A systematic review on the association between inflammatory genes and cognitive decline in non-demented elderly individuals
June 17, 2022•26,561 words
A systematic review on the association between inflammatory genes and cognitive decline in non-demented elderly individuals
Available online 11 December 2015
Abstract
Cognitive impairment, or decline, is not only a feature of Alzheimer׳s disease and other forms of dementia but also normal ageing. Abundant evidence from epidemiological studies points towards perturbed inflammatory mechanisms in aged individuals, though the cause–effect nature of this apparent relationship is difficult to establish. Genetic association studies focusing on polymorphism in and around inflammatory genes represent a viable approach to establish whether inflammatory mechanisms might play a causal role in cognitive decline, whilst also enabling the identification of specific genes potentially influencing specific cognitive facets. Thus, here we provide a review of published genetic association studies investigating inflammatory genes in the context of cognitive decline in elderly, non-demented, samples. Numerous candidate gene association studies have been performed to date, focusing almost exclusively on genes encoding major cytokines. Some of these studies report significant cognitive domain-specific associations implicating Interleukin 1β (IL1β) (rs16944), Tumour Necrosis Factor α(TNFα) (rs1800629) and C-reactive protein (CRP) in various domains of cognitive function. However, the majority of these studies are lacking in statistical power and have other methodological limitations, suggesting some of them may have yielded false positive results. Genome-wide association studies have implicated less direct and less obvious regulators of inflammatory processes (i.e., PDE7A, HS3ST4, SPOCK3), indicating that a shift away from the major cytokine-encoding genes in future studies will be important. Furthermore, better cohesion across studies with regards to the cognitive test batteries administered to participants along with the continued application of longitudinal designs will be vital.
1. Introduction
Cognitive impairment is a defining feature of Alzheimer׳s disease (AD) and multiple other neurodegenerative diseases that preferentially affect elderly populations (Bettcher and Kramer, 2014). Further, cognitive decline is also characteristic of normal ageing, which not only has an adverse impact on quality of life in elderly populations but the extent of decline is considered a potential marker of prospective pathophysiology (Eshkoor et al., 2015). Although the biological processes and molecular mechanisms mediating cognitive decline in aged populations are not fully understood, abundant evidence highlights immune and inflammatory mechanisms as potential candidates (Li et al., 2014, McAfoose and Baune, 2009, McAfoose et al., 2009). Numerous epidemiological studies have found that peripheral inflammatory markers such as Interleukin 1β (IL1β), IL6, and Tumour Necrosis Factor α (TNFα) are increased in the elderly (Fuchs et al., 2013, Karim et al., 2014, Yaffe et al., 2003), and that this is exaggerated in AD and other dementia patients relative to age-matched healthy controls (Faria et al., 2014, Yarchoan et al., 2013). Likewise, these observations also hold true when sampling from both cerebrospinal fluid and post-mortem brain tissue (Alcolea et al., 2014, Monson et al., 2014, Sudduth et al., 2013). However, these findings are merely correlational and so the cause–effect relationship is not clear.
Although it is not possible to directly manipulate inflammatory mechanisms in humans to investigate subsequent effects on cognitive abilities, Mendelian randomisation represents a viable strategy to explore cause–effect relationships (Smith et al., 2005). By taking advantage of naturally occurring genetic variation known to influence the levels or functioning of inflammatory genes, it becomes possible to determine whether altered inflammatory mechanisms might be causative of cognitive decline whilst also highlighting specific genes in the process – this is of course the implicit premise, or assumption, of genetic association studies (Smith et al., 2005). Thus, in this review we aim to harness findings from published genetic association studies (both candidate-based and genome-wide) investigating inflammatory gene variants and cognitive decline in aged, non-demented, populations in order to glean information concerning cause–effect relationships, particularly as they might relate to different facets of cognition.
We begin by providing a brief primer outlining the “cytokine model of cognitive function” (McAfoose and Baune, 2009) and applying this specifically to cognitive decline in the non-demented elderly. We then provide a systematic review of all published genetic association studies investigating immune genes in elderly populations, where we summarise the major findings and offer critical appraisal to assess reliability and validity. Finally, we highlight some of the major limitations of this research whilst making recommendations as to further research directions in this area.
2. Primer on the “cytokine model of cognitive function” and possible relevance for cognitive decline in non-demented elderly
Although the precise nature of the relationship between immune/inflammatory mechanisms and elderly cognitive decline is not known, there is ample evidence to indicate that inflammatory molecules – particularly the major cytokines – are certainly capable of influencing cognitive function through the regulation of various biological substrates of cognition. We refer the reader to McAfoose and Baune (2009) for a more comprehensive accounting of this evidence, though here we provide a brief primer. The vast majority of research in this area has focused on hippocampal-dependent learning and memory using murine models of perturbed inflammatory genes or mechanisms. For example, intracranial injections of IL1β and IL1RN in rats has been shown to impact memory performance as evidenced by altered performance in classic behavioural tests such as the Morris water maze and the inhibitory avoidance task (Brennan et al., 2003, Depino et al., 2004, Pugh et al., 2001, Yirmiya et al., 2002). Likewise, studies utilising transgenic and knock out mouse models have also implicated TNFα and IL6 as additional regulators of memory and learning performance (Aloe et al., 1999, Baune et al., 2008b, Fiore et al., 2000).
In accord with these behavioural studies, numerous molecular studies in animals have highlighted a role for these same cytokines in various biological processesunderlying cognitive function. In particular, IL1β, TNFα, and IL6 have been heavily implicated in the regulation of two opposing forms of classical Hebbian synaptic plasticity, particularly in hippocampal tissue: long-term potentiation (LTP) and long-term depression (LTD) (Balschun et al., 2004, Butler et al., 2004, Schneider et al., 1998). Furthermore, evidence from animal studies also suggests that neurogenesisin the hippocampus is under the influence of inflammatory cytokines such as IL1β and IL6 (Koo and Duman, 2008, Vallières et al., 2002), whereas TNFα has been shown to be important for a non-Hebbian form of synaptic plasticity called synaptic scaling, which functions to facilitate Hebbian synaptic plasticity by promoting homoeostasis during periods of prolonged synaptic inactivity or hyperactivity (Stellwagen and Malenka, 2006). Thus, based on these lines of evidence, McAfoose and Baune (2009) have outlined a “cytokine model of cognitive function” whereby the pleiotropic cytokines IL1β, IL6, and TNFα – amongst other cytokines – are proposed to play a major role in memory, learning, and relevant underlying mechanisms.
Despite this animal evidence clearly indicating “causal” capability, it remains to be determined how the cytokine model of cognitive function might relate to cognitive decline as a consequence of normal ageing. Indeed, as mentioned above, the up-regulated inflammatory markers observed in aged populations (Fuchs et al., 2013, Karim et al., 2014, Yaffe et al., 2003) and dementia patients (Faria et al., 2014, Yarchoan et al., 2013) may in fact be a consequence of cognitive decline as opposed to a “cause” or an exacerbating factor, which again underscores the importance of assessing genetic association studies in this field. Further, even if we assume inflammatory mechanisms do exert a “causal” effect, it is not clear as to which biological substrate – or substrates – outlined within the cytokine model of cognitive function might be affected, or whether ageing- and pathology-specific factors such as amyloid plaque toxicity might in fact be more applicable.
Nevertheless, in the context of the current work, a basic, logical extension of the cytokine model of cognitive function assumes that the majority of non-demented elderly individuals are able to tolerate the increased cytokine activity associated with normal ageing without significant cognitive decline. However, when this increased cytokine activity occurs on a background of already up-regulated inflammatory mechanisms due to some genetic predisposition, this subsample of elderly individuals will be susceptible to deficits in synaptic plasticity, synaptic scaling, and/or neurogenesis, thereby leading to significant cognitive decline. Again, a comprehensive assessment of published genetic association studies investigating immune genes in the context of non-demented elderly cognitive decline represents a first step in validating this hypothesis.
3. Experimental procedures
Our search through databases PubMed, Medline, Scopus, PsycInfo and Web of Science was targeted to identify studies investigating genetic associations between cognitive ageing and immune genes using the following key words: ‘genetic variants’, ‘polymorphisms’, ‘immune genes’, ‘cognitive aging’, ‘age-related cognitive decline’, ‘association study’, ‘GWAS’. We defined cognitive ageing as cognitive performance of samples with a mean age >60 years old who did not fulfil clinical criteria for dementia. However, we included Mild Cognitive Impairment (MCI) as it is defined as a transitional state between healthy cognition and clinically diagnosable dementia. We defined immune genes as genes known to play a role in the regulation of the immune system. The search was not limited by date of publication; therefore we cover earlier studies as well as very recent publications. The search was limited to articles written in English. Following these criteria we identified 25 candidate-gene association papers and 3 GWAS publications, which collectively highlight a total of 19 immune genes; 9 encoding inflammatory cytokines, 2 encoding proteins involved in the complement system, and 8 encoding a miscellaneous class of proteins we refer to as “inflammatory regulators”. See Table 1for a list of these genes along with some basic descriptive and functional information in the context of immunity and the biological substrates of cognition.
| Gene symbol | Protein name | Chromosome | Immune function | Innate or adaptive immunity? | Cognition-related mechanism |
|---|---|---|---|---|---|
| IL1α | Interleukin 1α | 2q13 | Cytokine, pro-inflammatory | Innate | Facilitates memory formation (Depino et al., 2004) via enhancing neurogenesis (Greco and Rameshwar, 2007) |
| IL1RN | Interleukin 1 receptor antagonist | 2q13 | Cytokine, pro-inflammatory | Innate | Implicated in hippocampal-dependant learning and memory (del Rey et al., 2013) |
| TNFα | Tumour Necrosis Factor α | 6p21.33 | Cytokine, pro-inflammatory | Innate | Promotes homoeostasis during periods of prolonged synaptic inactivity or hyperactivity (synaptic scaling) (Stellwagen & Malenka, 2006) |
| IL6 | Interleukin 6 | 7p15.3 | Cytokine, pro- and anti-inflammatory | Innate | Physiological role in fine tuning of memory consolidation (Balschun et al., 2004), possible role in termination of LTP (Yirmiya and Goshen, 2011). Long-term over-expression of IL-6 reduces hippocampal neurogenesis (Vallieres et al., 2002) |
| IL10 | Interleukin 10 | 1q32.1 | Cytokine, anti-inflammatory | Innate | Prevents reactive oxygen species production by overwriting the IL-1β-induced inhibition of glutamate release during LTP (Kelly et al., 2001), therefore plays a protective role in memory formation |
| IL15 | Interleukin 15 | 4q31.21 | Cytokine, pro-inflammatory | Innate, adaptive | Involved in normal hippocampal activity by regulating GABA transmission (He et al., 2010) |
| IL18 | Interleukin 18 | 11q23.1 | Cytokine, pro-inflammatory; interferon-gamma inducing factor | Innate, adaptive | Plays a physiological role in learning and LTP (del Rey et al., 2013); in pathology increases AD-associated Aβ production (Sutinen et al., 2012) |
| TGFβ1 | Transforming growth factor β | 19q13.2 | Cytokine, pro- and anti-inflammatory | Innate | Plays a physiological role in synaptic transmission and memory formation (Caraci et al., 2015) |
| CRP | C-reactive protein | 1q23.2 | Complement system | Innate | Association between elevated level of peripheral CRP, poorer memory, and decreased medial temporal brain volume suggests detrimental role of CRP on cognitive functioning (Bettcher et al., 2012) |
| IL1βCE | Interleukin 1β-converting enzyme | 11q22.3 | Cleaves pro-IL1β to produce mature IL1β, up-regulates IL-1α production | Innate | Can exert indirect effects on cognitive functioning through the regulation of IL-1α and IL-1β production |
| LGALS3 | Galectin 3 | 14q22.3 | Antimicrobial activity against bacteria and fungi | Innate, adaptive | Association between elevated serum level and MMSE scores in AD patients suggesting a neuropathological role of Gal-3 in cognitive functioning (Wang et al., 2013) |
| MX1 | Myxovirus resistance protein 1 | 21q22.3 | Interferon-induced GTP-binding protein | Innate, adaptive | Unknown |
| HS3ST4 | heparan sulphate (glucosamine) 3-O-sulfotransferase 4 | 16p12.1 | Plays a role in herpes simplex virus type 1 pathogenesis | Innate, adaptive | Unknown |
| SPOCK3 | Sparc/osteonectin, Cwcv and Kazal-like domains proteoglycan (Testican) 3 | 4q32.3 | Regulates cytokine secretion in response to smallpox vaccine | Innate, adaptive | Unknown |
| PDE7A | Phosphodiesterase 7A | 8q13.1 | Mediates expression of pro-inflammatory cytokines during T-cell activation | Innate, adaptive | Unknown |
| CLU | Clusterin, apolipoprotein J | 8p21.1 | ? | Innate | Association with elevated blood levels and cognitive decline in AD patients suggesting the CLU role in neuropathological signalling cascade (Schrijvers et al., 2011) |
| PICALM | Phosphatidylinositol binding clathrin assembly protein | 11q14.2 | ? | ? | Being involved in endocytic trafficking, PICALM modulates autophagy and alters clearance of tau in AD pathology (Moreau et al., 2014) |
| CR1 | Complement receptor 1 | 1q32.2 | Complement system | Innate | May inhibit neurogenesis in dentate gyrus (Moriyama et al., 2011) |
4. Results
4.1. Inflammatory cytokines
4.1.1. Interleukin 1β (IL1β)
IL1β has been one of the most prominent candidates featuring in genetic association studies of cognitive ageing in non-demented population studies. In a cross-sectional study utilising a population sample of n=385 Caucasians, consisting of n=172 females (mean age: 72.2±4.4) and n=197 males (mean age: 73.1±4.4), Baune et al., 2008a, Baune et al., 2008b found a highly significant association between the IL1β SNP rs16944 and episodic memory (p=0.003) (Baune et al., 2008a). Specifically, individuals homozygous for the major C allele exhibited better episodic memory performance than the CT/TT group. However, no significant associations were found between rs16944 and attention/processing speed or motor function, suggesting that IL1β effects on cognition may be memory-specific.
