The role of TNF and its receptors in Alzheimer’s disease
Introduction
Tumor necrosis factor (TNF) is one of the main proinflammatory cytokines that plays a central role in initiating and regulating the cytokine cascade during an inflammatory response [53], [89]. Expression of TNF can be induced by bacterial lipopolysaccharide (LPS), mitogens and viruses [149]. It participates in local and systemic events involving inflammation. Along with interferon gamma (IFN-γ), TNF is a potent paracrine stimulator of other inflammatory cytokines, including interleukin-1 (IL-1)-1), IL-6, IL-8, and granulocyte-monocyte colony-stimulating factor. They in turn continue and amplify the response in various ways, such as activating T and B cells and stimulating acute-phase protein synthesis, including colony-stimulating factor [26], [124].
TNF’s effect on vascular endothelial cells include, morphological changes, modulation of expression of surface antigens, and elaboration of procoagulant activity [see 10 for review]. TNF increases the expression of adhesion molecules on the vascular endothelium which can allow leukocytes and immune cells, such as neutrophils and macrophages, to be attracted to areas of tissue damage and infection [6], [48]. For example, neutrophils migrate into the intravascular space and release biologically active substances like lysozyme and hydrogen peroxide which leads to degranulation [11]. LPS activated mononuclear phagocytes are the major producer of TNF and TNF-activated phagocytes engulf and clear infectious agents and cellular debris [55], [149]. TNF is the most abundant product of activated macrophages [11]. In the presence of IFN-γ, the synthesis of TNF by LPS activated macrophages is further enhanced causing them to differentiate and activate nitric oxide synthase which, in turn induces nitric oxide production that effects the killing of microorganisms. In the absence of IFN-γ, TNF stimulates macrophages to differentiate along a pathway that results in the production of insulin-like growth factor-I [161]. Higher levels of TNF are generally related to the severity of the response, but whether greater TNF production causes more severe inflammation or whether more severe inflammation elicits increased TNF synthesis is unclear [55].
However, TNF is also important in limiting and terminating inflammation, enhancement of the repair of the damage, and angiogenesis. This is illustrated in TNF knockout mice. When homozygous TNF gene knockout mice were infected with the bacterium, Corynebacterium parvum, there was little or no initial response but the mice went on to develop a severe and fatal inflammatory reaction [91]. Heterozygous mice for the TNF gene were more susceptible to endotoxin-induced shock and to certain infections while normal mice developed an inflammatory response that resolved. Thus, TNF appears to have a dual role of being pro-inflammatory in the early phase of a response to an infection and an anti-inflammatory function in the later phases of the response in order to limit the extent or duration of the inflammation and to promote repair [89], [91].
Expression of TNF mRNA appears to be present at a low levels or absent in the normal brain [see 152 for review]. In the normal rat brain it has been detected in the hypothalmus, hippocampus, cortex, cerebellum and brainstem, however most studies using in situ hybridization methods could not detect transcripts in the brains of mice or rats. The protein also appears to be localized to the same brain regions and at low or undetectable levels. In humans, TNF mRNA has been detected in the basal ganglia, cortex, and deep and sub-cortical white matter regions [156].
Because of its’ low levels of expression it has been difficult to determine what its’ precise role is in the normal brain. In vitro studies indicate it can suppress activity of glucose-sensitive neurons of the hypothalmus, alter presynaptic a2-adrenergic receptor responsiveness in the median eminence, and modify ion channel permeability in rat hippocampal neurons [152]. TNF is produced by neurons, microglia, and astrocytes, although the latter two may be in response to pathological stimuli [20]. In an inflammatory or diseased state, TNF along with a variety of pro-inflammatory mediators and neurotoxic substances are produced by activated microglia [107]. TNF, IL-1, and IL-6 are the primary cytokine mediators of inflammation that are produced by these cells and they tend to induce each other [124]. TNF also activates NF-κB [98], [109] which stimulates the production of many substances, including survival factors such as manganese superoxide dismutase (MnSOD) [107] and the transcription of other cytokines [151], including TNF itself [144], [145]. In astrocytes, TNF, along with other substances, is a strong inducer of IL-6 [148].
