Thanks to the blood-brain-barrier (BBB)—a structure made from specialised endothelial cells that prevents the passage of immune cells from the blood into the extracellular fluid of the brain and spinal cord (central nervous system)—the brain enjoys a special immunological privilege. However, research over the past few decades has now established that not only can this barrier be compromised and peripheral immune cells be let in, but the brain has its own immune environment consisting of resident immune cells such as microglia, which release inflammatory mediators such as cytokines and chemokines during an acute immune response. It is when this acute immune response is maintained that chronic neuroinflammation can occur – which has now been linked to several neurodegenerative brain diseases such as Alzheimer’s, Parkinson’s, and multiple sclerosis. This opens the possibility for a reorientation of the search for novel therapies in the direction of immunomodulatory drugs, and the possibility of saving those affected from diseases that can be debilitating, depersonalising, and often fatal.
Acute inflammation: characterising a typical neuroimmune response
It is important to establish that although the term ‘neuroinflammation’ is now associated with chronic and neurotoxic inflammation, this neurotoxicity arises from a prolongation of valuable acute, short-term inflammation that occurs during an immune response.
The cells with the largest role to play during this acute immune response are the microglia. These are the CNS analogue of peripheral innate immune cells, and constitute approximately 5-20% of glial cells (non-neuronal cells) in the CNS  (see Fig. 1. for a summary of glial cells in the CNS). They can be characterised as the sentinels of the neuroimmune system, being the first responders to infection or injury.
Interestingly, based on the signals from cytokines, microglia can be phenotypically ‘polarised’ according to the function they are to perform: one phenotype promotes ‘classical’ proinflammatory activation (this phenotype is named M1), while the other phenotype promotes ‘alternative’ anti-inflammatory and pro-healing activation (M2). Research has found, however, that the term ‘polarised’ is a little misleading, given that M1 and M2 are not in fact binary states; rather, the phenotype of any microglial cell will sit on a spectrum between the two (see Fig. 2.). 
In a typical neuroimmune response, microglia release cytokines such as interferon gamma (IFN𝛾), interleukin 1 beta (IL1β), tumour necrosis factor alpha (TNFα), and reactive oxygen species (ROS) (as seen in Fig. 2), which then cause microglia to become polarised to the M1 state – an instance of autocrine signalling (acting on the cells that produced the signalling molecules) and paracrine signalling (acting on nearby cells). This M1 state allows microglia to perform functions such as antigen presentation and the destruction of intracellular pathogens. This also allows microglia and astrocytes (another type of glial cell, described in Fig. 1.) to upregulate the pro-inflammatory cytokines that cause them to polarise in the first place, some of which can induce apoptotic cell death in neurons via complex intracellular signalling pathways. These cytokines will then go on to polarise more microglia - this self-propagating cyclicity characterises many cases of chronic neuroinflammation, which are detailed with reference to specific diseases below. This proinflammatory state therefore helps to kill pathogens and remove dead or damaged cells - ‘classical’ (M1) activation. Yet after this initial inflammatory response, the M1 state is shifted to the ‘alternative’ M2 state by the upregulation of cytokines such as interleukin 4 (IL-4), interleukin 10 (IL-10), interleukin 13 (IL-13) and transforming growth factor beta (TGFβ) (as seen in Fig. 2). This state promotes anti-inflammatory functions by upregulating various enzymes and proteins that, promote the clearance of debris and the formation of new blood vessels. It is this M2 switch that brings the immune response to a resolution, and thereby characterises it as acute. 
Chronic inflammation: the link to neurodegeneration
Acute neuroinflammation serves the purpose it evolved to perform: protection, damage control, and repair that is immediate and short-lived.
Yet it is when these microglia do not get polarised to M2, and are instead constantly stimulated by inflammatory cytokines that a sustained, chronic inflammatory response occurs. This phenomenon has been found to play a causative role in many brain diseases. In many of these cases, however, it is difficult to separate cause from consequence and to establish a linear progression of events, since disease begets neuroinflammation which in turn begets disease, and so on – resulting in the vicious cycle that characterises all of the diseases described below.