In accordance with this, another cross-sectional study focusing on rs16944 in n=161 elderly (mean age: 78.7±4.3) Han Chinese men found a significant association with working memory as measured by the backward Wechsler digit span task (WDST) (p=0.004) (Tsai et al., 2010). However, the authors found no significant associations between rs16944 and the forward WDST (p=0.621) or the short-term memory subscale of the Cognitive Abilities Screening Instrument (CASI), whereas they did find a significant association between rs16944 and the abstraction and judgement subscale of the CASI (p=0.010) suggesting the role of rs16944 in cognition may in fact be more generalised. Nevertheless, in both of the significant (p<0.05) associations revealed by Tsai et al., major C homozygotes performed best in corresponding cognitive tasks, which is in accordance with Baune et al. suggesting the T allele of rs16944 exerts a detrimental effect on cognition in the elderly.
Complementing the Tsai et al. study above, Sasayama et al. (2011) utilised a sample of n=99 elderly (mean age: 65.0±3.8) Japanese women. Rather than focusing solely on rs16944 they adopted a haplotype-tagging approach for SNP selection, which yielded five SNPs for genotyping including rs16944. Although no significant associations were found with rs16944, the authors did find significant associations for both rs1143634 (p=0.0037) and rs1143633 (p=0.010) with a composite score measuring verbal intelligence quotient (IQ), but not with a performance IQ composite. Interestingly, rs1143644 was significantly associated with the digit span task component of the verbal IQ composite (p<0.05), which again highlights a possible role for IL1β in memory processes. Furthermore, a longitudinal study utilising a sample of n=481 elderly (mean age at baseline: 72.47±5.59, 63% female) African American (AA) participants has reported significant associations for two IL1β SNPs, rs1143627 (p=0.006) and rs1143634 (p=0.016), with global cognition at baseline (but not rate of decline) (Benke et al., 2011). There is however two additional studies featuring IL1β SNPs that found no significant associations with either global or domain-specific cognition in elderly samples (Lau et al., 2012, Marioni et al., 2010).
Finally, a functional magnetic resonance imaging (fMRI) study investigating the IL1βSNP rs1143627 has implicated IL1β in resting state brain activity, which may mediate the apparent effects of this gene on cognition (Zhuang et al., 2012). Indeed, using a sample of healthy elderly controls (n=33) (mean age: 72.85±3.39, 45% female) and amnestic mild cognitive impaired (a-MCI) participants (n=47) (mean age: 71.96±4.78, 40% female), the authors found significant genotype and genotypegroup effects on resting-state functional magnetic resonance imaging (fMRI) data across numerous brain regions within the frontal, parietal, and occipital lobes (p<0.05). However, the authors found no significant associations between rs1143627 and various measures of cognitive performance in their aMCI group, though this may be due to the very small sample size (n=47).
4.1.2. Interleukin 1α (IL1α)
A prospective association study utilising a sample of n=2091 Scottish Caucasians (mean age: 67.2±6.51; 72.7% female) found nominally significant associations between two IL1α SNPs, rs2856838 and rs3783654, with mental flexibility (p<0.05) (Marioni et al., 2010). However, the authors found no significant associations with executive function, processing speed, non-verbal reasoning, or delayed/immediate memory. In a case-control study comparing subjective memory complainers (SMC; n=226) (mean age: 63.51±0.69; 64.2% female) and non-complainers (NMC; n=167) (mean age: 62.37±0.67; 70.7% female), Lau et al. (2012)found no significant difference in genotype frequencies for the IL1α SNP rs1800587 between these groups. Likewise, neither was there evidence of a significant association with global cognitive performance when the SMC and NMC groups were combined. Interestingly, in the above-mentioned longitudinal study by Benke et al. (2011), who also utilised a Caucasian sample of n=3575 elderly (mean age at baseline: 72.67±5.42, 57% female) in addition to their AA sample, the authors found a significant association between the IL1α SNP rs17561 (p=0.016) and annual rate of decline in memory performance. This is despite the fact that no significant associations were evident for this SNP in cross-sectional analyses at baseline, thereby highlighting the importance of the longitudinal design in cognitive ageing research.
4.1.3. Interleukin 1 receptor antagonist (IL1RN)
In this same longitudinal study, Benke et al. (2011) also found significant associations between three IL1RN SNPs, rs17042917 (p=0.016), rs4251961 (p=0.0027), and rs931471 (p=0.0032), with global cognition in their Caucasian sample at baseline. To date this is the only study focusing on IL1RN SNPs within the context of cognitive decline specifically in non-demented elderly samples, and so there is clearly a need for replication in independent samples.
4.1.4. Tumour Necrosis Factor α (TNFα)
Baune et al., 2008a, Baune et al., 2008b found a highly significant association (p=0.004) between the TNFα SNP rs1800629 and attention/processing speed (Stroop test), whereby the GA/AA group performed better than the GG group. In contrast, no significant associations were found with either episodic memory or motor function, which, as was the case with IL1β, suggests the role of this SNP might be domain-specific. However, in a longitudinal study utilising n=112 Octogenaraints (60% female) Danish Caucasians with baseline measurements at 80 years of age and follow-up at 85, rs1800629 was significantly (p=0.04) associated with 5-year decline in overall IQ (Krabbe et al., 2009), indicating this SNP/gene may in fact influence global cognition. Furthermore, the direction of this association was also discordant with Baune et al., whereby decline was greater in heterozygotes (GA) relative to major G homozygotes. Unfortunately this study did not report any attention- or processing speed-specific findings, so it is not possible to determine whether this global effect may have been driven primarily by domain-specific decline.
In the aforementioned study comparing SMCs with NMCs, Lau et al. (2012) found no significant associations with global cognition for either rs1800629 or rs1799724 upstream of TNFα. Likewise, in another cross-sectional study, Gajewski et al. found no significant association between rs1800629 and global cognition in 131 aged (mean age: 70.5±4.5, 61.8% female) German participants, though they did observe a significant association with short-term memory performance (p<0.010) (Gajewski et al., 2013). The direction of this association was in accordance with Krabbe et al. (2009), whereby GG homozygotes performed better than the GA/AA group. Interestingly, in addition to standard cognitive testing, Gajewski et al. (2013) also administered an electroencephalogram (EEG) to measure event-related potentials (ERPs) whilst participants performed a visual search task, which is widely considered to be a valid intermediate phenotype for attention and executive function. They found multiple significant (p<0.05) associations between rs1800629 and ERP latency and amplitude across various brain regions, indicating the A allele of rs1800629 confers attentional dysfunction. This was also supported by behavioural data from the visual search task where significantly slower reaction times and a higher rate of missed targets were observed in the GA/AA group.
4.1.5. Interleukin 6 (IL6)
To date, four genetic association studies have investigated IL6 SNPs in the context of cognitive ageing. Baune et al., 2008a, Baune et al., 2008b found a near significant (p=0.05) trend for rs1800796 (putative IL6 promoter SNP) with motor function, which clearly requires replication in an independent sample. The aforementioned study by Marioni et al. (2010) did not reveal any significant associations between any of the IL6 SNPs they genotyped (rs2069832, rs2069840, rs1800795) with either global or domain-specific cognition (Table 1.), though it is worth noting that the authors did not include a measure of motor function so a direct comparison with Baune et al. is not possible. Likewise, the three IL6 SNPs Marioni et al. selected were not in appreciable linkage disequilibrium with rs1800796 (r2<0.52) in Caucasians. Furthermore, two additional studies utilising the above-mentioned Danish Octagenarians did not reveal any significant associations of rs1800795 with verbal IQ, performance IQ, overall IQ (Krabbe et al., 2009), or global cognition (Dato et al., 2010).
4.1.6. Interleukin 10 (IL10)
Krabbe et al. (2009) and Dato et al. (2010) also looked at IL10 SNPs in this same sample of octogenarians. After performing both cross-sectional and longitudinal analyses using cognition data at 80 and 85 years, the study by Krabbe et al. revealed no significant (p<0.05) associations for either of the IL10 SNPs genotyped, rs1800872 and rs1800896, with measures of verbal IQ, performance IQ, or full IQ. However, in contrast the study by Dato et al., which focused solely on rs1800896, did reveal significant associations with both the MMSE (p=0.01) and a composite measure of global cognition based on five cognitive tests (p=0.01) in females, but not males (p=0.96 and p=0.22, respectively). For both the MMSE and the composite measure in females, better performance was found in carriers of the A allele of rs1800896 relative to the T allele. Thus, these findings suggest the role of IL10 may be sex-specific.
To date, there has been one additional study focusing on IL10, which utilised a sample of n=138 elderly (mean age: 80.37±5.93) Caucasians with mild cognitive impairment (MCI) along with n=63 AD patients and n=63 healthy controls matched for age and sex (Arosio et al., 2010). Like Dato et al. (2010), this was a single SNP study focusing solely on rs1800896. Initially, the authors found a near significant (p=0.05) difference in genotype frequencies between MCI, AD, and healthy controls, along with a significant (p=0.02) difference in allele frequencies. Specifically, Arosio et al. found a greater frequency of A alleles in AD patients relative to both MCI subjects and healthy controls, with MCI subjects forming an intermediate. Furthermore, in secondary analyses comparing MCI subjects stratified on the basis of whether they presented as amnestic MCI (a-MCI; n=30) or multiple impaired cognitive domain MCI (mcd-MCI; n=108), the authors found significant differences in both genotype (p<0.05) and allele (p=0.02) distributions, with an apparent overrepresentation of A alleles in the a-MCI group. Thus, these secondary analyses indicate a possible role for the A allele of rs1800896 specifically in memory decline, though this is discordant with the overall direction observed by Dato et al. above.
4.1.7. Interleukin 15 (IL15) and Interleukin 18 (IL18)
To date, IL15 and IL18 have only been investigated in the aforementioned Danish octogenarians. Dato et al. (2010) genotyped four SNPs in and around IL15 and found significant associations of rs2254514 (p=0.02) and rs2322262 (p=0.03) with a composite global cognition score at baseline (i.e., 80 years) based on five cognitive tests in females only, not males (p=0.73 and p=0.64, respectively). Furthermore, providing partial support for this the authors also found a near significant association between the former SNP, rs2254514, and the mini-mental state examination (MMSE) in females (p=0.06) but not males (p=0.26).
For IL18, Krabbe et al. genotyped two SNPs, rs1946518 and rs187238, and performed haplotype analyses. They found significant associations between this two-SNP haplotype and performance IQ in cross-sectional analyses at both 80 (p=0.03) and 85 (p=0.02) years, though longitudinal analyses did not reveal any significant differences in rate of decline over five years. Thus, the stability of this apparent IL18 effect on cognition suggests it may have occurred at an earlier time point. In contrast to Krabbe et al., Dato et al. (2010) performed single SNP analyses focusing solely on rs187238 at baseline (i.e., 80 years). They found a near significant association between this SNP and their composite cognitive score in males (p=0.05) but not females (p=0.55), though no significant associations were evident with the MMSE.
