A nonsteroidal antiinflammatory drug (NSAID) with nitric oxide-releasing capabilities dramatically reduced amyloid burden in a transgenic mouse model, according to a report in today's Journal of Neuroscience. More surprisingly, it did so while increasing the activity of microglia.

This study adds fuel to several debates, the first being whether NSAIDs should be used to help fight Alzheimer's disease. There is mounting evidence that NSAID use lowers the risk of AD, but that these drugs also can cause gastrointestinal bleeding and ulceration. Nitric oxide (NO)-donating NSAIDs have been developed in an attempt to circumvent this problem.

Marcia Gordon, Paul Jantzen, and colleagues at University of South Florida and elsewhere found that NCX-2216, an NO-releasing derivative of the NSAID flurbiprofen, dramatically reduced the heavy Aβ loads and Congo Red staining found in transgenic mice expressing both Swedish mutant AβPP and mutant presenilin-1. In contrast, ibuprofen produced only modest Aβ reduction and no reduction in staining, the latter result suggesting that ibuprofen only reduced non-fibrillar Aβ deposits but not plaques. A cyclooxygenase-2-selective NSAID reduced neither of these markers.

The study also speaks to the debate on whether the chronic inflammatory response seen in later stages of AD it a good or a bad thing. One side argues that microglia release reactive oxygen and nitrogen species, and other proinflammatory molecules that harm cells. The other side points to evidence that activated microglia can scavenge and clear potentially toxic Aβ deposits.

Jantzen et al. found that the NO-releasing NSAID dramatically increased the activation of microglia in the doubly transgenic mice. In normal mice, however, the drug did not generate activated microglia. Conversely, the same NSAID lacking the NO-releasing function did not activate microglia in the amyloid-burdened transgenic mice. "These data are consistent with the argument that activated microglia can clear Aβ deposits. We conclude that the NO-generating component of NCX-2216 confers biological actions that go beyond those of typical NSAIDs," the authors write.—Hakon Heimer

See related Live Discussion

Q & A with co-author David Morgan, University of South Florida, Tampa:
Q: What is the status of the NO-donating NSAID used in this trial? Is it used in humans, experimental, or slated for clinical trials anywhere?
A: The drug is presently not being used in humans. A closely related compound, NO-naproxen is in clinical trials by AstraZeneca and Nicox for arthritis. Certainly these data encourage further investigation of its potential use in humans.

Q: Do you have any idea what sort of NO-linked mechanism might be at work?
A: It's too early to speculate. We need to identify which components of the molecule are responsible for which actions. We first thought it unlikely the NO at such a low molarity would have much effect on brain, but our data thus far suggest it is responsible for the microglial-activating effects.

Q: The paper appears to suggest that, initially, microglial activation is good, but becomes detrimental when chronic. Given that the Elan trial of an AD vaccine was aborted due to cerebral inflammation, is this question testable in humans?
A: There are many scenarios where activation of microglia is thought to be beneficial. The facial nerve transection model is perhaps best studied in this regard. However, clearly when associated with an inflammatory reaction, microglial activation is believed detrimental (Patrick McGeer refers to this action as "autotoxicity".) My guess is the Aristotelian mean applies here as to almost everything else; a little is good (or can be), too much is bad. The inverted U shaped function is hard to avoid in biology. The question is how to titrate microglia activation.

I think it best to await the longer-term outcomes in the Elan vaccine trial before concluding what the problems are. I have yet to read precisely what the conditions exhibited by the patients were, and have heard rumors ranging from latent herpes virus eruptions to multiple sclerosis caused by the adjuvant, not the Aβ vaccine. Elan has released two somewhat cryptic descriptions. I have every confidence they will provide us with more details once they have fully characterized the problem (reported so far in less than five percent of the patients).

I believe work from other groups and our own lab demonstrating exaggerated astrocyte and microglia activation in aged brain following injury, implies the aged brain is in a proinflammatory state. Bob Floyd has taken to calling this "smoldering inflammation." Hence, I think any agent that might provoke the eruption of a CNS conflagration needs to be evaluated carefully, and prior knowledge of counteragents demonstrated to quell such activation should be available, before administration to humans.


  1. This paper presents data suggesting that chronic use of a NO-releasing NSAID attenuates Aβ plaque load in an animal model of Aβ neuropathology. Additionally, the data indicate that the efficacy of this NO-releasing NSAID was mediated by microglia activation. Based on this observation, the authors suggested that NO-releasing NSAIDs could be used for treatment of Alzheimer's disease.

    It is extremely interesting to again see that microglia appear to be involved in amyloid plaque clearance. The observation is of high scientific interest. However, while beneficial to amyloid scavenging, NO has been universally involved in promoting neuronal death (Irvani et al. 2002). Chronic exposure to elevated NO levels may provoke unwanted effects.

