Association of Interleukin-1 Gene Polymorphisms with Alzheimer’s Disease

Abstract

Interleukin-1 (IL-1) is markedly overexpressed in Alzheimer’s disease. We found the IL-1A 2,2 genotype in 12.9% of 232 neuropathologically confirmed Alzheimer’s disease patients and 6.6% of 167 controls from four centers in the United Kingdom and United States (odds ratio, 3.0; controlled for age and for ApoE [apolipoprotein E] genotype). Homozygosity for both allele 2 of IL-1A and allele 2 of IL-1B conferred even greater risk (odds ratio, 10.8). IL-1 genotypes may confer risk for Alzheimer’s disease through IL-1 overexpression and IL-1– driven neurodegenerative cascades.

Alzheimer’s disease (AD), the most common cause of dementia, is estimated to affect 5% of those more than 65 years of age and 40% of those more than 80 years of age, or a total of approximately 20 million people worldwide. The neuropathological characteristics of AD include (β-amyloid plaques that are diagnostic when associated with dystrophic neurites,¹ neurofibrillary tangles, loss of neurons and synapses, and proliferation of glial cells. There has been substantial progress, in the last few years, in unraveling the genetic influences in AD. A small proportion of AD cases are inherited as an autosomal dominant trait and are attributable to point mutations in genes that encode the β-amyloid precursor protein (βAPP),² presenilin 1,³ or presenilin 2.⁴ A major genetic risk factor for the much more common sporadic AD is possession of the apolipoprotein E (ApoE) ε4 allele.⁵ However, possession of ApoE ε4 is neither necessary nor sufficient for the development of AD, leaving scope for other potential genetic or environmental influences.

Interleukin-1 (IL-1) is a potent proinflammatory cytokine that is markedly overexpressed in the brains of patients with AD, predominantly in microglia,⁶ which suggests a role for inflammatory processes in AD pathogenesis.⁷ This idea has received support from epidemiological studies that show that the use of anti-inflammatory agents, in particular nonsteroidal anti-inflammatory drugs, is associated with delayed onset or slowed progression of disease.⁸ IL-1 has two structurally distinct forms, IL-1α and IL-1β, encoded by separate genes (IL-1A and IL-1B, respectively) located in a cluster on the long arm of chromosome 2 that also includes the IL-1 receptor antagonist gene.⁹ Common polymorphisms have been described in both genes and there is evidence that they have functional significance. A polymorphism of the IL-1B gene (+3953), for instance, which introduces a TaqI restriction site, results in two alleles, designated allele 1 and allele 2.¹⁰ Homozygosity for allele 2 of IL-1B is associated with a fourfold increase in production of IL-1β compared with homozygosity for allele 1.¹⁰ A polymorphism in the 5′ regulatory region of the IL-1A gene (a C-to-T transition at position −889 relative to the start site of transcription) again results in two alleles, also designated allele 1 and allele 2.¹¹ Both of these IL-1 polymorphisms have been associated with inflammatory diseases. For instance, IL-1A allele 2 has been associated with juvenile rheumatoid arthritis.¹¹ In view of the evidence that suggests a role for inflammatory processes, and for IL-1 in particular in AD, we hypothesized that these IL-1 polymorphisms modulate susceptibility to AD.

Materials and Methods

Case Selection

Tissues and DNA samples from a total of 233 patients with clinical histories of dementia (mean age, 81.5 [SD, 7.8] years), that satisfy CERAD (Consortium to Establish a Registry for Alzheimer’s Disease)¹ criteria for the postmortem neuropathological diagnosis of AD, were available from four participating centers (Table 1). Age at death was not recorded in 4 of the patients with AD; 169 aged nondemented controls, with a mean age of 74.4 (SD, 9.6) years, without significant AD pathology, were available from the same departments.

Table 1.

