The complex role of inflammation and gliotransmitters in Parkinson’s disease

Abstract

Our understanding of the role of innate and adaptive immune cell function in brain health and how it goes awry during aging and neurodegenerative diseases is still in its infancy. Inflammation and immunological dysfunction are common components of Parkinson’s disease (PD), both in terms of motor and non-motor components of PD. In recent decades, the antiquated notion that the central nervous system (CNS) in disease states is an immune-privileged organ, has been debunked. The immune landscape in the CNS influences peripheral systems, and peripheral immunological changes can alter the CNS in health and disease. Identifying immune and inflammatory pathways that compromise neuronal health and survival is critical in designing innovative and effective strategies to limit their untoward effects on neuronal health.

Introduction

Parkinson’s disease (PD) is the second most common neurodegenerative disease worldwide and is associated with extensive motor impairment and loss of quality of life, but the underlying etiology of PD remains enigmatic(Poewe et al., 2017). In some instances, genetic variants conferring increased risk of PD have been identified, but familial PD represents a small subset (>10%) of PD cases, with most cases still considered of unknown origin or idiopathic (Bandres-Ciga et al., 2020; Lesage and Brice, 2009). Regardless of origin, inflammation and immunological dysfunction are common components of PD, both in terms of motor and non-motor components of PD(Barnum and Tansey, 2012; Kim et al., 2022; Lindqvist et al., 2012). In recent decades, the antiquated notion that the central nervous system (CNS) in disease states is an immune-privileged organ, has been debunked(Brioschi and Colonna, 2019; Harris et al., 2014; Louveau et al., 2015). At large, the field agrees that the immune landscape in the CNS influences peripheral systems, and that peripheral immunological changes can alter the CNS in health and disease and for PD pathogenesis, the contribution of peripheral organs has gained much attention in recent years.

The routes by which neuroimmune communication occurs continues to be debated, with many groups having made headway in this area(Abbott et al., 2018; Goldeck et al., 2016; Iliff et al., 2012; Kortekaas et al., 2005; Mogensen et al., 2021; Ransohoff and Brown, 2012; Ransohoff et al., 2003). But the consensus remains that neuroinflammation and peripheral inflammation occurs in PD. CNS glial cells, namely astrocytes and microglia, are frontline responders to brain changes and have been extensively studied in PD. Similarly, peripheral organs and in specific the role of peripheral blood innate and adaptive immune cells have more recently received extensive coverage in the context of neuroinflammation and peripheral inflammation in PD. In this review, we summarize key findings in CNS inflammation, effects on various neuronal populations, as well as accompanying peripheral immune changes which are implicated in the view of PD as a disease of chronic inflammation.

Microglia and astrocytes in neuroimmune communication and neurodegenerative disease

Neuronal functions in healthy and disease states are modulated by both neuronal factors and via cell-to-cell communication with CNS-resident microglia and astrocytes. Emerging clinical and experimental evidence suggests that normal glial function is critical to normal brain function and neuronal homeostasis, and that dysfunction of these cells plays key roles in the development and/or progression of CNS diseases(Frost and Schafer, 2016; Schwartz et al., 2013; Sofroniew and Vinters, 2010; Wyss-Coray and Mucke, 2002). The ability of glia to recognize and respond to disruptions in proteostasis, such as alpha synuclein aggregates, and glial or immune expression of mutant genes implicated in PD(Booth et al., 2017; Jeon et al., 2020; Zhang et al., 2016), have led to extensive investigation of glia in PD. Indeed, immuno-modulation by glia in preclinical PD models continues to be investigated. Microglia, CNS myeloid cells which are the self-renewing tissue resident macrophages, are heterogenous and survey the CNS parenchyma at steady state via a network of tiny processes. At rest, microglia are termed ramified, and phenotypically characterized by a small cell body, with fine processes extending outwards(Hristovska and Pascual, 2015; Lier et al., 2021; Morioka et al., 1993; Streit, 1993; Streit, 2002; Streit and Graeber, 1993). During insult or injury, microglia become active phagocytes and exhibit an activated, macrophage-like ameboid phenotype(Gopinath et al., 2020b; Shaerzadeh et al., 2020; Streit, 2002). Basics of microglial morphology, activation and subsequent functional changes are investigated extensively; for this information we refer readers to Streit et al 2002, Kettenman et al 2011 among many others(Badanjak et al., 2021; Beynon and Walker, 2012; Donat et al., 2017; Kettenmann et al., 2011; Norden et al., 2016; Pike et al., 2022b; Sarkar et al., 2020; Sarlus and Heneka, 2017; Streit, 2002; Streit et al., 2004).

