate identification of the alleles.

The alleles are in particular polyadenine (poly A) alleles, specifically those identified as alleles A1-A10 in the table herein. The presence of alleles A7-A10, or other alleles longer than the A1-A6 alleles, indicates an increased risk of schizophrenia. As noted, identification is suitably by using primers flanking the poly-A repeat tracts, labelling, amplifying the labelled primers by polymerase chain reaction, and determining the size of the alleles.

Specifically, identifying the presence of alleles A7-A10 and possibly higher alleles will be of value in the following situations :-

1. In patients with recognised psychotic behaviour but whose diagnosis is not certain. The presence of one of the alleles with a higher number of repeats than the A1-A6 alleles serves as a specific indicator of schizophrenia in the diagnosis.

2. In individuals manifesting disturbed behaviour for the first time and in whom the diagnosis of schizophrenia may be difficult and readily confused with a variety

of other conditions. The presence of one of the above alleles will aid in confirmation of a diagnosis of schizophrenia. Hence, more effective planning of treatment and a more accurate prognosis may be achieved. 3. In children and adults who are not themselves schizophrenic but who have a family history of schizophrenia, or who are displaying schizophrenic-like behaviour, or who simply wish to know whether or not they are at risk of schizophrenia, identification of the alleles present will be of value. Such identification will allow preventive measures to be initiated which may reduce incidence of carriers with the risk-bearing alleles from developing clinical, phenotypic schizophrenia. Such preventive measures may include social and behavioural measures, drug treatment and eventually specific nutritional and pharmaceutical treatments which may completely prevent the emergence of the clinical disorder. Given the costs of schizophrenia to the individual patient, to their families and to society, the advantages of such early diagnosis and prevention are obvious.

Prophylaxis

The invention further extends to the prevention of schizophrenia, based on the suggestion that in schizophrenia there is excessive activity of cPLA2. There is other evidence pointing in the same direction. Elevated circulating levels of cPLA2 have been detected in schizophrenic patients, also elevated levels of phospholipid breakdown products have been detected in the brains of schizophrenic individuals by nuclear magnetic resonance scanning, and in the red cell membranes of schizophrenics there appear to be reduced phospholipid concentrations of two important essential fatty acids, docosahexaenoic acid (DHA) and arachidonic acid (AA), these two essential fatty acids (EFAs) being removed from phospholipids by cPLA2, suggesting increased activity of this enzyme. Since normal membrane composition is required for normal nerve cell function, this excessive activity of cPLA2 may be responsible for abnormal nerve cell

function and hence for the abnormal behavioural and other manifestations of schizophrenia.

One way of reducing the impact of excessive cPLA2 activity is to make available increased amounts of the two main EFAs which are removed from cell phospholipids by cPLA2, arachidonic acid and docosahexaenoic acid: the presence of increased amounts of these two fatty acids increases the chance of incorporation of the AA and DHA into membranes tending to normalise them and thus reduce the impact of the presence of excess cPLA2 activity.

We propose that in individuals who are not yet diagnosed as phenotypically schizophrenic but in whom the abnormal alleles are demonstrated, schizophrenia may be prevented or at least the risk of development of it reduced by the administration of adequate levels of A A and DHA. The invention therefore extends to diagnosis as above followed by prophylactic administration of effective amounts of AA and DHA in free or combined form, particularly together in the form of a triglyceride or diol derivative, and to other aspects of prophylaxis, or preparation of medicaments therefore, as set out in the claims herein.

The dose of each of these fatty acids per day may be in the range of from 10 mg to 100 g, preferably 100 mg to 10 g, and very preferably 500 mg to 5 g. The fatty acids may be presented orally, parenterally or topically in any form which may be biologically assimilated into the body. Possible forms of administration may include free fatty acids, lithium, sodium, potassium or other salts, any esters, any amides, and mono-, di- or triglycerides, any phospholipids, or any other appropriate form. Particularly desirable because of their ease of oral administration without generating gastro-intestinal upset, and their ease of intravenous administration in the form of non-toxic emulsions, are triglycerides. A mix of triglycerides containing DHA and AA, or a triglyceride containing at least one molecule of AA and one of DHA, with the other fatty acid on the triglyceride being either AA, DHA or another fatty acid, or a triglyceride prepared from a fatty acid mix containing at least 10% and preferably at least 20% or very preferably at least 30% each of DHA and AA are particularly desirable variants of the invention.