    Considering the fiasco associated with rushing of the vaccination clinical trial by Elan Pharmaceuticals (based on the similar mechanism of scavenging amyloid through stimulating microglia activities), it is necessary for a critical analysis of any treatment before we make the transition to human trials. The chronic use of NO-releasing drugs may pose serious hazards even among current users, especially within an aging population at high risk for multiple neurodegenerative disorders. Further, given that treatment with certain NSAIDs is currently the leading cause of iatrogenic illness, we need to proceed cautiously and diligently. We strongly urge the systematic and careful monitoring of current chronic users of NO-releasing drugs for signs of neurological and possibly gastro-intestinal problems.


    . Involvement of inducible nitric oxide synthase in inflammation-induced dopaminergic neurodegeneration. Neuroscience. 2002;110(1):49-58. PubMed.

  2. Jantzen and colleagues have carried out a thorough and provocative study on the comparative effects of NSAIDs in reducing the amyloid burden in transgenic mice. The selective COX-2 inhibitor (Celecoxib) was ineffective, the mixed inhibitor (ibuprofen) was modestly effective, while the NO-NSAID NCX-2216 was highly effective. NO-NSAIDs have great clinical promise as COX-inhibiting agents because they retain the antiinflammatory properties of their parent NSAIDs while having greatly reduced gastrointestinal (GI) toxicity (Wallace et al., 1994). The GI tract is innervated by a system of cholinergic and nitric oxide neurons, which apparently govern the propulsion of food (Aimi et al. 1993). The NO-NSAIDs evidently exert their beneficial effects by stimulating nitric oxide receptors. NCX-2216 is an unusual member of the NO-NSAID family in that ferulic acid, an antioxidant primarily employed as a food preservative, is used as the linker to flurbiprofen.

    High doses of NSAIDs were administered in the Jantzen et al. study, being 4-10 times above therapeutic doses in humans. The most encouraging aspect of the study was that NCX-2216 had so little GI toxicity under these circumstances. It should therefore be an excellent candidate for a clinical trial in Alzheimer's disease.

    The appearance of more MHC-2 positive microglia in NCX-2216 treated transgenic mice compared with non-treated transgenic mice was unexpected, but this may not be paradoxical. COX inhibitors are known to reduce neurotoxicity of stimulated microglia in vitro (Klegeris et al., 1999), but there are no data to indicate that such treatment reduces their phagocytic capability. This needs to be explored. The data of Jantzen et al. indicates that NCX-2216 actually enhances such phagocytic capability, an aspect that also deserves further exploration. It needs to be determined whether this is a special property of NCX-2216 or a general property of NO-NSAIDs. It needs further to be determined whether there are different effects of NO-releasing NSAIDs on human microglia, which have low iNOS activity, compared with mouse microglia, which have abundant iNOS activity (Walker et al., 1995). Finally, it needs to be determined how much of the intact molecule reaches the brain at the doses used in the Jantzen et al. study compared to the lower doses that would be used in humans. This is because esterases carry out substantial cleavage of NO-NSAIDs in peripheral organs and, under most circumstances, only the parent NSAID will reach the brain.

    David Morgan is absolutely correct in his comments that microglial activation is a beneficial phenomenon up to the level where host cells are damaged. He poses the critical question: how to titrate microglial activation? The titration level appears to be very different in transgenic mouse models than in Alzheimer's disease. β transgenic mice generate huge overloads of human β, a foreign protein to mice. But in humans, β is a normal metabolite of βPP, which is produced in peripheral organs as well as brain (See ARF news item). It accumulates late in life only in those brain areas that have a poor clearance capacity. But when it consolidates, β vigorously activates complement because of its affinity for human C1q (Rogers et al., 1992). The microglial titration level is high. Rodent C1q binds poorly to human β (Webster et al., 1999). The microglial titration level is therefore low. As a result, vaccination of transgenic mice with β can be a useful strategy. Antibodies against the locally produced foreign protein can enhance complement-mediated phagocytosis. The same strategy should be counterproductive in Alzheimer's. One reason is that overactive complement activation already exists in Alzheimer's, leading to autodestruction by the membrane attack complex (Webster et al., 1999). The titration level is already too high. Another reason is that a general autoimmune disorder may be induced in humans because antibodies can attack the multiple sites in the body where β is produced.

    Transgenic mouse models expressing β are, at best, a partial model of Alzheimer's disease. The limitations of such models must be clearly understood for useful therapeutic strategies in Alzheimer's disease to be developed.


    . Novel nonsteroidal anti-inflammatory drug derivatives with markedly reduced ulcerogenic properties in the rat. Gastroenterology. 1994 Jul;107(1):173-9. PubMed.

    . Histochemical localization of nitric oxide synthase in rat enteric nervous system. Neuroscience. 1993 Mar;53(2):553-60. PubMed.