Microglial dysfunction and senescence occur with age, with some groups indicating microglial function or morphology may vary in key regions affected by PD(De Biase et al., 2017; Shaerzadeh et al., 2020; Streit, 2002; Streit et al., 2004). With age as a primary risk factor for PD, it is possible that the aging process confers increased risk of dopamine neuron degeneration that could be linked to senescence of supporting microglia(Shaerzadeh et al., 2020; Streit, 2002; Streit et al., 2004). The concept of microglial activation and senescence is complex. Early studies conducted by McGeer et al. found that microglia showed increased HLA-DR expression (the human equivalent of major histocompatibility complex 2) nearby to dopamine neurons in the midbrains of PD patients postmortem(McGeer et al., 1988). Subsequent studies showed increased MHC-II expression in experimental models not only in the midbrain but in other brain regions as well, suggesting that microglial antigen presentation may play a critical role in PD progression(Imamura et al., 2003; Kurkowska-Jastrzębska et al., 1999; McGeer et al., 1988). These findings were supported by GWAS studies of PD patients relative to healthy controls, in addition to studies showing that peripheral innate immune cells in individuals with PD and a high-risk genotype in a common genetic variant in HLA-DRA implicated in late-onset PD expressed markedly increased MHCII in response to IFNγ stimulation (Hamza et al., 2010; Hill-Burns et al., 2011; Kannarkat et al., 2015; Nalls et al., 2011). In addition, microglia have also been implicated in cell-to-cell seeding of alpha-synuclein aggregates(George et al., 2019) and evidence from postmortem studies and experimental models of PD-like degeneration have shown that microglial activation is also associated with onset and progression of PD-like pathology, with activation in close association with neuroinflammation(Béraud et al., 2013; Block and Hong, 2007; Harms et al., 2018; Kim and Joh, 2006; Lavisse et al., 2021; Lull and Block, 2010; Muzio et al., 2021; Nagatsu and Sawada, 2005; Ouchi et al., 2009; Ouchi et al., 2005; Smajić et al., 2022). In some preclinical models of PD, microglia-mediated inflammation exacerbates dopamine neuron degeneration via inflammasome activation and subsequent cyclical inflammation(Lee et al., 2019; Pike et al., 2022a; Pike et al., 2022c). Nevertheless, the alternative hypothesis that microglia activation is a response to neuronal injury or cell death that initiates a runaway cycle of neuroinflammation, and a feed-forward cascade of neuronal demise should be considered. It is possible that early immune activation in PD may play protective roles, whereas chronic cycles of inflammation resulting from chronically activated microglia could be maladaptive and lead to the demise of selectively vulnerable neuronal populations. Taken together, these findings suggest microglial activation, antigen presentation via MHC-II, abnormal protein accumulation and subsequent inflammatory responses may play a role in the etiology and pathophysiology of PD.

Studies of microglial activation in PD were quickly followed by studies in which microglial activation was inhibited to attenuate the degeneration of dopamine neurons. In vitro studies showed that antagonists of TLR4 and TLR2, both involved in innate immune defenses against pathogens and the associated inflammatory response, promote dopamine neuron survival, but have yet to be studied in vivo in preclinical PD models(Gorecki et al., 2021; Hughes et al., 2019). Examination of microglia as therapeutic targets for PD are ongoing(Liu et al., 2019). As a note of caution, in other diseases processes such as amyotrophic lateral sclerosis (ALS), early preclinical evidence suggested microglial inhibition might prove an important therapeutic target. However, these results have failed to translate into human clinical trials, indicating that preclinical data from disease models should be interpreted with caution when extrapolating to human subjects(Dupuis et al., 2012; Gordon et al., 2007; Wobst et al., 2020). In the investigation of inflammation in PD, relationships between chronic anti-inflammatory drug use (aspirin, ibuprofen, and other non-steroidal anti-inflammatory drugs; NSAIDs) and the risk of developing PD has been investigated, with multiple analyses suggesting an association between inflammation, NSAID use and PD risk. However, the results were mixed with some studies suggesting that some NSAIDs, but not aspirin, confer decreased risk of developing PD, while other studies show NSAIDs, on the whole, do not alter the risk for PD if taken later in life(Gagne and Power, 2010; Poly et al., 2019; Rees et al., 2011; Samii et al., 2009). Results by Hernan et al. in 2006 suggested possible sex-specific effects of NSAIDs on PD risk, possibly explaining the contrasting results from other studies on the topic and warranting further examination(Hernán et al., 2006) during ongoing investigations of the links between inflammation and PD.

Astrocytes regulate synaptic transmission in addition to innate and adaptive immune responses, depending on timing and context of stimuli present during an inflammatory response (Coulter and Eid, 2012; Kirischuk et al., 2016; Newman, 2003). Astrocyte activation can be beneficial or detrimental for tissue repair, depending on the nature of the inflamed milieu. Astrocytes are in close contact with other CNS-resident cells such as microglia, oligodendrocytes, neurons, and blood vessels that maintain the blood brain barrier (BBB) integrity and its permeability. Astrocytes’ proximity to blood vessels in the CNS enables them to regulate immune cell trafficking via secretion of cytokines and chemokines that can activate adaptive or innate immune responses(Phatnani and Maniatis, 2015; Sofroniew, 2015a; Sofroniew, 2015b; Sofroniew and Vinters, 2010). Growth factors such as glial derived neurotrophic factor (GDNF), brain-derived neurotrophic factor (BDNF) and others are generated and released by astrocytes and are critical to the health and maintenance of dopamine neurons(Lein et al., 2007; Lin et al., 1993; Pöyhönen et al., 2019). In the presence of neuroinflammation, astrocytes become ‘reactive’, secreting a variety of bioactive molecules, including chemokines, growth factors and neurotrophins and altering their production of neuron-supporting growth factors(Linnerbauer et al., 2020; Sofroniew, 2015a; Sofroniew, 2015b). In their role as key regulators of the BBB, astrocyte end-feet surround and enclose CNS vasculature and are well positioned to interact with and influence any peripheral immune cells patrolling the CNS environment(Abbott et al., 2018; Abbott et al., 2006; Linnerbauer et al., 2020; Sofroniew, 2015a; Sofroniew, 2015b). During neuroinflammation, astrocytes are simultaneously exposed to an array of stimuli originated from periphery and CNS that can activate different intracellular signaling pathways and responses. Thus, astrocyte responses are key determinants of the net result of CNS and peripheral immune perturbation. Indeed, astrocytes under inflammatory conditions, are capable of upregulating MHC-II, and possibly presenting CNS antigens to patrolling peripheral immune cells, underscoring the importance of CNS-to-peripheral immune system connections in PD and contributions of peripheral immune cells in the context of PD-associated neuroinflammation(Abbott et al., 2006; Sofroniew, 2015a; Sofroniew, 2015b). The identification of key pathways in neuroinflammation together with the timely reconstruction of the inflamed milieu deserves further investigation. Understanding the nature and duration of immune insult(s) that stimulate or inhibit the production of pathogenic or protective astrocytes are critical for the development of new therapeutic approaches in neuroinflammatory and neurodegenerative disorders.

