Evidence of Inflammation in Parkinson’s Disease and Its Contribution to Synucleinopathy

An activated innate immune response was found in the postmortem PD brain

Early studies demonstrated that innate immune cells were found in the postmortem PD brain. Both microglial and astrocytic activation was observed, with consistent microglial activation. Microglia constitute approximately 5%–12% of CNS cells and are capable of self-renewal independent of hematopoietic stem cells [22]. In the CNS, microglia present as major immune mediators; in particular, they perform the functions necessary for the recruitment of the immune system [22]. In PD postmortem studies, microglial activation was observed to varying degrees in different brain regions. Mirza et al. [23] described the activation of microglia labeled with CR3/43 and ferritin in the substantia nigra (SN) but not in the putamen of the PD brain. Moreover, MHC-II-expressing microglia are found at significantly higher levels in the SN, putamen, hippocampus, transentorhinal cortex, cingulate cortex, and temporal cortex as well as in the lymphatic system in PD than in the brain regions of normal controls [24]. Doorn et al. [25] showed differential microglial expression patterns in the PD brain and age-matched control brain: CD68-stained amoeboid microglia were identified in the SN and hippocampus, especially in the hippocampal CA2 region, whereas Iba1-stained microglia were significantly increased in the SN of the PD brain. However, such microglia were not found in the subregions of the hippocampus in age-matched control brains [25]. In another study, Kouli et al. [26] characterized microglial responses in various brain regions by human leukocyte antigen–DR isotype (HLA-DR) and Iba1 staining. They found significantly higher microglial activation in the amygdala of demented PD patients than in agematched controls [26]. Consequently, although microglia exhibited different activation patterns in terms of microglial phenotypic markers and PD brain regions, microglial activation was consistently observed in the postmortem PD brain.

In addition to microglia, astrocytes are the most abundant CNS cell type. They demonstrate diverse morphological and functional characteristics dependent on specific brain areas [27]. Astrocytes are known to contribute to various physiological functions, including maintenance of neurons, formation of the BBB, and regulation of synapse functions [28]. Astrocytic activation in postmortem PD brains was thought not to exist because the differences in the distribution and cellular density of astrocytes stained for glial fibrillary acidic protein (GFAP) or metallothioneins I and II were not observed in the SN and putamen of PD patients [23]. However, astrocytes containing aggregated αSyn have been found in the PD brain, suggesting that astrocytes play a role in taking up abnormal αSyn [29,30]. Recently, reported data show that GFAP-positive astrocytes are significantly increased in the SN but are not found in other brain regions of PD subjects relative to age-matched controls [26]. Therefore, human autopsy studies demonstrate that microglial and astrocytic activation is involved in the PD process. However, the data obtained from postmortem brains might reflect the findings of advanced disease status or confounding terminal status affecting inflammation at death. Consequently, the in vivo status of PD must also be investigated.

In vivo evidence of inflammation in PD patients

In vivo evaluation can reflect the various stages of PD, and compared with autopsy data, it can better exclude the confounding factors affecting inflammation. Positron emission tomography (PET) imaging using ligands targeting microglial or astroglial activation has provided some insights into this respect.

Activated microglia largely exist in two polarized states, namely, the M1 phenotype and the M2 phenotype [31]. The M1 phenotype is associated with the release of proinflammatory cytokines, whereas the M2 phenotype is accompanied by the production of anti-inflammatory molecules [31]. Current PET imaging using 18 kDa translocator protein (TSPO), which is the main target for the development of ligands, can detect activated microglial status because this protein is expressed increasingly when microglia are activated. However, it is unclear why none of the current TSPO ligands are specific tracers of M1 or M2 microglia.

