Could amyloid-beta fibrils—which form the hallmark plaques of Alzheimer’s disease (AD)—actually protect the brain? That intriguing possibility caught scientists’ attention last year when Lawrence Steinman’s group at Stanford University in Palo Alto, California, reported that peripheral injection of Aβ peptides calmed neuroinflammation and relieved motor symptoms in several mouse models of multiple sclerosis (MS). Now, Steinman and colleagues have identified a sequence of six amino acids within Aβ and other amyloid-forming proteins that is sufficient to mediate the protection. The findings appeared online April 3 in Science Translational Medicine, and made it onto National Public Radio’s Science Friday . While it is premature to extrapolate results in MS mice to human AD, some say the research raises concerns about potential long-term consequences of amyloid-lowering therapies and adds to recent work highlighting the importance of inflammatory processes in AD.

Analyzing MS autopsy specimens several years ago, Steinman and colleagues found an abundance of Aβ, tau, and other amyloid-forming proteins in brain lesions and damaged axons (Han et al., 2008). To explore how Aβ relates to disease, the researchers injected synthetic Aβ40 or Aβ42 peptides into several MS animal models, including mice induced to develop MS-like brain inflammation known as experimental autoimmune encephalomyelitis (EAE). Much to their surprise, the treatment relieved paralysis (ARF related news story on Grant et al., 2012). Moreover, EAE symptoms were worse in mice lacking Aβ’s parent molecule, amyloid precursor protein (APP), and in mice deficient in other amyloidogenic proteins including prion protein (PrP) and tau. All these proteins have non-amyloid functions in neurons.

The amyloid-forming proteins have common structural features known as dry steric zippers, which are thought to be involved in protein-protein binding. Using computational approaches, scientists identified segments as short as six amino acids that can form these structures (Goldschmidt et al., 2010; Thompson et al., 2006; Eisenberg and Jucker, 2012). In the current study, first author Michael Kurnellas and colleagues synthesized hexameric peptides from various amyloid proteins—including Aβ, tau, and PrP—and injected them into EAE mice with early clinical symptoms. “Again, the animals got better,” Steinman said. When they stopped the peptide injections, the animals’ motor problems returned. As seen in the 2012 study, treated mice had reduced serum levels of pro-inflammatory cytokines—most notably interleukin-6 (IL-6), as well as IL-2, transforming growth factor-β (TGF-β), and IL-17. Therapeutic outcome depended on how well the hexamer could bind pro-inflammatory molecules and form fibrils.

Could the findings in MS mice have implications for AD? Some scientists think the diseases differ too much to draw clear parallels. Whereas the EAE mouse is an “acute system where the disease symptoms (paralysis) appear within days and animals respond to the treatment on the same time scale,” AD is “a chronic disorder requiring months to manifest in animal models,” noted Sanjay Pimplikar of Cleveland Clinic in Ohio. “The etiopathogenesis of MS and AD, and the course of the respective diseases, are too different to make a direct valid comparison.” (See full comment below.)

Richard Ransohoff, also of Cleveland Clinic, said the demonstration that Abeta peptides suppress myeloid cell inflammatory responses may be “good” for EAE but “bad” for AD since Aβ clearance is also likely blunted in the latter. “The concept that neuroinflammation in AD is a simple myeloid cell overreaction to Aβ peptides and fibrils, based on in vitro work or brain microinjection, is overblown. Instead, the microglia in AD brains are hypo-reactive due to the effects of Aβ peptides,” Ransohoff wrote in an email to Alzforum. Research published April 8 in PLoS ONE lends support to this view, by showing that impairment of microglial motility and phagocytic activity coincides with Aβ deposition in AD mice, and that Aβ vaccination restored these capacities (Krabbe et al., 2013). Broadly speaking, the role of microglia in AD remains controversial (see ARF related news story; ARF related news story; ARF Venusburg conference series).

Furthermore, scientists lack a clear understanding of what the various Aβ species do biologically, said Terrence Town of the University of California, Los Angeles. “We know amyloid fibrils are a defining feature of AD, but what they do at the molecular level is an open question.” While the prevailing view casts Aβ fibrils as villains, some research points to the idea that fibrils are biologically inert (see Martins et al., 2008), or even beneficial (see Kinghorn et al., 2006; ARF news story on Fowler et al., 2005).

Uncertainty about amyloid-beta’s biological function is one reason Steinman worries about intervening early in AD. Though amyloid-forming molecules are seen as culprits, “we have genes that express amyloid-beta and various other amyloid proteins in every cell of our bodies,” he said. “I think that under certain circumstances they may have guardian functions, and impairing these when people are unafflicted by AD could be detrimental.”

Pimplikar said the work by Steinman’s group “unequivocally demonstrates” that peripheral Aβ levels have significant consequences for CNS immune responses, underscoring the possibility for unintended consequences of amyloid-lowering approaches in therapeutic trials.