Immunological mediators in neurotransmission

Immunological mediators, originating from microglia, astrocytes, or peripheral immune cells, can act on non-immune cells in the CNS including neurons. A wide range of studies have found neurons express receptors for multiple types of cytokines, including TNF, IL1b, IL4, and IL10, among others(Friedman, 2001; Herz et al., 2021; Huang et al., 2011; Probert, 2015; Zhou et al., 2009). Although cytokines are most studied for their neuroprotective or pathological effects in disease states, cytokines can modulate the fundamental property of neurons—their electric signaling have apoptotic effects on vulnerable populations. Cytokines can modulate the conductance of several channels across a wide variety of neuronal subtypes in both health and disease. This section reviews some examples of cytokine modulation of neural activity and how this may relate to multiple aspects of the pathophysiology in PD.

While the exact effect a given cytokine has on neuronal activity varies by neuronal subtype, the conserved phenomenon of cytokine modulation of neuronal conductance represents a prime example of the integration between immune and neural signaling. Cytokine regulation of neuronal activity in neurodegeneration, especially PD, represents an open field and compelling area of future study.

Perhaps one of the most well-studied areas for cytokine modulation of neural activity is in the context of pain and sensation. The immune system plays a critical role in the development of chronic pain, and recent studies have attempted to highlight the immune system as a potential therapeutic target(Kavelaars and Heijnen, 2021). Intuitively speaking, crosstalk between the immune system and pain-sensing nociceptive neurons seems like a critical aspect of sensing external threats and physical harm. The immune system detects tissue damage or microbial invasion and communicates to nociceptors which, in turn, relay this information to the CNS to register as pain. Inflammatory cytokines such as TNF, IL1, CCL2, IL33, and IL17, increase the activity of sensory neurons. For example, TNF can increase the amplitude of TRPV1 currents in dorsal root ganglia (DRG) neurons, making them more sensitive to cognate stimuli(Nicol et al., 1997; Sommer et al., 1998). IL17 has a similar effect, but this was only observed in females(Luo et al., 2021). Other pro-inflammatory cytokines increase neuronal excitability through their action on voltage-dependent conductance. Both acute and long-term IL1b can increase voltage-gated sodium (NaV) currents and can also increase hyperpolarized cyclic nucleotide (HCN) and voltage-gated calcium (CaV) currents while decreasing voltage-gated potassium (Kv) currents such as K(Ca)_1.1 collectively increasing DRG excitability(Noh et al., 2019; Stemkowski et al., 2021; Stemkowski et al., 2015). Yet another mechanism that has been reported for increased DRG activity in inflammatory settings is through synaptic events. Specifically, CCL2 increases excitatory synaptic event frequency and amplitude(Gao et al., 2009). Thus, cytokines directly modulate pain circuitry through diverse mechanisms involving intrinsic excitability, synaptic connectivity, or both.

Cytokine-mediated neuromodulation has been studied in the context of learning and memory given that cytokines can act as gliotransmitters. Multiple groups have consistently shown that immune-depletion paradigms, either general or targeted, impair performance on the Morris Water Maze, Y-Maze, or Contextual Fear Conditioning tasks (reviewed in(Salvador et al., 2021)). Traditionally, glia has been hypothesized as necessary mediators for the immune system’s effect on cognitive behaviors, but immune signals can directly act on central neurons to influence their excitability and/or connectivity. TNF was reported play a role in synaptic scaling, a form of learning and memory and to increase excitatory transmission in hippocampal neurons by increasing AMPAR membrane levels and decreasing GABA-AR surface levels, but IL-1β, another pro-inflammatory cytokine increased GABAergic post-synaptic potentials and decreased mEPSC frequency, suggesting there is no convergent effect of inflammation on hippocampal excitability(Beattie et al., 2002; Hellstrom et al., 2005; Stellwagen et al., 2005; Stellwagen and Malenka, 2006; Tomasoni et al., 2017). Furthermore, IL-17−/− mice were reported to have increased synaptic transmission in the dentate gyrus in slice preparation and increased intrinsic excitability in vitro(Ribeiro et al., 2019). Contrary to these findings, a separate report found that IL-17−/− mice had decreased synaptic transmission in the hippocampal CA1 region after a short-term memory task(Liu et al., 2014). Additionally, blocking IL-17α in an experimental encephalitis model partially rescues LTP deficits(Di Filippo et al., 2021). Similarly, anti-inflammatory cytokines, such as IL-4 and IL-10 have divergent effects on intrinsic hippocampal excitability by either decreasing or increasing K(Ca) currents, respectively(Chen et al., 2020; Levin et al., 2016; Nenov et al., 2019). Together these reports highlight the regional and context specificity that cytokine signaling can have in the CNS at steady and inflammatory states. Thus, in both branches of the nervous system, a motif appears: cytokines act as critical integrators of environmental conditions to neuronal computation. Across neuron subpopulations, cytokine modulation is a conserved phenomenon at the steady state, albeit with divergent effects. This knowledge has already been applied to pain research and shows promising signs of potential clinical translation, but our understanding of this phenomenon in neurodegeneration is limited. There are a few reports of the effect of cytokines in neuroinflammatory conditions but in the context of Alzheimer’s disease (AD) and PD, this is an area that deserves more effort(Di Filippo et al., 2021). Recognizing that cytokine signaling can affect memory formation at baseline, they may play a role in the cognitive decline of AD. Surprisingly, there is a dearth of knowledge on cytokine modulation of midbrain dopamine neurons, representing a critical area for future studies.