Alzheimer’s disease (AD)
Alzheimer’s disease is a neurodegenerative disease and the most common form of dementia – a global loss of cognitive faculties, particularly memory – that affects 25% of those aged 85 or older . It is characterised by two defining pathological features: amyloid-beta (Aβ) plaques and neurofibrillary tangles (NFTs) . Aβ plaques are extracellular deposits of Aβ, which are derived from amyloid precursor protein (APP), while neurofibrillary tangles are intracellular conglomerations of abnormal tau protein.
It is a multifactorial disease that has been shown to have both genetic and environmental risk factors that vary from high fat intake to head injuries . While none of these risk factors act as the cause of AD, what has been found to be a common feature of all cases studies is chronic neuroinflammation. Thirty years ago, microglia were found to localise to Aβ plaques , and since then, much research has been done on the relationship between AD and neuroinflammation. It has been repeatedly found that microglia are polarised to an M1 phenotype when localised to plaques, performing all the cytokine-releasing functions of classical activation, while occurring fairly early in the disease . Post-mortem AD brains show an abundance of these activated microglia, and an experiment in mice which were transgenic (containing artificially introduced DNA) for an AD mutation of APP found that astrocytes and microglia expressing the proinflammatory cytokines IL-1β, IL-6 and TNF were found to surround Aβ plaques . Ironically, the presence of inflammatory cytokines reduces the ability of microglia to phagocytose Aβ, thus hindering the resolution of the disease . It may even be the case that inflammatory cytokines such as IL-1β directly upregulate the expression of APP, thereby upregulating the production of Aβ  (this has been disputed, however, by studies in which IL-1β signalling was blocked in mice with AD, and Aβ deposits remained unaffected ). This means that neuroinflammation causes the accumulation of these plaques instead of their clearance, which continues to stimulate microglia and keep on releasing the proinflammatory cytokines that inhibit them from destroying the plaques. And so the cycle continues.
Knowledge of the role of neuroinflammation in AD has allowed for certain inflammatory mediators to be targeted for drug development. Cyclooxygenase 2 (COX2) is an inflammatory enzyme that is upregulated in an AD brain, and this strong upregulation is associated with neurotoxic mechanisms such as ischemia (inadequate blood supply) and excitotoxicity (excess glutamate over-stimulating NMDA receptors and causing neuron death). Non-steroidal anti-inflammatory drugs (NSAIDs) that target COX2 have therefore been shown to reduce the risk for AD by reducing microglial activation. NSAIDs may also have another mechanism of decreasing the proinflammatory response: acting as an agonist (or, an activator) for the nuclear transcription factor (protein that controls the rate of transcription from DNA to mRNA) peroxisome proliferator-activated receptor gamma (PPAR𝛾), which, as can be seen in Fig. 2C, upregulates those cytokines that cause microglia to polarise to the M2 state, thereby also causing microglia to act in an anti-inflammatory manner. 
Whilst the longest epidemiological study to date has indeed found that long-term NSAID use (lasting over 5 years) is neuroprotective against AD , clinical trials of NSAIDs for those patients already with AD have found inconclusive and varied results [13, 14, 15], demonstrating the need for further research into curative, as opposed to prophylactic, treatment.
Neuroinflammation in AD is not all bad, however. Whilst IL-1β has been implicated in the upregulation of APP (perhaps falsely, as mentioned above), it also upregulates the expression of tumour necrosis factor-α converting enzyme (TACE), which decreases Aβ production . C3, a key inflammatory protein that may be necessary for plaque clearance, is also activated during neuroinflammation [17, 18]. This ambiguity as to what is ‘good’ or ‘bad’ is emblematic of a greater need to sort through the complexity of interactions involved in neuroinflammation to identify specific protective pathways in the neuroimmune response, and exploit them for future therapies.
Parkinson’s Disease (PD)
Parkinson’s disease (PD) is the second most common neurodegenerative disease after AD. It involves the gradual loss of motor control involving tremor, rigidity, bradykinesia (slowness of movement), and sometimes dementia and depression . Its pathology is characterised by the death of dopaminergic neurons in a region of the midbrain called the substantia nigra pars compacta (SNpc) due to the presence of Lewy bodies - accumulations of the protein α-synuclein. Much like AD, a complex smorgasbord of genetics and environment have been shown to be risk factors (including head injuries and exposure to certain pesticides ), whilst the cause of the disease itself is still unknown. Like with AD again, however, neuroinflammation has been found to be a common hallmark of the disease, with studies of postmortem PD brains showing the presence of inflammatory mediators in the SNpc .