They may be made by methods per se known, for example those of the inventor's published EP-A-0609001 and EP-A-0611569 to which reference may be made. Alternatively the 1,3-propane diol esters of the applicant's co-pending PCT filing no. GB96/01053 may be used, particularly to deliver the AA and DHA.

If cPLA2 activity is excessive, it may also be appropriate to administer with the fatty acids an inhibitor of cPLA2. This could be a drug such as mepacrine or a glucocorticoid, but more desirable would be a fatty acid which itself had activity in inhibiting cPLA2. cPLA2 can be physiologically inhibited by increasing the amounts of cyclic AMP (cAMP) in cells. Cellular cAMP concentrations may be enhanced by the presence of prostaglandin El (PGE1) which is a powerful stimulator of cAMP formation. The levels of PGE1 may be increased by its immediate precursor, dihomogammalinolenic acid (DGLA) or by the immediate precursor of DGLA, gammalinolenic acid (GLA). Thus, GLA or DGLA may be given in any appropriate form and amount as described above for A A and for DHA. However, two forms of GLA and/or DGLA may be particularly appropriate. One is the derivative of ascorbic acid, ascorbyl-GLA or ascorbyl-DGLA. This is because, as described in previous patent applications by the inventor, published as European patent application no's. 95 301316.6 (EP-A-0675120) and 95 304498.9 (EP-A-0694302) to which reference may be made, the presence of ascorbic acid greatly stimulates the conversion of GLA or DGLA to PGE1 and hence will greatly increase the formation of cAMP. The other preferred form is in a triglyceride or triglyceride mix in which AA and DHA are also present. Thus, a triglyceride containing approximately equal amounts of AA, DHA and either GLA or DGLA may be used, such triglyceride being present in a triglyceride mix, preferably at a concentration of above 10% and very preferably at a concentration of above 20%. Also alternatively a triglyceride mix may be prepared from an approximately equimolar mix of the three fatty acids, AA, DHA and/or GLA or DGLA, the total concentration of the three fatty acids in the mix being greater than 40%, preferably greater than 50% and very preferably greater than 80%.

Detailed Application

The alleles present may be identified by any method known to those skilled in the art at present or by any method which may be developed in the future. One method of doing so, based on the work of Tay et al, is in summary to use primers and the polymerase chain reaction as described by Tay et al, to amplify the relevant region of the cPLA2 gene and prepare appropriate gels. The different alleles are identified by reading the gels and noting the number of repeats in the relevant region. Patients may thus be identified as homozygous or heterozygous and those who have A7-A10 alleles identified as at risk of developing schizophrenia.

Tay et al collected 20 ml of blood from 50 unrelated healthy donors and extracted genomic DNA by standard established techniques. Primers were designed to flank the cytosine-adenine (CA) and adenine (A) repeat tracts. These primers were end-labelled by incubation with 10 UT4 polynucleotide kinase (Boehringer Mannheim) and 50 microcuries of 22p_ATP (6000 Ci/mmol, Amersham). The labelled primers were then used to amplify the genomic DNA using the polymerase chain reaction. The amplified DNA was extracted and loaded on to a sequencing gel and analysed. The size of each allele was determined by comparison of the PCR products to a sequencing ladder. Full technical details of the method are published in Genomics 26: 138-141, 1995.

Four alleles were determined for the CA repeat sequence, but 96% were of one type, CA _j g, and so there was little individual variation at this gene locus. In contrast, at the polymorphic A locus, ten alleles of different repeat lengths were found (designated Al to AIO, Table 1), with relatively widely distributed frequencies in a normal population.

Table 1

Gene frequencies in 50 normal individuals at the poly (A) locus in the cPLA2 promoter region of chromosome 1 as determined by Tay et al, and in 50 schizophrenics as determined by the inventors. The gene frequencies in 10 schizophrenics who failed to flush in response to niacin are also shown. Statistical comparison using a chi square test showed the schizophrenics to be different from controls at p < I ^Q~ ^.