A surge of studies that cytokines directly modulate neuronal activity in health and disease have emerged, but there are still outstanding questions on cell-type specific mechanisms and relevance to behavior. Some groups have reported the intracellular signaling cascades responsible for increased or decreased electrical properties of the neurons following stimulation by cytokines, however this information is likely specific to certain neuronal types (Belkouch et al., 2011; Obreja et al., 2002; Schäfers et al., 2003). Given the growing appreciation for neuronal heterogeneity even within a single brain region, it is unknown how some cytokines may have divergent effects on membrane conductance and neuronal excitability(Blum et al., 2021; Kamath et al., 2022; O’Leary et al., 2020; Poulin et al., 2016). Whether cytokines affect expression, trafficking, or gating, these mechanisms must be elucidated to inform clinical implementation of cytokine-targeting therapies. Indeed, IL-4R expression on GABAergic neurons has been shown to decrease learning in a fear conditioning paradigm by altering dentate gyrus synaptic transmission, and, in a similar contextual learning paradigm, CCR5 was shown to modulate ensemble formation in dCA1 thereby affecting learning(Herz et al., 2021; Shen et al., 2022). Collectively these and subsequent studies will generate a more comprehensive atlas of cytokine-mediated modulation of behavior in the steady and inflammatory states, informing the field regarding cytokines as a therapeutic target in PD.

Midbrain dopamine neurons exhibit a canonical pacemaking property, tonically firing at 1–8Hz depending on the system(Grace and Bunney, 1984; Lin et al., 2021; Otomo et al., 2020). Properly regulated firing and neurotransmitter release by Substantia Nigra pars compacta dopamine neurons is critical in initiating goal-directed movement and is lost in PD. Dopamine neurons, particularly in the Substantia Nigra pars compacta, rely on the coordinated activity of a number of voltage and ligand-gated ion channels, including voltage-gated sodium (NaV), voltage-gated potassium (Kv), voltage-gated calcium (Cav), and hyperpolarized cyclic nucleotide-gated (HCN) channels to maintain their pacemaking. Additionally, ionotropic ligand-gated receptors, such as n-methyl-D-aspartate (NMDA) receptors are expressed on dopamine neurons and in the striatum where they regulate bursting activity and play crucial roles in synaptic plasticity. For a more comprehensive review on these channels in dopamine neurons see work by Gantz and colleagues (Gantz et al., 2018). Together, these conductances culminate in the tonic and phasic dopamine activity responsible for movement.

Alterations in channel expression and localization have been noted in PD animal models and post-mortem tissue. CaV channels, specifically CaV1.3, have long been thought to be critical in PD progression, as their inhibition rescued some animal models(Chan et al., 2007; Guzman et al., 2018; Kang et al., 2012; Lee et al., 2014; Liss and Striessnig, 2019; Marras et al., 2012; Schapira et al., 2014; Surmeier et al., 2012; Surmeier et al., 2017). However, it should be noted, that a clinical trial of CaV1.3 inhibitor failed, and a recent animal model suggests that this channel may be downregulated early in disease(2020; González-Rodríguez et al., 2021). Nevertheless, increased calcium influx is a major suspect of bioenergetic imbalances thought to belie dopaminergic vulnerability(González-Rodríguez et al., 2021; Guzman et al., 2010; Lieberman et al., 2017; Sanchez-Padilla et al., 2014). Other pacemaking related channels, particularly potassium channels, have also been implicated in PD(Chen et al., 2018). Specifically K(Ca) channel SK and A-type potassium channels have shown altered expression or potential therapeutic use in a variety of animal models including the A53T mouse model, 6OHDA rat models, and MPTP models in rat and non-human primate, and even in post-mortem tissue(Aidi-Knani et al., 2015; Alvarez-Fischer et al., 2013; Chen et al., 2014; Doo et al., 2010; Haghdoost-Yazdi et al., 2011; Kim et al., 2011; Lu et al., 2019; Mourre et al., 2017; Subramaniam et al., 2014). Again, disrupted pacemaking activity is thought to exacerbate bioenergetic stresses in dopamine neurons potentially leading to their demise. Furthermore, circuit dysfunction is a major cause of motor symptoms as well as L-DOPA-related complications (reviewed in (McGregor and Nelson, 2019)), and changes in NMDA receptor levels in the striatum are thought to contribute to these symptoms(Calabresi et al., 2000a; Calabresi et al., 2000b; Hallett and Standaert). Whether by altering basal ganglia circuitry or by potentially contributing to dopamine neuron vulnerability, dysregulation of membrane conductance represents an important aspect of PD pathophysiology.