The sequence of events in PD is a little easier to define than in AD. The inflammatory response is initiated by the activation of microglia against α-synuclein aggregates, which triggers the usual release of proinflammatory cytokines including TNF, inducing neuronal death. Yet this neuron death causes the accumulated α-synuclein to be released from within the cell body, again triggering the inflammatory response, leading to a feedback loop in which microglia are constantly polarised to M1 and astrocytes remain reactive (see Fig. 3) . The involvement of the peripheral immune system exacerbates this, with T cells crossing a BBB compromised by these proinflammatory cytokines to release even more proinflammatory cytokines such as IFN𝛾 and TNFα (see Fig. 3) . However, microglial cells are also able to phagocytose extracellular α-synuclein due to the presence of toll-like receptor 4 (TLR4) on their cell surface, the ablation of which was shown to ‘augment motor disability’ and resulted in increased TNF levels . Like in AD, this is yet another example of the inflammatory response being a complex mixture of beneficial and detrimental pathways, and the need to isolate those which are useful and harness them for therapeutic use.
Multiple Sclerosis (MS)
MS is known as the “neurological disease of young people,” being most commonly diagnosed in people in their 20s and 30s . It is a neurodegenerative disease characterised by the destruction of the myelin sheath surrounding the axons of neurons (demyelination), which then causes inflammation and disseminating lesions that are visible on MRI scans. This is most likely due to a protein associated with myelin becoming an autoimmunogen (a target of the immune system that is not foreign, but ‘self’) . The symptoms expressed as a result of this depend on the regions of the brain in which these lesions form, and can include problems with vision, movement, and balance .
Historically, MS has always been perceived as an autoimmune disease, initiated by T cells crossing a compromised BBB from the peripheral immune system into the CNS, and, alongside macrophages and B cells, targeting and destroying myelin, causing degeneration that further fuels the immune response. However, this ‘outside-in’ theory as to the primary stimulus for MS has been challenged with its antithesis: the ‘inside-out’ theory (see Fig. 4.). This proposes that the initial stimulus comes from a yet unknown cytodegeneration within the CNS itself, perhaps one that targets the oligodendrocytes that form the myelin sheath (see Fig. 1.), leading to the release of antigenic debris, which the immune system then reacts to, causing further degeneration, and so on. This paints a picture of the pathology of MS that is much like that of AD and PD, in that it is not the immune system that is defective, but a fault within the CNS itself (in AD, this comes in the form of Aβ plaques, and in PD, this comes in the form of α-synuclein aggregates). 
In any case, the immune system becomes activated, and T cells such as Th1 and Th17 cells become potent exacerbators of pathology due to the IFN𝛾 cytokine they release. As mentioned above, this polarises microglia to the M1 state (see Fig. 2.), and the consequential neuroinflammation produces cytokines that are neurotoxic to the oligodendrocyte-myelin complex, and inhibit oligodendrocyte precursor cell (cells that differentiate into oligodendrocytes) proliferation and maturation . This exposes neurons to damage by, reactive oxygen species (ROS) , leading to further degeneration.
Yet again, there have been found to be beneficial sides to neuroinflammation in MS. For example, astrocytes can infiltrate T cells clusters to induce apoptosis, and are also responsible for producing neurotrophic factors, which reduce degeneration and stimulate regeneration . Counter-intuitively, remyelination (a process involving the differentiation of oligodendrocyte precursor cells into oligodendrocytes, which then synthesise new myelin) is also impaired after the depletion of TNFα and IL-1β (two proinflammatory cytokines) acting on oligodendrocyte precursor cells . This may indicate that the proinflammatory cytokines released by microglia actually stimulate regeneration, although not enough regeneration to overcome the extent of neuroinflammation and neurodegeneration. Isolating these pathways will prove to be vital when considering a cure.