Compared to other cytokine genes, TNF is highly polymorphic; this may be partly due to its’ location between the highly variable class II and class I regions of the MHC or it could also reflect environmental selection due to its importance in many biologic processes [55], including the critical role in inflammation that is described above. The TNF gene is located in the class III region of the MHC at chromosome 6p21.32, is 2,676 bp long and contains 4 exons and 3 introns [70], [89]. TNF is produced as a membrane-bound 26 kDa precursor molecule and is cleaved by the enzyme TNF-α converting enzyme (TACE) to produce a soluble 17 kDa active form of TNF [12], [103]. Two polymorphisms are located in intron 1 and one polymorphism is located in intron 2, however a total of 8 single nucleotide polymorphisms (SNPs) are located in ∼1 kb of the 5′ UTR and promoter region [70].
Some of the promoter SNPs appear to affect transcriptional activation. Reporter gene constructs containing the A allele of the -308 SNP appear to have higher transcriptional activity than with the -308G allele [19], [82], [157], [163] and some in vitro studies have demonstrated higher levels of TNF are released from cells with the -308A allele of TNF [17], [85]. Other groups do not report any differences in activation between the two alleles [21], [60], [69], [136]. These conflicting results appear to be the result of the use of different transcription factors and cell lines; in addition various protocols and reporter constructs were used. One study reported the TNF -308 polymorphism affected TNF transcription in both a cell-type and stimulus-specific manner [81]. Although there is binding of a transcription complex to this region in all cell types, it is not the only factor governing whether the promoter region containing the -308A allele results in elevated expression in a given cell type since additional cell-type specific nuclear factors binding to this area are likely to be involved in expression.
One study has found an association between the G allele of the TNF -238 polymorphism and higher TNF production [69], but this was not demonstrated by a previous study [119]. Methodological differences, such as isolated cells vs. whole blood cultures, the amount of endotoxin used to stimulate the cells and differences in study size, could explain these conflicting results. Another explanation could be the presence of a third TNF promoter SNP located at -376. One group reported the -376 SNP was aligned with the binding site for the transcription factor OCT-1 and the -376A allele was found to bind the OCT-1 protein while the G allele did not [78]. The authors reported that the regions located at -238 and -308 did not bind any protein. Since all three polymorphisms are in linkage disequilibrium, the -238 and -308 SNPs effect on expression of TNF may be due solely to the functional -376 polymorphism or a functional site elsewhere.
There is a cluster of five highly polymorphic microsatellites (TNFa-e) surrounding the TNF gene [73], [106], [146] that have also been associated with altered TNF production. One In vitro study reported the TNFa2 and c2 alleles were associated with higher TNF production while the TNFa6 and TNFc1 alleles were associated with lower TNF production [120]. However, another study found the a2 and a6 alleles were associated with lower TNF synthesis than the a4 and a11 alleles [37]. Different methodologies using different agents for stimulation may explain these results. The TNFa2, b1, c1 alleles seem to be part of an HLA-DR4 extended haplotype DRB1∗0401 with HLA-B62 that appears to be associated with high TNF production while another HLA-DR4 extended haplotype, DRB∗0401, TNFa6, b5, c1, HLA-B44 is usually associated with low TNF production. This relationship between the microsatellites and TNF production is probably due to linkage disequilibrium [55].
If the promoter polymorphisms and the microsatellites lead to or are associated with increased TNF expression levels, they could contribute to the course and/or severity of diseases, especially those involving inflammation. Inflammatory conditions that have been reported to be have a positive association with the -308A allele include coeliac disease [35], [97], primary sclerosing cholangitis [10], and primary biliary cirrhosis [55]. Interestingly, the -308A allele seems to be protective against ulcerative colitis [17] while the -308G allele is a risk factor [55]. Additionally, the TNF a2, b1, c2, d4, and e1 microsatellite alleles have been found to be positively associated with ulcerative colitis [118], although HLA haplotypes were not included.
A variety of infectious diseases of different etiologies have been associated with the -308A allele. They include, mucocutaneous leishmaniasis [24], trachoma [32], meningococcal meningitis [105], leprosy [127], septic shock resulting from bacterial infections [100], brucellosis [23], and malaria [96], [153]. The A allele of the -376 SNP has also shown to be a risk factor for malaria, independent of the -308A allele [78]. Altered TNF expression such as increased or long-term exposure could cause more damage and interfere with the appropriate response to an infectious agent just as a reduction in TNF production could compromise a full immune response. In the case of malaria, the picture is further complicated by the fact that the aforementioned adhesion molecules on the vascular endothelium are also receptors for Plasmodium falciparum so that an increase in their expression, due to TNF activation, could preferentially select for the parasite-binding phenotypes that cause malaria [78].