Using 11C-PK11195 PET imaging, increased microglial activation was demonstrated in the pons, basal ganglia, frontal cortex, and temporal cortex in PD [32]. With the same ligand, microglial activation in the temporal and occipital regions of PD patients was demonstrated to be increased by 25%–30% compared with microglial activation in these regions in age-matched controls [33]. Increased microglial activation is also shown in the SN and putamen of patients with dementia with Lewy bodies (DLBs) as well as those with PD [34]. Similarly, compared with healthy controls, patients with idiopathic rapid-eye-movement sleep behavior disorder showed microglial activation in the SN [35]. These results indicate that microglia may be activated from the very early or preclinical stages of PD. In addition to 11C-PK11195 PET, microglial activation has been identified by other PET imaging ligands with improved binding qualities, such as 11C-DPA713, 18F-FEPPA, and 18F-DPA714 [36-38]. In a study using 11C-DPA713, a significant increase in uptake was shown in the occipital, temporal, and parietal cortex of the brains of PD patients [36]. Significantly higher microglial activation was also shown with 18F-DPA714 in the midbrain, frontal cortex, and putamen of PD patients relative to controls [38]. In summary, microglial activation has been evidenced in vivo by PET studies in the premotor or clinical stages of PD. It should be noted that different activation patterns occur in various brain regions as an effect of the ligand used, which is a limitation of such studies.

From another perspective, increased astrocytic activation in the cortex and brain stem of early PD patients was identified using ¹¹C-BU9908 PET imaging, which is a marker of the imidazoline 2 binding site and binds reactive astrocytes [39]. Therefore, in vivo PET imaging has provided further evidence of activated inflammation of innate immune cells in PD along with autopsy data.

Activation of peripheral immune cells in PD

In addition to the activation of innate CNS immune cells, peripheral lymphocyte infiltration has also been observed in the PD brain, which further demonstrates the immune association in PD [26,40-44]. Increased CD4 and CD8+ T cells but not B cells or natural killer (NK) cells have been observed in the SN of the postmortem PD brain [45]. In a recent study, CD4+ T cell infiltration was identified in postmortem brain sections, especially in the perivascular space and vessel wall directly in contact with astrocytes expressing MHC-II [41]. This suggested that astrocytes were capable of expressing the factors required for T cell infiltration and activation during PD progression [41]. In another study, a significant increase in CD4+ T lymphocytes in the SN was reported in demented PD patients, whereas a correlation existed between the number of CD4+ T cells and activated microglia in the amygdala [26]. In another study, a significant increase in CD8+ T cells but not CD4+ T cells was observed in the SN of PD patients compared with that in the SN of the control group, according to Lewy body staging [44]. The presence of infiltrating lymphocytes in specific brain regions affected by PD, such as the SN and amygdala, suggests that they reach these locations through targeted extravasation rather than through a random event.

In addition, the peripheral blood of PD patients presents altered lymphocytic patterns [46-48]. Increased monocyte [46] or NK cell activation has also been reported in the blood of PD patients [47-50]. Furthermore, Niwa et al. [47] reported an increase in NK cells but not regulatory T cells and type 17 helper T cells in the blood of PD patients. Similarly, a significantly high percentage of NK cells with a lower percentage of CD3+ T cells and CD4+ T cells was identified in PD patients with clinical features relative to healthy controls [48].

In addition to the blood, the CSF of PD patients has been shown to have higher activation of immune cells, such as T lymphocytes and monocytes, than that of control subjects [51]. Additional evidence of the increased number of monocytes in the CSF of PD patients was recently reported by Pillny et al. [52] These findings indicate that peripheral inflammatory cell activation occurs in the blood and CNS of PD patients and that peripheral immune cell activation may be involved in PD pathogenesis.

However, it is less known how inflammation contributes to selective neuronal vulnerability in PD, but it is postulated that genetic predisposition combined with environmental factors may increase selective nigral dopaminergic degeneration during chronic inflammation [53].

Changes in inflammatory mediators in PD

In addition to the activation of innate CNS immune cells and peripheral immune cells, the levels of various inflammatory cytokines, which are produced by activated immune cells, have been investigated in PD [44,54-56]. The released cytokines may cause endogenous αSyn to misfold and aggregate. Aggregated αSyn can induce an inflammatory cell response that continues to release cytokines, which accounts for the continued presence of inflammation and progressive αSyn pathology [43].

Altered cytokine expression has been detected in the postmortem PD brain [26,44,57]; cytokine expression was altered in the frontal cortex, with IL-2 levels significantly higher and IL-13 and IL-8 levels lower in PD brains than in controls [57]. In another study, a significant increase in the expression of proinflammatory IL-1β cytokine was observed in the SN and frontal cortex, whereas cytokine levels in other brain regions did not differ in PD patients and healthy controls [26].

