Phenotypic profile of alternative activation marker CD163 is different in Alzheimer’s and Parkinson’s disease

Microglia, immune cells specific to the brain, have been the focus of considerable interest by virtue of their association with the pathological hallmarks of Alzheimer’s Disease (AD), namely senile plaques consisting of mostly A-beta peptide (Aβ) and neurofibrillary tangles (NFT) composed of hyperphosphorylated tau. Activated microglia are also seen in the substantia nigra (SN) and the striatum in Parkinson’s Disease (PD). The varied effects of microglia in AD and PD depending on their stimuli e.g. Aβ [1–3], NFT [2, 4], neuromelanin [5, 6] and α-syn [7], indicate a multi-faceted role, both neurotoxic and neuroprotective [8, 9], but it is still uncertain if they play a key role in the pathogenesis of AD [10] and PD [11, 12]. It is still unknown as to whether they are initiators of the disease process, mediators of disease progression or a mixture of the two [13].

Microglia are involved in surveillance and maintaining homeostasis of the CNS, e.g. providing trophic support to neurons, antigen presentation and synaptic stripping [14]. Much like their peripheral macrophage counterparts, they also act as a first line of defense against infection, and are capable of destroying pathogens by triggering a pro-inflammatory state via the so-called “classical” activation pathway. While vital in protecting the tissue against foreign invasion, a consequence of classical activation can be bystander damage to the surrounding neurons, by releasing pro-inflammatory cytokines and reactive oxygen species. Although these cytokines are essential for normal functioning and are supportive of cell survival at low levels, chronic and unregulated secretion will ultimately lead to neuronal death [15]. These microglia are able to change their phenotype [16], reverting to a tissue repairing and wound healing role. This is known as the anti-inflammatory phase, whereby the microglia assume an alternative activated and/or acquired deactivation state [17]. In an ideal situation, this would spell the end of an infection episode and the restoration of normal functions. However in the case of a chronic illness like AD, the evidence suggests that there are microglia in both classical and alternative activated states [18, 19]. In the central nervous system, an extremely vulnerable environment that has limited capacity for regeneration, persistent activation of microglia can bring about irreversible injury to the surrounding tissue.

A variety of phenotypic markers have been used to identify both classical activation e.g. TNFα, IL-1α and ß, IL-6, nitric oxide synthase and alternative activation states e.g. IL-4, IL10, IL13 and TGFß in AD [18, 20, 21]. The ways in which alternative activation may contribute to Parkinsonian neuropathogenesis remains obscure. Moreover, possible pathogenic functions of these macrophage and microglial markers, including their stimuli for increased expression and their effects remain to be determined. Whether they are beneficial or detrimental to the viability of the neurons, cannot be assumed by virtue of their classification states. For example, pro-inflammatory cytokines might be associated with cytotoxicity, loss of ability to phagocytose Aβ plaques e.g. via downregulation of Aβ receptors [22] and neuronal degeneration, but TNF-α is protective due to its pre-conditioning effect on metabolic excitotoxicity, Aβ toxicity and ischaemia [23]. Alternatively, anti-inflammatory cytokines like TGF-ß are associated with alternative activation and hence are viewed as a form of immunosuppression over aberrant cytotoxicity. However its presence might also indicate the ongoing process of synaptic stripping [24].

Phagocytosis, as one of the key processes associated with alternative activation, is crucial for clearance of Aβ plaques. It has been reported that brain and blood-derived macrophages play an important role in facilitating the drainage of waste products from the brain, and are able to clear Aβ plaques much more efficiently than resident brain microglia [25–28]. However this is not known to take place efficiently in AD brains, with Aβ plaques and cerebral amyloid angiopathy [29, 30] standing testament to this fact. CD163 (Figure 1) is a phagocytic marker, of which expression is thought to be exclusive to perivascular (PVM) and meningeal macrophages [31–34]. It is a glycoprotein belonging to class B of the scavenger receptor cysteine rich superfamily. It functions as a membrane bound scavenger receptor for clearing extracellular haptoglobin-hemoglobin (Hp-Hb) complexes [35]. It is also able to recognize and bind Gram-negative and Gram-positive bacteria, and may have a role to play in host defense [36]. It has an immunoregulatory function, and is associated with the resolution phase of inflammation and the alternative activation of macrophages. This marker has been studied in diseases with a prominent inflammatory component, including HIV encephalitis, SIV encephalitis, multiple sclerosis and head injury. However, it has yet to be fully characterized in AD and PD tissue, where there is prolific glial activation.

