Twenty years ago it was recognized that inflammatory mechanisms play a pathogenic role in multiple sclerosis, but for other disorders the brain was generally considered to be "immunologically privileged." This view began to change with the discovery that microglia in the Alzheimer's disease (AD) cortex were activated (e.g., immunoreactive for the major histocompatibility complex),1,2 similar to peripheral macrophages at sites of inflammation. We now know that cytokines, complement, reactive oxygen and nitrogen species, and other markers of inflammation are highly upregulated in a wide range of neurodegenerative disorders, particularly AD.3 This should not be surprising because, by definition, tissue damage occurs in neurode-generative disorders, and inflammation is a normal response to tissue damage wherever it occurs in the body. In addition, the chronic presence of highly inert, highly insoluble deposits is another classic inflammatory stimulus, and such deposits are present by the millions in the AD brain as amyloid P peptide (AP) plaques and neu-rofibrillary tangles.
Parkinson's disease (PD) also commonly features highly insoluble, highly inert deposits, the Lewy body,4,5 and prion plaques have been reported to stimulate inflammation.6-8 Given so many potential causes of brain inflammation in neurodegenerative disorders, the only real surprise is how long it took to recognize them!
In addition to the fact that the brain is not immunologically privileged, there are three other fundamental concepts that need to be understood before one can begin to make sense of the thousands of AD, PD, and other neuroscience papers that have been written about inflammation and neurodegenerative disorders. The first is that inflammatory mechanisms are highly interactive. The classical and alternative complement pathways include over 40 proteins and regulators, all of which depend on each other and can have a multitude of effects, including the induction of cytokines, chemokines, and inflammation-related growth factors. In turn, cytokines such as interleukin-1 (IL-1), IL-6, and tumor necrosis factor-a (TNF-a) induce each other, induce chemokines, induce growth factors, and induce complement. It is therefore important to examine carefully any paper that claims to have discovered the primary, all-important inflammatory mechanism in AD or some other neurological disease. Inflammation seldom, if ever, works that way.
Another fundamental concept is that inflammation includes both destructive and beneficial mechanisms. To fail to appreciate this fact can have serious consequences. For example, a recent clinical trial attempted to enhance inflammation in the AD brain by immunizing against A^. This did have the desired beneficial effect of promoting A^ clearance, but it also injured several patients and killed another, presumably through additional inflammatory mechanisms that were simultaneously invoked.9,10
A third and final fundamental concept of inflammation in neurodegenerative disease is that inflammation is unlikely to be an etiology, a root initial cause, of any brain disorder. When we speak of the inflammatory response, we do so for a reason: inflammation almost always arises as a response to some other, more primary pathogen. This does not mean, however, that inflammation is unimportant. Especially in the brain, inflammation can often cause more damage than the agent that engendered it, as is typically the case in head trauma and encephalitis. Wounds and infections of the brain are etiologies, but the secondary inflammatory responses to them are often what kill the patient.
A mature view of inflammation in neurodegenerative disorders therefore requires that we recognize four simple but frequently unaddressed concerns: (1) that inflammatory mechanisms are likely to occur in any neurodegenerative disorder, (2) that these mechanisms are highly complex and interactive, (3) that they have both beneficial and destructive consequences, and (4) that they likely will require as much clinical attention as the etiology that gave rise to them. With this fundamental background, we can then begin to piece together the many basic research and clinical findings about inflammation in such diseases as AD and PD, and we can begin to develop rational therapies.
Most of the basic research on AD inflammation has centered on A^ deposits and microglia. This follows from the initial discovery that microglia cluster within A^ deposits and are activated there.11-13 By going back and forth from in situ tissue sections of the postmortem AD brain to in vitro assays, much knowledge about these processes has been gained. For example, it is now known that microglia possess receptors that are responsive to A^, including the macrophage scavenger receptor,14 the receptor for advanced glycation endproducts (RAGE)15 and the formyl peptide receptor.1618 A^ activation of the formyl peptide receptor, among other things, stimulates microglial secretion of IL-1,17 and A^ activation of RAGE, among other things, stimulates microglial secretion of M-CSF.19 IL-1, and M-CSF, in turn, promotes the expression of many other inflammatory mediators through such transcription factors as NF-kB20 and C/EBP.21 C/EBP is, in fact, upregulated in the AD cortex.22 There are many other A^-receptor interactions, as well as the secretory products and transcription factors they induce, that activate microglia and account for the colocaliza-tion of activated microglia with A^ deposits.
Microglia and inflammation have also been implicated in PD pathogenesis.23-27 Studies by Herrera and colleagues,28,29 among others,30-32 demonstrate that administration of the classic inflammatory stimulus lipopolysaccharide (LPS) causes a selective loss of dopamine in the substantia nigra, accompanied by massive activation of microglia, which cluster around deteriorating dopamine neurons.
These data have resonance with a number of other findings:
1. LPS is a common bacterial toxin, and bacterial exposure in utero may be a factor predisposing to PD.33
2. Many of the putative etiologies of PD, including head trauma,3435 influenza with encephalitis,36,37 and environmental toxins such as pesticides and MPTP,38-40 all have in common their ability to stimulate inflammation.
3. Substantia nigra neurons appear exquisitely sensitive to oxidative and nitric oxide insult,41-43 and both are major weapons for inflammatory attack.