While there are several immune-based, progressive, animal models of PD, none have been studied for how immune signaling modulates neural activity in model progression. All the channels above have shown some capacity to be modulated by cytokines. For example, A-type currents in sensory neurons are suppressed by IL-33(Wang et al., 2022b). CaVs are regulated by a host of cytokines—TNF increases L-type currents in hippocampal neurons, IL-1β increases CaV currents in sensory neurons, and IL-17α decreases N-type currents in sympathetic neurons(Chisholm et al., 2012; Furukawa and Mattson, 1998; Noh et al., 2019; Stemkowski et al., 2015). Similarly, ionotropic glutamate receptors, including NMDA receptors, exhibit a diverse sensitivity to various cytokines. Low dose IL-1β increases NMDA-mediated Ca-influx in hippocampal neurons(Viviani et al., 2003), but decreases mEPSCs at a higher dose(Tomasoni et al., 2017). Thus, channel modulation by cytokines exhibits significant cell-type specificity. Currently, there is a lack of information on cytokine modulation of mid-brain dopaminergic neurons. Studies focusing on these neurons specifically will be required before examining immune-mediated changes to conductances in PD models. Nevertheless, such investigations may reveal additional indications for cytokine-targeted therapies in PD.

Peripheral myeloid cells in Parkinson’s disease

While the involvement of peripheral immunity in neuroinflammation has garnered a lot of attention, investigating whether changes in the peripheral immune system are cause or consequence of degeneration of dopamine neurons or other selectively vulnerable populations that degenerate in PD, recent data support the interpretation that loss of catecholaminergic neurons in PD alters immunophenotype of peripheral myeloid and lymphoid cells (Chen et al., 2021; De Virgilio et al., 2016; Gardai et al., 2013; Gopinath et al., 2021; Gopinath et al., 2020b; Gopinath et al., 2020c; Gopinath, 2022; Grozdanov et al., 2019; Kustrimovic et al., 2019; Kuter et al., 2020; Öberg et al., 2021; Sonninen et al., 2020; Sulzer et al., 2017; Tan et al., 2020; Troncoso-Escudero et al., 2018; Vila et al., 2001). These data strongly support the existence of a compensatory immune response in the peripheral immune cells, in PD. Within the myeloid compartment, PD patients exhibit pronounced increase in inflammatory monocytes that is associated with a decrease in resolution-phase immune cells(Gopinath, 2022; Grozdanov et al., 2014). The inflammatory characteristics of these cells underpins an altered dynamic between inflammatory monocytes (that are normally upregulated by inflammatory mediators such as TNF and IL-1β) and resolution-phase monocytes that display a tissue-repairing phenotype.

Macrophages and monocytes are of particular interest among myeloid cells, as these cells are dominant first responders to inflammation(Kapellos et al., 2019; Merah-Mourah et al., 2020; Wijeyekoon et al., 2018; Wong et al., 2012). Blood derived monocytes and macrophages are known to enter the CNS under acute inflammatory conditions in response to CNS lesions and secrete inflammatory factors which can contribute to neuroinflammation and degeneration(Cheng et al., 1998; Harms et al., 2018; Varvel et al., 2016). The recent discovery that parenchymal border macrophages (PBMs), a specialized kind of perivascular and leptomeningeal macrophages, regulate CSF flow and clear toxic aggregates suggests a complex interplay between multiple myeloid cell types in maintaining CNS homeostasis(Drieu et al., 2022; Mazzitelli et al., 2022). Unlike chronic conditions such as PD, acute CNS inflammation in multiple sclerosis (MS) is associated with extensive peripheral immune infiltration into the CNS and subsequent inflammatory damage(Forbes and Miron, 2022; Mikita et al., 2011). Indeed, peripheral myeloid cells acquire and present CNS antigens to peripheral adaptive immune cells, directly contributing to demyelination in MS(Riedhammer and Weissert, 2015; van Langelaar et al., 2020). Whether or not peripheral myeloid cells participate directly in neurodegenerative processes in PD remains to be determined. Recent data from Sulzer and colleagues support the idea that alpha-synuclein-specific T-cells develop in PD patients and may correlate with markers of clinical progression as described in more detail below(Dhanwani et al., 2022; Garretti et al., 2019; Sulzer et al., 2017). Finally, in addition to peripheral myeloid cells, recent reports of reservoirs of myeloid cells in the meninges derived from the skull and vertebral bone raise the possibility that these could play important role in maintaining brain health, and may become dysfunctional with age and contribute to PD (Cugurra et al., 2021).

Dendritic cells (DCs), which arise from myeloid lineage cells in the bone marrow, are uniquely positioned among myeloid cells as professional antigen presentation cells (APCs)(Hilligan and Ronchese, 2020). Foremost among DC roles in immunity is to present peptides from inflammatory or pathogenic states to other immune cells, including B and T cells (adaptive immune system), to activate a systemic immune response. In the context of CNS inflammation, DCs can migrate from the meninges and glymphatic vessels after exposure to CNS antigens, trafficking to nearby lymph nodes (primarily cervical lymph nodes) and present CNS antigens to peripheral lymphoid cells(De Laere et al., 2018; Laman and Weller, 2013; Mohammad et al., 2014; Sagar et al., 2012). Subsequently, peripheral adaptive immune cells can home into the inflamed CNS compartment, potentially contributing to inflammation and neurodegeneration.