Allele No. of As in the Gene frequencies in: poly-A tract Controls All Non-Flushing Schizophrenics : Schizophrenics

Al 41 0.05 (5%) 0.04 (4%) 0.0 (0%)

A2 43 0.31 (31 %) 0.02 (2%) 0.0 (0%)

A3 45 0.01 (1 %) 0.03 (3%) 0.0 (0%)

A4 48 0.19 (19%) 0.03 (3%) 0.0 (0%)

A5 49 0.04 (4%) 0.06 (6%) 0.0 (0%)

A6 50 0.32 (32%) 0.20 (20%) 0.0 (0%)

A7 55 0.01 (1 %) 0.19 (19%) 0.05 (5%)

A8 57 0.02 (2%) 0.22 (22%) 0.65 (65%)

A9 58 0.04 (4%) 0.16 (16%) 0.20 (20%)

AIO 60 0.01 (1 %) 0.05 (5%) 0.10 (10%)

A1-A6 - 0.92 (92%) 0.38 (38%) 0.0 (0%)

A7-A10 - 0.08 (8%) 0.62 (62%) 1.0 (100%)

In the present work we obtained blood samples from 50 schizophrenic individuals. They were diagnosed as schizophrenic on the basis of the Structured Clinical Interview for DSM-IHR published by the American Psychiatric Association. Among this group, ten were further defined as failing to show facial flushing in response to 100 mg or more of oral niacin. Niacin given at this dose causes marked facial flushing in normal individuals. It does so by stimulating the conversion of arachidonic acid (AA) to prostaglandin D2, a powerful cutaneous vasodilator. If AA has been depleted from membrane phospholipids as a result of sustained overactivity of PLA2, then there may not be available enough AA for niacin to produce a flushing response. A failure to flush

in response to niacin in some schizophrenics may therefore identify a group whose PLA2 activity is particularly abnormal.

Table 1 shows the allele frequencies in the 50 normal controls described by Tay et al. We used exactly the same techniques as described by Tay et al and as summarised above to extract DNA from the blood samples obtained from schizophrenics and to identify the alleles present at the polymorphic A locus. The results for the overall 50 schizophrenic populations and for the 10 who did not flush in response to niacin are also shown in table 1. As can be seen, the normal population has 92% A1-A6 alleles with only 8% as A7-A10. In contrast, the 50 schizophrenics have only 38% A1-A6 alleles and 62% as A7-A10. The schizophrenics who are further defined as being non-flushers have none of their alleles as A 1-A6, all being A7-A10. Thus in these non-flushing schizophrenics both of the poly-A alleles derived from the mother and father must be A7-A10.

These findings indicate that the presence of at least one of the two alleles in the A7-A10 region is a strong marker for schizophrenia. The presence of two alleles in this region may be expected to indicate an even greater risk. The presence of a few normal individuals carrying one of these alleles is consistent with what is known about schizophrenia where the normal population contains "carriers" who are not affected by the disease. At present the numbers of patients who have been studied are relatively small and it is possible that in the future other alleles with even larger numbers of A repeats may be identified as carrying a risk of schizophrenia.

We have now slightly modified the Tay et al technique and a detailed description of the cunent protocol for analysis follows.

I. PCR (Polymerase Chain Reaction) protocol for CPLA2 poly(A) polymorphism

Reference:

Tay, A. et al (1995) Cytosolic phospholipase A2 gene in human and rat chromosomal localisation and polymoφhic markers. Genomics 26, 138-141 as cited earlier.

Solutions required: lOμM primer cPLA2-a (CCT CCT TTC TAG AAA TTC AG) lOμM primer cPLA2-b (CAG AGC TTC AGT GAG CCA)

10X One-Phor-All Buffer Plus (Pharmacia)

³² γ-P-ATP [lOμCi/μl]

T4-Polynucleotide Kinase (Pharmacia) deionized H2O (sterile)

10X PCR buffer II (Perkin-Elmer)

25mM MgCl2 (Perkin-Elmer) lOmM dNTP mix (dATP, dGTP, dCTP, dTTP)

Taq DNA Polymerase (Perkin-Elmer)

Equipment:

0.5-10 μl micropipettor and sterile tips 10-100 μl micropipettor and sterile tips Eppendorf microcentrifuge DNA Engine PTC-200 (MJ Research) 0.2 ml PCR tubes

Note : All items used should be sterile, i.e. autoclaved to eliminate contamination of the sample with DNases, which will degrade the DNA. This includes tubes, pipets pipet tips etc.

Method:

Note: This method will nominally make up enough reaction mix for 30 reactions. However, allow at least 10% (i.e. 3 reactions) extra for pipet enor, therefore, this will set up 27 reactions for PCR.

A. Label primer cPLA2-a:

1. Set up the following reaction in a 0.2ml tube:

10 μM primer cPLA2-a 9.0

10X One-Phor-All Buffer Plus 4.0

³² γ-P-ATP [lOμCi/μl] 10.0

T4-Polynucleotide Kinase 2.0 deionized H2O (sterile) 15.0

Final Volume: 40.0μl

Note: The addition of ³² P label must be done in the radioactive lab, behind proper shielding.