With regard to the inflammatory response in the peripheral system, Eidson et al. [55] found that the levels of two proinflammatory cytokines, namely, TNFα and IFNγ, were decreased in the serum of PD patients compared with that in the serum of control subjects. Recently, a significant increase in plasma proinflammatory cytokines, i.e., TNFα, IL-6, and IL-1β, was also found in PD patients [54,58,59]. However, variation in serum cytokine levels over the diurnal cycle must be cautiously interpreted, and various confounding factors, such as the diurnal cycle, drugs, and other medical conditions, must be controlled [55].

The inflammasome is a cytoplasmic multiprotein complex that controls the inflammatory response and coordinates antimicrobial host defense, for which the essential components are a sensor protein, ASC adapter protein, and the inflammatory protease caspase-1. These components are assembled by the detection of pathogenic signals; inflammatory caspases are activated, and proinflammatory cytokines, i.e., IL-1β and IL-18, are released to induce cell death [60]. Inflammasome activation has been observed in the peripheral blood of PD patients. Expression of inflammasome NLRP3, a type of sensor protein, was found to be significantly increased in the peripheral blood mononuclear cells (PBMCs) of PD patients, whereas other types of inflammasomes (e.g., NLRP1 and NLRP4) did not differ between PD and control subjects [54]. Likewise, NLRP3 plasma levels were significantly higher in PD patients than in healthy controls in another study [56]. However, a recent study that used monocytes from peripheral blood did not find NLRP3 activation despite the detection of altered cytokine levels [61].

Overall, the determination of cytokines or the inflammasome in synucleinopathy might provide new possible evidence of the association of inflammation and αSyn and could reveal potential biomarkers for PD.

Pathological αSyn triggers inflammation

PD animal models can provide evidence of immune cell activation during pathological αSyn introduction and its propagation [62-67]. After fibrillar αSyn injection into the brains of mice, thriving microglial and astrocytic activation was observed with αSyn propagation in these brains (Figure 2) [63,68,69]. Moreover, Lee et al. [70] reported that αSyn accumulation in glial cells produces proinflammatory cytokines and chemokines. In addition, fibrillar αSyn has been shown to induce an inflammatory response through synthesis of IL-1β through toll-like receptor (TLR) 2 interactions, which thereby leads to the activation of the NLRP3 inflammasome, followed by increased reactive oxygen species production in human monocytes [71]. This suggests that fibrillar αSyn encounters the TLR2 receptor on the surface of immune cells and triggers a downstream signaling pathway with the subsequent production of cytokines and inflammatory molecules. The roles of TLR4 and TLR2 in relation to pathological αSyn have been identified in several studies [72-75]. The involvement of other immune receptors in pathological αSyn has also been reported, e.g., cellsurface lymphocyte activation gene 3 protein (LAG3) [76,77] and complement receptor 4 (CR4) [78], indicating that αSyn triggers an inflammatory response through interactions with various immune receptor-mediated pathways. A recent study identified the upregulation of microglial voltage-gated potassium channel Kv1.3 in response to aggregated αSyn in primary microglial cultures [79].

Figure 2.

Fibrillar αSyn injection induced inflammation in the mouse brain. Iba1-positive microglia and GFAP-positive astrocyte immunoreactivity were significantly increased in the striatum of fibrillar αSyn-injected mice compared to monomeric αSyn-injected and PBS-injected mice at 7 days after injection. DAB staining immunohistochemistry. Scale bar: 100 μm. αSyn, alpha-synuclein; GFAP, glial fibrillary acidic protein; PBS, phosphate buffered saline; DAB, 3,3’-diaminobenzidine.

In a recent study by Li et al. [80], the authors investigated microglial phenotypic polarization against αSyn and found that monomeric αSyn modulated microglia toward an anti-inflammatory phenotype after exposure, whereas oligomeric αSyn led microglia toward a proinflammatory phenotype. This suggests that the pathologic form of αSyn may contribute to the aggravation of inflammation, unlike monomeric αSyn.

