Viable mouse gene ablations that robustly alter brain Aβ levels are rare

Abstract

Background: Accumulation of amyloid-β (Aβ) peptide in the brain is thought to play a key pathological role in Alzheimer's disease. Many pharmacological targets have therefore been proposed based upon the biochemistry of Aβ, but not all are equally tractable for drug discovery.

Results: To search for novel targets that affect brain Aβ without causing toxicity, we screened mouse brain samples from 1930 novel gene knock-out (KO) strains, representing 1926 genes, using Aβ ELISA assays. Although robust Aβ lowering was readily apparent in brains from a BACE1 KO strain, none of the novel strains exhibited robust decreases in brain Aβ, including a GPR3 KO strain, which had previously been proposed as an Aβ target. However, significantly increased Aβ was observed in brain samples from two KO strains, corresponding to genes encoding the glycosylphosphatidylinositol mannosyl transferase PIGZ and quinolinate phosphoribosyltransferase (QPRT).

Conclusions: Thus, gene ablations that are permissive for mouse survival and that also have a robust effect on Aβ levels in the brain are rare.

Figure 1

Outline of screening tiers. Aβ40 in four brain samples (sagittal left halves) from 1930 different mouse KO strains were determined. Of these, 77 KO strains were selected for further Aβ40 determinations, and in some cases Aβ42 determinations, of the corresponding right brain halves. The gene information for four of the KO strains is reported in the current study. Prioritization for possible drug discovery efforts was then based upon multiple considerations, including the robustness of Aβ inhibition, tractability of the target for small molecule inhibition, and relevance of the target to AD.

Figure 2

Primary screen of Aβ40 levels. Panel A - The median values of Aβ40 relative to baseline for each of 1930 KO strains is shown in red. Assay plate values for BACE1, BACE2 double KO brains are shown in blue, and values for Aβ1-14 synthetic peptide-blocked wild type brains are shown in green. The results for the PIGZ KO and the QPRT KO are indicated by arrows. Panel B - Same as panel A, except mean values of Aβ40 are plotted.

Figure 3

Aβ levels in left and right brain halves of selected KO strains. Panel A - Median values of Aβ40 in right brain halves were plotted against median values in left brain halves. An ellipse highlights the results for UBE2R2 and ADRM1. The results for the PIGZ KO and the QPRT KO are indicated by arrows. Panel B - Same as panel A, except that mean Aβ values were plotted. Panel C - Relative values of Aβ40 and Aβ42 from left and right brain halves in individual animals is shown for four named KO strains. Note that left and right are not directly comparable because different extraction and assays procedures were used (see Methods). Bars represent the mean value.

Figure 4

Effect of GPR3 KO on brain Aβ levels in young mice. Panel A - Aβ42 was assayed in brain extracts from ten GPR3 wild type (GPR3+/+), ten heterozygous (GPR3+/-), and ten homozygous GPR3 KO (GPR3-/-) mice. Brain Aβ42 was also determined in BACE1, BACE2 homozygous double KO (B1/2 dKO) and wild type (BACE+/+) control mice. Bars represent the mean values. Panel B - the same brain extracts were assayed for Aβ40. Panel C - the same brain extracts were assayed for Aβ1-x.

Figure 5

Effect of GPR3 KO on brain Aβ levels in aged mice. Aβ40 and Aβ42 were determined in one year aged mice for 13 GPR3 wild type (G3+/+) and 13 homozygous GPR3 KO (G3-/-) mice. For comparison, Aβ40 and Aβ42 for five young wild type (B1/2+/+) and four young BACE1, BACE2 homozygous double KO (B1/2 dKO) are shown. Values are expressed as percentage of mean value in the GPR3 wild type mice. One individual GPR3 wild type mouse with a value of 624% was excluded from the analysis.