In today’s PLoS One, Robert Moir and others at Massachusetts General Hospital in Charlestown formally publish a finding they had earlier presented at the 9th International Conference AD/PD held last spring in Prague. Far from being merely a catabolic byproduct, these scientists assert, the amyloidogenic peptide Aβ serves the physiological function of fighting off microbes as part of the body’s innate immune system. First author Stephanie Soscia and colleagues report experiments comparing Aβ’s activity against a range of clinically important microorganisms. For a detailed story of the context, clues from the microbiology literature, and this in vitro data, see ARF conference news. In the paper, the authors add new results from experiments comparing antimicrobial activity in high-Aβ temporal lobe samples and low-Aβ cerebellar samples from people who had had Alzheimer disease versus age-matched controls.—Gabrielle Strobel.


Make a Comment

To make a comment you must login or register.

Comments on News and Primary Papers

  1. Very interesting.... As a matter of fact, we have been observing significant extracellular toxicity of oligomeric Aβ compared to fibrillar in yeast (1).

    However, it is rather difficult to imagine that Aβ would be secreted to kill pathogenic cells, as it is a cyto/neurotoxic protein by itself, even though you may consider Aβ more antimicrobial than neurotoxic.

    View all comments by Prashant Bharadwaj
  2. This interesting concept could begin to explain the widespread amyloid response in AD and in prodromal AD. It may also begin to shed some light as to why Aß vaccination has not been a successful treatment approach.

    View all comments by Elliott Mufson
  3. What good news that microbes are becoming acceptable as possible agents in AD. Regarding the proposal of a generic mechanism, “not one based on any particular organism,” surely a prerequisite for the participation of a putative agent in AD, a disease of long duration, is that 1) the agent actually resides permanently in many aged brains and 2) in infected cells it causes the formation of the biomarker molecules of AD brains.

    Both herpes simplex virus type 1 (Jamieson et al., 1991; Wozniak et al., 2007; Wozniak et al., 2009a; Wozniak et al., 2009b) and Chlamydia pneumoniae (Balin et al., 1998; Little et al., 2004) meet these criteria. Furthermore, in the case of HSV1, its remains do indeed lie almost entirely “inside every plaque in AD brains” (Wozniak et al., 2009b). As to other microbes, in normal circumstances very few comply even with the first criterion, let alone the second.

    Perhaps it is time for AD researchers to understand that certain microbes can remain long-term in the body, causing slow cumulative damage, i.e., that not all cause merely acute infections and then do a runner (and also that “controls,” too, can harbor a microbe asymptomatically).

    View all comments by Ruth Itzhaki
  4. The issue covered in this article is very interesting and reconciles many previous data on amyloids in general and on Aβ peptides in particular.

    In general, it has been reported that a variant of SAA (serum amyloid A), whose aggregates are involved in a severe form of systemic amyloidosis, aggregates into annular assemblies able to kill bacterial cells following increase of its serum levels in chronic infections (the SAA itself is an acute-phase protein) (1). In addition, recent findings from Charlie Glabe's lab clearly show that the same Aβ raised against amyloid oligomers recognizes oligomers formed by bacterial pore-forming toxins, further underscoring the analogies between amyloids and natural protein oligomers evolved to kill target cells (2).

    Apart from this, the possible physiological role of specific amyloids is clearly emerging not only in the microbial world (see the curli, tofi, chaplins, and hydrophobin stories) but also in mammalian cells, as it has clearly been shown by the Balch group in the case of Pme17-favored melanogenesis (reviewed in 3). In this context, the repeatedly reported and experimentally supported significance of the Aβ peptides as vessel sealants alternative to the fibrinogenic pathway in the CNS is of significance (reviewed in 4). Therefore, it is not surprising, but yet highly interesting, that Aβ peptides may be evolved to provide to the organism (possibly not only to the CNS) a further tool to help fight microbial invasion.

