The studies by Meyer-Luehman et al. extend insights into the in vivo formation of amyloid deposits by amyloid "seeds" that may be hetero- and/or homo-amyloidogenic inducers of amyloid fibrillization. This is significant because these types of studies will lead to the clarification of the perplexing conundrum of why there is a frequent co-occurrence of multiple different types of amyloids in neurodegenerative disorders characterized by brain amyloidosis. Indeed, double and triple neurodegenerative brain amyloidoses appear to far exceed in incidence and prevalence any neurodegenerative brain amyloidosis linked to a single amyloidogenic protein or peptide, and this enigma demands clarification if we are to develop more effective therapies for these disorders.

For example, with respect to Aβ deposits, these may occur by themselves as pathological signatures of single brain amyloidoses, such as cerebral amyloid angiopathy (CAA), which most commonly manifests clinically as stroke. This notwithstanding, CAA is more commonly an incidental finding in neurologically normal...  Read more

The studies by Meyer-Luehman et al. extend insights into the in vivo formation of amyloid deposits by amyloid "seeds" that may be hetero- and/or homo-amyloidogenic inducers of amyloid fibrillization. This is significant because these types of studies will lead to the clarification of the perplexing conundrum of why there is a frequent co-occurrence of multiple different types of amyloids in neurodegenerative disorders characterized by brain amyloidosis. Indeed, double and triple neurodegenerative brain amyloidoses appear to far exceed in incidence and prevalence any neurodegenerative brain amyloidosis linked to a single amyloidogenic protein or peptide, and this enigma demands clarification if we are to develop more effective therapies for these disorders.

For example, with respect to Aβ deposits, these may occur by themselves as pathological signatures of single brain amyloidoses, such as cerebral amyloid angiopathy (CAA), which most commonly manifests clinically as stroke. This notwithstanding, CAA is more commonly an incidental finding in neurologically normal individuals, suggesting that Aβ deposits may not be sufficient in and of themselves to cause a neurodegenerative dementing disease. In contrast, neurodegenerative dementia linked to Aβ amyloidosis is commonly a double or triple brain amyloidosis with coexistent tau amyloid (in the form of neurofibrillary tangles, or NFTs), and this dementia is, of course, known as Alzheimer disease (AD).

However, most familial and sporadic cases of AD are actually triple brain amyloidoses since α-synuclein amyloid also is deposited in Lewy bodies (LBs) together with NFTs and senile plaques, and this disorder is known as the LB variant of AD (LBVAD). Notably, LBVAD is the most common subtype of AD. Recent studies from the Lee/Trojanowski group have begun to dissect out mechanisms whereby α-synuclein and tau can cross-seed the fibrillization of each other to form amyloid fibrils, and further research on this may help clarify the common co-occurrence of LBs and tangles in the same patient (1).

Additionally, further studies are needed to understand how tau and Aβ might cross-fibrillize or promote the fibrillization of each other. Much earlier studies by our group demonstrated potential avenues to explore these issues further—using model systems that antedated the development of transgenic mouse models of brain amyloidoses—by injecting purified PHFtau into rat brains, which induced deposits of Aβ associated with the PHFtau injection sites (2,3). However, transgenic mouse models of brain amyloidoses resulting from the overexpression or regulatable expression of mutant or wild-type tau and α-synuclein or the Aβ precursor protein are much more powerful model systems in which to explore these questions further. This is demonstrated very elegantly in the studies by Meyer-Luehman et al. Further dissection of the cross-seeding by hetero- and/or homo-amyloidogenic inducers of amyloid fibrillization could extend these elegant studies by immunodepletion of AD and LBVAD extracts to remove not only Aβ, but also tau amyloid and α-synuclein amyloid to understand the differential contributions of each of these three amyloidogenic proteins in the induction of Aβ.