Lymphoid cells in Parkinson’s disease

Peripheral lymphoid cells enter the CNS compartment both under steady state and in disease state(Balasa et al., 2021; Cabarrocas et al., 2003; Gross et al., 2020; Kleine and Benes, 2006; Strazielle et al., 2016). Most notably, in PD, T cells are detected near dying dopamine neurons, suggesting T cells may play a role in early degeneration or PD progression(Ambrosi et al., 2017; Brochard et al., 2009; Galiano-Landeira et al., 2020; Garretti et al., 2019; Sulzer et al., 2017). However, whether T cells are responsible for directly damaging dopamine neurons remain less understood. In neurodegeneration and neuroinflammation, T cells which recognize specific antigens (antigen-specific T cells) are of particular interest. CD8 positive T cells, known as cytotoxic T cells, and CD4 positive T cells, known as helper T cells, both can recognize antigens. The investigation of antigen specificity in T cells of PD patients have shown variable results, reflecting the heterogeneity of human PD and the limited interpretation of T cell involvement in the neurodegenerative process(Garretti et al., 2019; Singhania et al., 2021; Sulzer et al., 2017). For example, Sulzer and others have shown that antigen specific T cells in PD patients can recognize alpha synuclein, implicated in familial PD(Bandres-Ciga et al., 2020; Oliveira et al., 2015; Singleton et al., 2003; Sulzer et al., 2017). However, the epitopes recognized by T cells derived from PD patients did not reveal any single antigen dominating the peripheral T cell landscape suggesting while T cells may be involved in PD, a single clonally expanded population of antigen-specific T cells is unlikely to directly underlies degeneration of dopamine neurons(Singhania et al., 2021). Of interest, is that T cells are primary sources of inflammatory cytokine interferon gamma (IFNy) in the periphery, which has been repeatedly shown to play a role in PD and PD-like neurodegeneration(Barcia et al., 2011; Ferrari et al., 2021; Kambayashi et al., 2003; Murray et al., 2002; Panagiotakopoulou et al., 2020; Perussia, 1991).

B cells are also key players in the adaptive immune system but have not been directly detected in the brains of PD patients, PD mice nor in proximity to degenerating neurons(Brochard et al., 2009; McGeer et al., 1988; Wang et al., 2022a). However, given that PD patients experience persistent peripheral inflammation, the link between altered B cells and PD is theorized(Campos-Acuña et al., 2019; Ferrari and Tarelli, 2011; Gopinath et al., 2021; Gopinath et al., 2022; Kustrimovic et al., 2019; Qin et al., 2016; Süβ et al., 2021; Williams et al., 2022). Consistent with this hypothesis, decreased naïve B cells and an increase in memory B cell subsets are reported in PD patients, suggesting an immunological activation state in response to inflammation(Wang et al., 2022a). B cells undergo a process called somatic hypermutation, during which highly specific antigen-specific antibodies are produced(Bräuninger et al., 2003; Küppers, 2003; Wang et al., 2022a). A requirement for somatic hypermutation is an immunological trigger, such as inflammation(Beltrán et al., 2014; Brink and Phan, 2018; Pierre et al., 1997). Since both CNS and peripheral inflammation is thought to occur in PD, both peripheral and CNS locations for B cell affinity maturation have been investigated(Jain and Yong, 2021; Lu and Robinson, 2014). While still uncorroborated, meninges could function as a lymphoid interface in which antigen specific T cells meet and induce B cell somatic hypermutation(Hartlehnert et al., 2021; Serafini et al., 2004). Consistent with this idea, a subset of meningeal B cells were recently shown to derive from a local bone marrow source (the calvaria) and become antigen-experienced in the CNS(Brioschi et al., 2021). Potential roles of B cells in PD warrants further exploration, as antibodies specific to components of dopamine neurons have been detected in the blood of PD patients at higher levels than in control subjects(Besong-Agbo et al., 2013; Carvey et al., 1991; Han et al., 2012; Kannarkat et al., 2013).

Inflammatory mediators in Parkinson’s disease

Beyond antigen presentation and canonical immune functions, both CNS and peripheral immune cells are involved in the inflammatory response, and therefore are thought to be central to neuroinflammation in PD. Mounting evidence suggests that PD involves a chronic inflammatory state, supported by the fact that inflammatory cytokines such as TNF, IL-6, IL-1β and other cytokines are increased in both brain and peripheral immune system(Barnum and Tansey, 2012; Imamura et al., 2003; Kim et al., 2022; Lindqvist et al., 2012; Magnusen et al., 2021; Mogi et al., 1994; Nagatsu et al., 2000; Nagatsu and Sawada, 2005; Sawada et al., 2006). Whether cytokines from the CNS can reach peripheral circulation and vice versa in PD is unclear. Microglia and astrocytes are potent sources of CNS cytokines including TNF and IL6, with activated microglia secreting inflammatory cytokines and upregulating expression of MHC-II(Ezcurra et al., 2010; Hickey and Kimura, 1988; Imamura et al., 2003; Imamura et al., 1990; Perry et al., 1985; Perry et al., 2010). Microglia are particularly dense in the midbrain region(Kim et al., 2000), with different percentages of microglia relative to dopamine neurons in the dopaminergic midbrain(Shaerzadeh et al., 2020). Whether or not microglia are the main drivers of neurodegeneration, dying dopamine neurons release damage-signal molecules such as neuromelanin, ATP, alpha-synuclein, driving microglial activation and subsequent inflammatory cytokine secretion(Davalos et al., 2005; Gao and Hong, 2008; Perry et al., 2010; Wilms et al., 2003; Zhang et al., 2005). In fact, microglial release of inflammatory molecules may influence not only dopamine neuron death but also induce further cyclic microglial activation, potentially leading to a runaway cycle of neuroinflammation(Block and Hong, 2007). While still debated and lacking clinical evidence, the literature supports a role for cytokine signaling in PD.