    View all comments by Massimo Stefani
  5. Reply to Anonymous by Rob Moir and Rudy Tanzi
    Our identification of Aβ as an antimicrobial peptide strongly supports the idea that AD is, in fact, a disease mediated by the innate immune system. However, we urge caution in drawing unsubstantiated conclusions based on our findings. In particular, we feel it is far too premature to consider untried and highly speculative treatment strategies based around the hypothesis that pathogens infecting the brain cause AD. AD is a terminal illness and available treatment options will only modestly slow disease progression, at best. We therefore empathize with the incentive patients and their families feel to embrace a new albeit unproven therapy.

    At the same time, there are good scientific arguments why it is clearly premature to consider antimicrobial treatments for AD given our present level of knowledge. Firstly, simply because Aβ is an AMP does not necessarily mean it is accumulating in AD to fight a microbial pathogen. Aβ could be induced in the brain as part of the innate immune system in response to a chronic persistent infection, to a past transient infection, or to non-pathogenic insults that are unrelated to microorganisms. There are more than 30 known autoimmune diseases. Most of these involve inappropriate immune responses but do not appear to involve microbial pathogens.

    Second, even if a microbial pathogen is involved, it may only be involved in “jump-starting” the amyloid cascade. The infection may be long gone by the time the clinical symptoms of AD manifest, in which case antimicrobial drugs would have no effect on the disease's progression. Third, if a pathogen is involved, it is unclear what antimicrobial drug would be appropriate since the possible pathogenic agent could be a virus, bacteria, yeast, or other parasite. We don't know which is most relevant at this time, but this knowledge is key if any antimicrobial treatment strategy is to have any chance of success. Fourth, few of the antibiotics and antiviral agents presently available can actually cross the blood-brain barrier.

    In summary, with no clear microbial pathogen to target and highly limited numbers of brain-penetrating drugs in any case, it is premature to attempt antimicrobial-based therapies to treat AD. Moreover, antimicrobial treatments based on “best guess” ideas are far more likely to be harmful than beneficial to AD patients since all drugs, and particularly deliberately cytotoxic ones such as antimicrobials, have side-effects. With this said, we strongly believe that the identification of Aβ as an antimicrobial peptide will open up bright new avenues of research. Once the central questions generated by our findings begin to be answered, novel treatments will hopefully arise.

    View all comments by Robert Moir
  6. I have a friend who is showing signs of familial Alzheimer's at age 40. He commented that upon taking Keflex for an injury, his cognitive function improved greatly. I am not a researcher, but am wondering: as an antimicrobial, is this drug in any way similar to clioquinol, and could any antibiotic have a similar effect?

    View all comments by A. Anonymous
  7. I have proposed that Aβ may block HPV16 cell entry due to the fact that it depletes dynamin 1. I refer to my hypothesis, A role for the HPV16E7 oncogene in the pathology of DS and AD. ALS Therapy Development Institute, Feb 17, 2009.

    View all comments by Mary Reid
  8. We would like to respond to a few points in the comments by Moir and Tanzi.

    They state that
    1. “...the (microbial) infection may be long gone by the time clinical symptoms of AD manifest, in which case antimicrobial drugs would have no effect on the disease’s progression.” However, in the case of HSV1, we have very good evidence from three different techniques (solution PCR of DNA extracted from human brain, in situ PCR of brain sections, and an immunological investigation of intrathecal antibodies) that not only is this virus present as a persistent, latent infection in the elderly brain, but also that it reactivates, leading to an acute infection (1-3).

    2. “...few antiviral agents…can actually cross the blood-brain barrier.” However, acyclovir, or its biodrug, valacyclovir, was used to treat MS patients in two studies (based on the putative role of another herpes virus in MS), and CSF measurements showed that acyclovir did indeed cross the blood-brain barrier and none of the patients demonstrated a damaged blood-CSF barrier (4).

    3. “...antimicrobial treatments are far more likely to be harmful than beneficial to AD patients since all…have side effects.” However, acyclovir is used routinely to treat the rare brain disease HSE caused by acute HSV1 infection; it prevents viral replication and has thereby enabled most sufferers to survive what was previously a fatal disease. Further, acyclovir has relatively few side effects even at very high doses (except in those with renal impairment); e.g., in one of the MS studies, a dose of 3g per day for two years caused no detectable ill-effects (5).