4.1.8. Transforming growth factor β1 (TGFβ1)
Despite the established immunosuppressive role of TGFβ1 (Kobie and Akporiaye, 2003), this gene has featured in just one candidate gene association so far in the context of cognitive ageing. Using a similar experimental design to the above study investigating IL10, Arosio et al. compared the genotype and allele frequencies of two TGFβ1 SNPs, rs1800470 and rs1800471, between MCI (n=48), AD (n=193), and healthy control (n=198) participants (Arosio et al., 2007). In initial analyses the authors found a significant difference in both genotype (p=0.025) and allele (p=0.009) frequencies of rs1800470 between AD patients and healthy controls, but not MCI participants. Specifically, this difference consisted of a ~10% increase in C alleles in the AD group relative to healthy controls. No significant differences were found for rs1800471. In secondary analyses after a four-year follow-up, Arosio et al. also found a significant difference in both genotype and allele frequencies of rs1800470 between stable MCI (n=18) participants and those that progressed to AD (n=21), with a ~20% increased in C alleles in the newly diagnosed AD group. Thus, this study highlights rs1800470 as a risk factor for AD, though it is unclear from these findings alone whether it also involved in cognitive decline as a result of normal ageing. Replication is clearly required in a large population sample.
4.2. C-reactive protein (CRP)
Beyond inflammatory cytokines, another major component of the immune system and the inflammatory response is the complement system. C-reactive protein (CRP) is a particularly prominent molecule of this system, and has featured in two genetic association studies focusing on cognitive decline in the non-demented elderly. The first such study utilised four distinct Caucasian population samples, each with genotype data for at least three CRP SNPs (Marioni et al., 2010). In a sample of n=2091 elderly (mean age: 67.2±6.5, 73% female) individuals, out of a total of eight independent cognitive measures the authors found significant associations between both rs1130864 (p=0.016) and rs1417938 (p=0.017) with a cognitive test measuring mental flexibility, the latter of which was also significantly associated with general cognitive ability (p=0.034). In another sample of n=1066 elderly (mean age: 67.9±4.2, 48% female) individuals with data from ten independent cognitive measures, Marioni et al. found significant associations between rs1130864 and working memory (p=0.012), attention (p=0.030), and emotion processing (p=0.017). There was no overlap with the previous sample in terms of significant associations, though it should be acknowledged that there was no genotype data for rs1417938 in this latter sample. Further, no significant associations were found in the other two samples (n=534, n=1091), both of which featured seven independent cognitive measures.
The second study featuring CRP utilised a longitudinal design with a total sample size of n=5680 elderly (mean age at baseline: 75.3±3.3, ~50% female) Caucasian individuals from the PROspective Study of Pravastatin in the Elderly at Risk (PROSPER) sample (Mooijaart et al., 2011). Three CRP SNPs – rs1417938, rs1800947, rs1205 – were selected based on previous evidence showing they influence peripheral CRP levels, and by performing haplotype analyses the authors did observe some nominally significant associations with annual rate of change in delayed (p=0.044) and immediate (p=0.039) memory recall. Thus, together with the above study by Marioni et al. (2010) there is evidence from both cross-sectional and longitudinal analyses of nominally significant associations between CRP and cognitive impairment in the elderly (Table 2).
| Author | Study design | Samplecharacteristics Ethnicity | Cognitive Phenotype | Genotype Gene SNPs | Key findings | Replication |
|---|---|---|---|---|---|---|
| Arosio et al. (2007) | Prospective | Clinical N=439 (198 Controls, 48 MCI; 193 AD); age range 63–99yrs; 66.5% female Caucasian (Northern Italy) | Global cognition | TGFβ1 | Cross-sectionally: rs1800470 difference in both genotype and allele frequencies between AD and controls, but not MCI; Longitudinally: difference between MCI and those that progressed to AD; no differencies for rs1800471 | No |
| rs1800470 | ||||||
| rs1800471 | ||||||
| Cognitive measures: MMSE | ||||||
| Case-control | ||||||
| MCI (Petersen criteria) | ||||||
| 4-year follow-up | ||||||
| Arosio et al. (2010) | Prospective | Population-based N=264 (138 MCI (30 a-MCI, 108 mcd-MCI), 63 AD, 63 Controls); mean age±SD 80.4 ± 5.9 Caucasians (Northern Italy) | Global cognition | IL10 | near significant difference in genotype frequencies between MCI, AD, and healthy controls, along with a significant difference in allele frequencies; overrepresentation of A alleles in the a-MCI group | No |
| Cognitive measures: | rs1800896 | |||||
| Case-control | a-MCI; mcd-MCI (Petersen criteria) | |||||
| Baune et al. (2008a) | Cross-sectional | Population-based N=369; mean age±SD 72.7±4.4; 46.6% female | Global cognition episodic memory attention and processing speed motor function cognitive measures: MMSE, three word recall, Stroop, Purdue Pegboard | IL1β | rs16944 (IL1B)associated with episodic memory (CC allele performed better than T/TT group); rs1800629 (TNFa)associated with attention/processing speed (GA/AA group performed better than GG group); rs1800796 (IL6)near-significant trend for motor function | No |
| rs16944 | ||||||
| TNFα | ||||||
| Caucasians | rs1800629 | |||||
| IL6 | ||||||
| rs1800796 | ||||||
| Benke et al. (2011) | Prospective ≥3 follow-ups | Clinical N=4056 (3575 Caucasians, 481 AA); Mean age±SD at baseline: Caucasians 72.67±5.42, AA | Global cognition Processing speed Cognitive measures:MMSE 3MS DSST VFT | IL1β | Association for rs1143627 and rs1143634 (IL1β) with global cognition at baseline (AA); rs17042917, rs4251961 and rs931471 (IL1RN)associated with baseline performance on 3MS and rs17561 (IL1α) associated with annual rate of memory (Caucasians) | No |
| IL1α | ||||||
| IL1RN | ||||||
| 31 SNP | ||||||
| 72.45±5.59; Gender (% female): Caucasians 57, AA 63 | ||||||
| Dato et al. (2010) | Prospective | Population-based N=1,651; mean age±SD 93.1(±0.3); 72% female; Danish Caucasians | General cognitive functioning Cognitive measures: MMSE VFT | IL6 | rs1800896 (IL10)association with global cognition in females, but not males; rs2254514 and rs2322262 (IL15) association with global cognition at baseline in females only; rs187238 (IL18) near significant association with cognitive score in males, but not females | Yes |
| IL10 | ||||||
| IL15 | ||||||
| IL18 TNFα | ||||||
| 9 SNPs | ||||||
| Gajewski et al. (2013) | Cross-sectional | Population-based N=131; mean age±SD 70.5±4.5, 61.8% female Caucasians | Global cognition (5 domains) Cognitive measures: MMSE, MWT-B, RWT, ROSF, TMT, VLMT, DSST, Stroop, WDST | TNFα | association with short-term memory, GG performed better than GA/AA; No association with global cognition or any other specific domains | No |
| rs1800629 | ||||||
| Krabbe et al. (2009) | Prospective 5 years follow-up | Clinical N=112; Age 80 and 85; 59.8% female Danish Caucasians | IQ: ‘Fluid’ and ‘Crystallised’ intelligence Cognitive measure: WAIS | IL6 | IL18 haplotype (rs187238 and rs1946518 is an independent risk factor of poor performance IQ; rs1800629 (TNFα)associated with 5-year decline in overall IQ (decline greater for GA relative to major GG); IL6 no association with IQ | No |
| rs1800795 | ||||||
| IL10 | ||||||
| rs1800896 | ||||||
| rs1800872 | ||||||
| IL18 | ||||||
| rs187238 | ||||||
| rs1946518 | ||||||
| TNFα | ||||||
| rs1800629 | ||||||
| Lau et al. (2012) | Cross-sectional case-control | Community-based | subjective memory complainers (SMC), non-complainers (NMC) | IL1α | no effect of any of the genetic factors investigated on cognitive performance | No |
| N=393 (226 cases, | rs1800587 | |||||
| 167 controls); | IL1β | |||||
| Age range 34–94; | ||||||
| 66.9% female | Cognitive measures:CAMDEX-R | rs1143634 | ||||
| Australians | TNFα | |||||
| CAMCOG-R | ||||||
| rs1800629 | ||||||
| rs1799724 | ||||||
| Marioni et al. (2010) | Prospective 10 years follow-up | Population-based ±SD 67.2±6.5; 72.7% female Scottish Caucasians | General intelligence factor ‘g’ | IL1α | rs2856838 and rs3783654 (IL1α)nominally associated with mental flexibility; | No |
| IL1β | ||||||
| Cognitive measures: | no associations between any of the IL6 or | |||||
| MMSE, VFT, DSST, RAVENS, RAVLT, TMT, MHVS | IL6 | IL1β SNPs with either global or domain specific cognition | ||||
| 8 SNPs | ||||||
| Persson et al. (2014) | Prospective | Population-based | 10 brain regions | IL1β | rs16944 (IL1β)(carriers of two T alleles) | No |
| N=167 at baseline, 90 at follow-up; | Brain activity measure: | rs16944 CRP | showed increased parahippocampal gyrus shrinkage; | |||
| age range 19–79; USA (metropolitan area) | MRI | rs3091244 | rs3091244 (CRP)- no effect | |||
| Sasayama et al. (2011) | Cross-sectional | Community-based | Global cognition | IL1β | rs1143634 and rs1143633 associated with verbal IQ score, strongest performance in GG and TT;rs16944 no association with IQ | No |
| N=99; | rs2853550 | |||||
| mean age±SD 65±3.8;Females only | Cognitive measure: WAIS-R | rs1143634 | ||||
| Japanese | rs1143633 | |||||
| rs1143630 rs16944 | ||||||
| Tsai et al. (2010) | Cross-sectional | Community-based N=161; | Global cognition | IL1B | Association with backward WDST; no association between rs16944 and forward WDST or short term memory scale (CASI); association between rs16944 and abstraction and judgement scale (CASI), CC performed best | No |
| Working memory | rs16944 | |||||
| mean age±SD 78.7±4.3 Males only | ||||||
| Cognitive measures: | ||||||
| Chinese | MMSE | |||||
| CASI | ||||||
| WDST | ||||||
| Zhuang et al. (2012) | Prospective Case-control | Clinical | Brain spontaneous activity | IL1β | Genotype and genotypegroup effects on resting-state fMRI across frontal, parietal, occipital lobes; | No |
| N=80 (47 a-MCI, 33 Controls) | rs1143627 | |||||
| Global cognition | ||||||
| Mean age±SD: 72.5±4.1 | Cognitive measures:Resting-state fMRI; | No association between rs1143627 and cognition in a-MCI group | ||||
| Gender (% female: a-MCI 40.4, Controls 45.5) Chinese | ||||||
| MMSE, RAVLT, TMT, DSST, WDST, Clock drawing test |
Abbreviation list:
AA – African–Americans, AD – Alzheimer׳s disease, a-MCI – amnestic Mild Cognitive Impairment (impairment in memory performance), CAMCOG-R – Cambridge Cognitive Examination-Revised (screening dementia in stroke patients), CAMDEX-R – Cambridge Mental Disorders of the Elderly Examination-Revised (differentiates dementia from normal cognitive aging), CASI – Cognitive Abilities Screening Instrument, DSST – The Digit Symbol Substitution Test (psychomotor attention and processing speed), fMRI – functional Magnetic resonance imaging, IQ – Intelligence quotient, MHVS – Mill Hill vocabulary scale, MCI – Mild Cognitive Impairment (assessment of early, but abnormal, state of cognitive impairment), mcd-MCI – multiple impaired cognitive domains Mild Cognitive Impairment (impairment in in at least two cognitive domains), MRI – Magnetic resonance imaging, 3MS – The Modified mini-mental state examination (overall cognitive function), MMSE – Mini Mental State Examination (overall cognitive function), MWT-B – Mehrfachwahl-Wortschatz-Test (premorbid intelligence), RAVENS – Raven׳s standard progressive matrices (non-verbal reasoning), RAVLT – Rey Auditory Verbal Learning Task (immediate and delayed memory), ROCF – Rey-Osterrieth Complex Figure test (visuospatial memory), RWT – Regensburg Word Fluency Test (divergent thinking), SD – standard deviation, SNP – single nucleotide polymorphism, Stroop - Stroop Colour World Test (mental flexibility/executive function), TMT – Trial Making Test (mental flexibility), VLMT – Verbal Learning and Memory Test (verbal memory), VFT – Verbal Fluency Test (executive function), WAIS – Wechsler Adult Intelligence Scales, WAIS-R – Wechsler Adult Intelligence Scales-Revised, WDST – Wechsler Digit Span Test (memory performance)
4.3. Other inflammatory regulators
In addition to the above-mentioned studies focusing directly on cytokine-encoding genes [and CRP], three studies have focused on less direct regulators of the inflammatory response. Indeed, utilising the PROSPER sample (n=5680) that featured in Mooijaart et al. (2011), one such study investigated a possible role for the gene encoding the IL1β-converting enzyme (IL1βCE) in cognitive decline (Trompet et al., 2008). IL1βCE mediates the post-translational cleavage of the inactive precursor of IL1β to produce mature, active, IL1β, and so this gene is known to play a significant role in regulating IL1β levels and activity. In cross-sectional analyses at baseline with four IL1βCE haplotype-tagging SNPs (rs554344, rs488992, rs1977989, rs580253), Trompet et al. found highly significant associations between three of these four SNPs and global cognition, whilst two of them (rs554344, rs580253) were also associated (p<0.05) with attention, processing speed, delayed (but not immediate) memory, as well as composite scores for executive function and memory function. In addition, longitudinal analyses revealed significant (p<0.05) associations between all four SNPs and a change over time in attentional processes, whilst SNPs rs554344 and rs580253 were also significantly associated (p<0.05) with a change over time in processing speed as well as composite scores for global cognition and executive function.