What all three of the above neurodegenerative diseases have in common is chronic inflammation and stimulation of microglia, so that they remain in the classically activated M1 state. Research into M2-inducing factors may therefore provide therapies that could potentially be applied to all three, and perhaps even more neuroinflammation-associated neurodegenerative diseases.
The anti-inflammatory cytokine IL-4, for example, has been tested in animal models of MS, with overall success. A group using an MS animal model called experimental autoimmune encephalitis (EAE) found that transduction (transfer of genetic material) with a viral vector expressing IL-4 reduced the symptoms of EAE [34, 35], most likely due to the induction of M2 microglia, which produce neurotrophic mediators such as activin-A  that support oligodendrocyte differentiation, remyelination and regeneration.
In AD, the induction of M2 microglia is also significant, because while M1 microglia cannot clear Aβ, M2 microglia can. Treatment with IL-4 can block inhibition of Aβ phagocytosis , lowering the pH of the phagosome (the vacuole into which the Aβ is engulfed) to allow for better degradation of Aβ. Injections of only 100 ng of IL-4 decreased Aβ levels in a few days, correlating with an increase in pro-phagocytic and degradative enzymes . Another group showed the longer-term results of sustained IL-4 expression using a viral vector, which yielded a reduction in gliosis, decreased Aβ, and improved spatial memory .
IL-4 treatment could be administered indirectly in the form of an already FDA-approved drug for relapsing-remitting MS called glatiramer acetate (GA), which works by shifting the Th1 state of T cells to Th2, and inducing production of IL-4. The latter is probably essential to its function, since the phenotype of microglia after treatment with GA was similar to the phenotype of microglia after treatment with IL-4 .
To the knowledge of the author, IL-4 studies have not been performed in PD patients, but may perhaps yield similar positive results. An M2 inducer that has been tested in PD is the PPAR𝛾 agonist (also the mechanism by which NSAIDs may work, as detailed above), a type of drug which was found to be neuroprotective in animal PD models, and works by upregulating anti-inflammatory cytokines that induce an M2 microglial state . In MS, PPAR𝛾 activation has also been demonstrated to decrease T cell proliferation and increase T cell apoptosis , thereby generally decreasing the autoimmune response. Larger scale trials are currently underway to assess the therapeutic potential of PPAR𝛾 agonists in the treatment of MS.
Examples were provided for AD, PD, and MS as to the ways in which classical M1 activation is not always deleterious, and, in fact, may involve pathways that are essential for neuroprotection or regeneration, or mediators that bring about both negative and positive consequences. In the same way, M2 microglia are not always beneficial – one study involving IL-4 injections into AD mouse models found that amyloid pathology was in fact exacerbated . Indeed, while the usual functions of M2 microglia are associated with wound repair and regeneration, some studies have shown that they can trigger inflammation  and recruit macrophages . Much like how M1 and M2 states are not as binary as they seem, the effects of the M1 and M2 states are similarly complex and non-polarisable.
Fundamental similarities have been drawn between many different diseases beyond those described here. Neuroinflammation can play a primary or secondary role in Huntington’s, ALS, stroke, traumatic brain injury, and spinal cord injury; this role can be characterised by shifts between a reparative, acute inflammatory response to one that is chronic, exacerbating the vicious cycle of inflammation and degeneration. The three examples simply put these principles into context, using some of the most recent research to demonstrate how these seemingly isolated diseases share a universal milieu. It is this universality that allows one treatment – whether that be a type of drug such as PPAR𝛾 agonists or a specific cytokine such as IL-4 – to target the fundamental problem: the inability of microglia to switch to the anti-inflammatory M2 state.
Nevertheless, this issue becomes so much more complex when one realises that the labels M1 and M2 are simply labels. These are labels which are purely abstract, and generate the misconception that microglia can either be one way or another, good or bad, when they really describe complex, intertwined processes that can sometimes go against our expectations. Attempts to characterise, while they generally apply, ultimately fail. Neuroinflammation and its mediators are neither friend nor foe – rather, they are completely indifferent. It is only by closely examining the specific pathways that are productive or counterproductive that researchers can isolate them and begin to bring them together, in what could one day be a one-pill-solves-all for degenerative disease.
This author has not yet uploaded a bio.