In the brains of PD patients, expression of IL-1, IL-2, IL-1β, TNF, IL6, TGFβ, and IFNγ have all been associated with gliosis, dopamine neuron loss or both (Collins et al., 2012; Gopinath et al., 2021; Kang et al., 2021; Karpenko et al., 2018; Marogianni et al., 2020). Both CSF and blood of PD patients show increased levels of TNF, IL-1β and IL-6, with increased TGFβ largely restricted to CSF(Imamura et al., 2003; Marogianni et al., 2020). Most compelling are the data indicating that inhibition of cytokine signaling, particularly TNF, in PD models resulted in partial rescue of dopamine neurons, suggesting cytokine signaling originating in the CNS may play a role in PD pathophysiology and etiology(Elyaman and Khoury, 2017; McCoy et al., 2006; McCoy and Tansey, 2008; Mirza et al., 2000). Epidemiological data also support the role of TNF in PD pathogenesis; specifically, individuals with inflammatory bowel disease (IBD) have higher risk for late-onset PD unless treated with anti-TNF biologics (Peter et al., 2018).

Strong evidence in the literature, both experimental and clinical studies, support the notion that TNF is increased in both CNS and periphery in PD(McCoy et al., 2006; McCoy and Tansey, 2008; Mogi et al., 1994; Nagatsu et al., 2000; Qin et al., 2016). While many cells express TNF-receptors, immune cells in both CNS and periphery are the primary source for TNF(Aggarwal, 2003; Bazzoni and Beutler, 1996; Kalliolias and Ivashkiv, 2016; Kraft et al., 2009; MacEwan, 2002; Veroni et al., 2010). Myeloid derived cells such as peripheral monocytes and CNS-resident microglia are particularly sensitive to inflammatory stimuli and secrete TNF upon immune stimulation(Eissner et al., 2000; Gregersen et al., 2000; Hanisch, 2002; Kraft et al., 2009; Segueni et al., 2016). In the CNS, during inflammation, microglia are the main source for TNF(Gregersen et al., 2000; Welser-Alves and Milner, 2013). Whereas in the periphery, monocytes, and other myeloid cells as well as endothelial tissue produce TNF under inflammatory conditions(Gane et al., 2016; Imaizumi et al., 2000; Krishnaswamy et al., 1999; Mandal et al., 2020; Schröder et al., 2018). In addition, recent studies have shown that PD patients’ peripheral immune cells are marked by altered expression of tyrosine hydroxylase, the rate limiting enzyme in dopamine synthesis typically studied in dopamine neurons, with a direct link to increased TNF, suggesting that CNS-originating cytokines such as TNF may influence peripheral immunophenotype and modulate the inflammatory immune landscape in PD(Gopinath et al., 2021).

PD postmortem studies show increased expression of reactive oxygen species (ROS) from mitochondrial stress and inducible nitric oxide synthase, which contribute to increased ROS and can damage dopamine neurons(Anzalone et al.). The significance of oxidative stress is supported by many preclinical and clinical studies revealing the impact of mitochondrial ROS on dopamine neuron degeneration(Dias et al., 2013; Ekstrand and Galter, 2009; Good et al., 2011; Hao et al., 2010; Meredith and Rademacher, 2011; Morais et al., 2009; Simola et al., 2007; Trist et al., 2019; Weng et al., 2018). ROS is considered fundamental to the degenerative process, where the substantia nigra (SN) of PD patients exhibit mitochondrial dysfunction, disturbed calcium homeostasis, and reduced antioxidant molecules(James et al., 2015; Reeve et al., 2014; Trist et al., 2019; Venkateshappa et al., 2012). PD is associated with a significant increase in iron in the degenerating SN dopamine neurons that is measurable in living PD patients and in post-mortem PD brain. Dopamine neurons experience high baseline levels of oxidative stress due to their canonical pace-making activity; dopamine release underlie the higher energy demands of dopamine neurons than other neurons(Dagra et al., 2021; Dias et al., 2013; Lin et al., 2021; Pissadaki and Bolam, 2013). Therefore, dopamine neurons are uniquely vulnerable to ROS-mediated damage(Surmeier et al., 2012). Mitochondria are a potent source of ROS at steady state and during degeneration of dopamine neurons(Brand et al., 2004). During PD progression, extracellular ROS produced by glia may be an additional source of ROS. In healthy states, mitochondrial enzymes counterbalance excess ROS production, while mutations in ROS-scavenging mitochondrial enzymes confer a PD phenotype in PD animal models and in PD patients. Induced radical species are expressed on demand in myeloid cells in all tissue compartments in both CNS and periphery(Anzalone et al.; Brigelius-Flohé and Maiorino, 2013; Karnati et al., 2013; Kawamata and Manfredi, 2010; Ruszkiewicz and Albrecht, 2015). During systemic inflammation, both microglia and peripheral myeloid cells are abundant sources of oxygen radicals(Kraft and Harry, 2011; Kumar et al., 2014; Taylor et al., 2003). In animal models of PD, including 6-hydroxydopamine (6-OHDA) and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), microglial oxygen radicals accelerate the toxin-mediated dopamine neuron lesion(Gao et al., 2003a; Gao et al., 2003b; Zhang et al., 2005). Though microglial-originating radical species are unlikely to reach peripheral circulation in sufficient concentrations to influence peripheral immune cells, cytokines produced by activated microglia may enter peripheral circulation via glymphatic drainage and/or subarachnoid granulations and alter the inflammatory state of peripheral myeloid cells(Norden and Godbout, 2013). Since cytokines are BBB permeable, peripheral cytokines can influence the CNS neuroinflammatory landscape in PD(Norden and Godbout, 2013). This bidirectional mechanism may be central to a feed-forward CNS-to-periphery loop of runaway systemic inflammation and by extension, neurodegeneration(Gopinath, 2022). Microglial and peripheral immune responses to inflammatory challenges are also exaggerated in aged subjects compared to healthy adult subjects, aligning with the overarching hypothesis in the PD field that the disease results, at least in part, from the intersection of age and inflammation in brain and periphery (Buchanan et al., 2008; Rosczyk et al., 2008; Sierra et al., 2007). We therefore speculate that a combination of oxidative stress, lysosomal and/or mitochondrial dysfunction in combination with an inflammatory environment related to aging, lifestyle choices, etc., create the “perfect storm” for PD pathogenesis and progression (Tansey et al., 2022).