In another study performed by the same group again utilising the PROSPER sample, Trompet et al. switched focus to the gene encoding Galectin-3 (LGALS3) (Trompet et al., 2012), which has previously been implicated in both acute and chronic inflammation as evidenced by studies investigating galectin-3-deficient mice (Henderson and Sethi, 2009, Hsu et al., 2000). Using data from three haplotype-tagging SNPs (rs4644, rs4652, rs1009977), the authors again found multiple significant associations. In cross-sectional analyses all three SNPs were significantly (p<0.05) associated with processing speed and immediate memory, whilst one of these SNPs, rs1009977, was also significantly (p<0.05) associated with attention and delayed memory. In longitudinal analyses all three SNPs were significantly (p<0.05) associated with rate of decline in attention processing, whilst a nominally significant association was observed between rs4644 and rate of decline in immediate memory (p=0.043). However, no significant associations were found for rate of decline in either processing speed or delayed memory performance, whilst no significant associations were found in either cross-sectional or longitudinal analyses for global cognition.
The third study utilised a longitudinal design focusing on the gene encoding the myxovirus resistance protein 1 (MX1), which is an interferon-induced antiviral protein, in Chinese participants (Ma et al., 2012). Although this study was primarily interested in AD, they also included n=316 (mean age at baseline: 79.3±7.5, 86% female) age-matched healthy controls, to whom the authors administered cognitive tests at baseline and then again at a two-year follow-up. Out of 10 haplotype-tagging SNPs and a further two candidate SNPs, Ma et al. found five of them were significantly (p<0.05) associated with global cognitive decline. Thus, given these three studies report largely positive findings, collectively they highlight the importance of looking beyond the major cytokine-encoding genes and focusing on other biological candidates such as cytokine-regulating genes in further studies of cognitive decline (Table 3).
| Author | Study design | Sample characteristics ethnicity | Cognitive phenotype | Genotype gene SNPs | Key findings | Replication |
|---|---|---|---|---|---|---|
| Carrasquillo et al. (2015) | Prospective | Population-based N=2,262; | Cognitive decline (logical memory) | CLU | rs11136000-G (CLU)associated with worse baseline memory and incident MCI; | No |
| 3 years follow-up | PICALM | |||||
| Age range at baseline 49–98; | CR1 | |||||
| rs610932-C | ||||||
| rs3851179-G (PICALM) had an unexpected protective effect on incident MCI | ||||||
| 9 SNPs | ||||||
| Cognitive measures: MCI (Petersen criteria) WMS-R | ||||||
| 66% female | ||||||
| Caucasians | ||||||
| Chibnik et al. (2011) | Prospective | Community-based N=1,666; | Rate of cognitive decline (episodic memory, global cognition) | CR1 | A-allele of rs6656401 (CR1) associated with a faster rate of decline in global cognition as well as multiple cognitive domains including episodic memory and perceptual speed | Yes |
| rs6656401 | ||||||
| mean age±SD at baseline: 78.4±7.1; 69.5% female | ||||||
| CLU | ||||||
| rs11136000 | ||||||
| No association with CLU; | ||||||
| Nominal association with PICALM | ||||||
| Cognitive measures: | ||||||
| Non-Hispanic white | CERAD | PICALM | ||||
| WMS-R | rs7110631 | |||||
| Keenan et al. (2012) | Prospective | Community-based | Episodic memory decline | CR1 | rs4844609 (A-allele) influences episodic memory decline | Yes |
| N=1,709; | 41 SNPs within the block LD containing rs6656401 | |||||
| 3.2 years mean follow-up | mean age±SD at enrolment 78.4±7.1;69.2% female | Cognitive measures: | ||||
| Word list memory, | ||||||
| WMS-R | ||||||
| Non-Hispanic white | RAVLT | |||||
| Ma et al. (2012) | Prospective | Clinical N=316; mean age±SD | Rate of cognitive decline | MX1 10 tagged and 2 candidate SNPs | Carrier of minor alleles of five MX1 SNPs (rs457274, rs2071430, rs461093, rs469083, rs1557372) associated with faster global cognitive decline | No |
| case-control | ||||||
| 2 years follow-up | 79.3±7.5; | Cognitive measures: | ||||
| 86% female | CDR | |||||
| Chinese | ||||||
| ADAS-Cog | ||||||
| WMS-R | ||||||
| Marioni et al. (2010) | Case-control | 3 Population-based samples | General cognitive factor ‘g’ | CRP | rs1130864 and rs1417938 associated with mental flexibility; | Yes |
| rs1205 | ||||||
| N=4,782; | Cognitive measures: | rs1130864 | rs1130864 associated with working memory, attention and emotional processing | |||
| mean age±SD 69.7±4.2; | VFT, DSST, RAVENS, RAVLT, TMT, FACES, LM, LNS, MR, VFT, WMS-R | rs1800947 | ||||
| rs1417938 | ||||||
| rs3093077 | ||||||
| 59.6% female | rs3093068 | |||||
| Scottish Caucasians | ||||||
| Mengel-From et al. (2011) | Prospective | Population-based | General cognition | CLU | rs11136000 T-allele (CLU)associated with cognitive composite score (association stronger in females); | Yes |
| N=1,380; | Cognitive measures: VFT WMS-R | rs11136000 | ||||
| 92–93 years at the enrolment, | PICALM | |||||
| rs3851179 A-allele (PICALM)associated with better cognitive performance (in males) | ||||||
| rs3851179 | ||||||
| 69% female Danish Caucasians | CR1 | |||||
| rs6656401 | ||||||
| Mengel-From et al. (2013) | Prospective | Population-based | Cognitive composite measure | CLU | rs11136000 T-allele associated with better baseline cognitive performance; haplotype (rs11136000, rs1532278, rs9331888) performed better than non-carriers | Yes |
| N=2,224; | rs11136000 | |||||
| Cognitive measures: | ||||||
| 73–93 years | ||||||
| at the enrolment; Danish Caucasians | VFT | rs9331888 | ||||
| WMS-R | rs1532278 | |||||
| rs9331908 | ||||||
| Mooijaart et al. (2011) | Prospective | Clinical N=5,680; mean age±SD 75.3±3.3; 51.7% female | General cognition memory attention processing speed | CRP | Haplotype analyses (all three SNPs) revealed nominally significant associations with annual rate of change in delayed and immediate memory recall | No |
| Cognitive measures: MMSE, Picture-Word Recall test, Stroop, LDST | ||||||
| Caucasians (Scotland, Ireland, Netherlands) | ||||||
| rs1417938 | ||||||
| rs1800947 | ||||||
| rs1205 | ||||||
| Thambisetty et al. (2013) | Prospective eight annual follow-ups | Population-based | Rate of change in memory performance | CLU | Association with faster rates of memory decline relative to noncarriers in the presymptomatic stages of disease progression | No |
| N=694; | rs11136000 | |||||
| mean age±SD | ||||||
| Cognitive measures: | ||||||
| 69.3±7.3; | BVRT | |||||
| 41% female | CVLT | |||||
| Caucasian | ||||||
| African American | ||||||
| Trompet et al. (2008) | Prospective | Clinical N=5,680; mean age±SD 75.3±3.3; 51.7% female | Global cognition Processing speed | IL1βCE | Cross-section analysis: associations between three of the four SNPs (not rs1977989) with global cognition; Longitudinal analyses: associations between all four SNPs and a change over time in attentional processes | No |
| rs554344 | ||||||
| Attention | rs488992 | |||||
| Memory | rs1977989 | |||||
| Caucasians (Scotland, Ireland, Netherlands) | ||||||
| Cognitive measures: | rs580253 | |||||
| MMSE, LDT, Stroop, 15-PLT | ||||||
| (Trompet et al. 2012) | Prospective | Clinical N=5,680;mean age±SD 75.3±3.3;51.7% female | Global cognition | LGAL33 | Cross-sectional analysis: all three SNPs associated with processing speed and immediate memory; rs1009977 associated with attention and delayed memory; | No |
| Processing speed | rs4644 | |||||
| Attention | rs4652 | |||||
| Memory | rs1009977 | |||||
| Caucasians (Scotland, Ireland, Netherlands) | Cognitive measures: | |||||
| MMSE, LDT, Stroop, 15-PLT | Longitudinal analysis: all three SNPs | |||||
| associated with rate of decline in attention | ||||||
| Verhaaren et al. (2013) | Prospective | Population-based N=5,171; age range 45–99; 56.4% female | Global | CLU | In non-demented people, there is only marginal joint effect of these SNPs on cognition | No |
| cognitionExecutive | PICALM | |||||
| function | CR1 | |||||
| Memory | 11 SNPs | |||||
| Processing speed | ||||||
| Cognitive measures: | ||||||
| Caucasians (Netherland) | MMSE, GMSS, RAVLT, Stroop, LDST, VFT |
Abbreviation list:
ADAS-Cog – Alzheimer׳s Disease Assessment Subscale-Cognitive subscale, BVRT – Benton Visual Retention Test (visuospatial memory), CDR – Clinical Dementia Rating, CERAD – Consortium to Establish a Registry for Alzheimer’s Disease, CVLT – California Verbal Learning Test (verbal memory), DSST – The Digit Symbol Substitution Test (psychomotor attention and processing speed), FACES – Faces and family pictures subtest (non-verbal memory, immediate and delayed memory), GMSS – Geriatric Mental State Schedule (general cognition in the elderly), LD – linkage disequilibrium, LDST – The Letter Digit Substitution Test (processing speed and attention), LM – Logical Memory (from Wechsler Memory Scale III), LNS – Letter Number Sequencing (working memory, non-verbal reasoning, processing speed), LTD – Letter–Digit Coding Test (processing speed), MCI – Mild Cognitive Impairment (assessment of early, but abnormal, state of cognitive impairment), MMSE – Mini Mental State Examination (overall cognitive function), MR – Matrix Reasoning (working memory, non-verbal reasoning, processing speed), 15-PLT – 15-Picture Learning test (immediate and delayed recall), RAVENS – Raven׳s standard progressive matrices (non-verbal reasoning), RAVLT – Rey Auditory Verbal Learning Task (immediate and delayed memory), SD– standard deviation, SNP – single nucleotide polymorphism, Stroop – Stroop Colour World Test (mental flexibility/executive function), TMT – Trial Making Test (mental flexibility), VFT – Verbal Fluency Test (executive function), WMS-R – Wechsler Memory Scale-Revised
4.4. Evidence from genome-wide association studies
As an alternative to selecting additional biological candidates beyond the major inflammatory cytokines for candidate gene association studies, genome-wide association studies potentially represent a more efficient means of highlighting less obvious, less direct, regulators of inflammatory responses for further study. To date, three GWAS have been performed focusing on cognition in non-demented elderly individuals (Davies et al., 2015, De Jager et al., 2012, Debette et al., 2015). Importantly, all three studies were concordant in highlighting the APOE locus with p-values exceeding the generally accepted genome-wide significance threshold (p<510−08), which arguably serves as a positive control considering the wealth of previous literature implicating this locus in cognition and dementia.