CD163 is generally considered to be a specific PVM marker [31, 40], and indeed in our study of brains from neurologically normal controls, CD163 expression was restricted to such macrophages. However, this was not the case in our study of AD and PD, where in addition to an intense staining of PVM, CD163 was also expressed by activated microglial in the parenchyma. Foamy macrophages and microglia in the brain parenchyma have previously been found to express CD163 in HIV-encephalitis [41], multiple sclerosis, and head injury tissue [42]. In concordance with the general consensus, these are inflammatory disorders, and the observation can be attributed to breakdown of the BBB and infiltration of peripheral monocytes [31, 33, 40]. Therefore the reason for significant increases of CD163 immunoreactive parenchymal microglia in AD could be an immune response specific to the neuropathology of AD [1, 19]. Although to a lesser degree, we also observed CD163 immunoreactive microglia in PD, but both the numbers of immunoreactive microglia and intensity of their immunoreactivity were less than that in AD. CD163 immunoreactive microglia were most prominent in the brainstem of PD cases, coinciding with the regions affected in the earlier stages of the disease. While some of the CD163 immunoreactive microglia coincided with plaques both in AD and PD, focal aggregations of such cells were also seen in the absence of plaques. There was no correlation between the number of Aβ plaques and the number of CD163 immunoreactive microglia across the cases. Such an observation might suggest that these microglia are reacting both to neuronal debris and extracellular abnormal aggregates of protein, including Aβ and occasional Lewy bodies. The reason for CD163 expression in the parenchymal microglia of AD and PD patients cannot be solely attributed to Aβ plaques.

In AD, microglia are thought to be classically activated, with numerous sources from postmortem tissue in humans and transgenic animals to in vitro cell cultures providing evidence that Aβ plaques stimulate pro-inflammatory reactions [19, 34, 36, 43, 44]. Classically activated microglia are plentiful and widespread throughout the brain in PD and such activation has been linked to the destruction of dopaminergic neurons in the SN and striatum [12, 45, 46]. Recently, the alternative activation state of microglia in AD has garnered attention. Colton et al. demonstrated the presence of mRNAs of alternative activation genes in postmortem specimens and cell cultures and transgenic mice models, although Walker et al. 2009 [47] did not detect the mRNA of IL-4 and IL-13 in most of the brain samples examined. Most importantly, microglia have been shown to be capable of phagocytosis through immunization with Aβ₄₂ peptides in the Elan clinical trials. In these patients, microglia were found to increase their expression of CD68 and phagocytosed Aβ was found within lysosomes [48].

CD163 is a marker for acquired deactivation and phagocytosis [16, 20]. Its expression is induced/enhanced by interleukins (IL)- 6, IL-10, macrophage colony stimulating factor (M-CSF), and glucocorticoids [49–52]. The promoter region of CD163/HbSR gene contains potential binding sites for glucocorticoid receptor and transcription factors for myeloid differentiation [53]. Production of IL-6 and IL-10 actually occurs simultaneously with upregulation of CD163, and act as intermediates in boosting CD163 expression [54]. Though touted as an anti-inflammatory marker, it is not induced by IL-4 or IL-13. It is downregulated by lipopolysaccharide (LPS), TNF-α, TGFβ and IFN-γ [55–57].