Pritam Das of Mayo Clinic in Jacksonville, Florida, thinks the new data “support a potential beneficial role of amyloid-beta peptides in the CNS and in the periphery,” possibly through binding and clearance of inflammatory mediators in response to injury and aging. However, the rapid relief of symptoms in EAE mice suggests that peripheral injection of amyloidogenic peptides may affect multiple immune pathways specifically or non-specifically, Das noted. “More data on the mechanism of this phenomenon is still warranted,” he wrote in an email to Alzforum.

The findings add to other recent work highlighting the importance of inflammatory responses in AD—namely, the discovery that rare mutations in the TREM2 microglial receptor confer almost as much AD risk as the leading risk gene APOE4 (ARF news story on Guerreiro et al., 2012 and Jonsson et al., 2012), and the association of TREM2 variants with elevated spinal fluid levels of tau and phospho-tau (ARF news story on Cruchaga et al., 2013). —Esther Landhuis

Q&A With Lawrence Steinman. Questions by Esther Landhuis.

Q: What motivated the current study?

A: We found amyloid in multiple sclerosis (MS) lesions. We studied this a number of different ways—most compelling was laser capture dissection and proteomics on lesions in autopsy specimens (Han et al., 2008). There was plenty of amyloid, plenty of tau, crystallin (HSPB5), etc. We wanted to know what they might be doing.

We injected amyloid-β peptides peripherally into mice induced to develop an experimental version of MS, expecting it to do nothing or make their disease worse. Lo and behold, the animals got better (ARF related news story on Grant et al., 2012).

Continuing the work, we noticed commonalities in the chemical structures of other infamous proteins including tau and prion protein (PrP). David Eisenberg, a chemist at UCLA, had an algorithm for selecting protein segments that will fold into β-sheets and form amyloid structures (Eisenberg and Jucker, 2012). We wanted to see what happens if you give hexapeptides of β amyloid (or tau, or PrP, or crystallin, or other amyloid-forming proteins) to EAE mice. Again, the animals got better.

We also found, from our and others’ studies, that when these amyloidogenic proteins are knocked out, the symptoms are worse.

Q: You found that the chaperone function of the injected hexamers correlated with therapeutic benefit. Is it the extent of binding in general, or which proinflammatory molecules are being bound, that matters most?

A: It probably is important which proinflammatory molecules are being bound, but it does not seem very specific. As long as the hexamer could form amyloidogenic structures, it did not seem to matter which molecule was injected.

Q: Why do you think specificity is not required?

A: This is a new concept. The chaperones are binding a specific structure via their capacity to form steric zippers with other β pleated sheets. This is not the type of interaction that occurs with receptor ligand interactions with, for example, immunoglobulin and cognate antigen.

Q: Do you think there is something unique about this autoimmune model that explains why any amyloid rescues?

A: As I pointed out, amyloid-forming proteins are therapeutic not only in EAE, but in models of stroke, brain trauma, optic nerve ischemia, myocardial infarction, and brain trauma. Colin Masters' group published on brain trauma, and over the years we have shown that αB-crystallin, which forms amyloid, is therapeutic in EAE, stroke, myocardial infarction, and optic nerve ischemia. We reference all these studies in the paper on hexamers in Science Translational Medicine. And in EAE, stroke, and brain trauma, knocking out amyloid-forming proteins worsens disease. This is not merely a phenomenon restricted to EAE, as others have argued.

Q: Have you tested these hexamers in models of other neurodegenerative or inflammatory diseases?

A: We are testing now the hexamers in other models of neurodegeneration and inflammation.

Q: Your findings run counter to the prevailing view in Alzheimer’s disease that Aβ can be toxic. Do you think your results in a mouse model of MS, in which neuroinflammation is prominent, would apply to AD, where the role of inflammation is less clear?

A: It’s a stretch to extrapolate from studies on MS/EAE to AD. However, studies like this make us think about anomalies in the data.

For example, in research published last month in Biological Psychiatry, PET scans revealed less brain amyloid in people with a CR1 allele that increases AD risk compared to those without the risk allele (Thambisetty et al., 2012, and Gandy et al., 2013).

Maybe it is possible that low amyloid correlates with early cognitive decline. Maybe under certain conditions, we should be saying, “I’m forgetting something because I’m having a low amyloid moment.”

Q: But obviously a single paper cannot overturn the amyloid hypothesis….

A: The amyloid hypothesis of AD does have a good foundation. Maybe the hypothesis is right. But right now there are a lot of problems, the negative clinical trials being most prominent. Lots of trials of amyloid-lowering approaches have been done, lots of money spent, but little success.

