LRRTM3 promotes processing of amyloid-precursor protein by BACE1 and is a positional candidate gene for late-onset Alzheimer's disease

Abstract

Rare familial forms of Alzheimer's disease (AD) are thought to be caused by elevated proteolytic production of the 〈␤42 peptide from the ␤-amyloid-precursor protein (APP). Although the pathogenesis of the more common late-onset AD (LOAD) is not understood, BACE1, the protease that cleaves APP to generate the N terminus of A␤42, is more active in patients with LOAD, suggesting that increased amyloid production processing might also contribute to the sporadic disease. Using high-throughput siRNA screening technology, we assessed 15,200 genes for their role in 〈␤42 secretion and identified leucine-rich repeat transmembrane 3 (LRRTM3) as a neuronal gene that promotes APP processing by BACE1. siRNAs targeting LRRTM3 inhibit the secretion of A␤40, A␤42, and sAPP␤, the N-terminal APP fragment produced by BACE1 cleavage, from cultured cells and primary neurons by up to 60%, whereas overexpression increases A␤ secretion. LRRTM3 is expressed nearly exclusively in the nervous system, including regions affected during AD, such as the dentate gyrus. Furthermore, LRRTM3 maps to a region of chromosome 10 linked to both LOAD and elevated plasma A␤42, and is structurally similar to a family of neuronal receptors that includes the NOGO receptor, an inhibitor of neuronal regeneration and APP processing. Thus, LRRTM3 is a functional and positional candidate gene for AD, and, given its receptor-like structure and restricted expression, a potential therapeutic target.is the most common form of dementia and a debilitating neurodegenerative disease. In the majority of cases, the disease presents after age 65 because of largely unknown causes. However, in familial AD (FAD), the likely pathogenic trigger is release of amyloidogenic A␤ peptides from amyloid-precursor protein (APP), a transmembrane domain protein present in neuronal and other cells (1). A␤, the primary constituent of amyloid plaques, is generated from APP by proteolysis of the extracellular domain by BACE1 (Fig. 1A, 2), followed by intramembranous cleavage within its residual transmembrane domain by ␥-secretase [composed of presenilin-1 (PSEN1) or -2 (PSEN2) plus nicastrin (NCSTN), aph-1 (APH1), and pen-2 (PSENEN) (3)]. ␥-Secretase cleaves at multiple sites and yields mostly the 40-aa A␤40 peptide along with the more amyloidogenic 42-aa A␤42 (1). Alternatively, APP is cleaved by ␣-secretase within A␤, resulting in nonamyloidogenic metabolites. Mutations in APP or the presenilins account for most cases of FAD, and all increase A␤42 production from APP (1, 4, 5). These data are the basis of the amyloid hypothesis of AD.In late-onset (LO)AD, the etiology is not understood, but the major genetic risk factor is APOE4, which promotes A␤ deposition in the brains of mice (6), and BACE1 protein and activity are increased in LOAD (7-9). Furthermore, LOAD (10-12) and circulating A␤42 (13) levels were independently linked to chromosome 10q21-22, suggesting that genetic modifiers of LOAD risk also regulate A␤ metabolism. These data raise the possibility that genes regulating A␤ metabol...

LRRTM1 on chromosome 2p12 is a maternally suppressed gene that is associated paternally with handedness and schizophrenia

“…The LRRTMs are a four-member gene family first described in 2003, and it is interesting to note that LRRTM3 has recently been proposed as a susceptibility factor for late-onset Alzheimer's disease. 50 We recommend, therefore, that the whole LRRTM gene family is investigated in relation to psychiatric and neurological disorders.…”

High-Throughput Screening by RNA Interference: Control of Two Distinct Types of Variance

“…With an n = 1 per gene, it is impossible to determine the error of measurement, and then in turn to determine how much of the total variance is due to differences in the effects of genes on the pathway being studied. In a recently published paper, 12 individual siRNAs were found to have a coefficient of variance of roughly twenty (i.e., standard deviation roughly 20% of the mean levels for all readouts tested) calculated from multiple runs of positive controls and non-silencing negative controls. This fact, combined with the large number of genes tested, leads one to the conclusion that in a genome-wide siRNA screen for effects on a given biological process, the majority of hits are expected to be false positives.…”

“…Consider a screen with twenty thousand genes tested, searching for a reduction in the production of protein X. Assume an error of measurement as seen in Majercak et al, 12 and 0.1% of genes actually decreasing protein production by 40% or more. If 5% of the remaining 19,980 genes fall two standard deviations from the mean (as expected by chance), we should see almost 500 genes showing a reduction in protein X levels of 40% or more due to chance alone, resulting in roughly 95% false positives at this level.…”

“…For example, consider the study cited above. 12 This screen was designed to detect genes which affect the processing of the amyloid-precursor protein (APP), especially production of the Ab42 metabolite which has been implicated in the pathogenesis of Alzheimer's disease. Hit selection cutoffs based strictly upon biological variance are shown in Figure 1A.…”

“…This approach has been quickly embraced in both academic and industrial research. [7][8][9][10][11][12] High-throughput siRNA screens combine the issues particular to both the analysis of HTS and to the study of complex biological systems, and as a result, the researcher has two forms of variance to contend with: measurement and biological. Most experimental scientists are faced with biological variance on a daily basis, but are less concerned with error of measurement as this is often small relative to that found between the organisms in question.…”