Number, with Mean Age ± SD, of Patients and Subjects from Four Centers: Little Rock, AR (US), and Glasgow, Bristol, and Oxford (UK)

	Glasgow	Little Rock	Bristol	Oxford
Alzheimer’s disease patients
n	61	27	92	53
Age (mean ± SD), yr	82 ± 8.7	76.9 ± 7.9	82.2 ± 7.7	81.9 ± 6.1
Control subjects
n	39	36	52	42
Age (mean ± SD), yr	77.2 ± 9	70.4 ± 8.1	77 ± 8.9	72 ± 10.6

Genetic Analysis

Genotyping of the IL-1A (−889) and IL-1B (+3953) polymorphisms was performed, blind to the diagnostic group, as described previously.^10,11 ApoE genotyping was also performed as described previously.¹² IL-1 genotypes could not be determined in 3 subjects (1 AD patient and 2 control subjects).

Statistical Analysis

Logistic regression analysis was used to identify the factors predictive of AD. Age, center, possession of the ApoE ε4 allele, IL-1A allele 2, IL-1B allele 2, both IL-1A allele 2 and IL-1B allele 2, IL-1A 2,2 genotype, IL-1B 2,2 genotype, and both IL-1A 2,2 and IL-1B 2,2 genotypes were considered possible explanatory variables. Odds ratios are for the development of AD in the presence of the factor relative to its absence. For age, the odds ratio is relative to the preceding.

Results

IL-1A and IL-1B genotype and allele frequencies and combinations in AD and control subjects are shown in Table 2. The IL-1A 2,2 genotype was possessed by 12.9% of the AD patients, compared with 6.6% of the controls. For the IL-1B 2,2 genotype, the corresponding results are 7.3% and 4.8%, respectively. The previously described composite genotype that comprises IL-1A allele 2 plus IL-1B allele 2 is present in 36% of AD patients and 30% of controls. Of the 399 patients in the study with complete data, 17 were homozygous for allele 2 of both IL-1A and IL-1B and 15 (88%) of 17 of these patients had AD. A univariate logistic regression analysis showed that age (p < 0.001), possession of the ApoE ε4 allele (p < 0.001), IL-1A 2,2 (p = 0.03), both IL-1A 2,2 and IL-1B 2,2 (p = 0.01) genotypes, and the participating center (p = 0.03) were significantly associated with AD. A stepwise multiple regression based on the above significant variables showed that each of the following was sufficient to confer increased risk for AD: age, possession of the ApoE ε4 allele, possession of the IL-1A 2,2 genotype, and possession of the IL-1A 2,2 and IL-1B 2,2 genotypes (Table 3). No association of AD with participating center was seen by this analysis, and the relationship of the center with AD in the univariate analysis was attributable to differences in the patient age distribution among the different centers. These results, taken together with accumulating evidence for a role of IL-1 in AD pathogenesis,^6,7,13 strongly support a role for these specific IL-1 gene polymorphisms in the modulation of susceptibility to AD but do not rule out the possibility that these polymorphisms are in dysequilibrium with other genes on chromosome 2.

Table 2.

IL-1A, IL-1B, and ApoE Genotype and Allele Frequencies for Alzheimer’s Disease and Control Subjects

Gene	Genotype or Allele	AD (n = 233)	Controls (n = 169)
IL-1A	1,1	103 (44.4%)	82 (49.1%)
	1,2	99 (42.7%)	74 (44.3%)
	2,2	30 (12.9%)	11 (6.6%)
	Allele 1	305/464 (66%)	238/334 (71%)
	Allele 2	159/464 (34%)	96/334 (29%)
IL-1B	1,1	147 (63.4%)	111 (66.5%)
	1,2	68 (29.3%)	48 (28.7%)
	2,2	17 (7.3%)	8 (4.8%)
	Allele 1	362/464 (78%)	270/334 (81%)
	Allele 2	102/464 (22%)	64/334 (19%)
IL-1A 2 and IL-1B 2		83 (35.8%)	50 (29.9%)
IL-1A 2,2 and IL-1B 2,2		15 (6.5%)	2 (1.2%)
ApoE	2,2	0 (0%)	2 (1.2%)
	2,3	8 (3.4%)	23 (13.6%)
	2,4	9 (3.9%)	3 (1.8%)
	3,3	78 (33.5%)	107 (63.3%)
	3,4	114 (48.9%)	31 (18.3%)
	4,4	24 (10.3%)	3 (1.8%)
	ε2	17/466 (3.6%)	30/338 (8.9%)
	ε3	278/466 (59.7%)	268/338 (79.3%)
	ε4	171/466 (36.7%)	40/338 (11.8%)

Data are not available for IL-1 genotypes for 1 AD patient and 2 controls; see Materials and Methods.

