The two identified cell surface receptors for the complement activation peptide, C5a, designated CD88 and C5L2, have been reported to be upregulated in mouse models of several neurodegenerative (inflammatory) diseases, suggesting a role in either disease pathology or modulation or both (reviewed in
[22, 35]). The results presented here show increased CD88 and C5L2 protein expression in the brain of human AD cases as compared with age-matched controls. CD88 and C5L2 were associated with NFTs (mainly intraneuronal), DNPs, and NTs in the hippocampus and frontal cortex of AD cases. While CD88 and C5L2 were detected in the endothelial cells lining vessel walls, CD88 and C5L2 receptors were predominantly colocalized with both early (AT8) and mature (PHF) tangle markers associated with degenerating neurons.
The detection of a quantitative enhancement of CD88 in AD brain differs from previous studies that showed either similar neuronal expression of CD88 receptor between control and AD samples
, or a decrease of the CD88 receptor in neuron populations of AD brains compared to controls
. These quantitative differences observed in IHC studies could be attributed to differences in tissue processing or in the epitopes recognized by the antibodies. The antibodies in our study were raised against an N terminus epitope similar to, but not identical with, that used by O’Barr and colleagues
. In our study, the results were confirmed with an antibody raised against a C terminal CD88 epitope which showed similar staining patterns. In addition, Western blot analysis using brain lysates from the same cases used for IHC also showed an overall increase in reactivity of both the CD88 37 k Mr and 45 k Mr bands in the AD samples compared to controls, thereby validating our IHC data.
The C5L2 receptor distribution at the protein level in the CNS of human samples has not been previously reported, although Northern blot analysis demonstrated mRNA expression in human frontal cortex, hippocampus, hypothalamus and pons
. In rat brain, C5L2 was reported to be present constitutively in neurons and astrocytes and also upregulated by noradrenaline correlating with a hypothesized anti-inflammatory role
. In a rat model of amyotrophic lateral sclerosis (ALS) C5L2 receptor was upregulated at early stages of the disease in motor neurons
[26, 37] and was shown to be colocalized with ubiquitinated intracellular aggregates which are characteristic of ALS
. In the present work, C5L2 was associated with NFTs, prevalent in AD brain. Protein levels were increased in hippocampal lysates from AD brains compared to controls. These results were validated by utilizing two distinct cohorts of AD and control samples as well as antibodies to two different C5L2 epitopes. Preabsorption of the anti N1-23 antibody with the N1-23 peptide (N-terminus of C5L2) showed a loss of immunostaining in cortex and hippocampus and a blockade of reactivity of the 37 k Mr band in Western blots confirming that the reactivity detected was indeed against C5L2.
The role of C5a receptors in neurons, microglia and astrocytes in neuroinflammatory diseases (for review see
) is not clear. Some reports have suggested that C5a induces neuroprotective pathways in some scenarios
[12, 38], and enhanced recruitment of phagocytes to plaques could be postulated to facilitate clearance of plaques and cell debris
[25, 39]. While both beneficial and detrimental roles are not mutually exclusive, support for an overriding detrimental role for CD88 has been provided by the improvement in pathology and clinical symptoms obtained in animals models of AD, Huntington’s disease (HD) and ALS treated with specific antagonists for this receptor (PMX53 and PMX205)
[25–27], as well as in ischemic stroke
. In murine AD models, PMX205 was effective in reducing plaques, reactive glia, and improving neuronal integrity as well as cognition
. While the exact mechanism is not known, the protective effects of the antagonist could be due to the blockade of C5a binding to CD88 in glia, although an interaction of PMX205 with a neuronal CD88 is also possible. In the Tg2576 and Arc48 mouse models of AD, CD88 receptors were shown to be present in microglia polarized towards plaques and were upregulated in parallel with plaque and glia pathology
. The lack of glial CD88 detection in our study with human tissue could be due to the differential ability of the antibodies used to detect glial CD88 in fixed tissue and/or a lower degree of inflammation in the human AD brain at the time of death relative to that in the inflamed brains reported by others
The presence of C5aR in neurons suggest their involvement in functional roles that may differ from those in glia and may also be different according to neuronal cell types and/or stages of development. Several putative roles have been attributed to the neuronal C5aR, such as cytoskeletal plasticity, induction of adhesion molecules or neurotrophins and clearance of anaphylotoxins (for reviews see
[22, 35, 42]). CD88 has also been proposed to have a role in the development of cerebellum
 and to act as a guidance cue for granule cells
. In contrast, C5aR might be involved in apoptosis in neurons (15). Recently, C5a was shown to be generated by CNS neurons and to induce neuronal apoptosis upon ligation of CD88 in murine systems
. However, the role of CD88 and C5L2 prior to their colocalization accumulated in NFTs (as seen in our study) is unknown. CD88 and C5L2 have been previously shown to colocalize in human neutrophils, with CD88 mainly expressed extracellularly and C5L2 mainly intracellularly. After ligand binding, the CD88 receptors were internalized, and both (CD88 and C5L2) colocalized and associated with ß arrestin. It was postulated that the C5L2 receptor can negatively modulate neutrophil CD88 receptor signaling
 reflecting an anti-inflammatory function. C5L2 has also been postulated to have other roles, such as decoy receptors and could even exert a proinflammatory function (for a review
). However, at the stage where both C5L2 and CD88 colocalize with NFTs, these receptors are probably in a non-functional state.
The accumulation of the C5a receptors observed in AD could be due to increased synthesis and/or decreased degradation. TNFα, a proinflammatory cytokine released by microglia and astrocytes
, is increased in AD brains
 and has been shown to mediate the upregulation of CD88 receptor mRNA and protein in neurons in vivo. Region specific increases in CD88 mRNA that reached statistical significance during aging and with AD were shown by microarray analysis of a relatively large cohort of young, older and AD individuals
. However, the observed increased levels and association of both CD88 and C5L2 with tangles is also consistent with a common altered pathway of degradation. Seven transmembrane receptors, like CD88, recruit ß-arrestin, and, upon internalization, the receptor and ß-arrestin are ubiquitinated leading to degradation (reviewed in
). In PHF’s, tau is also ubiquitinated
[51, 52]. In AD a decreased activity of the ubiquitin proteasome system (UPS) has been observed, and PHF-tau was shown to inhibit UPS mediated protein degradation
. Thus, it is possible that the proteasome dysfunction observed in AD can cause tau, C5L2 and CD88 accumulation. In addition, deficits in autophagy, another mechanism of protein clearance important for removal of aggregates and misfolded proteins, are associated with AD
[54, 55]. Such defects could also contribute to decreased turnover of CD88 and C5L2.
In summary, levels of the receptors for the complement activation fragment C5a, CD88 and C5L2 are elevated in AD brain. Their colocalization with NFTs suggests that this accumulation could predominantly be a consequence of altered turnover of these receptors, rather than an increase in synthesis to contribute to, or compensate for, an inflammatory environment. While these findings are also consistent with a role of these receptors in neuron degeneration, they do not rule out a role for glial CD88 in AD pathology. Given the potential therapeutic value of inhibiting CD88 function to prevent or slow progression of AD as demonstrated in murine models of AD, further study of these receptors in the brain is warranted.