    4. “Aβ could be induced in the brain as part of the innate immune system….” We agree; as we said in our response to comments on our Journal of Pathology paper, Aβ production (and AD-like tau, which is formed, too [6]) might well be protective—but only initially: we have preliminary data suggesting that the cell might produce Aβ in an abortive attempt to combat the virus, but this eventually results in overproduction and toxicity of the peptide products.

    View all comments by Ruth Itzhaki
  9. It is good news that Aβ is a broad-spectrum antimicrobial peptide that shows bacteriostatic activity against clinically important organisms at concentrations similar to those of LL-37. I think we must understand the interaction with the other drugs, herbal consumption, and the contraindications. Information about other drug interaction, herbal interaction, must be informed by research altogether, especially for Alzheimer's. Bravo for you!

    View all comments by Jenry Simanjuntak
  10. It is interesting to link this idea (i.e., the involvement of Aβ as a functional protein from the evolutionarily ancient innate immune system) to a 10-year-old publication by Coulson et al.

    This publication provides evidence that the normal function of APP among evolutionarily distinct species is to regulate cell-cell or cell-substrate interactions.

    Perhaps the simple adhesion capabilities of Aβ42 could serve as an evolutionarily conserved, old immune system mechanism with a neurodegenerative “side effect.” Interesting also that some mammals do not accumulate plaques in age in contrast to some others that do (see Johnstone et al., 1991). Notably, normal mice are among the former.

    View all comments by Ivan Goussakov
  11. Comment by Ruth Itzhaki, Curtis Dobson, and Matthew Wozniak
    This study provides convincing experimental data for a role of Aβ as an antimicrobial peptide performing a normal function in the innate immune response [1]. It is reminiscent of the hypothesis of Bishop and Robinson [2] suggesting that β amyloid (Aβ) might protect the brain against pathogens. As Soscia et al. state, their results might explain why several micro-organisms cause Aβ accumulation in cell culture and/or in mouse models.

    In the case of the bacteria Chlamydia pneumonia and Borrelia burgdorferi, spirochetes are present in a high proportion of Alzheimer disease (AD) brains [3,4], and both can cause AD-like changes [5,6]. Similarly, the virus, herpes simplex virus type 1 (HSV1), which is a risk for AD when present in the brains of ApoE4 carriers [7], leads to AD-like changes in cell cultures and mouse brains [8,9]. Significantly, HSV1 DNA is very specifically located in AD amyloid plaques [10] throughout the plaque, not just on the “sticky” surface, as shown in sections that are much thinner than the average plaque diameter.

    Tests of Aβ’s antimicrobial action against the above microorganisms, which are directly implicated in AD, would be of particular interest. Many antimicrobial peptides show antiviral activity [11], but any apparent reduction in infectivity has to be distinguished from any cytotoxic effects on the mammalian host cells, which is obviously difficult with cytotoxic agents like Aβ, as we reported previously [9].

    The findings of Soscia et al. may also relate to our observations that the receptor binding region of human apolipoprotein E—the protein encoded by ApoE, the type 4 allele of which is a major genetic susceptibility factor for AD—has broad antibacterial and antiviral activity [12], and can generate a family of broad-action potent antimicrobial peptides [13]. Possibly Aβ and ApoE have additive and/or complementary antimicrobial activities.

    It is intriguing that HSV1 causes Aβ accumulation, as the virus decreases synthesis of most host cell proteins. Possibly the peptide is needed for the synthesis of progeny virus. Interestingly, Aβ fibrils enhance infection of several enveloped viruses, including HSV1 [14]. The alternative—as suggested by Soscia et al.—that the increase in Aβ occurs as part of the cells’ defense response, with Aβ acting as an antimicrobial peptide, is supported by the fact that HIV [15], HSV2 (unpublished observations), and the bacteria mentioned above, and also chemicals (such as hydrogen peroxide [16] and mercury [17]), increase Aβ production.

