This paper attempts to address two important questions that stem from the “amyloid cascade” model of Alzheimer pathology: why is AD age-dependent, and which specific forms of the β amyloid peptide (Aβ) are responsible for the toxicity? The experimental model employed in this study are transgenic Caenorhabditis elegans worms engineered to constitutively express human Aβ1-42 in muscle cells. These transgenic worms accumulate intracellular Aβ and show a progressive paralysis that begins in adulthood. Cohen et al. manipulated two (likely interacting) stress response pathways in C. elegans and examined the effect on Aβ toxicity and the accumulation of Aβ species. One response pathway, which is controlled by the HSF-1 transcription factor, regulates the response to heat shock, and is the major pathway by which cells detoxify misfolded proteins (which are a primary result of heat shock). The other stress response pathway, which is controlled by DAF-2 (a homolog of insulin/IGF receptor) and the downstream transcription factor DAF-16, predominantly controls the worm response...  Read more

This paper attempts to address two important questions that stem from the “amyloid cascade” model of Alzheimer pathology: why is AD age-dependent, and which specific forms of the β amyloid peptide (Aβ) are responsible for the toxicity? The experimental model employed in this study are transgenic Caenorhabditis elegans worms engineered to constitutively express human Aβ1-42 in muscle cells. These transgenic worms accumulate intracellular Aβ and show a progressive paralysis that begins in adulthood. Cohen et al. manipulated two (likely interacting) stress response pathways in C. elegans and examined the effect on Aβ toxicity and the accumulation of Aβ species. One response pathway, which is controlled by the HSF-1 transcription factor, regulates the response to heat shock, and is the major pathway by which cells detoxify misfolded proteins (which are a primary result of heat shock). The other stress response pathway, which is controlled by DAF-2 (a homolog of insulin/IGF receptor) and the downstream transcription factor DAF-16, predominantly controls the worm response to starvation, although activation of this pathway broadly contributes to stress resistance. A byproduct of activating DAF-16 (or inactivating DAF-2, a negative regulator of DAF-16), is a significant increase in worm lifespan.

The results from inactivating HSF-1 by RNA interference (RNAi) are reasonably straightforward. HSF-1 inactivation increases the rate of paralysis of the Aβ worms, and leads to the accumulation of higher-molecular-weight species of Aβ. Extracts of Aβ worms with decreased HSF-1 levels are better at seeding formation of amyloid fibrils and worse at disaggregating amyloid fibrils in vitro. These results are consistent with a model in which activation of HSF-1 in Aβ worms leads to the expression of protective chaperone proteins, which counter both the formation of higher molecular forms of Aβ and also protect the worms from Aβ toxicity. Chaperone proteins are known to be induced by and co-localize with Aβ in this worm model (Fonte et al., 2002), and unpublished experiments from my lab have demonstrated that forced expression of some of these chaperone proteins can suppress Aβ toxicity in C. elegans. The results of the HSF-1 RNAi experiments are all consistent with the idea that chaperone proteins play an important role in modulating the accumulation of toxic intracellular Aβ species.

The results of experiments using DAF-2 and DAF-16 RNAi are more complicated and harder to interpret. Knockdown of DAF-2 by RNAi (activation of the DAF-2/DAF-16 stress response pathway) protects from Aβ-induced paralysis, and knockdown of DAF-16 by RNAi (inactivation of this stress response pathway) increases paralysis, consistent with expectations. However, DAF-2 RNAi suppression of toxicity is associated with an increase in higher-molecular-weight Aβ species and in vitro fibril seeding, while DAF-16 RNAi decreases levels of higher-molecular-weight Aβ and in vitro fibril seeding capacity. These results strongly argue that the higher-molecular-weight Aβ species are not the toxic component, which is perhaps not surprising, given that these AD model worms are born with fibrillar amyloid deposits (Link et al., 2001) but do not show paralysis until they are adults. Conversely, Aβ can induce paralysis and protein oxidation in C. elegans independent of the formation of detectable amyloid (Drake et al., 2003). So, if the higher-molecular-weight Aβ species are not relevant, what are the HSF-1 and DAF-2/DAF-16 pathways doing to suppress Aβ toxicity? By altering their extraction protocol, these researchers were able to identify a ~16 kD Aβ species that correlated with levels of paralysis in the RNAi experiments, implying that this could be the key toxic species regulated by the HSF-1 and DAF-2/DAF-16 pathways.

What does this study tell us about the age-dependence of Alzheimer disease? Perhaps that it is the decay of protective cellular systems with age that “unleash” toxic Aβ species. If this speculation is true, then perhaps AD risk alleles will be identified that function in stress response pathways analogous to those investigated in this study. What is the toxic Aβ species in this worm model? This is trickier to answer, because there are some unproven assumptions underlying this study. One assumption is that these stress response pathways operate at the level of Aβ accumulation/multimerization, not somewhere downstream in the toxic process. Accumulation of polyglutamine repeat proteins in a similar C. elegans model (Gidalevitz et al., 2006) have been shown to generally perturb protein homeostasis, so it is possible that these pathways actually act on non-native forms of endogenous worm proteins whose folding has been secondarily perturbed by Aβ accumulation. In this case, the forms of Aβ that accumulate when these response pathways are inhibited may be irrelevant. It is also important to point out that correlation of any specific form of Aβ with toxicity does not prove causality. Although many current studies have pointed to oligomeric Aβ species as being directly involved in Aβ toxicity, it is actually quite difficult to show that a specific oligomeric form of Aβ visualized on a denaturing gel is really the toxic form in vivo. Nevertheless, this study does add to the growing body of evidence that high-molecular-weight forms of Aβ (e.g., Aβ fibrils) are not the key toxic species in numerous AD model systems.