In addition to the inflammatory response through glial cells, the adaptive immune response through lymphocytes may be affected by αSyn. Following fibrillar αSyn injection into the CNS parenchyma, a significant increase in peripheral immune cell infiltration, such as B lymphocyte (CD19+), T helper lymphocyte (CD4+), T cytotoxic lymphocyte (CD8+), activated myeloid cell (CD11+), and NK cell infiltration, was observed in the CNS compared with the effects in a monomer-injected mouse model [63]. Likewise, an alteration of peripheral immune cells was identified in the blood and peripheral organs, including the spleen and lymph nodes, following fibrillar αSyn injection into the brain [63]. These data suggest that pathological αSyn triggers immune cell infiltration and activation in the brain as well as in the peripheral system. Moreover, Harms et al. [81] reported that αSyn overexpression induced MHC-II-mediated antigen presentation on microglia to CD4+ T cell expression; thus, the interaction between microglia and CD4+ T cells led to potent cytokine production. In agreement with this, a recent study demonstrated that αSyn triggers incremental MHC-II expression associated with an accompanying T cell response; therapeutic targeting of T cell disruption in animal models was found to reduce MHC-II expression as well as CNS myeloid cells in relation to αSyn expression [82].

To understand the differential contribution of various peripheral immune cells in synucleinopathy, Iba et al. [83] evaluated T lymphocyte populations in 10–11-month-old transgenic mice overexpressing human αSyn, which demonstrated considerable accumulation of αSyn in the cortical and subcortical regions. The authors found an increase in the number of CD3 and CD4+ T cells in mouse brains, and CD3+ T cells were detected in close proximity to glial cells (Iba1 and GFAP) and human αSyn (SYN211)-positive neurons in the neocortex, hippocampus, and striatum, which supports previous observations in the human brain [40,41]. Furthermore, an accompanying increase in IFNγ and CD1d-restricted NK T cells was identified in this animal model [83]. In another study, when αSyn oligomer was increased using an injection of adeno-associated virus coding αSyn in the T cell-competent rat brain, the presence of CD4+ and CD8+ T cell infiltration was correlated with the upregulation of MHC-II-expressing microglia. However, this correlation was not observed in T cell-deficient rats [69]. Taken together, these studies imply that pathological αSyn contributes to the recruitment of lymphocytes as well as the activation of innate immune cells.

Inflammation opposes αSyn toxicity

One of the primary roles of microglia and astrocytes is their macrophage-like function [84,85]. Lee et al. [86] explored the efficiency of clearing extracellular αSyn in different brain cell types, including neurons, astrocytes, and microglia, and found that microglia exhibited the highest rate of αSyn clearance. Supporting evidence by Loria et al. [87] also indicated that astrocytes were able to efficiently clear fibrillar αSyn more efficiently than neurons. A recent study by Tsunemi et al. [88] also emphasized the importance of lysosomal function in astrocytes for αSyn degradation during the pathogenesis of the disease. Although neurons and astrocytes can clear αSyn, microglia are the major immune cells that quickly engulf and degrade αSyn to maintain brain health by changing their morphological state to M1 (the proinflammatory phenotype) or M2 (the antiinflammatory phenotype) [89]. Choi et al. [74] suggested that the neuroprotective function of microglia occurs through ingestion and degradation of neuron-released αSyn through a selective autophagy mechanism. Therefore, a better understanding of microglial function in relation to αSyn will likely contribute to the identification of therapeutic targets for inflammation in synucleinopathies.

In studies of the cellular mechanism underlying αSyn clearance, in vitro research has identified that αSyn alters microglial function through interactions with TLR [72,73,75,90]. Stefanova et al. [72] demonstrated that TLR4 deficiency in αSyn-overexpressing transgenic mice led to significantly increased levels of αSyn in the midbrain and forebrain compared with αSyn-overexpressing transgenic mice with intact TLR4 expression, indicating that impairment of TLR4 influences extracellular αSyn clearance through microglia. In accordance with this study, Fellner et al. [73] indicated that TLR4 is required for the αSyn-dependent activation of glial cells. In addition, TLR4 was recently linked to selective autophagic degradation of αSyn in synucleinopathy [74], whereas oligomeric αSyn interacts with TLR2 to induce microglial activation [90]. Therefore, the modulation of the TLR response to αSyn can be a possible treatment target for the improvement of αSyn propagation.