    In fact, Bishop and Robinson [2] implied that Aβ might initially entomb the agent, thereby preventing further damage to the host, but eventually, Aβ overproduction might result in toxicity via oligomer formation. However, our previous investigation into the effect of Aβ on HSV1 revealed no antiviral activity [9]; possibly in vitro tests of Aβ for antiviral or antibacterial activity might yield different results depending on its conformation or extent of aggregation; hence, it would be important to test the antiviral activity of Aβ prepared by the method used by Socia et al.

    View all comments by Curtis Dobson
  12. My 2006 publication (excerpt below) explicitly described Aβ as an antimicrobial peptide based on its multiple similarities to melittin.

    In addition, the described antimicrobial action was explicitly cited as a basis for Aβ to be considered a component of the innate immune system, likely as a rapid response to infections by enveloped viruses such as HSV.

    Excerpt from Kammerman et al., 2006:

    "Numerous groups have reported that Aβ42 can disrupt lipid membranes by creating pore-like holes (ion channels) within the membranes. This property of Aβ42 appears to be related to an antimicrobial function; nature is replete with examples of peptide antimicrobials that effect their function through membrane disruption. In fact, some of these peptides show strong activity against HSV-1. For example, melittin, which has an α-helical amphipathic structure similar to Aβ42, has anti-HSV activity. We assert that β amyloid, by virtue of its similar molecular shape and size, is a peptide antimicrobial component of the innate immune system which can neutralize enveloped viruses such as HSV-1. We contend accordingly that ion-channel pore formation in the HSV-1 envelope generated by Aβ42 is virucidal."

    View all comments by Eli Kammerman
  13. A surprising report has been published in PLoS ONE (1). Amyloid-β is now thought to be an antimicrobial peptide. If this report is right, the present amyloid treatment strategies are completely wrong and should be stopped to develop new ones.

    The amyloid peptide is toxic to bacteria, not neuronal cells, from the outside. However, if amyloid-β accumulates inside the cell, its antimicrobial immune effect turns into a toxic one for that cell. Now, this toxicity will be driven by homocysteic acid (HA). In the presence of HA, the activated immune system, which produces more amyloid than dormant immune systems, will be a risk factor for Alzheimer disease.

    However, why does the Alzheimer’s brain activate the immune system? Infection is one possibility, but it is unlikely that all Alzheimer’s patients suffered from infections. Another possibility that we have proposed is HA (2). HA is also well known as an activator for the immune system (3). Consequently, HA might activate the immune system in Alzheimer disease, inducing amyloid production.

    If so, anti-inflammatory treatments should not be conducted because we need the antimicrobial amyloid. Instead, an HA vaccine could treat the activated immune system and cure cognitive impairment.

    A Chinese article (4) reported that a high level of HA was observed in cerebral infarction with dementia, and this high level of HA paralleled the severity of dementia. At present, it is thought that an infarction induced the dementia through not amyloid pathology, but through oxidative stress, or radical formation, etc. If HA toxicity is not based on amyloid pathology, that indicates that Alzheimer’s dementia induced by HA is not due to amyloid pathology. From these data, I propose that Alzheimer disease is composed of two pathologies. One is amyloid pathology, and the other is the functional damage due to HA toxicity.

    Thus, on the one hand, it could be that the activated immune system induces amyloid production, and that amyloid is not toxic but an antimicrobial peptide. While on the other hand, HA toxicity is pathological and induces neurodegeneration. If so, Alzheimer disease should be treated with an HA vaccine, instead of anti-amyloid treatments.

    View all comments by Tohru Hasegawa


News Citations

  1. Prague: Aβ Rehabilitated as an Antimicrobial Protein?

Further Reading

Primary Papers

  1. . The Alzheimer's disease-associated amyloid beta-protein is an antimicrobial peptide. PLoS One. 2010;5(3):e9505. PubMed.