Part two of a four-part story. See part one.
BACE1 is the major β-secretase for amyloid precursor protein, making it a prime target for drug developers. However, because this protease cleaves many substrates, scientists worry about adverse effects of BACE inhibition (see part one). The BACE1 homolog BACE2 is thought not to play a role in AD, but it, too, cleaves many substrates in tissues outside the brain and most inhibitors known to date block both BACE1 and 2. What are those substrates and what functions might BACE inhibitors compromise? Those questions dominated BACE Proteases in Health and Disease, a meeting hosted by Stefan Lichtenthaler, Technical University of Munich, on October 6–8, 2013. The meeting attracted a Who's Who of BACE researchers from academia and industry.
Long a backwater to γ-secretase, recent high-profile discoveries and the entry of pharmaceutical companies have invigorated BACE research. While BACE1 knockout mice initially were reported to be mostly normal, subtle phenotypes soon emerged. Axonal guidance defects, retinal pathology, loss of dendritic spines and muscle spindles, seizures, hyperactivity, and memory defects are among the faults that have been reported in these mice. Some of this is developmental, but not all (see part three).
Researchers continue to identify potential BACE substrates, with the number now reaching dozens. Robert Vassar, Northwestern University, Chicago, one of the original discoverers of the secretase in 1999, summed up the million-dollar question facing the field: How much BACE inhibition is needed to benefit Alzheimer's disease, and will that amount be safe? Vassar cited the recent discovery of the A673T mutation that protects APP against BACE cleavage, and their carriers against AD (see July 2012 news story). "That suggests about 20 percent decrease of Aβ over production a lifetime might be sufficient to prevent AD," said Vassar, suggesting that 50 percent inhibition of BACE might be warranted for AD prevention, and more if amyloid has already accumulated. "To address the potential side effects, we need to sort out which of the symptoms in knockout mice are due to developmental versus later blockage of BACE," he said.
Substrates and Developmental Phenotypes
To understand what BACE1 does, Vassar’s group has localized the protease in the adult brain using immunohistochemistry and electron microscopy. They concluded that BACE1 abounds in presynaptic terminals, where it seems to be required for axon guidance (see Kandalepas et al., 2013). In BACE1 knockouts, olfactory sensory neuron axons fail to find their way to glomeruli in the olfactory bulb; also, infrapyramidal bundle mossy fibers projecting from the dentate gyrus turn porematurely toward their terminal area in the CA3 region of the hippocampus. In rodents, shorter infrapyramidal bundles compromise memory, noted Vassar.
Work from the lab of Mark Albers, Harvard Medical School, Boston, adds a twist to this role of BACE. Albers and colleagues reported that enhanced BACE1 cleavage of human APP contributes to miswiring in olfactory glomeruli (see Aug 2012 news story). In Munich, Albers reported that in the absence of APP or APP-like protein 2 (APLP2), another BACE substrate, olfactory neuron axons fail to target to the correct glomeruli. This mimics the axonal guidance deficits seen in BACE1 knockouts, suggesting that failure to process APP/APLP2 drives that phenotype.
Vassar believes that the unprocessed neural cell adhesion molecule CHL1 could underlie defects in BACE knockouts. CHL1 knockouts have almost identical phenotypes to BACE1 nulls, including axon guidance defects (see Hitt et al., 2012). Other researchers at the meeting also fingered CHL1 as a BACE substrate. Using proteomics approaches, researchers at Bart De Strooper’s lab at KU Leuven, Belgium, and Peer-Hendrik Kuhn and colleagues at Lichtenthaler’s lab identified CHLI and the related L1 as BACE1 substrates (see Jun 2012 news story). De Strooper said that BACE inhibitors cause full-length L1 and CHLI to accumulate in synaptosomes. He also noted that without CHL1 signaling, somatosensory thalamic neurons fail to project to their proper location in the brain, again suggesting guidance defects.
Bastian Dislich, Technical University of Munich, uncovered CHL1 with a mass spectroscopy-based proteomic approach that Vassar called a tour de force. Dislich raised three generations of BACE1 knockouts on chow laced with lysine containing heavy isotopes. He took brain extracts from those animals at postnatal day three, and mixed them with extracts from wild-type, age-matched mice that had been raised on regular food. By examining the ratio of heavy to light species in the mass spec, he identified peptides that were more abundant in the BACE1 knockouts. They would be likely substrates for BACE1. The strategy identified APP, APLP2, and a number of other membrane-bound proteins, including CHL1. Dislich said that about 30 percent of those hits function in axon guidance or neurite outgrowth. He plans to check these with Western blots to see whether the full-length proteins, or the ectodomains that would be cleaved by BACE1, accumulate.
