The inflammation component of Alzheimer's Disease

The inflammation component of AD

Imperial College London

April 24th-25th

Alejo Nevado-Holgado,

Simon Lovestone’s team

University of Oxford

What is the inflammatory component?

• Inflammation: It is the so-called unspecific part of the immune system,

which responds in the same manner to all health threads such as injure,

infection, toxins…

• Link to AD: A central role for inflammation in AD has long been suggested

(Akiyama2000). Evidence coming from multiple sides:

• Tissues: Seen in the behaviour of microglia (Prinz et al. 2011), cytokines

(Swardfager et al. 2010), chemokines (Cartier et al. 2005) and complement

molecules/receptors (Harold et al. 2009; Lambert et al. 2009) during AD

• GWAS: Inflammatory genes are among those showing the highest odds

ratio for AD (Jones et al. 2010, Hardy 2015).

• Epidemiology: Prior history of inflammatory drugs (Wang et al. 2015) and

diseases (McGeer et al. 1996, Wallin et al. 2012) affect AD risk later in life.

Tissue studies

• Microglia are the main brain white cells. They have an important role on Aβ

and tau clearance (Prinz et al, Nat. Neurosci. 14, 1227-1235).

• Chemokines are signalling proteins used by the immune system. Their

receptors are localised in microglia and other white cells, and they can

trigger cell death (Cartier et al, Brain Res, 48, 16-42).

• The complement system are a group of small proteins that potentiate the

immune reaction. GWAS studies very often point to this system (Nat. Genet.

41, 1088-1093).

GWAS studies

• GWAS studies find which SNPs (1 letter mutations) appear most frequently

in people with AD. The genes most frequently found are from: (1) vesicle

recycling; (2) lipid metabolism; and (3) inflammatory system. (Jones et al. 2010

PLoS One 5, e13950; Hardy et al, J. Internal Med. 275, 296-303).

Epidemiology studies

• People regularly taking NSAIDs earlier

in life, have fewer risk of developing AD

(J Alzheimer's Dis. 44, 385-396).

• The longer and earlier taking NSAIDs,

the better.

• However, using NSAID as a therapy is

not effective (Schneider et al, J Internal Med,

275, 251-283).

Epidemiology studies

• Rheumatoid arthritis late in life decreases AD risk.

• But rheumatoid arthritis earlier in life may increase AD risk.

Citation Design ADs / controls OR

Graves et al 1990 Case control of AD 130 / 130 1.18

French et al 1985 Case control of AD 78 / 156 0.62

Heyman 1984 Case control of AD 40 / 80 1.19

Broe et al 7 Case control of AD 170 / 170 0.56*

Li et al 8 Case control of AD 70 / 140 0.16*

Can Health 10 Case control of AD 201 / 468 0.54*

Breitner et al 11 Case control of AD 50 / 50 0.64

Jenkinson et al 14 Case control of AD 96 / 92 0.17*

Lindsay et al 2002 Random population based 194 / 3894 0.6

Tyas et al 2001 Random population based 36 / 658 0.81

McGeer et al 4 Random population based 29 / 7461 0.15

Myllykangas et al 16 Random population based 2 / 1613 0.23*

Beard et al 15 Random population based 23 / 498 -

Wallin et al 2012 Random population based (20 yrs) ? / 1449 2.46*

Epidemiology studies

• Gout has also been studied, but

much less than rheumatoid arthritis

or NSAIDs.

• A matched population study with

Taiwan data found no Gout-AD

effect (989 patients, 3956 controls)

(Shin-Wei et al, 2013, Int J Geriatr

Psychiatry, 28, 1205-1206).

• A matched control study with UK

data found Gout-AD effect (2251

patients, <250000 controls, odds

ratio 0.76) (Na Lu et al, 2015, Ann

Rheuym Dis).

Our project

• The presented results make very clear that there is an inflammation-AD

link, and that this link has a crucial role, but they say very little about the

nature of this link.

• To learn about the nature of the link, we would need to answer:

• What other inflammatory diseases, drugs and genes alter AD?

• Through which mechanism do they produce this alteration in AD?

Our project

Investigating the inflammation link

• Three levels approach:

diseases, drugs and genes

• Using two types of data: EHRs

and -omics

• EHRs: ideal for diseases and

drugs, but more difficult to

access

• Omics: The only ones for

genes, and easy to access

(ArrayExpress & GE-omnibus).

• We can also address some

aspects of drugs and diseases

with the omics data.

Fig. NSAID data power in Rotterdam cohort (Veld 2001

New Eng J Med) ANM and UK biobank.

Our project

• EHRs: ideal for diseases and

drugs, but more difficult to

access.

Our project

• Omics: The only ones for

genes, and easy to access

(ArrayExpress & GE-

omnibus).

• We can also address

some aspects of drugs

and diseases with the

omics data.

Fig (↗). Over-respresentation analysis ran

on the genetic multi-signatures. The multi-

signatures capture which genes may

potentially link AD with each drug.

Our project

• We can run our analysis with different prior inflammatory drugs or

diseases, and with different tissues of AD patients:

• Blood data: 600 samples. Looking for more.

• Brain data: 300+600 samples. More are accessible.

• Next, we would like to compare these results with epidemiological data.

Fig (←). Inflammation link in

AD brain

Fig (→). Inflammation link in

AD blood

Simon’s team:

Jessica Ash

Ben Liu

Alison Braid

Elena Ribe

Corinne Prescot

Laura Thei

Sarah Westwood

And of course, Simon Lovestone

The EMIF team for Inflammation:

Usha Gungabissoon

Donal Skelly

Angelika Wientzek

Mark Gordon

Meritxell Sabido

Sandra Guedes

Many thanks to: