Impaired Innate Immunity Mechanisms in the Brain of Alzheimer’s Disease

Among environmental factors likely associated with Alzheimer’s disease (AD), persistent virus infections, and age-related progressive decline of immune competence might play a pivotal role. However, AD antimicrobial brain immune responses are poorly investigated. The present study focused on genes involved in antimicrobial defenses, especially against virus infections, in the AD brain. In particular, mRNA levels of IRF7, MED23, IL28B, and IFN-α genes were analyzed in hippocampus and temporal cortex brain samples from AD and non-demented controls. All subjects were also genotyped for APOE ε, IRF7, MED23, and IL28B gene polymorphisms. Most AD patients showed decreased mRNA levels of all investigated genes in the hippocampus and temporal cortex. However, a small group of AD patients showed increased hippocampal mRNA expression of MED23, IL28B, and IFN-α. mRNA levels of MED23, IL28B, IFN-α from the hippocampus and those of MED23 from the temporal cortex were further decreased in APOE ε4 allele AD carriers. Moreover, rs6598008 polymorphism of IRF7 was significantly associated with decreased hippocampal expression of IRF7, MED23, IL28B, and IFN-α. These findings suggest that AD brains show impaired innate antimicrobial gene expression profiles, and individual genetic makeup, such as positivity for the APOE ε4 and IRF7 A alleles, might affect brain immune efficiency.


Introduction
Alzheimer's disease (AD) is a chronic neurodegenerative disease and the most frequent form of dementia in the elderly [1]. It is of interest that declining immunity during aging is often associated with persistent antigen stimulation and peripheral chronic inflammation [2]. Moreover, neuroinflammation has emerged as a relevant component of AD brain pathology [3].
Pathogens, such as viruses of the herpes family, through frequent cycles of reactivation and latency, constantly trigger the immune system, which is not able to completely eradicate these microbes. Therefore, persistent neurotropic pathogens might play a role in microglia activation in the brain of genetically susceptible elderly and contribute to neurodegenerative processes [4,5].
The amyloid-β (Aβ) peptide, which is associated with neurodegenerative processes of AD, showed antimicrobial activity against eight common and clinically relevant microorganisms [6]. Aβ peptide In the present study, the mRNA expression of genes involved in antimicrobial responses, such as IRF7, MED23, IL28B or IFN-λ3, and IFN-α, in hippocampus and temporal cortex specimens from human controls and AD patients was studied. Moreover, the influence of apolipoprotein (APOE) ε 4, IRF7, MED23, and IL28B gene polymorphisms upon mRNA levels of IRF7, MED23, IL28B, and IFN-α genes was also analyzed.
Our findings showed that impaired mRNA levels of IRF7, MED23, IL28B, and IFN-α were present in AD hippocampus and temporal cortex samples. APOE ε4 and IRF7 A alleles negatively affected mRNA levels in AD hippocampus.
Antimicrobial defense mechanisms of innate immunity appear to be impaired in the AD brain, and such alterations might contribute to neurodegeneration.

Results
Demographic and clinical features, such as age, gender, clinical diagnosis, disease duration (DD), post-mortem interval (PMI), brain weight, cause of death, and brain area, of patients with AD have been reported in Table 1. AD neuropathological diagnosis, Braak and Thal amyloid deposition scores have been summarized in Table 2.  Levels of IRF7, MED23, IL28B, and IFN-α mRNA in AD brain hippocampus and temporal samples are reported in Figure 1. The majority of AD brains showed decreased hippocampus mRNA levels of IRF7 (n = 28), MED23 (n = 20), IL28B (n = 21), and IFN-α (n = 19), whilst a small AD group had increased mRNA levels of MED23 (n = 11), IL28B (n = 12), and IFN-α (n = 9).
A similar mRNA expression pattern was detected in AD temporal cortex samples, since a larger AD group showed downregulation of IRF7, MED23, IL28B, and IFN-α genes ( Figure 1). A minority of AD showed normal or slightly increased mRNA levels of the above immune factors ( Figure 1).   Brain samples were also genotyped for APOE ε polymorphism and gene variations of IRF7, MED23, and IL28B. The presence of APOE ε4 allele was associated with decreased mRNA levels of MED23, IL28B, and IFN-α in AD hippocampus samples and of MED23 in AD temporal cortex samples, as shown in Figure 2. Brain samples were also genotyped for APOE ε polymorphism and gene variations of IRF7, MED23, and IL28B. The presence of APOE ε4 allele was associated with decreased mRNA levels of MED23, IL28B, and IFN-α in AD hippocampus samples and of MED23 in AD temporal cortex samples, as shown in Figure 2. IRF7 gene polymorphism affected IRF7, MED23, IL28B, and IFN-α mRNA levels, and A allele carriers showed significantly decreased levels of the four immune factors in AD hippocampus samples ( Figure 3).

Figure 2.
Effect of APOE ε4 allele on antiviral immune gene expression of AD patients. q-PCR data showing relative expression (2 -∆Ct values using CYC1 and EIF4A2 as reference genes) of IRF7 (A,E), MED23 (B,F), IL28B (C,G), IFN-α (D,H) in hippocampus (A,B,C,D) and temporal cortex (E,F,G,H) of AD patients grouped in APOE ε4 noncarrier/or APOE ε4 carrier. Data from each group are shown as a box and whiskers plot with the ends of the whiskers represent the minimum and maximum data values, the horizontal line represents the median, and the "+" represents the mean relative expression values. *p < 0.05 **p < 0.01 (unpaired t-test).
IRF7 gene polymorphism affected IRF7, MED23, IL28B, and IFN-α mRNA levels, and A allele carriers showed significantly decreased levels of the four immune factors in AD hippocampus samples ( Figure 3). No statistically significant difference in mRNA levels of IRF7, MED23, IL28B, and IFN-α from temporal cortex samples between IRF7 A carriers and A noncarriers was observed (data not shown).
The presence of MED23, IL28B, IFN-α gene polymorphisms was not associated with mRNA levels of the four immune factors in both hippocampus and temporal cortex specimens (data not shown).
No relationship between IRF7, MED23, IL28B, and IFN-α mRNA levels and Braak and Braak or Thal scores, duration of the disease, and brain weight was found (data not shown). No statistically significant difference in mRNA levels of IRF7, MED23, IL28B, and IFN-α from temporal cortex samples between IRF7 A carriers and A noncarriers was observed (data not shown).

Discussion
The presence of MED23, IL28B, IFN-α gene polymorphisms was not associated with mRNA levels of the four immune factors in both hippocampus and temporal cortex specimens (data not shown).
No relationship between IRF7, MED23, IL28B, and IFN-α mRNA levels and Braak and Braak or Thal scores, duration of the disease, and brain weight was found (data not shown).

Discussion
The efficiency of immune responses declines with advancing age [24,25], and old age is a major risk factor for AD. However, the role of the immune system in the disease is still unclear. Genome-wide association studies performed in AD reported that several genes with immune regulatory functions were associated with differential risk of the disease [26,27]. However, a more defined role of genes regulating innate immunity with the pathogenesis and clinical history of AD remains to be assessed.
Infections by pathogens, such as the Herpes virus, have been suggested to play a role in the clinical progression of the disease [28]. It is known that even in healthy young persons, the immune system never completely eradicates these pathogens. In the elderly, repeated cycles of activation and latency, along with infective agent persistence, may further impair immune responses and accelerate the senescence of the immune system [2,25]. Moreover, herpes viruses, which are neurotropic, might directly infect and damage selected brain areas in genetically susceptible elderly, contributing to neurodegenerative mechanisms [5], and herpes DNA has been indeed found in AD brains [4,8,25]. [8,29,30]. Other pathogens have been implicated in the clinical history of AD [31], and chronic infections are emerging risk factors for the disease [32].