References:
1. Giasson BI, Forman MS, Higuchi M, Golbe LI, Graves CL, Kotzbauer PT, Trojanowski JQ, Lee VM. Initiation and synergistic fibrillization of tau and alpha-synuclein. Science. 2003 Apr 25;300(5619):636-40. Abstract

2. Shin RW, Bramblett GT, Lee VM, Trojanowski JQ. Alzheimer disease A68 proteins injected into rat brain induce codeposits of beta-amyloid, ubiquitin, and alpha 1-antichymotrypsin. Proc Natl Acad Sci U S A. 1993 Jul 15;90(14):6825-8. Abstract

3. Shin RW, Lee VM, Trojanowski JQ. Aluminum modifies the properties of Alzheimer's disease PHF tau proteins in vivo and in vitro. J Neurosci. 1994 Nov;14(11 Pt 2):7221-33. Abstract

View all comments by John Trojanowski

BACE1 is the principal β-secretase for generation of amyloid-β peptides. Since the identification of BACE1, several lines of BACE1 knockout mice have been made, which are viable and show no major behavioral and pathological abnormalities, suggesting that BACE1 is a safe therapeutic target for Alzheimer disease (AD). Notably, some BACE1 KO mice show premature lethality and subtle alterations in emotional response and locomotor activities. BACE1 KO neurons also display subtle changes in synaptic plasticity and sodium conductance. These deficits are not noted in all the reported mice, but similar discrepancies in behavioral phenotyping have been noticed in mice derived from different strain backgrounds and gene targeting vectors.

Willem and colleagues are the first to show a convincing neuropathological abnormality in BACE1 KO mice. An observation that the highest expression of BACE1 protein correlates with the onset of peripheral nerve myelination promotes them to examine the progression of myelination in the sciatic nerve of BACE1 KO mice. They find that axons of BACE1 KO mice...  Read more

BACE1 is the principal β-secretase for generation of amyloid-β peptides. Since the identification of BACE1, several lines of BACE1 knockout mice have been made, which are viable and show no major behavioral and pathological abnormalities, suggesting that BACE1 is a safe therapeutic target for Alzheimer disease (AD). Notably, some BACE1 KO mice show premature lethality and subtle alterations in emotional response and locomotor activities. BACE1 KO neurons also display subtle changes in synaptic plasticity and sodium conductance. These deficits are not noted in all the reported mice, but similar discrepancies in behavioral phenotyping have been noticed in mice derived from different strain backgrounds and gene targeting vectors.

Willem and colleagues are the first to show a convincing neuropathological abnormality in BACE1 KO mice. An observation that the highest expression of BACE1 protein correlates with the onset of peripheral nerve myelination promotes them to examine the progression of myelination in the sciatic nerve of BACE1 KO mice. They find that axons of BACE1 KO mice are hypomyelinated from early postnatal stages to adulthood. Interestingly, mice deficient in cell-cell signaling protein type III neuregulin 1 (NRG1-β3) and its receptor, ErbB, display a very similar hypomyelination in peripheral axons, indicating a cross-talk between BACE1 and the NRG1-β3/ErbB signaling pathway. In line with this notion, membrane-bound NRG1 full-length protein accumulated in BACE1 KO mice and exogenous expression of BACE1 increased the release of the NRG1 ectodomain in culture, suggesting that NRG1-β is a novel substrate for BACE1. The cleavage by BACE1, or in combination with TACE, may result in the release of the EGF-like domain of NRG-β from neurons. This domain interacts with the receptor tyrosine kinase ErbB at the surface of Schwann cells, promoting the myelination process. It appears that BACE1 cleaves NRG1-β at the stalk region, but the precise cleavage site has not been revealed.

There are two major questions to be addressed in the future research. The first question is whether the lack of BACE1-mediated cleavage of NRG-β is solely responsible for the hypomyelination in BACE1 KO mice. Willem and colleagues’ findings do not completely rule out the involvement of other BACE1 substrates. For example, it will be interesting to examine whether the myelination is altered in APP KO mice, which display decreased locomotor activity and forelimb grip strength. The second question is whether BACE1 is involved in other functions of NRG1 family signaling molecules. NRG1 plays many essential roles in the CNS, heart and other peripheral tissues. It is important to revisit BACE1 KO mice to examine potential pathology in these systems.