4. The substantia nigra is reported to have the highest density of microglia of any brain structure.44,45
Complement proteins are among the many inflammatory molecules produced by microglia.46,47 In AD, both the classical48-50 and the alternative51 pathways are activated in an antibody-independent fashion by A^ and by neurofibrillary tangles52 through mechanisms similar to those for coat proteins of RNA tumor viruses. This produces complement anaphylatoxins that further fuel the inflammatory reaction, as well as complement opsonins that target A^ for microglial phagocytosis.49 The opsonins appear to play a major role in A^ clearance, as inhibition of complement opsonin actions in transgenic mouse models of AD results in significantly increased A^ deposition.53
Antibodies to A^ can also opsonize A^ for phagocytosis by microglia through microglial Fc receptors, as shown by passive and active A^ immunization in human A^-overexpressing transgenic mice.5455 In addition to helping clear A^, C1q binding to A^ potently aggregates it into the form seen in parenchymal senile plaques.56-59 The terminal component of both classical and alternative complement pathway activation is the membrane attack complex (MAC), which forms on and lyses targeted and, sometimes, "innocent bystander" cells.60 MAC fixation on neurite processes in the vicinity of A^ deposits has been demonstrated in the AD brain.50 Lastly, AD deficits in the critical complement regulator CD59 (membrane inhibitor of reactive lysis) may be permissive for complement damage.61
Cytokines are also abundantly increased in pathologically vulnerable regions of the AD brain.362 They can be produced by microglia and astrocytes and, similar to complement, have multiple actions. Of particular importance, IL-1 appears to stimulate production of the A^ precursor protein.63 Moreover, the classic proinflammatory triad of IL-1, IL-6, and TNF-a, along with other cytokines and growth factors, may stimulate apoptosis and angiogenesis.3
Many of these processes can be modeled in neuron and glial cultures derived from rapid autopsies of AD and control patients.2 As in the AD cortex, astrocytes in culture take up periplaque positions. Neurons in culture die when seeded on synthetic A^ deposits, whereas neurons seeded outside the deposits extend neurites that either skirt the A^ or are retracted. These behaviors are also seen in AD tissue sections. Microglia, as in the AD brain, cluster around and within A^ deposits, where they increase their expression of a wide range of inflammatory mediators, including complement, cytokines, chemokines, growth factors, and nitric oxide.2,15,19,46,64,65
We have used in vitro culture models to investigate mechanisms of A^ immunization and the possibility that certain nonsteroidal anti-inflammatory drugs (NSAIDs) might be a useful therapeutic adjunct to A^ immunization. Indomethacin, in particular, is known to inhibit secretion of cytotoxic inflammatory factors that may have caused adverse reactions in A^-immunized patients. In addition to these potentially useful inflammation-inhibiting properties, our studies demonstrate that indometha-cin does not materially impede microglial chemotaxis to or phagocytosis of A^.66
It is generally the case that chronic inflammation exacerbates disease. Nonetheless, the pathogenic relevance of inflammation in AD and PD continues to be an issue. The evidence for a pathogenic role of inflammation in these disorders essentially rests on the following grounds:
1. There is selective upregulation of inflammatory mechanisms in pathologically vulnerable but not pathologically spared regions of the AD and PD brain.3 6768
2. The classical hallmarks of AD, A^ deposits and neurofibrillary tangles, have in common a unique ability to directly interact with and stimulate inflammatory mechanisms.3,48,49,63,67,69
3. Inflammation is a common denominator of many putative etiologies of PD (e.g., head trauma, influenza with encephalitis, bacterial infections, environmental toxins such as pesticides and MPTP).23 70
4. Complement fixation and lysis of neurites occur in the AD cortex.50
5. Complement opsonization of deteriorating dopamine neurons occurs in the PD substantia nigra.71
6. Polymorphisms of inflammatory genes may increase susceptibility to AD and PD.72-80
7. Elevated C-reactive protein levels predict dementia onset as much as 20 years later.81
8. A rise of inflammation markers immediately precedes synapse loss and cognitive deterioration in AD.82
9. Basic research studies show protective effects of common antiinflammatory drugs in animal models of PD.83-86
10. Nearly 20 epidemiologic studies suggest that common anti-inflammatory drugs may delay onset of AD.87-89
11. One small pilot trial has shown a beneficial effect of the NSAID indomethacin.90
Set against these findings, two large-scale clinical studies, one with the steroid antiinflammatory drug prednisone,91 and another with the nonsteroid anti-inflammatory drug hydroxyquinoline,92 failed to find a therapeutic benefit in patients with AD. Two critical questions therefore remain: are we using the right anti-inflammatory drugs, and are we using them in the right patients?
The steroid prednisone has a well-known adverse reaction profile that includes confusion and agitation at the clinical level93 and damage to hippocampal neurons at the basic science level.94,95 These are obviously not desirable properties for the treatment of AD patients. Conversely, several NSAIDs that have not yet been tested in AD are reported to have such desirable actions as inhibition of COX I (without inhibition of COX II), inhibition of A^ production, inhibition of lipoxygenase, and PPAR-y facilitation. Indomethacin, which did show therapeutic effects in a limited AD trial,90 is both a potent COX I96 and COX II inhibitor,97 an inhibitor of A^ production,98 a lipoxygenase inhibitor,99 and a PPAR-y agonist.100 It is taken worldwide by millions of arthritis patients. Nonetheless, its gastrointestinal and renal adverse effects, which may be somewhat higher than other NSAIDs,101 have apparently inhibited its further consideration in AD trials.
A second critical consideration in the application of anti-inflammatory drugs for AD is when and to whom they should be given. Anyone who has ever removed the brain of a patient who has suffered AD can attest that the already existing damage is unlikely to be reversed by any treatment now known or that will be known for the next 50 years. Moreover, AD inflammation, as previously noted, is a response to preexisting pathogens, such as A^ and neurofibrillary tangles. Thus, our best hope for treatment is to remove the initiating stimuli or to quench the inflammation they produce before significant damage is done to the brain. Clinical trials using NSAIDs to prevent or delay AD in susceptible but still normal elderly patients are now underway, and should provide answers about their efficacy in the next few years.
Was this article helpful?