Recent findings confirm and extend the association of herpes virus infection with neurodegeneration and AD
Recent data showed that amyloid-Aβ peptide protected against microbial infection in AD animal models [33]. For instance, Aβ oligomerization was necessary for its antimicrobial activity, and brain infection of 5XFAD mice by Salmonella Typhimurium bacterium resulted in an amyloid deposition surrounding the invading bacteria [33], and Aβ peptide has been related to immune defensive mechanisms of the human brain [34]. These data are compatible with the notion that the amyloid-Aβ peptide might be a component of the innate immunity against brain pathogens and virus brain chronic infection may induce its production.
Findings in the present article showed that gene expression of antimicrobial defense factors, such as IRF7, MED23, IL28B, and IFN-α, was impaired in AD brains.
Most AD brains showed decreased mRNA levels of these defensive factors, while a minority of AD patients had increased levels. Differential expression patterns might represent different AD clinical stages. Brains overexpressing the immune factors might actively respond to an active regional infection or tissue insult. After a partially successful immune response, a latent viral phase is induced, and these factors are downregulated. Such a decrement would decrease the concomitant activation of microglia and astrocyte, downregulate neuroinflammation, and mitigate neuronal damages. In other words, we suggest that different expression patterns of immune factors and cytokines may describe different reactivation and latency cycles of infectious agents.
Increased mRNA levels of MED 23, IL28B, and IFN-α were found only in the AD hippocampus, and our data suggest that different brain regions appear to be differentially involved in these chronic inflammatory responses. Replicating neurons in the hippocampus cortex [35] appear to be more susceptible to virus infections or microbial products/toxins. Therefore, persistent inflammation may induce accelerated neurodegenerative mechanisms in this brain area.
Possible regulatory mechanisms of the innate immunity genes upon amyloid-Aβ peptide expression in normal or AD brains have been poorly explored. However, it cannot be excluded that amyloid-Aβ peptide might function as an emergency defensive mechanism and compensate the impaired efficiency of other specialized immune defensive genes in the aging brain.
Here, we showed that MED23, IL28B, and IFN-α mRNA hippocampus levels in AD APOE ε4 carriers were further decreased. The temporal cortex from AD APOE ε4 carriers also showed the lowest values of MED23 and IFN-α mRNA levels.
It is known that APOE affects immunity since increased systemic pro-inflammatory states and altered immune responses have been found in APOE deficient mice [36,37]. Therefore, at least part of the increasing AD risk effect of the APOE ε4 allele might be mediated by a negative influence on the brain's immune efficiency.
Our data show that IRF7 allele A carrier status was associated with decreased levels of IRF7, MED23, IL28B, and IFN-α expression in the AD hippocampus. These findings reinforce the notion that individual genetic makeup affects brain immune efficiency.
Findings regarding IFN and brain functions are scanty. However, transgenic mice chronically overexpressing astrocyte IFN-α developed a progressive inflammatory encephalopathy and neurodegeneration [38]. Moreover, IRF7 deficient mice produced elevated levels of CXCL13 after virus infection and showed impaired regulation of microglia activity [39].
It is interesting to note that type I-IFNs also play a critical role in regulating tissue homeostasis and regeneration. Insufficient resilience, defined as impaired repair and regeneration of host tissues, rather than inefficient infectious agent clearance, may induce chronic neuroinflammation [40], and inefficient brain resilience over the years might contribute to neurodegeneration in preclinical and clinical AD.
Our results are in accordance with different genetic investigations from genome-wide association studies showing an association of several immune regulatory genes with AD pathogenesis [27]. In fact, subsequent complementary statistics and bio-informatic approaches showed that several single-gene polymorphisms in IFN genes increased AD risk [41]. Results from our investigation further support the relevance of IFN family genes in AD.
Our findings from post-mortem brain samples, however, do not rule out that the impaired expression of these genes might be a late event in the clinical progression of the disease, since, most AD patients showed high Braak and Thal scores.
Protein levels of these immune factors in AD brains were not measured, since many variables, such as post-mortem latency, disease stage or duration, may increase case protein variability, and this is a limitation of our investigation. Moreover, astrocyte or microglia classical neuropathological markers of activation were not investigated. Therefore, the contribution of these cell populations to brain mRNA levels of IRF7, MED23, IL28B, and IFN-α remains to be explored. However, astrogliosis, along with a decrease of neuron number, are classical neuropathology hallmarks of the AD brain [42] and increased or decreased mRNA levels might also be related to the brain area balance of these different cell populations.
Our data showed that different factors, such as brain area locations and individual genetic makeup, along with other undetermined factors, affect gene expression of innate immunity components and indirectly support the notion that impaired innate immune responses against brain insults, such as viruses, bacteria, fungi or their products, might accelerate neurodegenerative mechanisms in the elderly. A recent book elaborates in detail the role of infective agents in AD [43], and the present investigation agrees with the notion of infection association with AD.
In conclusion, the maintenance of efficient immune responses during aging might slow down neurodegenerative mechanisms associated with senile dementia and affect both the prevalence and incidence of AD. Further investigations regarding infections, immune defensive mechanisms, and AD progression might open new ways for AD prevention and therapy.