References:
Drake J, Link CD, Butterfield DA. Oxidative stress precedes fibrillar deposition of Alzheimer's disease amyloid β-peptide (1-42) in a transgenic Caenorhabditis elegans model. Neurobiol Aging. 2003 May-Jun;24(3):415-20. Abstract

Fonte V, Kapulkin V, Taft A, Fluet A, Friedman D, Link CD. Interaction of intracellular β amyloid peptide with chaperone proteins. Proc Natl Acad Sci U S A. 2002 Jul 9;99(14):9439-44. Epub 2002 Jun 27. Abstract

Gidalevitz T, Ben-Zvi A, Ho KH, Brignull HR, Morimoto RI. Progressive disruption of cellular protein folding in models of polyglutamine diseases. Science. 2006 Mar 10;311(5766):1471-4. Epub 2006 Feb 9. Abstract

Link CD, Johnson CJ, Fonte V, Paupard M, Hall DH, Styren S, Mathis CA, Klunk WE. Visualization of fibrillar amyloid deposits in living, transgenic Caenorhabditis elegans animals using the sensitive amyloid dye, X-34. Neurobiol Aging. 2001 Mar-Apr;22(2):217-26. Abstract

View all comments by Chris Link

Is there any evidence of pancreatitis prior to the development of AD which may explain the large weight loss which is reported to occur several years prior to AD? It's of interest that HSF-1 is activated in acute pancreatitis (1) and β-APP is activated by HSF-1 (2). Perhaps AD occurs as a result of a prolonged stress response to pancreatitis. The subsequent increased ornithine decarboxylase activity in response to increased APP expression (3) may be expected to reduce the arginine load; however, the trade-off is the development of AD if this response continues. The study by He and colleagues (4) finding that GSK-3β and ERK MAPK facilitate the inactivation of activated HSF-1 has me wondering whether we may exacerbate pancreatitis as a consequence of GSK-3β inhibition. It's interesting that lithostathine and pancreatitis-associated protein are involved in the early stages of AD (5). Also of interest is that pancreatitis-associated protein (PAP)-like protein is elevated in the early stages of scrapie infection (6).

The study by Wolozin and colleagues, soon to be the subject of...  Read more

Is there any evidence of pancreatitis prior to the development of AD which may explain the large weight loss which is reported to occur several years prior to AD? It's of interest that HSF-1 is activated in acute pancreatitis (1) and β-APP is activated by HSF-1 (2). Perhaps AD occurs as a result of a prolonged stress response to pancreatitis. The subsequent increased ornithine decarboxylase activity in response to increased APP expression (3) may be expected to reduce the arginine load; however, the trade-off is the development of AD if this response continues. The study by He and colleagues (4) finding that GSK-3β and ERK MAPK facilitate the inactivation of activated HSF-1 has me wondering whether we may exacerbate pancreatitis as a consequence of GSK-3β inhibition. It's interesting that lithostathine and pancreatitis-associated protein are involved in the early stages of AD (5). Also of interest is that pancreatitis-associated protein (PAP)-like protein is elevated in the early stages of scrapie infection (6).

The study by Wolozin and colleagues, soon to be the subject of an ARF live discussion, reporting a decreased risk of emphysema for those using statin drugs might implicate α1 antitrypsin. There are conflicting reports of α1 antitrypsin deficiency associated with pancreatitis. Are those using statins less likely to develop pancreatitis?

References:
1. O'Reilly DA, Roberts JR, Cartmell MT, Demaine AG, Kingsnorth AN. Heat shock factor-1 and nuclear factor-kappaB are systemically activated in human acute pancreatitis. JOP. 2006 Mar 9;7(2):174-84. Abstract

2. Dewji NN, Do C. Heat shock factor-1 mediates the transcriptional activation of Alzheimer's beta-amyloid precursor protein gene in response to stress. Brain Res Mol Brain Res. 1996 Jan;35(1-2):325-8. Abstract

3. Nilsson T, Bogdanovic N, Volkman I, Winblad B, Folkesson R, Benedikz E. Altered subcellular localization of ornithine decarboxylase in Alzheimer's disease brain. Biochem Biophys Res Commun. 2006 Jun 2;344(2):640-6. Epub 2006 Apr 7. Abstract

4. He B, Meng YH, Mivechi NF. Glycogen synthase kinase 3beta and extracellular signal-regulated kinase inactivate heat shock transcription factor 1 by facilitating the disappearance of transcriptionally active granules after heat shock. Mol Cell Biol. 1998 Nov;18(11):6624-33. Abstract

5. Duplan L, Michel B, Boucraut J, Barthellemy S, Desplat-Jego S, Marin V, Gambarelli D, Bernard D, Berthezene P, Alescio-Lautier B, Verdier JM. Lithostathine and pancreatitis-associated protein are involved in the very early stages of Alzheimer's disease. Neurobiol Aging. 2001 Jan-Feb;22(1):79-88. Abstract

6. Skretting G, Austbo L, Olsaker I, Espenes A. Cloning and expression analysis of an ovine PAP-like protein cDNA, a gene differentially expressed in scrapie. Gene. 2006 Jul 5;376(1):116-22. Epub 2006 Mar 17. Abstract

View all comments by Mary Reid