In addition to the APOE locus, Debette et al. (2015) (n=29,076; mean age: 63.6±7.0, 56% female) highlighted two additional genome-wide significant SNPs associated with verbal declarative memory performance: rs11074779 (p=3.1110−08) downstream of HS3ST4, and rs6813517 (p<2.5810−08) upstream of SPOCK3. Interestingly, both HS3ST4 and SPOCK3 represent potential candidates for involvement in the immune response, whereby HS3ST4 has been implicated in herpes simplex virus (HSV)-1 pathogenesis (Tiwari et al., 2005) whilst SPOCK3 has been found to be associated with variations in cytokine secretion in response to smallpox vaccine (Kennedy et al., 2014). Likewise, in De Jager et al. (2012)(n=3028; mean age: 75.85±6.55, 60% female) the next most highly significant SNP associated with global cognition after the APOE locus was rs10808746 (joint discovery and replication: p=2.3010−05) localised within an intron of PDE7A, which encodes a phosphodiesterase previously implicated in mediating the expression of proinflammatory cytokines during T cell activation (Kadoshima-Yamaoka et al., 2009, Yang et al., 2003). Although the third GWAS by Davies et al. (2015)(n=53,949, >45 years, 57% female) did not highlight any obvious inflammatory candidates, collectively these GWAS have identified three potential inflammatory genes as possible regulators of cognitive ageing, which may warrant further confirmatory research.
In addition to these three GWAS investigating cognitive decline there have been numerous GWAS of late-onset AD (LOAD), the findings from which could potentially be transferrable to normal cognitive ageing. Two particularly prominent examples, both with n>15,000, have highlighted three novel loci in the aetiology of LOAD at genome-wide significance: CLU (rs11136000), PICALM (rs7110631), and CR1 (rs6656401) (Harold et al., 2009, Lambert et al., 2009). Interestingly, not only have these three loci received considerable attention in subsequent replication studies in non-demented elderly samples, but they have also previously been implicated to one extent or another in the regulation of the inflammatory response; i.e., CR1 encodes Complement Receptor 1, which, like CRP, is involved in the complement system (Lambert et al., 2009).
The first replication study, by the De Jager group (Chibnik et al., 2011), utilised longitudinal data from a sample of n=1666 elderly (mean age: 78.25±7.0, 70% female) participants. The authors found the A allele of rs6656401 (CR1) to be significantly associated with a faster rate of decline in global cognition (p=0.011) as well as multiple cognitive domains including episodic memory (p=0.003) and perceptual speed (p=0.002). Accordingly, it was the A allele of rs6656401 that had previously been identified as the risk allele for LOAD (Lambert et al., 2009). However, Chibnik et al. were unable to replicate the CLU locus, whilst only achieving nominal significance for the PICALM locus. Although two subsequent studies were unable to convincingly replicate Chibnik et al. (Carrasquillo et al., 2015, Mengel-From et al., 2011), further work by the De Jager group suggests that rs6656401 may not be the best proxy for this CR1 signal and highlights a non-synonymous coding SNP in the same linkage disequilibrium block, rs4844609, as a potential candidate (Keenan et al., 2012). Thus, further replication studies with genotype data for rs4844609 are now required to further explore the role of the CR1 locus in cognitive ageing.
Concerning the CLU and PICALM loci, a study utilising n=1380 (92–93 years, 69% female) Danish Caucasians found significantly better global cognition associated with the T allele of rs11136000 (CLU) (p=0.016) and the A allele of rs3851179 (PICALM) (p=0.024), though the latter association was only evident in males (Mengel-From et al., 2011). Furthermore, another study by the same group focused solely on the CLU locus, adding additional SNPs beyond rs11136000 (Mengel-From et al., 2013). The authors found multiple significant genotype and haplotype associations with baseline cognition as well as rate of decline in the same Danish Caucasians as well as a slightly younger replication cohort (n=573; mean age: 78.8 (73–95)). Likewise, in a study investigating nine loci previously implicated in LOAD (including CR1, CLU, and PICALM) in n>2200 elderly (median age: 77 (49–98), 66% female) Caucasians, CLU was the only locus found to be significantly associated with logical memory performance at baseline (p=0.012), though not five-year change (p=0.32) (Carrasquillo et al., 2015). However, in a longitudinal study following n=694 elderly (mean age: 69.3±7.3, 41% female) individuals, the authors found that the CLU SNP rs11136000 was significantly (p<0.05) associated with faster rates of memory decline in those individuals who progressed to MCI or AD (n=95), though this was not the case in those individuals who remained cognitively normal (n=599) (Thambisetty et al., 2013). This suggests the CLU locus may in fact be specific to pathological as opposed to normal cognitive ageing.
Finally, two recent studies calculated polygenic risk scores using genotype data from either nine (Verhaaren et al., 2013) or eleven (Carrasquillo et al., 2015) SNPs previously implicated in LOAD (including CR1, CLU, and PICALM) along with the APOE locus. These studies found significant associations between these risk scores and logical memory (n>2000) (Carrasquillo et al., 2015), global cognition, memory function, executive function, and processing speed (n=5171) (Verhaaren et al., 2013). However, when polygenic risk scores were calculated in the absence of the APOE locus, in both studies these significant associations were lost. Thus, these risk score associations were clearly driven by the APOE locus as opposed to the additional LOAD risk loci included such as CR1, CLU, and PICALM (Table 4).
| Phenotype | Sample characteristics | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Study | Design | Ethnicity | Platform | # of SNPs | Cognitive functions | Cognitive measures | Sample size | Age | Featured genes | Replication |
| Davies et al. (2015) | Prospective | European | Illumina610-Quadv1 chip | 549,692 | General intelligence factor | MHT | N=53,949 | ≥45; 57% female | No association with inflammatory genes | 1,367 additional subjects from:SATSA Gender OCTO-Twin |
| RAVENS | (5 cohorts) | |||||||||
| CAGES | LM | |||||||||
| VFT | ||||||||||
| Debette et al. (2015) | Prospective Heart and Aging Research in Genomic Epidemiology consortium | European, African–Americans | Various Affymetrix and Illumina platforms | N/A | Verbal declarative memory | word list delayed recall (WL-dr), | N=29,076 (19 population-based cohorts) | mean age±SD 63.6 ± 7; 56% female | rs11074779 near | 10,617 additional participants of European descent, 3811 of AA ancestry |
| HS3ST4and of rs6813517 near SPOCK3associated with verbal declarative memory | ||||||||||
| paragraph | ||||||||||
| delayed recall (PAR-dr) | ||||||||||
| De Jager et al. (2012) | Prospective | European | Affymetrix GeneChip 6.0 | 672,266 | Global cognition Cognitive decline slope | 17 cognitive tests | N=3,028 | mean age±SD 75.9±6.6; 60% female | rs10808746 (PDE7A) associated with global cognition | 2,279 additional subjects for top 50 SNPs |
| ROS | non-Hispanic | |||||||||
Abbreviation list:
AA – African–Americans, RAVENS – Raven׳s standard progressive matrices (non-verbal reasoning), CAGES – Cognitive Ageing Genetics in England and Scotland project, Gender – The Sex Differences in Health and Aging Study, LM – Logical Memory (from Wechsler Memory Scale III), MHT – Moray House Test (verbal reasoning), OCTO-twin – The Study of Origins of Variance in the Oldest-Old, ROS – Religious Older Study, SATSA - The Swedish Adoption/Twin study of Aging, SD – standard deviation , SNP – single nucleotide polymorphism, VFT – Verbal Fluency Test (executive function)
5. Overarching limitations
Before drawing conclusions from the above-reviewed studies, it is important to consider the various limitations associated with these studies. There are of course numerous study-specific limitations that are beyond the scope of this review, though we will highlight some of the more pervasive here.
5.1. Power and the multiple testing problem
Two of the most obvious major limitations are: (i) power and (ii) the multiple testing problem. The only study reviewed here to provide information from a power calculation was Marioni et al. (2010), who utilised a sample of n=2091 Caucasians. In their discussion section the authors reported that, assuming a MAF of 0.23, which was the lowest they observed of the SNPs they genotyped, their study had just over 60% power for a two-tailed significance test (α=0.05) to detect 0.5% of the cognitive variance. Considering the maximum variance explained by any one of their SNPs was 0.3%, the study by Marioni et al. therefore had <60% power. Importantly, not counting the GWAS and the LOAD candidates reviewed here, the sample size in this study was only exceeded by the Benke et al. (2011) study (n=3575) and the three studies utilising the PROSPER cohort (n>5000) (Trompet et al., 2008, Trompet et al., 2012, Mooijaart et al., 2011). Thus, the majority of the candidate gene association studies reviewed here were clearly vastly underpowered, which indicates some of them may have yielded false-positive results. Meta- or mega-analyses combining data from existing studies represents the most obvious means of increasing power in future studies. The lack of cohesion across cognitive ageing studies with regards to the different cognitive testsadministered may prove problematic, though there are tests such as the Stroop testthat featured across multiple studies. Moving forward, the field in general would benefit from the formation of consortia to guide the continued collection of participants and standardise cognitive testing with the express intention of combining samples down the track.
Concerning the multiple testing problem, many of the studies reviewed here applied Bonferroni correction. In the context of genetic association studies, the Bonferroni method is widely regarded as being overly conservative (Hendricks et al., 2014), and as yet (to our knowledge) there is no well-validated method that is able to account for linkage disequilibrium between SNPs or intercorrelations between phenotypes (i.e., different cognitive tests). Thus, until a method such as this begins to gain traction, the use of current correction techniques to reduce the likelihood of false positive associations remains heuristic at best. This of course renders genetic association studies particularly prone to false negative findings, thereby making it difficult to determine with any real confidence whether resultant findings are indeed valid or spurious. This problem is further exacerbated by the obvious lack of studies within the cognition field aiming to perform like-for-like replication of previous genetic association studies using independent samples. There is therefore a vital need for future studies to specifically address this lack of like-for-like replication in order to assess the reliability of findings.
5.2. Distinguishing pathological from non-pathological cognitive decline
Finally, a potential limitation specifically impacting upon studies within this field is the possibility that elderly samples might be “contaminated” with individuals in the early stages of a dementing disease. This issue raises a conceptual question: are dementing diseases qualitatively or quantitatively different to normal ageing? Whilst it is true that LOAD pathology – such as amyloid-β plaques – is produced as a consequence of normal ageing (Fukumoto and Asami-Odaka, 1996), this is by no means evidence against a qualitative difference. Accordingly, the studies reviewed here have made efforts to control for this issue by applying an MMSE scorethreshold <24 to remove possible dementia cases before analysis (Sasayama et al., 2011, Tsai et al., 2010). However, cross-sectional studies are still prone to the inclusion of participants who may go on to develop dementia after the study has been completed, and it is unclear how this might impact the validity of findings. Thus, the longitudinal studies reviewed here (Benke et al., 2011, Trompet et al., 2008) most likely represent the “cleanest” cohorts as a result of multiple opportunities to identify and remove dementia cases as the studies progress.
6. Conclusions and discussion
In conclusion, numerous candidate gene association studies focusing on the major cytokine-encoding genes have indicated that these cytokines may play a causative role in the regulation of cognitive decline in the elderly; i.e., IL1β (rs16944) and TNFα (rs1800629). However, given the above-mentioned limitations concerning power and the multiple-testing problem, it is not possible to determine what proportion of these significant associations might constitute false positive results. In our opinion, the strongest evidence for a causal role comes from the longitudinal study by Trompet et al. (2008), which revealed strikingly consistent and significant associations between the inflammatory regulator IL1βCE (Figure 1A) and both global and domain-specific cognitive decline in a very large sample of elderly individuals (n=5804). Interestingly, given the well-established role for IL1βCE in the cleavage of pro-IL1β and consequent production of mature IL1β (Figure 1B), this arguably lends weight to the IL1β genetic association studies reviewed here (i.e., Baune et al., 2008a, Baune et al., 2008b; Tsai et al., 2010) assuming IL1βCE exerts its apparent effect on cognitive decline via this mechanism (Figure 1A). In partial support of this assumption, Trompet et al. demonstrated that the same IL1βCEpolymorphisms associated with global cognition were also significantly associated with mature IL1β production in the periphery of a subsample of their elderly non-demented sample. This lending of weight may also be extendable to the studies investigating IL1α and IL1RN (Benke et al., 2011, Marioni et al., 2010) as known regulators of IL1β, particularly IL1RN (Figure 1A).