In summary, Willem and colleagues reveal a novel function of BACE1 in myelination. The findings raise some concern about the safety of inhibiting BACE1 as a treatment for AD. Nevertheless, the generally healthy BACE1 KO mice still make BACE1 the best therapeutic target for the inhibition of Aβ production in AD.

View all comments by Huaibin Cai

It’s a Wrap; Axonal Myelination Is Regulated by the Alzheimer Disease Target, BACE
A fundamental developmental process has once again crossed paths with a major player in the pathogenesis of Alzheimer disease. Shortly after its discovery, BACE, via its interaction with neuregulin-1, has been implicated in the molecular neurobiology of central and peripheral axon myelination. Data from several labs have shown that specific members of the neuregulin-1 (NRG1) family of trophic factors are critical to Schwann cell differentiation, proliferation, survival, and now to the process of myelination itself. Whether axons are myelinated singly (and the number of myelin wraps required) or left unmyelinated and ensheathed in bundles, is governed by expression of the type III isoform of neuregulin-1 (Michailov et al., 2004; Taveggia et al., 2005; Chen et al., 2006; Ogata et al., 2004): It is the...  Read more

It’s a Wrap; Axonal Myelination Is Regulated by the Alzheimer Disease Target, BACE
A fundamental developmental process has once again crossed paths with a major player in the pathogenesis of Alzheimer disease. Shortly after its discovery, BACE, via its interaction with neuregulin-1, has been implicated in the molecular neurobiology of central and peripheral axon myelination. Data from several labs have shown that specific members of the neuregulin-1 (NRG1) family of trophic factors are critical to Schwann cell differentiation, proliferation, survival, and now to the process of myelination itself. Whether axons are myelinated singly (and the number of myelin wraps required) or left unmyelinated and ensheathed in bundles, is governed by expression of the type III isoform of neuregulin-1 (Michailov et al., 2004; Taveggia et al., 2005; Chen et al., 2006; Ogata et al., 2004): It is the expression level of NRG1 that communicates axon caliber to the Schwann cell.

Neuregulin-1 isoforms are ligands for heterodimeric combinations of ErbB receptor tyrosine kinases, specifically ErbB2, B3 and B4 (the EGF receptor is known as ErbB1). After binding to their cognate receptors, these ligands induce tyrosine phosphorylation and subsequent downstream activation of the PI3K (phosphatidyl inositol-3 kinase) pathway (Maurel and Salzer, 2000). The importance of neuregulin-1 cannot be overemphasized: pan-NRG1 as well as ErbB2, ErbB3, and ErbB4 KO mice are each embryonic lethals, and neuregulin-1 gene products in the CNS have been implicated in numerous other processes including neurotransmitter receptor regulation (see review by Falls, 2003).

Akt is the signaling node downstream of neuregulin and PI3K that is most implicated in survival and the specialization of Schwann cells and oligodendrocytes to form myelin (Flores et al., 2000; Li et al., 2001; Taveggia et al., 2005). Akt is already familiar to those studying neurodegeneration, motor neuron disease, and schizophrenia (Emamian et al., 2004; Humbert et al., 2002; Kaspar et al., 2003; Magrane et al., 2004). In Schwann cells, Akt transduces the neuregulin signal to inactivate the proapoptotic proteins Bad (Li et al., 2001) and GSK3β (Ogata et al., 2004), as well as to increase the expression of proteins that specify myelin differentiation: MAG (Ogata et al., 2004); P0; PMP22; and MBP (Chen et al., 2006). The transcription factors, Oct-6 and Krox-20 fill in the signal cascade (Taveggia et al., 2005).