Post-Mortem Human Brain Tissue
Post-mortem brain tissues of AD patients and age-matched nondemented controls (ctrl) were obtained from the Netherlands Brain Bank (NBB; Amsterdam, The Netherlands), in accordance with rules and regulations of the Ethical Code from BrainNet Europe. Patients or their next of kin gave written informed consent to NBB for brain autopsy and use of tissue and clinical information for research purposes. This study was approved by the NBB scientific committee and conditions for transferring and using brain samples regulated by the Material Transfer Agreement (MTA). Neuropathological evaluation and AD pathology staging followed Braak and Braak criteria for neurofibrillary tangles (NFTs) and Thal criteria for amyloid deposition [44,45]. AD brain tissues were stored in liquid nitrogen, and 20 µm thick slices of the hippocampus and temporal cortex were cut at −20 • C, collected in RNAse free Eppendorf vials, and stored at −80 • C for the following molecular analyses.
Twenty-nine AD hippocampus brain samples, nineteen AD temporal cortex brain samples, six hippocampus, and four temporal cortex samples from ctrl cases were included in this study. The final selection was based on the availability of DNA and RNA samples of good quantity and quality.

Genomic DNA Isolation
Genomic DNA was obtained from frozen samples and purified according to the Phenol:Chloroform:Isoamyl Alcohol (25:24:1) extraction's protocol (Sigma-Aldrich, St. Louis, MO, USA) after overnight incubation with proteinase K 10 mg/mL (Roche, Basel, Switzerland) and ATL buffer (Qiagen, Hilden, Germany). Absorbance measurements were made on a NanoDrop 1000 (Thermo Scientific, Wilmington, DE, USA), and the ratio of absorbance at 260 nm and 280 nm was used to assess the purity of DNA samples, further stored at −80 • C.

RNA Isolation
RNA extraction from frozen hemi-brain hippocampus or temporal cortex samples was performed using an RNA-Bee kit (AMSBIO, Cambridge, MA, USA) according to the manufacturer's instructions. Total RNA was purified according to phenol-chloroform standard extraction after overnight incubation with proteinase K. Absorbance measurements were made on a NanoDrop 1000 (Thermo Scientific, Wilmington, DE, USA), and the ratio of absorbance at 260 nm and 280 nm was used to assess the purity of RNA samples, further stored at −80 • C.
Quantitative PCR assay (q-PCR) for gene expression analysis was realized in a CFX96 Touch™ System instrument (Bio-Rad, Hercules, CA, USA), and all reactions were run in triplicate in 96-well optical plates. q-PCR data were analyzed by Bio-Rad CFX Manager™ Software (version 3.1, Bio-Rad, Hercules, CA, USA). Using the 2− ∆∆Ct method [47], the gene expression data were computed as the gene expression fold change after normalization to the two reference genes (CYC1 and EIF4A2).

SNPs Detection
TaqMan ® SNP Genotyping Assay (Applied Biosystems, Foster City, CA, USA) was used for genotyping AD patients according to the manufacturer's instructions. It included an unlabeled PCR primer pair to detect specific targeted SNP and two different Taqman ® probes for detecting two SNP alleles: one probe was labeled with VIC ® dye and the other one with 6-FAM ® dye. Allelic discrimination was performed by probe signal intensity from PCR (RT-PCR) using a CFX96 Touch™ System instrument (Bio-Rad, Hercules, CA, USA).
APOE ε allele (rs429358 and rs7412) was assessed by RT-PCR using Taqman ® probes according to the manufacturer's instructions. The upstream variant of IRF7 (rs6598008 A/G), MED23 (rs3756784 T/G), and IL28B (rs12979860 C/T) genes were also analyzed by RT-PCR using Taqman ® probes according to the manufacturer's instructions.

Statistical Analysis
Statistical analysis was performed using the Statistical Package for the Social Sciences (version 22.0; SPSS Inc, Chicago, IL, USA) and two-sided p-values are presented.
After careful quality control of the normalized data, a generalized linear model analysis (ANOVA) followed by Bonferroni post-test or unpaired t-test was used to analyzed differences in gene expression data between groups.

Conclusions
Our data indirectly support the notion that impaired innate immune responses against brain insults induced by microorganism infection might accelerate neurodegenerative mechanisms in the elderly.
Maintenance of efficient immune responses in the elderly might slow down neurodegenerative mechanisms associated with age-related cognitive decline and affect the prevalence and incidence of AD.
Funding: Research was supported by the University of Bologna, the Italian Ministry of Research and University, and by the Netherlands Brain Bank.