Figure 1. From immune gene variation to cognitive decline in the non-demented elderly. (A) A large genetic association highlights the interleukin 1β-converting enzyme (ICE) in the regulation of cognitive decline in non-demented elderly (Trompet et al., 2008), thereby also implicating other members of the IL1 family. Further, a large GWAS highlights two putative immune regulators, HS3ST4 and SPOCK3, in the regulation of non-demented elderly at genome-wide significance. (B) Non-coding SNPs in and around immune genes associated with cognitive decline in non-demented elderly most likely exert cis-acting effects on immune gene transcription, potentially by altering relevant transcription factor binding potential. The resultant change in mRNA levels is then assumed to translate to altered protein levels. (C) Altered immune protein levels in the periphery could potentially influence central mechanisms by crossing the blood–brain barrier (BBB). (D) Model outlining various central mechanisms by which immune gene variation in combination with elevated inflammatory markers as a consequence of normal ageing might promote cognitive decline in the non-demented elderly.It is also worth mentioning the very recent GWAS of cognitive decline in non-demented elderly by Debette et al. (2015), which highlighted two genes involved in the immune response (HS3ST4, SPOCK3), importantly, at genome-wide significance (p<5*10−08) (Figure 1A). These two genes therefore represent striking candidates worthy of follow-up functional work to determine how they might influence (i) cytokine activity and (ii) biological substrates of cognition, which, as was the case with IL1βCE, again highlights the importance of future research in this field focusing on inflammatory regulators as opposed to the major inflammatory cytokines themselves. Further, findings from GWAS investigating LOAD (i.e., Harold et al., 2009; Lambert et al., 2009) and other dementing disease should not be overlooked as possible sources of additional candidates, though care needs to be taken when extrapolating findings to cognitive decline in non-demented elderly individuals.
Many of the genetic association studies reviewed here included a functional component to determine whether the SNPs they studied were also associated with peripheral (typically serum or plasma) cytokine levels (i.e., Baune et al., 2008a, Baune et al., 2008b; Lau et al., 2012). In accordance with complex disease/phenotype genetics in general, the assumption here is that polymorphic loci associated with cognitive decline, particularly given their largely non-coding placement in the genome (i.e., intronic or intergenic), most likely exert a cis-acting effect on local transcription, which is then assumed to translate to altered proteinlevels (Figure 1C). Thus, coming back to IL1βCE, the effect of IL1βCE SNPs on peripheral IL1β production, as shown by Trompet et al. (2008), is assumed to be mediated by a cis-acting effect on IL1βCE transcription that translates to altered protein levels and therefore altered pro-IL1β-processing activity. In general, findings from the functional arms of these studies were mixed, with some showing evidence of significant associations (i.e., Lau et al., 2012) with altered cytokine levels whereas others did not (i.e., Baune et al., 2008a, Baune et al., 2008b). Of course, given the phenotype of interest is cognitive decline, from an intuitive standpoint, future studies should be looking for central as opposed to peripheral cis-acting effects where possible.
However, at the same time peripheral effects should not be ignored due to the potential for cross talk between central and peripheral inflammatory mechanisms, whereby inflammatory cytokines are widely known to cross the blood–brain barrier (BBB) by either passive or active transport (Banks et al., 1995). Thus, in this way it is conceivable that cis-acting effects on inflammatory genes in the periphery may impact indirectly on cognition-related mechanisms in the brain (Figure 1D) such as those proposed in the cytokine model of cognitive function (McAfoose and Baune, 2009) (Figure 1E). This raises some interesting clinical issues indicating it may in fact be possible to prevent or minimise cognitive decline in the non-demented elderly simply by intervening at the peripheral level using anti-inflammatory agents. Based on the findings from Trompet et al. (2008), although this study requires replication in independents samples, we propose that intervention specifically at the level of IL1β production in the periphery warrants investigation. Clearly from a clinical standpoint this would be preferable to intervention of the IL1β pathway further downstream, say by antagonism of IL1β receptors, as this would involve the use of agents capable of crossing the BBB.
Finally, there is little evidence from the literature reviewed here that convincingly indicates a causal effect of any specific inflammatory gene on domain-specific cognition, though this might be due to the serious lack of like-for-like replication in the field in general as discussed above. However, it is worth highlighting the SNP rs16944, located directly upstream of the IL1β gene, which was found to be associated with memory decline (albeit using different memory tests) in two independent non-demented elderly samples, one Caucasian (Baune et al., 2008a, Baune et al., 2008b) and the other Asian (Tsai et al., 2010). Although both samples were relatively small, both studies were concordant with regards to the direction of the observed effect whereby the T allele appeared to be associated with greater memory decline. Interestingly, a recent structural imaging genetics study investigating the relationship between candidate inflammatory SNPs and regional brain shrinkage over a two year period revealed a significant association between this same allele of rs16944 and increased parahippocampal shrinkage in n=167 participants (age range: 19–79 years) (Persson et al., 2014). Given that the parahippocampal region is known to be heavily involved in memory processes (Van Strien et al., 2009), future research is required to determine whether rs16944 exerts an effect on IL1β transcription in this region, and if so, whether this effect is region-specific or more global. Findings from this research may help to determine whether rs16944 does indeed exert memory-specific effects in the non-demented elderly.
To summarise, moving forward in this field we urge replication of the IL1βCEfindings from Trompet et al. (2008) in independent samples. We also propose there is a need for a general shift in focus away from the major cytokine-encoding genes more towards cytokine-regulating genes. The GWAS approach clearly represents an efficient means of highlighting potentially relevant cytokine- or inflammatory-regulating genes for further study, and there may be scope for GWAS meta- or mega-analyses combining samples from the various studies reviewed here in order to increase power. Thus, future studies focusing on a more comprehensive set of inflammatory-regulator genes, utilising large sample sizes comparable to Trompet et al., will enable better elucidation of the cause–effect relationship between inflammatory mechanisms and cognitive decline in non-demented individuals. Ultimately, given that causal mechanisms likely represent the best candidates for therapeutic intervention, this research will help to determine whether inflammatory mechanisms represent viable targets.
Role of funding source
This study is supported by the National Health and Medical Research Council Australia (APP1060524 to BTB). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Contributors
DS wrote a first draught, LC created the tables, BTB provided the study design, assisted with literature search and edited the draught multiple times. All authors contributed to the final version, edited and approved the final version.
Conflict of interest
All authors declare no conflict of interest.
Acknowledgements
This study is supported by the National Health and Medical Research CouncilAustralia (APP1060524 to BTB). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
References
- D. Alcolea, M. Carmona-Iragui, M. Suárez-Calvet, M.B. Sánchez-Saudinós, I.Sala, S. Antón-Aguirre, R. Blesa, J. Clarimón, J. Fortea, A. Lleó
Relationship between β-Secretase, inflammation and core cerebrospinal fluid biomarkers for Alzheimer׳s disease
J. Alzheimers Dis., 42 (2014), pp. 157-167
- L. Aloe, F. Properzi, L. Probert, K. Akassoglou, G. Kassiotis, A. Micera, M. Fiore
Learning abilities, NGF and BDNF brain levels in two lines of TNF-α transgenic mice, one characterized by neurological disorders, the other phenotypically normal
Brain Res., 840 (1999), pp. 125-137
- B. Arosio, L. Bergamaschini, L. Galimberti, C. La Porta, M. Zanetti, C. Calabresi, E. Scarpini, G. Annoni, C. Vergani
+10 T/C polymorphisms in the gene of transforming growth factor-beta1 are associated with neurodegeneration and its clinical evolution
Mech. Ageing Dev., 128 (2007), pp. 553-557
- B. Arosio, L. Mastronardi, C. Vergani, G. Annoni
Intereleukin-10 promoter polymorphism in mild cognitive impairment and in its clinical evolution
Int. J. Alzheimers Dis., 2010 (2010), pp. 1-5
- D. Balschun, W. Wetzel, F. Pitossi, H. Schneider, W. Zuschratter, H.O.Besedovsky, F. Campomar, B. Aires
Interleukin-6: a cytokine to forget
FASEB J., 18 (2004), pp. 1788-1790
- W.A. Banks, A.J. Kastin, R.D. Broadwell
Passage of cytokines across the blood–brain barrier
Neuroimmunomodulation, 2 (1995), pp. 241-248
- B.T. Baune, G. Ponath, M. Rothermundt, O. Riess, H. Funke, K. Berger
Association between genetic variants of IL-1beta, IL-6 and TNF-alpha cytokines and cognitive performance in the elderly general population of the MEMO-study
Psychoneuroendocrinology, 33 (2008), pp. 68-76
- B.T. Baune, F. Wiede, A. Braun, J. Golledge, V. Arolt, H. Koerner
Cognitive dysfunction in mice deficient for TNF- and its receptors
Am. J. Med. Genet. B. Neuropsychiatr. Genet., 147B (2008), pp. 1056-1064
- K.S. Benke, M.C. Carlson, B.Q. Doan, J.D. Walston, Q.L. Xue, A.P. Reiner, L.P.Fried, D.E. Arking, A. Chakravarti, M.D. Fallin
The association of genetic variants in interleukin-1 genes with cognition: findings from the cardiovascular health study
Exp. Gerontol., 46 (2011), pp. 1010-1019
- B.M. Bettcher, J.H. Kramer
Longitudinal inflammation, cognitive decline, and Alzheimer׳s disease: a mini-review
Clin. Pharmacol. Ther., 96 (2014), pp. 464-469