Disruption of the BBB in synucleinopathy has been observed in a PD animal model [91]; this process may allow the entrance of peripheral immune cells into the brain for pathological αSyn clearance. A recent study demonstrated the protective role of NK cells (granular lymphocytes of the innate immune system) in efficiently internalizing and degrading extracellular αSyn; the systemic depletion of NK cells in a PD mouse model exacerbated motor abnormalities and αSyn propagation compared with the effects in a control group [92]. In another study, reconstitution of T cell populations reduced αSyn accumulation and resulted in persistent microgliosis in the striatum of mice with an immune system impaired by αSyn preformed fibril injection [93]. These data support that NK cells, T lymphocytes, microglia, and astrocytes can play protective roles in the αSyn pathology of PD.

Inflammation contributes to αSyn propagation

Accumulating evidence now supports the case that glial activation contributes to αSyn propagation. Microglial exosomes also contribute to the progression of αSyn pathology through exosomal transfer of αSyn from cell to cell [17,19]. In addition, reactive microglia enhance the transmission of exosomal αSyn through TLR2 [94]. Recently, Izco et al. [95] investigated the temporal relationship between αSyn and inflammation in a mouse model of PD. They found that the peak expression of Iba1-positive microglia, GFAP-positive astrocytes, and IL-1β occurred before αSyn accumulation [95]. Furthermore, several researchers have been able to block or suppress microglial activation using pharmacological drugs and thereby have identified decreases in αSyn pathology in a mouse model [17,94].

There is also evidence that astrocytes actively transfer aggregated αSyn to healthy astrocytes through direct contact or tunneling nanotubes [41,96]. Ngolab et al. [97] demonstrated that DLB brain-derived exosomes trigger αSyn accumulation in astrocytes in the mouse brain, suggesting the participation of astrocytes in αSyn transmission. In addition, activated microglia have been shown to induce neurotoxic A1 astrocytes by secreting IL-1α, TNFα, and C1q complements [98]. Thus, Yun et al. [99] demonstrated that microglia-induced inhibition of A1 astrocytes led to a decrease in pathological αSyn in an animal model.

A recent study using intranasal infusion of endotoxin–lipopolysaccharide (LPS) to induce olfactory bulb (OB) inflammation in mice found that LPS triggered microglial activation, inflammatory cytokine expression, and phosphorylated αSyn accumulation in the OB and SN through IL-1β/IL-1 receptor 1-dependent signaling [100]. This finding suggests that inflammation, particularly IL-1, plays a role in mediating the initiation and propagation of αSyn pathology from the OB [100].

In relation to systemic inflammation, La Vitola et al. [101] developed a PD mouse model through systemic administration of LPS followed by intracerebroventricular injection of an αSyn oligomer as well as an αSyn-overexpression transgenic mouse administered LPS. The authors demonstrated that peripherally induced neuroinflammation influences the action of αSyn oligomers to potentiate a detrimental effect [101]. The influence of the adaptive immune system on αSyn has also been explored [102]. By crossing human αSyn-overexpressing transgenic mice and lymphocytelacking mice, Sommer et al. [102] determined that the presence of T lymphocytes modulated toward a proinflammatory M1 phenotype was associated with an increased number of αSyn aggregates in the SN and striatum. Their findings indicate that the inflammatory response contributes to αSyn transmission through various pathways.

Therapeutic strategies targeting the interaction between inflammation and αSyn in PD

As discussed above, extensive evidence suggests that inflammation plays a critical role in αSyn-associated PD pathogenesis (Figure 1). Therefore, researchers have explored treatment targets by modulating the immune response to ameliorate αSynassociated PD pathology (Figure 3). This has led to numerous preclinical and clinical trials, as presented in Table 2 [103-117].

Figure 3.

Ongoing therapeutic option targeting the interaction of inflammation and αSyn. Therapeutic targets currently under evaluation are described in the figure, but all are in the early stages. Efforts to explore new therapeutic targets based on this mechanism will be required. αSyn, alpha-synuclein; PAP, papaverine; LAG3, lymphocyte activation gene 3 protein; CR4, complement receptor 4; TLR, toll-like receptor; IL, interleukin; TNF, tumor necrosis factor; GLP-1R, glucagon-like peptide-1 receptor; SGK1, serum and glucocorticoid-regulated kinase 1 inhibitor; CSF, cerebrospinal fluid.

Table 2.

Therapeutic trial for inflammation and its interaction with αSyn in PD