What’s new in two recent papers is that type III neuregulin-1 has become the newest substrate of BACE-1, indicating this enzyme has an essential role to play in the decision to myelinate axons, and by how much. The studies by Hu et al. (2006) and Willem et al. (2006) have implicated BACE-1, the β-secretase enzyme for the amyloid precursor protein, as being a critical regulator of the levels of cleaved type III neuregulin-1. Hu et al. decided to examine myelination in BACE-1 deficient mice because BACE is transported into axons, along with APP, by a kinesin-1-dependent pathway. Willem et al. on the other hand noted that BACE-1 expression is highest in the developmental period corresponding to peripheral myelination. Both papers show that in the absence of BACE-1, myelination is reduced to the same degree as is seen in neuregulin hypomorphs (having only one copy of the NRG1 gene) or in conditional knockdowns of ErbB2. Measurement of the g-ratio (interior/exterior fiber diameter) shows a significant reduction in the number of myelin wraps associated with both central and peripheral axons when the BACE-1 gene has been inactivated. In the Hu et al. study, the decrease in myelination paralleled a reduction in expression of compact myelin proteins such as myelin basic protein (MBP) and the proteolipid protein (PLP). Willem et al. show through the added generation of BACE-1/2 compound mutants and BACE-2 mutant mice that BACE-1 activity alone is responsible for the hypomyelination phenotype. It is interesting that BACE-deficient mice also show abnormal bundling of small-diameter, unmyelinated axons (Willem et al.), in agreement with Taveggia et al., who use co-cultures of NRG1-/- neurites with Schwann cells to prove that NRG1 signaling is required for this critical function, too. Thus, normal ensheathment is not a default in the absence of NRG1 or BACE.

In both studies, the absence of BACE-1 leads to the accumulation of the inactive, full-length type III neuregulin-1 precursor and a corresponding decrease in the cleavage product or active ligand in brain. Presumably, BACE-1 cleaves the transmembrane precursor to release the soluble extracellular domain from the axons. Willem et al. produced a fusion between type III neuregulin-1 and secreted alkaline phosphatase to show that co-transfection with BACE-1 led to a direct increase in cleavage and release of the extracellular fragment. A likely direct interaction between BACE-1 and NRG1 was concluded. The ligand is then free to interact with and activate ErbB receptor tyrosine kinases on adjacent myelinating cells. How this diffusion takes place is presently unclear.

As expected from the loss of NRG1 cleavage in BACE-1-null mice, signaling in the PI3K/Akt pathway is reduced as revealed by the reduction in activated pAkt levels and drops in MBP and PLP in brain (Hu et al.). To prove that BACE-1-/- neurons cannot activate PI3K/Akt signaling in Schwann cells in a functional way might require some added work. This could be addressed, for instance, by showing that an axotomized nerve graft from a normal mouse can rescue myelination of a recovering host BACE-null axon (i.e., after transplantation into the transected defect of a peripheral nerve belonging to a recipient BACE-1-/- mouse) if prior gene transfer with myrAkt is attempted.

It should be emphasized that both authors found heterozygote BACE-1 mice to be phenotypically normal. Moreover, BACE-1 levels drop considerably in the adult state. These alone might suggest that partial inhibition of BACE-1, as advocated for AD therapy, would have no chance for adverse effects. Unfortunately, such mice still produce substantial amounts of β amyloid (Cai et al., 2001; Luo et al., 2001) and do not correct cognitive defects when crossed with APP/PS-1 transgenic AD mice (Laird et al., 2005). Thus, it will still be important to know the full impact of BACE-1 inhibition in the adult animal. A conditional KO model may be one approach to this question. However, there already exist cautionary signs. For instance, it is already known that BACE-1-deficient mice have a number of undesirable effects relating to impaired spatial reference memory and synaptic function (LTD-reversal) (Laird et al., 2005) as well as reduced pain threshold (Hu et al., 2006). BACE inhibition also has the theoretical effect of mitigating the protective back-signaling role of the NRG1-intracellular domain (Bao et al., 2003). Neuregulins may also have beneficial effects on APP metabolism and protection from Aβ (Rosen et al., 2003; Di Segni et al., 2005) that could be susceptible to BACE inhibition. Theoretically, remyelination could be impaired after certain central and peripheral injuries in BACE-1-inhibited patients. Ironically, loss of Notch signaling from inhibition of γ-secretase activity has somewhat dampened the enthusiasm of that approach to reduce Aβ production (Haass, 2004). With every “wrap” there seems to appear an interesting twist, which solves yet another puzzle but never the most important one—how do we treat Alzheimer disease?

View all comments by Henry Querfurth
View all comments by Kenneth Rosen