Much interest has been shown in Aβ and its removal from the brain via natural methods or vaccination [58]; CD163 immunoreactive microglia clusters around Aβ plaques support the idea that they might serve a similar purpose. As seen in our study, CD163 immunoreactive microglia did not react with NFT or intracellular Lewy bodies, only extracellular Lewy bodies- further supporting a role in phagocytosis. In spite of this, most of them remain in a ramified state, albeit with shortened and thickened processes, which raises doubt with regards to phagocytic function. One possibility might be that they are influenced by the chronic state of the disease and the persistent generation of Aβ plaques (in AD) [22]. Staining with MRC1 remains restricted to the perivascular spaces in AD and PD. MRC1 is a mannose receptor involved in adhesion, pathogen recognition and clearance, and a marker for PVM [59]. This is similar to other findings that no other MRC1 expression was found in the parenchyma, despite acknowledged BBB damage e.g. in multiple sclerosis; and also in excitotoxic damage, acute inflammation and other chronic neurodegeneration models [31, 59]. It demonstrates that although they might serve similar phagocytic functions, CD163 and MRC1 expression and consequently their targets might be different e.g. MRC1 is involved in the recognition and endocytosis of foreign pathogens [59, 60]. It also indicates that blood-derived macrophages can adjust quickly to the CNS environment.

There has been doubt as to whether the BBB is intact in AD [19, 61]. In this study most of the microglia were found near the meninges, which might indicate infiltration of PVM from the periphery. Furthermore, some of these microglia were found to coincide with leakage of fibrinogen around vessel walls, an indication that there is a breakdown in BBB. Migration of CD163 immunoreactive macrophages from the periphery into the parenchyma might stem from compromise in the vasculature. This also corresponds with the observation that many of the CD163 immunoreactive microglia were not immunopositive for Iba1. Iba1 is a calcium binding protein expressed in macrophages/microglia [62, 63], which labels all microglia and only certain types of PVM [29, 64–66]. Signaling or infiltration through the BBB could strengthen the idea that systemic inflammation plays a part in the pathology of AD and PD [67–69], and might explain the variations in amount of CD163 positivity within each disease. Presence of microglia double immunoreactive for Iba1 and CD163 can be explained by the capability of resident parenchymal microglia to upregulate CD163, induced by signaling from the periphery. Vascular risk factors that are associated with an increased chance of developing AD might result in either an infiltration of peripheral macrophages into the brain, or increased signaling through the BBB to cause an abnormality in protein expression and upregulation. With regards to histological techniques, there has been no marker singularly capable of distinguishing between blood derived macrophages and resident brain microglia. If these CD163 immunoreactive microglia are indeed derived from the blood, it would be interesting to further confirm their origins using markers specific for systemic macrophages. Animal models can also be used to track CD163 immunoreactive peripheral macrophages and their probable entry into the brain parenchyma with or without a breach of the BBB. More experiments can be done to delve deeper into the presence and function of CD163 in AD and PD, and the subset of microglia/macrophages that express CD163 in these diseases.

In this novel study we have shown CD163 immunopositivity in parenchymal ramified microglia in all the AD and PD cases we tested. Many were found clustered around neuritic plaques; apposition with CD68 indicates the possibility for phagocytic function, and that Aβ is a stimulant for microglia. Further investigation is needed into the molecular interactions between CD163 and amyloid plaques, and whether these microglia co-express other receptors for Aβ. Amyloid plaques do not necessarily associate with CD163 immunoreactive microglia and vice versa, suggesting that the real reason for upregulation of CD163 by microglia is not a response to increased Aβ deposition. They were also found at a higher density around compromised blood vessels. AD had a much more florid reaction compared to PD, suggesting either a reaction specifically towards AD related pathology, or a more apparent BBB breakdown in AD than PD. Our study is the first to make a large-scale investigation on the expression of CD163 in AD and PD. Increase in CD163 immunoreactivity in AD was more significant than in PD, and this might be attributed to a more obvious vascular breakdown in AD, as well as the tendency for CD163 immunoreactive microglia to react with extracellular protein pathology.