We have amyloid-β (and other amyloid-forming proteins) in every cell of our body. They’re turned on. Why do we have these genes? It’s a big enigma. I think that under certain circumstances they may have guardian functions—and impairing that process in people unafflicted by AD could be detrimental. We will find out in future trials, but I worry.

This is all speculative. I still teach the amyloid hypothesis to my medical students. That’s my professional responsibility until it is proven or disproven.

Q: Many would argue that amyloid results from aberrant misfolding in many cases, and that these proteins have other primary functions, many of which are known, for example, for tau, superoxide dismutase, transthyretin. Do you see amyloid formation as a primary function for these proteins?

A: All these proteins form amyloid, but I would not attempt to argue what the "primary function" of a molecule is. Molecules obey the laws of chemistry, not the strange and wondrous pathways of human cognition.

Q: There are some indications now from clinical trials that early treatment with anti-Aβ immunotherapies can slow decline, and now trials are looking to treat even earlier. Does that bode well for proving the amyloid hypothesis, and how would it fit with the idea that amyloid is protective?

A: Well, I worry that really early treatment will be problematic, as these molecules may be there for physiologic reasons.

Q: Do you know if anyone plans to test the hexameric peptides from your current paper in Alzheimer's mouse models?

A: I am unaware that anyone is trying. These models all overexpress Aβ to begin with. It would be worthwhile for someone to try, but I think the deck is stacked in these AD “models.”

We plan to do so, if we can get funding. Study sections have been dismissive of this work. We are truly ‘outside the box.’

Q: What are you doing in follow-up?

A: We are looking at amyloid more carefully in MS patients with early cognitive decline. We want to see if amyloidogenic peptides from the small heat shock protein αB-crystallin (HSPB5) can be useful as a therapeutic for MS.

Q: Thank you for this interview.


  1. This fascinating paper by Kurnellas et al. follows up on their previous work showing immunomodulatory properties of amyloid peptides (Aβ40 and 42) and αB-crystallin in ameliorating experimental autoimmune encephalomyelitis (EAE), a model of multiple sclerosis.

    This paper provides an in-depth mechanistic analysis of this phenomenon, suggesting that the amyloidogenic proteins and peptides affect the immune response by either mediating chaperone-like activity and/or binding and “scavenging” potential inflammatory mediators including ApoE, complement, etc., from the plasma. In terms of Alzheimer’s disease, while it is well documented that various assemblies of amyloid-β can be toxic to neurons and affect synaptic activity, these data support a potential beneficial role of Aβ peptides in the CNS and in the periphery. To date, it is not known and is still debated what the normal physiological function of the amyloid peptide is. Indeed, Aβ peptides bind ApoE, and this interaction is involved in the clearance/degradation pathways of Aβ peptides from the brain. Thus, based on the data in this paper, it is tempting to speculate that this interaction can also then affect the binding and clearance of other proteins, including potential inflammatory mediators in the brain or periphery in response to injury and aging.

    We know now that with age, the amyloid peptide clearance pathways are reduced/not efficient. Does this dysfunction in amyloid clearance lead to increased inflammatory reactions in the brain with age? Or will anti-Aβ treatment strategies in AD affect the inflammatory response in the brain? At this point, more data regarding the mechanism of this phenomenon are still warranted. The fact that this treatment strategy results in a rapid reduction in EAE disease progression suggests potential alternative mechanisms of action, or that peripheral injection of amyloidogenic peptides simply affect multiple pathways of the immune system either specifically or non-specifically.

    View all comments by Pritam Das
  2. Together with their previous study (1), this paper from the Steinman group leaves little doubt that Aβ peptides (both 40 and 42) exert beneficial effects in the EAE mouse model of MS; they do so by reducing neuroinflammation and attenuating paralysis. The current study also points to the potential underlying mechanisms (chaperon/chelating activity) and identifies hexameric peptides from the C-terminal end of Aβ40 and Aβ42 (among others) as therapeutic agents. The authors suggest that a similar mechanism could be operative in conditions such as stroke and brain trauma where inflammation is seen. This raises a question: What about AD? What implications do these findings have regarding our view of AD pathogenesis and the anti-amyloid therapeutic trials currently underway?

    Based on these findings, it might be tempting to think of Aβ as being "good" in AD. Atwood and colleagues had previously suggested Aβ to be beneficial based on its ability to chelate redox metal ions and thereby exert antioxidant effects (2). Incidentally, oxidative stress induces inflammation, whereas antioxidants reduce inflammation. The possibility that Aβ can be beneficial is obviously controversial and runs against the very essence of the amyloid hypothesis.

    However, I do not think the current investigations can be directly compared to AD or interpreted as supportive of the beneficial role of Aβ in Alzheimer's. The EAE model, by its very nature, is an "acute" system where the disease symptoms (paralysis) appear within days, and animals respond to the treatment on the same time scale. AD, by contrast, is a chronic disorder, and AD-like pathology needs months to manifest in animal models. Etiopathogenesis of MS and AD, and the courses of the respective diseases, are too different to make a direct, valid comparison.