2. Mix well. Plase reaction in DNA Engine, use LABEL programme to label primer.

B. Preparation of DNA samples:

1. Genomic DNA samples should be diluted to 50ng/μl.

2. Aliquot 200ng (4μl) of DNA sample into a 0.2ml tube.

3. Cover tubes with caps until reaction mix is prepared (Part C).

C. Preparation of PCR reaction mix:

1. In a 0.2 ml tube combine:

10X PCR buffer II 30.0

25mM MgCl2 18.0 lOmM dNTP mix 24.0 lOμM primer cPLA2-b 18.0 primer cPLA2-a, 32p labelled 40.0 (part A) deionized H2O (sterile) 50.0

Final Volume 180.0μl

Note: The addition of primer cPLA2-a, 32p labelled, as well as the remaining steps, must be done behind proper shielding.

2. Mix well. Aliquot 6μl reaction mix into each tube containing DNA sample.

3. Place samples in DNA Engine and run HOTSTART programme.

4. In a 0.2ml tube combine:

Taq DNA polymerase [5U/μl] 6.0μl deionized H2O (sterile) 24.0

Final Volume 30.0μl, [lU/μl]

5. When HOTSTART programme reaches 80°C, add lμl Taq DNA Polymerase (IU) per reaction. Final volume of reaction = 1 lμl.

6. Run cPLA2-l programme (PCR reaction).

Note: PCR programmes stored in the DNA Engine

HOTSTART: 95°C for 5 minutes (denature)

80°C for 15 minutes (for addition of Taq)

LABEL: 37°C for 30 minutes (labelling reaction)

55°C for 5 minutes (deactivate enzyme)

CPLA2-1: 95°C for 30 seconds )

58°C for 30 seconds ) = one cycle 72°C for 30 seconds ) Repeat cycle 30 times.

DENATURE: 95°C for 5 minutes 4°C indefinately

π. Analysis of PCR products by polyacrylamide gel electrophoresis

Reference:

Tay, A. et al (1995) as above.

Biorad Laboratories Instruction Manual for Sequi-Gen GT Nucleic Acid Electrophoresis Cell.

Solutions and Chemicals:

Acrylamide: Bisacrylamide (19: 1, 30%) (Biorad)

10X TBE Buffer (0.89M Tris Base, 0.89M Boric Acid, 20mM EDTA) (Biorad)

IX TBE Buffer (Biorad)

Urea (Biorad)

25% Ammonium Persulfate (25% APS) (Biorad)

TEMED (Biorad) deionized water (d.i. H2O)

DNA markers (3 ² P-labelled):ρBR322/Mspi (New England Biolabs)

Lambda/Hindlll/EcoRl (IBI) BioMax-MS-1 film (Kodak)

BioMax-MR-1 film (Kodak)

Loading Buffer #2: 95% formamide, 0.5% bromophenol blue, 0.5% xylene cyanol

Equipment:

Sequi-Gen GT nucleic acid electrophoresis cell and gel casting kit (Biorad) DNA Engine PTC-200 (MJ Research) 0.45μm Millex-HA syringe filter (Millipore)

Method:

A: Preparation of 8% polyacrylamide gel (denaturing)

1. Prepare an 8% acrylamide gel solution.

2. Clean glass plates and assemble the electrophoresis unit.

3. Use a 20-well comb (0.4mm thick) to form the wells for sample loading.

4. Inject acrylamide gel solution between the glass plates of the electrophoresis unit. Allow 45 minutes to set the gel.

B: Preparation of gel for electrophoresis run

1. Remove gel electrophoresis unit from casting tray, and place into lower buffer chamber.

2. Warm 600 ml IX TBE buffer in the microwave, 2 minutes, Power = 10.

3. Pour warmed buffer into upper buffer chamber, pour 400ml IXTBE buffer into lower buffer chamber.

4. Attach electrodes to the electrophoresis unit and to the power supply.

5. To pre-warm the gel, set the power supply at 65Watts, 45 minutes.

C: Preparation and loading of the samples

1. Add lOμl 10X loading buffer #2 (with formamide) to each PCR sample, and to lOμl of each DNA marker.

2. Place the PCR samples and tubes containing the two DNA markers on the DNA Engine.

3. Use the DENATURE programe to denature the samples and the markers.

4. Place on ice.

5. When gel electrophoresis has completed its pre-warming run, take off the top electrodes, and top up the buffer if necessary.