Dopamine transporter as a nexus for CNS-periphery inflammation in PD

Recent studies show dopamine, and the associated synthesis, storage and clearance mechanisms are dysregulated in PD(Gopinath et al., 2021; Gopinath, 2022; Mackie et al., 2022). Over the last three decades, PD diagnosis has relied on the loss of dopamine transporter (DAT) in the brains of PD patients(Brooks, 2010; Haavik and Toska, 1998; Murtomäki et al., 2022; Nagatsu et al., 2019; Nutt et al., 2004; Tabrez et al., 2012). Beyond serving as a PD marker, in general, DAT and vesicular monoamine transporter 2 (VMAT2) in the CNS collectively regulate dopamine homeostasis. Free extracellular dopamine quickly oxidizes and forms dopamine radicals such as quinolones, damaging dopamine neurons(Anderson et al., 2011; Miller et al., 1996; Terland et al., 1997). DAT rapidly clears extracellular dopamine and transports it into the intracellular space for repackaging into synaptic vesicles via VMAT2(Cheng et al., 2015; Pramod et al., 2013; Vaughan and Foster, 2013). Thus, the clearance and storage mechanisms protect the neurons from the oxidative stress. Peripheral immune cells such as monocytes and macrophages express functional dopaminergic proteins, indicating that in peripheral immune cells dopaminergic proteins modulate the activity of immune cells(Ambrosi et al., 2017; Cosentino et al., 2002; Gaskill et al., 2012; Gopinath et al., 2021; Gopinath, 2022; Kustrimovic et al., 2019; Mackie et al., 2018; Mackie et al., 2022; Matt and Gaskill, 2020; Qiu et al., 2004; Reguzzoni et al., 2002). For example, following TNF-induced inflammation an immune phenotype shift occurs where the number of Tyrosine hydroxylase (TH) expressing monocytes are increased. This is a reversible immune phenotype. Selective soluble TNF blockade, via XPro-1595, reverses the immuno-phenotype to control levels(Barnum et al., 2014; Gopinath et al., 2021; Karamita et al., 2017; MacPherson et al., 2017). Since TNF is increased in the CNS, CSF, and blood of PD patients, these data have significant clinical implication(Barnum et al., 2014; Gopinath et al., 2021; Karamita et al., 2017; MacPherson et al., 2017). Importantly, DAT in human myeloid cells also exhibits an immunosuppressive role(Mackie et al., 2022). In the presence of the inflammogen LPS, DAT works in reverse transport mode, transporting dopamine out of monocytes leading to attenuation of LPS-induced responses(Goodwin et al., 2009; Khoshbouei et al., 2003; Mackie et al., 2018; Mackie et al., 2022; Saha et al., 2014; Sambo et al., 2017). Subsequent studies provided a paradigm shift on the concept of CNS-to-periphery neuroimmune communication in PD. These studies revealed that in PD patients, the number of peripheral immune cells expressing DAT is increased. We posit that this phenotype change is triggered by the loss of CNS dopamine neurons, representing a compensatory mechanism to attenuate peripheral inflammation(Gopinath et al., 2020a; Gopinath et al., 2022).

Summary and Conclusions

Our understanding of the role of innate and adaptive immune cell function in brain health and how it goes awry during aging is still in its infancy. However, there is consensus that inflammatory mediators derived from brain-resident sentinels in charge of protecting neurons may be produced early in the disease process to produce neurotrophic factors for vulnerable neuronal populations that are sending signals that they are in trouble and becoming dysfunctional. In this manner, early immune activation in PD may play protective roles, whereas chronic cycles of inflammation resulting from chronically activated microglia and infiltrating immune cells from the periphery could be maladaptive and lead to the demise of selectively vulnerable neuronal populations. The challenge for the field is to a) identify the immune and inflammatory pathways that compromise neuronal health and survival to enable design of innovative and more effective strategies to limit their effects on neuronal health; b) use this knowledge to design biomarker-directed trials using immunomodulatory neuroprotective drugs to monitor target engagement and patient responsiveness to treatment.