Possible involvement of the SNP at -238 in Hepatitis B and C has been reported [64], [65]. Chronically affected individuals were more likely to have the -238A allele than those who had eliminated the virus. The TNF genotype could influence the initial response to viral infection, such as viral neutralization and clearance, or it might affect the long-term outcome of infection [55].
Although the -308A allele has been reported to be associated to several autoimmune diseases or conditions [34], [36], [46], [63], [120], [125], [138], [158], [159], [160], most of these associations appear to be due to linkage disequilibrium with HLA haplotypes. The same is true for the microsatellite associations with autoimmune diseases in that they seem to occur as part of extended MHC haplotypes [55], [56], [57], [58], [102], [104], [120], [140] although the TNFa2 and c1 associations may be independent of HLA genes [56], [104].
Section snippets
TNF and Alzheimer disease
The above introduction lays the background as to how TNF seems to play a role in various inflammatory and infectious diseases. We will present results from our laboratory that provide genetic evidence implicating a polymorphic haplotype of TNF in late onset Alzheimer disease (AD). Initially, we reported that a chromosome 6 genome screen detected a putative AD associated region near the TNF gene at chromosome 6p21.3 [28], [30], [52], which has been confirmed in other genomic scans [49], [113].
TNF receptors
The biologic effects of TNF are mediated by binding to its’ two main receptors, the p55 TNF receptor (TNFR1) and the p75 TNF receptor (TNFR2). The TNFR1 gene is localized to chromosome 12p13.31, contains 10 exons, and codes for a 55/60 kDa membrane receptor [44], [70], [84]. The TNFR2 gene is located at chromosome 1p36.22, contains 10 exons, and codes for a 75/80 kDa membrane receptor [70], [128], [129], [134]. Both receptors belong to a superfamily of transmembrane receptors that are defined
TNF receptors and Alzheimer’s disease
One study reported that patients with dementia of Alzheimer type [DAT] were found to have more of both types of receptors on T lymphocytes than controls [16]. This could indicate a systemic immune activation in DAT patients as compared with healthy controls. A separate study has investigated a polymorphism in another receptor in this superfamily of receptors, TNFR6 which encodes FAS, a cell-surface receptor involved in apoptosis initiation [42]. They found the promoter SNP in this gene along
Discussion
The role of TNF and inflammation in AD has been much discussed and debated [see 106 for review]. There is a lack of immune cell mediation in the brain and there are none of the classic features of inflammation such as edema, swelling, and vascular proliferation. However, the upregulation of many inflammatory mediators including the complement proteins, the primary proinflammatory cytokines mentioned above, and other acute phase proteins have been demonstrated in the AD brain. These mediators
Conclusion
TNF plays a pivotal role in the general inflammatory response throughout the body. It is not only involved in the activation of other inflammatory cytokines and the surrounding events during the initial immune response, it is also important in limiting and terminating inflammation to prevent further tissue damage. It is not clear what its role is in the normal brain, but it is clearly upregulated when damage to neuronal cells has occurred. In AD it is also upregulated in and involved with
References (165)
- et al.
Neuroinflammation and Alzheimer’s disease: critical roles for cytokine/Aβ-induced glial activatin, NF-κB, and apolipoprotein E
Neurobiology of Aging
(2000) - et al.
Physical mapping and genomic structure of the human TNFR2 gene
Genomics
(1996) - et al.
Association of TNF polymorphism with primary sclerosing cholangitis
J Hepatol
(1999) - et al.
Gene polymorphism at position -308 of the tumor-necrosis-factor-alpha (TNF-alpha) in multiple sclerosis and it’s influence on the regulation of TNF-alpha production
Neurosci Lett
(1996) - et al.
TNFs protect neurons against excitotoxic/metabolic insults, and promote maintenance of calcium homeostasis
Neuron
(1994) - et al.
Celiac disease and TNF promoter polymorphisms
Hum Immunol
(2000) - et al.
No primary association between the 308 polymorphism in the TNF alpha promoter region and insulin-dependent diabetes mellitus
Hum Immunol
(1996) - et al.
Amyloidogenicity of β/A4 and β/A4-bearing amyloid protein precursor fragments by metal-catalyzed oxidation
J Biol Chem
(1992) - et al.
β-amyloid fragment potentiates IL-6, and TNF-a secretion by LPS in astrocytes but not in microglia
Cytokine
(1997) - et al.
Structure of the human TNF receptor 1 (p60)gene (TNFR1) and localization t chromosome 12p13
Genomics
(1992)