In the study, the four first authors, Yun-wu Zhang, Shijie Liu, Xue Zhang, and Wu-Bo Li, combined elegant molecular, cell, and animal approaches to define the new pathway. They first undertook a functional screen for genes that decrease Aβ production in mouse cells using retroviral-mediated gene disruption. The insertion vector carries a strong promoter, which can drive transcription when it lands in or near a gene. Insertion downstream of genes results in a knockout phenotype via antisense transcription, while upstream insertion can result in a gene overexpression phenotype. After viral infection, N2a cells that express a mutated human amyloid precursor protein (APPSwe) were screened for γ-secretase inactivation by sorting for cells with increased cell surface expression of APP βCTF (C-terminal fragment, the substrate for γ-secretase).

The researchers focused on one clone that showed a high level of cell surface APP bCTF and a significant decrease in Aβ. They found the changes were due to the activation of a novel mouse gene by the viral insertion. This gene, Rps23r1, was unusual in that it originated by retroposition, a process by which mRNAs reverse transcribe and re-insert into the genome at random, creating new genes. Rps23r1 is a copy of a gene encoding a mouse ribosomal protein, but transcribed in the reverse orientation compared to its parent. That gives rise to a protein of entirely new sequence, and the Rps23r1 product proved to be a transmembrane protein mainly localized to the hippocampus, dentate gyrus, and cortex.

When the researchers overexpressed Rps23r1 in cells, they found it caused inhibition of GSK3α and β, two related kinases upstream of both Aβ production (see ARF related news story on Phiel et al., 2003; Su et al., 2004) and tau phosphorylation. The addition of the GSK3 inhibitor lithium to the cells did not further reduce kinase activity or Aβ production, suggesting that Rps23r1 has its effects via inhibition of GSK3. Tracing the actions of Rps23r1 farther upstream, the authors showed that the protein interacts with adenylate cyclase, boosts cAMP levels, and activates PKA. This results in inhibitory phosphorylation of GSK3, reduced Aβ production, and decreased tau phosphorylation.

To look at Rps23r1 action in vivo, the researchers made a transgenic mouse that overexpressed the protein in brain by twofold over endogenous levels. The mice had decreased GSK3 activity in brain, and increased CREB phosphorylation (an indicator of elevated PKA). When the mice were crossed with triple-transgenic mice expressing human mutated APP, tau, and presenilin 1 (Oddo et al., 2003), their offspring had increased levels of cAMP and PKA activity, and decreased GSK3 activation, tau phosphorylation, and Aβ levels. Levels of the synaptic marker PSD-95 and synapsin were both increased, the authors reported, suggesting that synaptic impairment may be ameliorated in the mice.

Despite their efforts, the researchers were unable to identify a human homolog of the mouse gene, but they did demonstrate that the Rps23r1 could regulate GSK3 and Aβ levels in human cells.

Scott Small of Columbia University, New York, calls the study an “important proof of principle.” Small, who was not involved in the work, told ARF, “Even though the new gene doesn’t clearly exist in humans, the findings do have immediate therapeutic implications because they suggest there could be a single drug that hits both pathologies simultaneously.” The study supports a “dual pathway” hypothesis for AD, as outlined by Small and Karen Duff in a review published last year (Small and Duff, 2008, and a related Live Discussion). The hypothesis suggests that Aβ and tau pathologies are linked through common upstream events, rather than viewing tau perturbation as a sequential consequence of Aβ overproduction (the amyloid hypothesis).

For his part, Xu speculates that the mouse-specific gene might explain the mystery of why mice do not get AD. While there are other possible explanations, the protection afforded by PKA activation and GSK3 inhibition could be a contributing factor. “Our biggest challenge now is to do a knockout,” Xu said. “We want to see if you knockout this gene, will you get Aβ buildup or an AD-type condition in the mice?”

The random targeting approach has yielded other interesting genes as well, which Xu says the group is characterizing now.—Pat McCaffrey.

Reference:
Zhang Y, Liu S, Zhang X, Li W, Chen Y, Huang X, Sun L, Luo W, Netzer WJ, Threadgill R, Wiegand G, Wang R, Cohen SN, Greengard P, Liao F, Li L, Xu H. A functional mouse retroposed gene Rps23r1 reduces Alzheimer's b-amyloid levels and tau phosphorylation. Neuron. 2009 Nov 12; 64:328-340. Abstract