6. Using a 20 cc. syringe filled with buffer, "swish" the urea out of the wells.

7. Load 5.0μl of each marker in the first two wells, then load 2.5μl of sample per well. In the last two wells, again load 5.0μl of each marker.

8. Re-connect the electrodes and set the power supply to 70Watts, 2 hours. Run the gel.

D. Post gel run

1. Dissassemble the apparatus. Carefully pry the glass plates apart.

2. Blot the gel onto filter paper (cut to 18 x 25cm).

3. Dry gel on gel dryer. Cover with plastic wrap. Programme: Cycle 2, 80°C, 45 minutes.

4. When dry, place in a film cassette, and expose to the appropriate film. Use BioMax-MS-1 film, for rapid detection or BioMax-MR-1 film for sensitive detection.

5. For BioMax-MS-1 film, use with an intensifying screen, and expose for one hour at -70°C. When using BioMax-MR-1 film, expose for 3 hours at -70°C (an intensifying screen is not necessary).

6. Develop the film: 5 minutes, GBX developer

30 seconds rinse with water 5 minutes, GBX fixer 5 minutes, rinse with water Let dry at room temperature.

E. Analysis of bands

1. Analyse the bands by comparison to DNA markers of known sizes in order to approximate their size in basepairs.

2. Using the pBR322/Mspl marker, the 147basepair band = 41 A's = Al allele. The 160bp band = 55 A's = A7 allele (within lbp).

3. Measure the distance between these two marker bands, and calculate how many millimeters each band should travel; e.g. 160-147 = 13 bases, distance between bands = 13mm; 13/13 = lbp/mm. Therefore increasing the DNA band by 1 base = a decrease of 1mm travelled down the gel.

4. For example, measured two unknown bands: band 1: 4mm; and band 2: 7mm from the 147 band of the DNA marker.

Calculate the bands' size: band 1: 4mm x lbp/mm = 4bp, 4bp+ 147bp = 151bp band 2: 7mm x lbp/mm = 7bp, 7bp+ 147bp = 154bp

5. In this PCR, the region sunounding the poly(A) region = 106 bp. So, to find the length of the poly(A) region (i.e. the number of A's), subtract this (106bp) from the size of the band.

Calculate the poly (A) length: band 1: 151-106 = 45 A's (assigned allele A3) band 2: 154-106 = 48 A's (assigned allele A4).

Example of Synthesis of a Triglyceride Containing A A, DHA and GLA

(i) A mixture of GLA (97%, l .Og, 3. όmmol), glycidol (280mg, 3.73 mmol) and tri-n-butylamine (20μl, O.Oδmmol) was heated at 85°C under nitrogen for 5h. The reaction was then cooled an purified by flash chromatography (5 % methanol/methylene chloride) to yield the monoacylglycerol as a colourless oil.

(ii) A solution of dicyclohexylcarbodiimide (DCC) (270mg, 1.31mmol) and 4- N,N-dimethylaminopyridine (DMAP) (160mg, 1.31mmol) in methylene chloride (5ml) was added to a solution of the monoacylglycerol (400mg, 1.14mmol) and DHA (98%, 350mg, l.Oδmmol) in methylene chloride (15ml) at 0°C under nitrogen. Tic analysis (3% methanol/methylene chloride) after 5h showed the reaction to be complete. Hexane

(30ml) was added to precipitate dicyclohexlyurea and the mixture was filtered, concentrated to dryness and purified by flash chromatography (2% methanol/methlene chloride) to yield a mixture of diacylglycerol positional isomers as a colourless oil.

(iii) A solution of DCC (80mg, 0.39 mmol) and DMAP (40mg, 0.35mmol) in methylene chloride (5ml) was added to a solution of the diacylglycerol (200mg, 0.3mmol) and AA (98%, 1 lOmg, 0.35mmol) in methylene chloride (10ml) at room temperature under nitrogen. As the reaction proceeded a precipitate of dicyclohexylurea formed. After 2 h tic analysis (8% ethyl acetate/hexane) indicated that the reaction was complete. Hexane (30ml) was added to precipitate more dicyclohexylurea and the reaction was filtered and concentrated to dryness. Purification by flash chromatography (5 % ethyl acetate/hexane) yielded the pure triglyceride as a colourless oil.