    Nonetheless, there are lessons to be learned for the AD field. First, if Aβ is immunosuppressive, then reducing its levels might enhance neuroinflammation and thereby negate any potential beneficial effects of reducing brain amyloid. Is that why the amyloid therapies have been ineffective in improving clinical outcomes? Second, these papers unequivocally demonstrate that levels of Aβ in the peripheral system have significant consequences in the CNS in terms of immune response. Thus, manipulating Aβ levels (by active/passive immunization or by pharmacological drugs) may have significant, unintended consequences, especially in the long term.


    . Reversal of paralysis and reduced inflammation from peripheral administration of β-amyloid in TH1 and TH17 versions of experimental autoimmune encephalomyelitis. Sci Transl Med. 2012 Aug 1;4(145):145ra105. PubMed.

    . Amyloid-beta: a chameleon walking in two worlds: a review of the trophic and toxic properties of amyloid-beta. Brain Res Brain Res Rev. 2003 Sep;43(1):1-16. PubMed.

    View all comments by Sanjay Pimplikar

Make a Comment

To make a comment you must login or register.


News Citations

  1. Aβ’s Good Side—Can It Improve Symptoms of Multiple Sclerosis?
  2. Microglia—Medics or Meddlers in Dementia
  3. Why Good Microglia Turn Bad—A Matter of Timing?
  4. Is It Good for You? Amyloid Shows New Side in Mammalian Cells
  5. Enter the New Alzheimer’s Gene: TREM2 Variant Triples Risk
  6. By Digging Tau, GWAS Unearths TREM (Again), Plus New Leads

Paper Citations

  1. . Proteomic analysis of active multiple sclerosis lesions reveals therapeutic targets. Nature. 2008 Feb 28;451(7182):1076-81. PubMed.
  2. . Reversal of paralysis and reduced inflammation from peripheral administration of β-amyloid in TH1 and TH17 versions of experimental autoimmune encephalomyelitis. Sci Transl Med. 2012 Aug 1;4(145):145ra105. PubMed.
  3. . Identifying the amylome, proteins capable of forming amyloid-like fibrils. Proc Natl Acad Sci U S A. 2010 Feb 23;107(8):3487-92. PubMed.
  4. . The 3D profile method for identifying fibril-forming segments of proteins. Proc Natl Acad Sci U S A. 2006 Mar 14;103(11):4074-8. PubMed.
  5. . The amyloid state of proteins in human diseases. Cell. 2012 Mar 16;148(6):1188-203. PubMed.
  6. . Functional impairment of microglia coincides with Beta-amyloid deposition in mice with Alzheimer-like pathology. PLoS One. 2013;8(4):e60921. PubMed.
  7. . Lipids revert inert Abeta amyloid fibrils to neurotoxic protofibrils that affect learning in mice. EMBO J. 2008 Jan 9;27(1):224-33. PubMed.
  8. . Neuroserpin binds Abeta and is a neuroprotective component of amyloid plaques in Alzheimer disease. J Biol Chem. 2006 Sep 29;281(39):29268-77. PubMed.
  9. . Functional amyloid formation within mammalian tissue. PLoS Biol. 2006 Jan;4(1):e6. PubMed.
  10. . TREM2 variants in Alzheimer's disease. N Engl J Med. 2013 Jan 10;368(2):117-27. Epub 2012 Nov 14 PubMed.
  11. . Variant of TREM2 associated with the risk of Alzheimer's disease. N Engl J Med. 2013 Jan 10;368(2):107-16. Epub 2012 Nov 14 PubMed.
  12. . GWAS of cerebrospinal fluid tau levels identifies risk variants for Alzheimer's disease. Neuron. 2013 Apr 24;78(2):256-68. Epub 2013 Apr 4 PubMed.
  13. . Effect of Complement CR1 on Brain Amyloid Burden During Aging and Its Modification byAPOEGenotype. Biol Psychiatry. 2012 Sep 27; PubMed.
  14. . CR1 and the "vanishing amyloid" hypothesis of Alzheimer's disease. Biol Psychiatry. 2013 Mar 1;73(5):393-5. PubMed.

Other Citations

  1. See Q&A with Lawrence Steinman below.

External Citations

  1. Science Friday

Further Reading


  1. . Reversal of paralysis and reduced inflammation from peripheral administration of β-amyloid in TH1 and TH17 versions of experimental autoimmune encephalomyelitis. Sci Transl Med. 2012 Aug 1;4(145):145ra105. PubMed.

Primary Papers

  1. . Amyloid fibrils composed of hexameric peptides attenuate neuroinflammation. Sci Transl Med. 2013 Apr 3;5(179):179ra42. PubMed.