Kuhn identified contactin-2, another cell-adhesion molecule, and seizure protein 6, which has ties to epilepsy. In BACE knockouts and in cells treated with BACE inhibitors, ectodomains of both of these proteins are poorly shed from cell surfaces. In Munich, Kuhn reported that BACE processes another cell membrane protein, delta/notch-like epidermal growth factor receptor, or DNER for short (Eiraku et al., 2002). DNER ectodomain shedding seems essential for proper brain development. Kuhn said that by processing DNER, BACE1 drives the differentiation of nestin-positive radial glial cells into neurons. In BACE1 knockouts and under BACE1 inhibition, these glia instead persist at postnatal day seven, disrupting the patterning of the neuronal cortex.
Riqiang Yan, Lerner Research Institute, Cleveland, who also discovered BACE at the same time as Vassar and others, described uncannily similar machinations are afoot with cell differentiation in the hippocampus of BACE1 knockouts. There he found increased astrogenesis at the expense of neurogenesis. “The knockout shifts the fate of neuronal stem cells,” he said. What drives that shift? In this case, enhanced Notch signaling seems to be the culprit. Yan reported that BACE1 cleaves the notch ligand Jagged1 from cell surfaces. In the absence of the protease, Jagged1, which abounds on granule cells of the hippocampus, accumulates and activates Notch on adjacent cells.
Scientists continue to uncover novel aspects to cleavage of neuregulin, a classic BACE substrate, as well. Christian Haass, Ludwig-Maximilians University, Munich, reviewed work his group published earlier this year showing that BACE1 cleavage helps release a soluble epidermal-like growth factor domain from neuregulin in neurons, which then signals through cell surface receptors on adjacent myelinating cells (see May 2013 news story). The soluble EGF domain can by itself rescue myelination defects in BACE1 knockout zebrafish, said Haass. His group found that this signaling pathway seems crucial for peripheral myelination only. Myelination in the central nervous system remain unaffected in BACE1 knockouts—at least in fish and mice.
Researchers at the meeting seemed in general agreement that CHL1 represents a bona fide BACE substrate, but many said it was too early to determine if other hits that have turned up in screens will prove to be true substrates. There could be indirect effects, for example, blocking cleavage of a real substrate could limit downstream processing of other proteins by other proteases. “Just because BACE seems to cleave a protein does not make that protein a substrate,” said De Strooper. The litmus test must be strict and must require that BACE1 colocalize with the protein and cleave it in a physiological setting, he said.
Most also agreed that it remains to be seen whether developmental phenotypes seen in BACE knockouts will have any bearing on what might happen in people treated with a BACE inhibitor. While some warned that developmental pathways often reactivate later in life, particularly during disease, most people at the meeting seemed to think that these are less relevant to BACE therapeutics than adult phenotypes. “We need a much better understanding of adult phenotypes as a way to predict side effects of BACE inhibitors,” concluded Lichtenthaler (see part three).—Tom Fagan
- Cloistered Retreat Takes the Pulse of BACE Research
- Blocking BACE—Do Adult Mouse Phenotypes Predict Side Effects?
- Protective APP Mutation Found—Supports Amyloid Hypothesis
- Aβ Scrambles Olfactory Wiring Before Plaques Form in Mice
- BACE Secrets: Newly Identified Substrates May Regulate Plasticity
- Paracrine Signal From BACE1-Clipped Neuregulin Rescues Myelin
- Kandalepas PC, Sadleir KR, Eimer WA, Zhao J, Nicholson DA, Vassar R. The Alzheimer's β-secretase BACE1 localizes to normal presynaptic terminals and to dystrophic presynaptic terminals surrounding amyloid plaques. Acta Neuropathol. 2013 Sep;126(3):329-52. PubMed.
- Hitt B, Riordan SM, Kukreja L, Eimer WA, Rajapaksha TW, Vassar R. β-Site Amyloid Precursor Protein (APP)-cleaving Enzyme 1 (BACE1)-deficient Mice Exhibit a Close Homolog of L1 (CHL1) Loss-of-function Phenotype Involving Axon Guidance Defects. J Biol Chem. 2012 Nov 9;287(46):38408-25. PubMed.
- Eiraku M, Hirata Y, Takeshima H, Hirano T, Kengaku M. Delta/notch-like epidermal growth factor (EGF)-related receptor, a novel EGF-like repeat-containing protein targeted to dendrites of developing and adult central nervous system neurons. J Biol Chem. 2002 Jul 12;277(28):25400-7. Epub 2002 Apr 11 PubMed.
No Available Further Reading