- B.M. Bettcher, R. Wilheim, T. Rigby, R. Green, J.W. Miller, C.A. Racine, K. Yaffe, B.L. Miller, J.H. Kramer
C-reactive protein is related to memory and medial temporal brain volume in older adults
Brain Behav. Immun., 26 (2012), pp. 103-108
- F.X. Brennan, K.D. Beck, R.J. Servatius
Low doses of interleukin-1β improve the leverpress avoidance performance of Sprague–Dawley rats
Neurobiol. Learn. Mem., 80 (2003), pp. 168-171
- M.P. Butler, J.J. O’Connor, P.N. Moynagh
Dissection of tumor-necrosis factor-alpha inhibition of long-term potentiation (LTP) reveals a p38 mitogen-activated protein kinase-dependent mechanism which maps to early-but not late-phase LTP
Neuroscience, 124 (2004), pp. 319-326
- F. Caraci, W. Gulisano, C.A. Guida, A.A.R. Impellizzeri, F. Drago, D. Puzzo, A.Palmeri
A key role for TGF-β1 in hippocampal synaptic plasticity and memory
Sci. Rep., 5 (2015), p. 11252
- M.M. Carrasquillo, J.E. Crook, O. Pedraza, C.S. Thomas, V.S. Pankratz, M. Allen, T. Nguyen, K.G. Malphrus, L. Ma, G.D. Bisceglio, R.O. Roberts, J.A. Lucas, G.E.Smith, R.J. Ivnik, M.M. Machulda, N.R. Graff-Radford, R.C. Petersen, S.G.Younkin, N. Ertekin-Taner
Late-onset Alzheimer׳s risk variants in memory decline, incident mild cognitive impairment, and Alzheimer׳s disease
Neurobiol. Aging, 36 (2015), pp. 60-67
- L.B. Chibnik, J.M. Shulman, S.E. Leurgans, J.A. Schneider, R.S. Wilson, D. Tran, C. Aubin, A.S. Buchman, C.B. Heward, A.J. Myers, J.A. Hardy, M.J. Huentelman, J.J. Corneveaux, E.M. Reiman, D.A. Evans, D.A. Bennett, P.L. De Jager
CR1 is associated with amyloid plaque burden and age-related cognitive decline
Ann. Neurol., 69 (2011), pp. 560-569
- S. Dato, K.S. Krabbe, M. Thinggaard, B.K. Pedersen, K. Christensen, H.Bruunsgaard, L. Christiansen
Commonly studied polymorphisms in inflammatory cytokine genes show only minor effects on mortality and related risk factors in nonagenarians
J. Gerontol. A: Biol. Sci. Med. Sci., 65 (2010), pp. 225-235
- G. Davies, N. Armstrong, J.C. Bis, J. Bressler, V. Chouraki, S. Giddaluru, E.Hofer, C.A. Ibrahim-Verbaas, M. Kirin, J. Lahti, S.J. van der Lee, S. Le Hellard, T.Liu, R.E. Marioni, C. Oldmeadow, I. Postmus, A.V. Smith, J.A. Smith, A.Thalamuthu, R. Thomson, V. Vitart, J. Wang, L. Yu, L. Zgaga, W. Zhao, R. Boxall, S.E. Harris, W.D. Hill, D.C. Liewald, M. Luciano, H. Adams, D. Ames, N. Amin, P.Amouyel, A.A. Assareh, R. Au, J.T. Becker, A. Beiser, C. Berr, L. Bertram, E.Boerwinkle, B.M. Buckley, H. Campbell, J. Corley, P.L. De Jager, C. Dufouil, J.G.Eriksson, T. Espeseth, J.D. Faul, I. Ford, R.F. Gottesman, M.E. Griswold, V.Gudnason, T.B. Harris, G. Heiss, A. Hofman, E.G. Holliday, J. Huffman, S.L.R.Kardia, N. Kochan, D.S. Knopman, J.B. Kwok, J.-C. Lambert, T. Lee, G. Li, S.-C.Li, M. Loitfelder, O.L. Lopez, A.J. Lundervold, A. Lundqvist, K.A. Mather, S.S.Mirza, L. Nyberg, B.A. Oostra, A. Palotie, G. Papenberg, A. Pattie, K. Petrovic, O.Polasek, B.M. Psaty, P. Redmond, S. Reppermund, J.I. Rotter, H. Schmidt, M.Schuur, P.W. Schofield, R.J. Scott, V.M. Steen, D.J. Stott, J.C. van Swieten, K.D.Taylor, J. Trollor, S. Trompet, A.G. Uitterlinden, G. Weinstein, E. Widen, B.G.Windham, J.W. Jukema, A.F. Wright, M.J. Wright, Q. Yang, H. Amieva, J.R. Attia, D.A. Bennett, H. Brodaty, A.J.M. de Craen, C. Hayward, M.A. Ikram, U.Lindenberger, L.-G. Nilsson, D.J. Porteous, K. Räikkönen, I. Reinvang, I. Rudan, P.S. Sachdev, R. Schmidt, P.R. Schofield, V. Srikanth, J.M. Starr, S.T. Turner, D.R. Weir, J.F. Wilson, C. van Duijn, L. Launer, A.L. Fitzpatrick, S. Seshadri, T.H.Mosley, I.J. Deary
Genetic contributions to variation in general cognitive function: a meta-analysis of genome-wide association studies in the CHARGE consortium (N=53949)
Mol. Psychiatry, 20 (2015), pp. 183-192
- P.L. De Jager, J.M. Shulman, L.B. Chibnik, B.T. Keenan, T. Raj, R.S. Wilson, L.Yu, S.E. Leurgans, D. Tran, C. Aubin, C.D. Anderson, A. Biffi, J.J. Corneveaux, M.J. Huentelman, J. Rosand, M.J. Daly, A.J. Myers, E.M. Reiman, D.A. Bennett, D.A. Evans
A genome-wide scan for common variants affecting the rate of age-related cognitive decline
Neurobiol. Aging, 33 (1017) (2012), pp. e1-15
- S. Debette, C.A. Ibrahim Verbaas, J. Bressler, M. Schuur, A. Smith, J.C. Bis, G.Davies, C. Wolf, V. Gudnason, L.B. Chibnik, Q. Yang, A.L. DeStefano, D.J.F. de Quervain, V. Srikanth, J. Lahti, H.J. Grabe, J.A. Smith, L. Priebe, L. Yu, N.Karbalai, C. Hayward, J.F. Wilson, H. Campbell, K. Petrovic, M. Fornage, G.Chauhan, R. Yeo, R. Boxall, J. Becker, O. Stegle, K.A. Mather, V. Chouraki, Q.Sun, L.M. Rose, S. Resnick, C. Oldmeadow, M. Kirin, A.F. Wright, M.K.Jonsdottir, R. Au, A. Becker, N. Amin, M.A. Nalls, S.T. Turner, S.L.R. Kardia, B.Oostra, G. Windham, L.H. Coker, W. Zhao, D.S. Knopman, G. Heiss, M.E.Griswold, R.F. Gottesman, V. Vitart, N.D. Hastie, L. Zgaga, I. Rudan, O. Polasek, E.G. Holliday, P. Schofield, S.H. Choi, T. Tanaka, Y. An, R.T. Perry, R.E. Kennedy, M.M. Sale, J. Wang, V.G. Wadley, D.C. Liewald, P.M. Ridker, A.J. Gow, A. Pattie, J.M. Starr, D. Porteous, X. Liu, R. Thomson, N.J. Armstrong, G. Eiriksdottir, A.A.Assareh, N.A. Kochan, E. Widen, A. Palotie, Y.-C. Hsieh, J.G. Eriksson, C.Vogler, J.C. van Swieten, J.M. Shulman, A. Beiser, J. Rotter, C.O. Schmidt, W.Hoffmann, M.M. Nöthen, L. Ferrucci, J. Attia, A.G. Uitterlinden, P. Amouyel, J.-F.Dartigues, H. Amieva, K. Räikkönen, M. Garcia, P.A. Wolf, A. Hofman, W.T.Longstreth, B.M. Psaty, E. Boerwinkle, P.L. DeJager, P.S. Sachdev, R. Schmidt, M.M.B. Breteler, A. Teumer, O.L. Lopez, S. Cichon, D.I. Chasman, F. Grodstein, B. Müller-Myhsok, C. Tzourio, A. Papassotiropoulos, D.A. Bennett, M.A. Ikram, I.J. Deary, C.M. van Duijn, L. Launer, A.L. Fitzpatrick, S. Seshadri, T.H. Mosley
Genome-wide studies of verbal declarative memory in nondemented older people: the cohorts for heart and aging research in genomic epidemiology consortium
Biol. Psychiatry, 77 (2015), pp. 749-763
- del Rey, A., Balschun, D., Wetzel, W., Randolf, A., Besedovsky, H. O., A cytokine network involving brain-borne IL-1beta, IL-1ra, IL-18, IL-6, and TNFalpha operates during long-term potentiation and learning, Brain Behav. Immun. 33, 2013, 15-23.
- A.M. .Depino, M. Alonso, C. Ferrari, A. del Rey, D. Anthony, H. Besedovsky, J.H.Medina, F. Pitossi
Learning modulation by endogenous hippocampal IL-1: blockade of endogenous IL-1 facilitates memory formation
Hippocampus, 14 (2004), pp. 526-535
- S.A. Eshkoor, T.A. Hamid, C.Y. Mun, C.K. Ng
Mild cognitive impairment and its management in older people
Clin. Interv. Aging, 10 (2015), pp. 687-693
- M.C. Faria, G.S. Gonçalves, N.P. Rocha, E.N. Moraes, M.A. Bicalho, M.T.Gualberto Cintra, J. Jardim de Paula, L.F. José Ravic de Miranda, A. Clayton de Souza Ferreira, A.L. Teixeira, K.B. Gomes, M.D.G. Carvalho, L.P. Sousa
Increased plasma levels of BDNF and inflammatory markers in Alzheimer׳s disease
J. Psychiatr. Res., 53 (2014), pp. 166-172
- M. Fiore, F. Angelucci, E. Alleva, I. Branchi, L. Probert, L. Aloe
Learning performances, brain NGF distribution and NPY levels in transgenic mice expressing TNF-alpha
Behav. Brain Res., 112 (2000), pp. 165-175
- T. Fuchs, J.N. Trollor, J. Crawford, D.A. Brown, B.T. Baune, K. Samaras, L.Campbell, S.N. Breit, H. Brodaty, P. Sachdev, E. Smith
Macrophage inhibitory cytokine-1 is associated with cognitive impairment and predicts cognitive decline – the Sydney memory and aging study
Aging Cell, 12 (2013), pp. 882-889
- H. Fukumoto, A. Asami-Odaka
Amyloid beta protein deposition in normal aging has the same characteristics as that in Alzheimer’s disease. predominance of A beta 42 (43) and association of A beta
Am. J. Pathol., 148 (1996), pp. 259-265
- P.D. Gajewski, J.G. Hengstler, K. Golka, M. Falkenstein, C. Beste
The functional tumor necrosis factor-α (308A/G) polymorphism modulates attentional selection in elderly individuals
Neurobiol. Aging, 34 (2013), pp. 2694.e1-2694.e12
- S.J. Greco, P. Rameshwar
Enhancing effect of IL-1alpha on neurogenesis from adult human mesenchymal stem cells: implication for inflammatory mediators in regenerative medicine
J. Immunol., 179 (2007), pp. 3342-3350
- D. Harold, R. Abraham, P. Hollingworth, R. Sims, A. Gerrish, M.L. Hamshere, J.S.Pahwa, V. Moskvina, K. Dowzell, A. Williams, N. Jones, C. Thomas, A. Stretton, A.R. Morgan, S. Lovestone, J. Powell, P. Proitsi, M.K. Lupton, C. Brayne, D.C.Rubinsztein, M. Gill, B. Lawlor, A. Lynch, K. Morgan, K.S. Brown, P.A. Passmore, D. Craig, B. McGuinness, S. Todd, C. Holmes, D. Mann, A.D. Smith, S. Love, P.G.Kehoe, J. Hardy, S. Mead, N. Fox, M. Rossor, J. Collinge, W. Maier, F. Jessen, B.Schürmann, R. Heun, H. van den Bussche, I. Heuser, J. Kornhuber, J. Wiltfang, M. Dichgans, L. Frölich, H. Hampel, M. Hüll, D. Rujescu, A.M. Goate, J.S.K.Kauwe, C. Cruchaga, P. Nowotny, J.C. Morris, K. Mayo, K. Sleegers, K. Bettens, S. Engelborghs, P.P. De Deyn, C. Van Broeckhoven, G. Livingston, N.J. Bass, H.Gurling, A. McQuillin, R. Gwilliam, P. Deloukas, A. Al-Chalabi, C.E. Shaw, M.Tsolaki, A.B. Singleton, R. Guerreiro, T.W. Mühleisen, M.M. Nöthen, S. Moebus, K.-H. Jöckel, N. Klopp, H.-E. Wichmann, M.M. Carrasquillo, V.S. Pankratz, S.G.Younkin, P.A. Holmans, M. O’Donovan, M.J. Owen, J. Williams
Genome-wide association study identifies variants at CLU and PICALM associated with Alzheimer׳s disease
Nat. Genet., 41 (2009), pp. 1088-1093
- Y. He, H. Hsuchou, X. Wu, A.J. Kastin, R.S. Khan, P.J. Pistell, W.H. Wang, J.Feng, Z. Li, X. Guo, W. Pan
Interleukin-15 receptor is essential to facilitate GABA transmission and hippocampal-dependent memory
J. Neurosci., 30 (2010), pp. 4725-4734
- N.C. Henderson, T. Sethi
The regulation of inflammation by galectin-3
Immunol. Rev., 230 (2009), pp. 160-171
- A.E. Hendricks, J. Dupuis, M.W. Logue, R.H. Myers, K.L. Lunetta
Correction for multiple testing in a gene region
Eur. J. Hum. Genet., 22 (2014), pp. 414-418
- D.K. Hsu, R.Y. Yang, Z. Pan, L. Yu, D.R. Salomon, W.P. Fung-Leung, F.T. Liu
Targeted disruption of the galectin-3 gene results in attenuated peritoneal inflammatory responses
Am. J. Pathol., 156 (2000), pp. 1073-1083
- K. Kadoshima-Yamaoka, M. Murakawa, M. Goto, Y. Tanaka, H. Inoue, H.Murafuji, A. Nagahira, Y. Hayashi, K. Nagahira, K. Miura, T. Nakatsuka, K.Chamoto, Y. Fukuda, T. Nishimura
ASB16165, a novel inhibitor for phosphodiesterase 7A (PDE7A), suppresses IL-12-induced IFN-gamma production by mouse activated T lymphocytes
Immunol. Lett., 122 (2009), pp. 193-197
- S. Karim, S. Hopkins, N. Purandare, J. Crowther, J. Morris, P. Tyrrell, A. Burns
Peripheral inflammatory markers in amnestic mild cognitive impairment