    . Toxicity of human THP-1 monocytic cells towards neuron-like cells is reduced by non-steroidal anti-inflammatory drugs (NSAIDs). Neuropharmacology. 1999 Jul;38(7):1017-25. PubMed.

    . Complement and cytokine gene expression in cultured microglial derived from postmortem human brains. J Neurosci Res. 1995 Mar 1;40(4):478-93. PubMed.

    . Complement activation by beta-amyloid in Alzheimer disease. Proc Natl Acad Sci U S A. 1992 Nov 1;89(21):10016-20. PubMed.

    . The mouse C1q A-chain sequence alters beta-amyloid-induced complement activation. Neurobiol Aging. 1999 May-Jun;20(3):297-304. PubMed.

  3. In humans, a substantial body of evidence suggests NSAIDs confer some protection from AD. Importantly, these clinical studies on NSAIDs shed no light on whether NSAIDS could have therapeutic efficacy in patients who already have AD. In addition, it is generally accepted that chronic inflammation is a feature of end-stage AD, and that this chronic inflammatory response contributes to the disease process. Recently, together with Eddie Koo and colleagues, we have shown that some NSAIDs can directly reduce production of the longer, more pathogenic Ab peptides (see ARF news item). Our findings are consistent with a previous report (Lim et al., 2000) and this current report showing that long-term treatment of animals with ibuprofen reduces Aβ load in βPP-transgenic mice. Based on this evidence, it is possible that NSAIDs could confer protection from AD by directly reducing production of Aβ42, acting as classical anti-inflammatory agents by directly inhibiting COX, influencing other potential pro-inflammatory targets (e.g NFkB), or a combination of these actions.

    In light of this background, the manuscript by Jantzen et al. raises more questions than it answers. In this case, NC-2216X administration to AbPP mice reduced Aβ load and, surprisingly, activated microglial cells. NC2216X was designed as an NSAID that would reduce gastric injury by release of NO. Thus, this study postulates that release of NO by nitro-flurbiprofen leads to activation of microglia, and this increased activation somehow results in increased Ab clearance. Such a scenario is plausible, but additional studies are needed to determine how NO-flurbiprofen or any other NSAID- like agent is acting. Such mechanistic insight will be essential for future therapeutic development.

    One of the major difficulties I have with the NO-releasing hypothesis in this case is that there is no evidence that intact NC2216X gets into the brain. My understanding (which may be incorrect) is that, in the acidic environment of the gut, the NO moiety is cleaved from NC2216X (hence its protection of the gut). Because of this concern, it would be useful if the authors follow this initial study up by actually determining what form of the drug gets into the brain. In this regard, we have data that shows that flurbiprofen can lower Aβ42 in AβPP mouse brain. It is therefore possible that NC2216X, or a metabolite thereof, can reduce Aβ by inhibiting Aβ42 production directly.

    Of note it is worth mentioning that most NSAIDs, although relatively safe, are "dirty drugs" with many targets besides the classically recognized cyclooxygenase enzymes. Moreover, many NSAIDs are extensively metabolized in the body, and it is conceivable that these metabolites might have unexpected actions. Epidemiologic studies on NSAIDs conferring protection from colon cancer have led to a similar debate regarding how NSAIDs confer protection from cancer. As with the studies on NSAIDs and AD, it appears that NSAIDs may confer protection against cancer through multiple mechanisms.

    In any case, NSAIDS and NSAID derivatives remain a very interesting class of potential therapeutic agents for AD prevention. Whether they will work in a therapeutic setting remains much less certain. Understanding the mechanism(s) by which they confer protection will be essential. This study and other recent studies on NSAIDs and other antiinflammatory compounds (see ARF news item on curcumin) serve to reinforce the notion that in order to realize the true potential of these agents, further studies dissecting contributions of potential mechanisms of action are needed.


    . Ibuprofen suppresses plaque pathology and inflammation in a mouse model for Alzheimer's disease. J Neurosci. 2000 Aug 1;20(15):5709-14. PubMed.

  4. This paper is an excellent follow up to the paper by Weggen and colleagues showing that a subset of NSAIDs lower Abeta42. This current paper provides the in vivo correlate of what Weggen and colleagues observed, and suggests that despite disappointing clinical studies with NSAIDs, particular NSAIDs might be truly beneficial in AD.


    . A subset of NSAIDs lower amyloidogenic Abeta42 independently of cyclooxygenase activity. Nature. 2001 Nov 8;414(6860):212-6. PubMed.

    View all comments by Benjamin Wolozin

Make a Comment

To make a comment you must login or register.


Other Citations

  1. related Live Discussion

Further Reading

Primary Papers

  1. . Microglial activation and beta -amyloid deposit reduction caused by a nitric oxide-releasing nonsteroidal anti-inflammatory drug in amyloid precursor protein plus presenilin-1 transgenic mice. J Neurosci. 2002 Mar 15;22(6):2246-54. PubMed.