The product of the above example has been subject to hplc triglyceride analysis and gc fatty acid analysis, by standard methods. In gc analysis the preparation of the methyl ester derivatives of fatty acids is well known. Boron trifluoride in methanol (12- 14% w/v) was used as catalyst. lOOmg of each of the triglycerides was transesterfied and analysed in a Hewlett Packard 5890 Series II equipped with a Supelcowax™ 10 capillary column (30m x 0.53mm x l.Oμm). Injector temperature set at 220°C and detector temperature at 250°C, the oven temperature was programmed starting at 180°C for 5 min after which is increased at a rate of 2°C/min until 210°C was reached and maintained at this temperature for a further 15 mins. lμl of each of the methyl esters of the fatty acids was injected using an autosampler, 7673 from Hewlett Packard. Each of the methyl esters of the fatty acids were identified by injecting standards. The results of the gc and hplc analyses are as follows: -

Retention Time Area % (peaks over 2.5%) gc:- 8.198 92.140

9.802 4.202 hplc:- 8.600 29.8217 (GLA)

14.187 31.2864 (AA)

23.350 35.0575 (DHA)

Example of Synthesis of 1,3-Propane Diol Diester with AA and DHA l-(z,z,z,z-eicosa-5,8, 11 , 14-tetraenoyloxy)-3-(z,z,z,z,z,z-docosa-4,7, 10, 13, 16, 19- hexaenoyloxy)propane.

(Diester ofAA and DHA with 1,3-propane diol).

Part i:

A solution of z,z,z,z-eicosa-5,8,ll,14-tetraenoic acid (150g) in methylene chloride (500ml) was added dropwise to a mixture of 1,3-dihydroxypropane (205g), 1,3- dicyclohexylcarbodiimide (130g) and 4-(N,N-dimethylamino)pyridine (87g) in methylene chloride (2500ml) at room temperature under nitrogen. When tic indicated that the reaction had gone to completion the reaction mixture was filtered. The filtrate was washed with dilute hydrochloric acid, water and saturated sodium chloride solution. The solution was dried, concentrated and purified by dry column chromatography to yield 1- (z,z,z,z-eicosa-5,8,l l , 14-tetraenoyloxy)-3-hydroxypropane as a pale yellow oil. Part 2:

A solution of 1,3-dicyclohexy .carbodiimide (23.7g) and 4-(N,N- dimethylamino)pyridine (15.9g) in methylene chloride (200ml) was added to a solution of l-(z,z,z,z-eicosa-5,8,l l,14-tetraenoyloxy)-3-hydroxypropane (33 .6g) and z,z,z,z,z,z-docosa-4,7, 10, 13,16,19-hexaenoic acid (30g) in methylene chloride (400ml) under nitrogen at room temperature. On completion of reaction as evidenced by tic analysis the solution was diluted with hexane, filtered, concentrated and purified by dry column chromatography to yield l-z,z,z,z-eicosa-5,8, ll,14-tetraenoyloxy)-3- (z,z,z,z,z,z-docosa-4,7, 10,13,16,19-hexaenoyloxy)propane as a pale yellow oil.

Examples of Prophylactic Compositions

1. A triglyceride mix prepared from a mix comprising 15-30% of each of three fatty acids, AA, DHA and GLA or DGLA is reacted together under appropriate chemical or enzymic conditions as well known in the art and shown in the applicant's published

patent applications refened to herein, to produce a triglyceride for oral or parenteral administration and administered in the amounts set out herein.

2. A triglyceride containing one molecule of AA, one of DHA and one of GLA or DGLA, present in a triglyceride mix at a concentration of at least 10%, preferably 20% and very preferably more than 30%, produced and administered as in Example 1.

3. As 1 and 2 but in which the AA and DHA are present without GLA or DGLA in the triglycerides, the GLA or DGLA being optionally present separately.

4. Triglyceride mixes as in 3, but in formulations in which GLA or DGLA are provided in the forms of ascorbyl-GLA or ascorbyl-DGLA, produced by the methods of the applicant's published patent applications refened to herein.

5. A 1,3-propane diol diester produced to contain A A and DHA and administered in the amounts set out herein.

6. As 5, with GLA or DGLA present separately, suitably as ascorbate.

All the above may be formulated in any appropriate dosage form, including soft or hard gelatin capsules, microcapsules presented alone or in the form of powders or other forms such as tablets, emulsions, suspensions and any other appropriate formulations for oral, enteral, parenteral or topical use such that the fatty acids may reach the body in a biologically assimilable form. Many such variations of presentation are known to those skilled in the art.