IL-1 = interleukin-1; ApoE = apolipoprotein E; AD = Alzheimer’s disease.

Table 3.

Logistic Regression Analysis Odds Ratio for Alzheimer’s Disease in the Presence of Factors Including Age and Possession of the ApoE ε4 Allele, and Possession of Either IL-1A 2,2 Genotype in Combination with IL-1B 2,2 (Model 1) or IL-1A 2,2 Genotype Alone (Model 2)

Factor	Odds Ratio	95% Confidence Interval	p
Model 1
ApoE ε4 allele	7.4	4.5–12.4	<0.001
Age (relative to preceding year)	1.11	1.07–1.14	<0.001
IL-1A 2,2 and IL-1B 2,2 genotype	10.8	2.1–56.8	0.005
Model 2
ApoE ε4 allele	7.2	4.4–12.0	<0.001
Age (relative to preceding year)	1.11	1.07–1.14	<0.001
IL-1A 2,2 genotype	3.0	1.3–6.9	0.011

ApoE = apolipoprotein E; IL-1 = interleukin-1.

Discussion

The findings of this study support our hypothesis that specific IL-1 gene polymorphisms confer increased risk for AD, perhaps through altered regulation of IL-1– driven neurodegenerative cascades, including synthesis and processing of βAPP.^6,7 We provide evidence to suggest that individuals with the IL-1A 2,2 genotype alone, and, in particular, with the combination of IL-1A 2,2 plus IL-1B 2,2, carry increased risk for AD. The associations of these IL-1 polymorphisms with AD remain statistically valid after controlling for age and for possession of an ApoE ε4 allele, the major genetic risk factor for sporadic, late-onset AD.⁵ The possibility that these IL-1 polymorphisms are in linkage dysequilibrium with other genes on chromosome 2 cannot be excluded by this study. For example, other genes that regulate IL-1 biology, including that which encodes the IL-1 receptor antagonist, are located close to the IL-1A and IL-1B genes.⁹

A key role for inflammatory processes in AD is suggested by the finding of markedly increased tissue levels of IL-1 in AD,^6,7 and by recent epidemiological evidence that anti-inflammatory medication may delay onset or slow the progression of the disease.⁸ We have proposed a glial–neuronal cytokine feedback loop that, once initiated, accounts for the progressive nature of AD.⁷ In this cytokine cycle, overexpression of IL-1 is proposed as the key agent that initiates and orchestrates neurodegenerative events that culminate in the neuropathologies changes of AD. IL-1 regulates all aspects of neuronal synthesis and processing of βAPP,^14–16 favoring continued deposition of β-amyloid and release of secreted fragments of βAPP (sAPP). These, in turn, promote microglial activation, an effect that is modulated by an isoform-specific influence of ApoE.¹⁷ Another of the consequences of elevated levels of IL-1 is astrocyte activation and up-regulation of S100β expression,¹³ favoring the growth of dystrophic neurites necessary for the conversion of β-amyloid deposits into the neuritic plaques diagnostic of AD. Neuronal stress induces βAPP synthesis, which promotes further microglial activation, with synthesis and release of IL-1 to produce feedback amplification of these events. Perpetuation of this cytokine cycle may explain the progressive nature of AD and also help to explain why traumatic brain injury, with consequent microglial activation and IL-1 synthesis,¹⁸ is the most consistently demonstrated environmental risk factor for AD.¹⁹

The IL-1B 2,2 genotype has already been associated with increased production of IL-1β,¹⁰ and the IL-1A gene polymorphism investigated in this study is located within the promoter region and may influence the expression of IL-1α. These IL-1 polymorphisms could therefore act by increasing the gain of the cytokine cycle. Further studies are required to determine if this might influence the rate of disease progression or influence responses to other genetic or environmental risk factors for AD (eg, trauma^18–20). In addition to potentially aiding in the differential diagnosis of patients with dementia and assessing the degree of risk for the development of dementia, identification of this genetic association should intensify the emphasis on the pursuit of anti-inflammatory therapy and pinpoint IL-1 as a potential therapeutic target for the protection against and treatment of AD.