Int. J. Geriatr. Psychiatry, 29 (2014), pp. 221-226
- B.T. Keenan, J.M. Shulman, L.B. Chibnik, T. Raj, D. Tran, M.R. Sabuncu, A.N.Allen, J.J. Corneveaux, J.A. Hardy, M.J. Huentelman, C.A. Lemere, A.J. Myers, A. Nicholson-Weller, E.M. Reiman, D.A. Evans, D.A. Bennett, P.L. De Jager
A coding variant in CR1 interacts with APOE-ε4 to influence cognitive decline
Hum. Mol. Genet., 21 (2012), pp. 2377-2388
- A. Kelly, A. Lynch, E. Vereker, Y. Nolan, P. Queenan, E. Whittaker, L.A. O’Neill, M.A. Lynch
The anti-inflammatory cytokine, interleukin (IL)-10, blocks the inhibitory effect of IL-1 beta on long term potentiation. A role for JNK
J. Biol. Chem., 276 (2001), pp. 45564-45572
- R.B. Kennedy, I.G. Ovsyannikova, V.S. Pankratz, H. Iana, R.A. Vierkant, G.A.Poland
Genome-wide analysis of polymorphisms associated with cytokine responses in smallpox vaccine recipients
Hum. Genet., 131 (2014), pp. 1403-1421
- J.J. Kobie, E.T. Akporiaye
Immunosuppressive role of transforming growth factor beta in breast cancer
Clim. Appl. Immunol. Rev., 3 (2003), pp. 277-287
- J.W. Koo, R.S. Duman
IL-1beta is an essential mediator of the antineurogenic and anhedonic effects of stress
Proc. Natl. Acad. Sci. USA, 105 (2008), pp. 751-756
- K.S. Krabbe, E.L. Mortensen, K. Avlund, H. Pilegaard, L. Christiansen, A.N.Pedersen, M. Schroll, T. Jørgensen, B.K. Pedersen, H. Bruunsgaard
Genetic priming of a proinflammatory profile predicts low IQ in octogenarians
Neurobiol. Aging, 30 (2009), pp. 769-781
- J.-C. Lambert, S. Heath, G. Even, D. Campion, K. Sleegers, M. Hiltunen, O.Combarros, D. Zelenika, M.J. Bullido, B. Tavernier, L. Letenneur, K. Bettens, C.Berr, F. Pasquier, N. Fiévet, P. Barberger-Gateau, S. Engelborghs, P. De Deyn, I.Mateo, A. Franck, S. Helisalmi, E. Porcellini, O. Hanon, M.M. de Pancorbo, C.Lendon, C. Dufouil, C. Jaillard, T. Leveillard, V. Alvarez, P. Bosco, M. Mancuso, F.Panza, B. Nacmias, P. Bossù, P. Piccardi, G. Annoni, D. Seripa, D. Galimberti, D.Hannequin, F. Licastro, H. Soininen, K. Ritchie, H. Blanché, J.-F. Dartigues, C.Tzourio, I. Gut, C. Van Broeckhoven, A. Alpérovitch, M. Lathrop, P. Amouyel
Genome-wide association study identifies variants at CLU and CR1 associated with Alzheimer’s disease
Nat. Genet., 41 (2009), pp. 1094-1099
- S. Lau, K.A. Bates, H.R. Sohrabi, M. Rodrigues, G. Martins, S.S. Dhaliwal, K.Taddei, S.M. Laws, I.J. Martins, F.L. Mastaglia, J.K. Foster, J.K. Phillips, R.N.Martins
Functional effects of genetic polymorphism in inflammatory genes in subjective memory complainers
Neurobiol. Aging, 33 (2012), pp. 1054-1056
- X.-M. Li, M.-T. Zhou, X.-M. Wang, M.-H. Ji, Z.-Q. Zhou, J.-J. Yang
Resveratrol pretreatment attenuates the isoflurane-induced cognitive impairment through its anti-inflammation and -apoptosis actions in aged mice
J. Mol. Neurosci., 52 (2014), pp. 286-293
- S.L. Ma, W. Huang, N.L.S. Tang, L.C.W. Lam
MxA polymorphisms are associated with risk and age-at-onset in Alzheimer disease and accelerated cognitive decline in Chinese elders
Rejuvenation Res., 15 (2012), pp. 516-522
- R.E. Marioni, I.J. Deary, G.D. Murray, F.G.R. Fowkes, J.F. Price
Associations between polymorphisms in five inflammation-related genes and cognitive ability in older persons
Genes Brain Behav., 9 (2010), pp. 348-352
- J. McAfoose, B.T. Baune
Evidence for a cytokine model of cognitive function
Neurosci. Biobehav. Rev., 33 (2009), pp. 355-366
- J. McAfoose, H. Koerner, B.T. Baune
The effects of TNF deficiency on age-related cognitive performance
Psychoneuroendocrinology, 34 (2009), pp. 615-619
- J. Mengel-From, K. Christensen, M. McGue, L. Christiansen
Genetic variations in the CLU and PICALM genes are associated with cognitive function in the oldest old
Neurobiol. Aging, 32 (2011), pp. 554.e7-554.e11
- J. Mengel-From, M. Thinggaard, R. Lindahl-Jacobsen, M. McGue, K.Christensen, L. Christiansen
CLU genetic variants and cognitive decline among elderly and oldest old
PloS One, 8 (2013), p. e79105
- N.L. Monson, S.J. Ireland, A.J. Ligocki, D. Chen, W.H. Rounds, M. Li, R.M.Huebinger, C. Munro Cullum, B.M. Greenberg, A.M. Stowe, R. Zhang
Elevated CNS inflammation in patients with preclinical Alzheimer׳s disease
J. Cereb. Blood Flow. Metab., 34 (2014), pp. 30-33
- S.P. Mooijaart, N. Sattar, S. Trompet, E. Polisecki, A.J.M. de Craen, E.J.Schaefer, S.E. Jahn, T. van Himbergen, P. Welsh, I. Ford, D.J. Stott, R.G.J.Westendorp
C-reactive protein and genetic variants and cognitive decline in old age: the PROSPER study
PloS One, 6 (2011), p. e23890
- K. Moreau, A. Fleming, S. Imarisio, A. Lopez Ramirez, J.L. Mercer, M. Jimenez-Sanchez, C.F. Bento, C. Puri, E. Zavodszky, F. Siddiqi, C.P. Lavau, M. Betton, C.J. O’Kane, D.S. Wechsler, D.C. Rubinsztein
PICALM modulates autophagy activity and tau accumulation
Nat. Commun., 5 (2014), p. 4998
- M. Moriyama, T. Fukuhara, M. Britschgi, Y. He, S. Villeda, H. Molina, B.T. Huber, M. Holers, T. Wyss-Coray
Complement receptor 2 is expressed in neural progenitor cells and regulates adult hippocampal neurogenesis
J. Neurosci., 31 (2011), pp. 3981-3989
- N. Persson, P. Ghisletta, C.L. Dahle, A.R. Bender, Y. Yang, P. Yuan, A.M.Daugherty, N. Raz
Regional brain shrinkage over two years: individual differences and effects of pro-inflammatory genetic polymorphisms
Neuroimage, 103 (2014), pp. 334-348
- R.C. Pugh, M. Fleshner, L.R. Watkins, S.F. Maier, J.W. Rudy
The immune system and memory consolidation: a role for the cytokine IL-1β
Neurosci. Biobehav. Rev., 25 (2001), pp. 29-41
- F.M. Ross, S.M. Allan, N.J. Rothwell, A. Verkhratsky
A dual role for interleukin-1 in LTP in mouse hippocampal slices
J. Neuroimmunol., 144 (2003), pp. 61-67
- D. Sasayama, H. Hori, T. Teraishi, K. Hattori, M. Ota, J. Matsuo, Y. Kawamoto, Y.Kinoshita, T. Higuchi, N. Amano, H. Kunugi
Association of interleukin-1β genetic polymorphisms with cognitive performance in elderly females without dementia
J. Hum. Genet., 56 (2011), pp. 613-616
- H. Schneider, F. Pitossi, D. Balschun, A. Wagner, A. del Ray, H.O. Besedovsky
A neuromodulatory role of interleukin-1β in the hippocampus
Proc. Natl. Acad. Sci. USA, 95 (1998), pp. 7778-7783
- E.M. Schrijvers, P.J. Koudstaal, S. Hofman, M.M. Breteler
Plasma clusterin and the risk of Alzheimer disease
JAMA, 305 (2011), pp. 1322-1326
- G.D. Smith, S. Ebrahim, S. Lewis, A.L. Hansell, L.J. Palmer, P.R. Burton
Genetic epidemiology and public health: hope, hype, and future prospects
Lancet, 366 (2005), pp. 1484-1498
- D. Stellwagen, R.C. Malenka
Synaptic scaling mediated by glial TNF-alpha
Nature, 440 (2006), pp. 1054-1059
- T.L. Sudduth, F.A. Schmitt, P.T. Nelson, D.M. Wilcock
Neuroinflammatory phenotype in early Alzheimer׳s disease
Neurobiol. Aging, 34 (2013), pp. 1051-1059
- E.M. Sutinen, T. Pirttilä, G. Anderson, A. Salminen, J.O. Ojala
Pro-inflammatory interleukin-18 increases Alzheimer׳s disease-associated amyloid-β production in human neuron-like cells
J. Neuroinflamm., 9 (2012), p. 199
- M. Thambisetty, L.L. Beason-Held, Y. An, M. Kraut, M. Nalls, D.G. Hernandez, A.B. Singleton, A.B. Zonderman, L. Ferrucci, S. Lovestone, S.M. Resnick
Alzheimer risk variant CLU and brain function during aging
Biol. Psychiatry, 73 (2013), pp. 399-405
- V. Tiwari, C.D. O’Donnell, M.-J. Oh, T. Valyi-Nagy, D. Shukla
A role for 3-O-sulfotransferase isoform-4 in assisting HSV-1 entry and spread
Biochem. Biophys. Res. Commun., 338 (2005), pp. 930-937
- L. Tong, G.A. Prieto, E.A. Kramar, E.D. Smith, D.H. Cribbs, G. Lynch, C.W.Cotman
Brain-derived neurotrophic factor-dependent synaptic plasticity is suppressed by interleukin-1β via p38 mitogen-activated protein kinase
J. Neurosci., 32 (2013), pp. 17714-17724
- S. Trompet, A.J.M. de Craen, P. Slagboom, J. Shepherd, G.J. Blauw, M.B.Murphy, E.L.E.M. Bollen, B.M. Buckley, I. Ford, A. Gaw, P.W. Macfarlane, C.J.Packard, D.J. Stott, J.W. Jukema, R.G.J. Westendorp
Genetic variation in the interleukin-1 beta-converting enzyme associates with cognitive function. The PROSPER study
Brain, 131 (2008), pp. 1069-1077
- S. Trompet, W. Jukema, S.P. Mooijaart, I. Ford, D.J. Stott, R.G.J. Westendorp, A.J.M. de Craen
Genetic variation in galectin-3 gene associates with cognitive function at old age
Neurobiol. Aging, 33 (2012), pp. 2232.e1-2232.e9
- S.-J. Tsai, C.-J. Hong, M.-E. Liu, S.-J. Hou, F.-C. Yen, C.-H. Hsieh, Y.-J. Liou
Interleukin-1 beta (C-511T) genetic polymorphism is associated with cognitive performance in elderly males without dementia
Neurobiol. Aging, 31 (2010), pp. 1950-1955
- L. Vallières, I.L. Campbell, F.H. Gage, P.E. Sawchenki
Reduced hippocampal neurogenesis in adult transgenic mice with chronic astrocytic production of interleukin-6
J. Neurosci., 22 (2002), pp. 486-492
- N.M. Van Strien, N.L.M. Cappaert, M.P. Witter
The anatomy of memory: an interactive overview of the parahippocampal-hippocampal network
Nat. Rev. Neurosci., 10 (2009), pp. 272-282
- B.F.J. Verhaaren, M.W. Vernooij, P.J. Koudstaal, A.G. Uitterlinden, C.M. van Duijn, A. Hofman, M.M.B. Breteler, M.A. Ikram
Alzheimer׳s disease genes and cognition in the nondemented general population
Biol. Psychiatry, 73 (2013), pp. 429-434
- X. Wang, S. Zhang, F. Lin, W. Chu, S. Yue
Elevated Galectin-3 Levels in the Serum of Patients With Alzheimer׳s Disease
Am. J. Alzheimers. Dis. Other Dement. (2013), pp. 1-4
- K. Yaffe, K. Lindquist, B. Penninx, E. Simonsick, M. Pahor, S. Kritchevsky, L.Launer, L. Kuller, S. Rubin, T. Harris
Inflammatory markers and cognition in well-functioning African–American and white elders
Neurology, 61 (2003), pp. 76-80
- G. Yang, K.W. McIntyre, R.M. Townsend, H.H. Shen, W.J. Pitts, J.H. Dodd, S.G.Nadler, M. McKinnon, A.J. Watson
Phosphodiesterase 7A-deficient mice have functional T cells
J. Immunol., 171 (2003), pp. 6414-6420
- M. Yarchoan, N. Louneva, S.X. Xie, F.J. Swenson, W. Hu, H. Soares, J.Q.Trojanowski, V.M.-Y. Lee, M.A. Kling, L.M. Shaw, A. Chen-Plotkin, D.A. Wolk, S.E. Arnold
Association of plasma C-reactive protein levels with the diagnosis of Alzheimer׳s disease
J. Neurol. Sci., 333 (2013), pp. 9-12
- R. Yirmiya, I. Goshen
Immune modulation of learning, memory, neural plasticity and neurogenesis
Brain Behav. Immun., 25 (2011), pp. 181-213
- R. Yirmiya, G. Winocur, I. Goshen
Brain Interleukin-1 is involved in spatial memory and passive avoidance conditioning
Neurobiol. Learn. Mem., 78 (2002), pp. 379-389
- L. Zhuang, X. Liu, X. Xu, C. Yue, H. Shu, F. Bai, H. Yu, Y. Shi, Z. Zhang
Association of the interleukin 1 beta gene and brain spontaneous activity in amnestic mild cognitive impairment
J. Neuroinflamm., 9 (2012), p. 263