Though far from clear-cut, connections between ApoE, Aβ, and LDLR have long captivated AD researchers. According to one theory, ApoE serves as a molecular chaperone, coaxing soluble Aβ to aggregate into amyloid (Wisniewski and Frangione, 1992). Consistent with this idea, studies by Holtzman and others have shown that lack of ApoE reduces amyloid deposition in AD mouse models (e.g., Bales et al., 1997; Holtzman et al., 2000; Cao et al., 2005). Work from the Holtzman lab put LDLR on the map, showing that it is the major receptor for uptake and degradation of ApoE in the mouse brain. In support of this, the researchers found increased levels of brain ApoE in LDLR-deficient mice (Fryer et al., 2005 and ARF related news story).

In the new study, first author Jungsu Kim and colleagues set out to reduce ApoE levels by ramping up LDLR expression in mice. They used a prion promoter to overexpress an LDLR transgene throughout the brain, and kept low-expressing (twofold) and high-expressing (11-fold) lines for further analysis. As expected, the scientists saw decreased levels of brain ApoE, measured by ELISA, in all LDLR transgenic lines. Relative to non-transgenic controls, mice with doubled LDLR expression had about half as much brain ApoE, and mice with fivefold and more LDLR had but 10 to 20 percent ApoE left. The ApoE-lowering effect held when the high-expressing LDLR transgenic line was crossed with APP/PS1 mice, an AD model that overexpresses mutant forms of amyloid precursor protein and presenilin-1.

Mirroring past findings of reduced amyloid deposition in ApoE-deficient AD mice, Holtzman’s team found fewer Aβ plaques in the cortex and hippocampus of APP/PS1/LDLR mice, relative to APP/PS1 littermates. This was measured using 3D6 anti-Aβ antibody or X-34 dye, which detects fibrillar amyloid. LDLR-overexpressing APP/PS1 mice also showed less neuroinflammation. Brain sections of these mice had a 70 percent decrease in CD11b-positive microglia, and only about half as much of the area staining positively for the astrocyte marker GFAP.

The new study also explores how LDLR overexpression drives down amyloid load. Using in vivo microdialysis, the researchers measured soluble Aβ levels in APP/PS1/LDLR (the high-expressing line) and APP/PS1 littermates at 2.5 months of age, when amyloid deposition has not yet begun. Steady-state Aβ levels in the brain interstitial fluid (ISF) were about 40 percent reduced in LDLR-overexpressing AD mice compared to littermates with wild-type LDLR levels. Then, by treating the mice with a γ-secretase inhibitor to block further Aβ production, the researchers were able to more accurately measure the rate of Aβ elimination from the ISF. They found that APP/PS1/LDLR transgenic mice were clearing Aβ about twice as quickly as the APP/PS1 mice.

Though previous reports have suggested that ApoE regulates clearance of brain Aβ (e.g., Bell et al., 2007; Deane et al., 2008), the current study does not definitively establish that ApoE is required for reduced amyloid load in APP/PS1/LDLR mice. “We showed LDLR overexpression decreases ApoE protein levels, and that LDLR overexpression increases Aβ clearance. What we haven’t done, but are currently working on, is show that the effect of LDLR on Aβ is mediated by ApoE,” Kim said.

Another looming question is whether the Aβ-reducing effects in the current study would also occur in mice expressing human ApoE, several scientists noted. Past research in AD mouse models suggests that ApoE’s effect on amyloid load differs markedly, depending on which form of ApoE is present. Here is the conundrum, in a nutshell. In PDAPP mice, knockout of ApoE reduces amyloid pathology (Bales et al., 1997). However, as Holtzman’s lab later showed, targeted knock-in of human ApoE into these mice does not re-establish normal amyloid load as one might expect, but instead further delays amyloid deposition (Holtzman et al., 1999 and ARF related news story). Holtzman told ARF that his lab has ongoing studies to address the effect of LDLR overexpression on AD mice that lack mouse ApoE but have the human ApoE gene knocked in. “I think LDLR will definitely lower ApoE3 and ApoE4,” he wrote in an e-mail, noting that both are strong ligands for LDLR. “I also think it is likely it will lower Aβ in a similar fashion as in our current experiments, even if human ApoE3 or E4 is present.”

Curiously, Aβ load in the current study differed greatly between the sexes: male APP/PS1 mice had less than half as much amyloid deposition as did their sisters. “If I didn't analyze females and males separately, I probably wouldn't have seen the difference (due to presence or absence of LDLR),” Kim told ARF. Female APP/PS1 mice also had more brain ApoE than males: Within the overall decrease, ApoE levels were up 50 to 60 percent in female versus male progeny of crosses to high-expressing LDLR mice, and up more modestly in females born of crosses to LDLR low expressors. Other groups have reported similar sex-skewed effects on Aβ aggregation in other APP transgenic models (Callahan et al., 2001; Wang et al., 2003), and these observations are consistent with epidemiological data suggesting women are more prone to develop AD (Andersen et al., 1999). Perhaps the elevated ApoE levels in female APP/PS1 mice relates to their increased amyloid load, the authors speculate, though they do not address this in the study.

Kim said the WashU scientists are extending their analyses to look at whether LDLR overexpression affects cognition, in addition to lowering Aβ, in the AD mice. Furthermore, he and colleagues are trying to see whether these effects can be triggered by several genes that have been shown to regulate the LDLR protein (Soutar and Naoumova, 2007).

The current data already suggest that boosting LDLR levels or activity could be an attractive approach for AD treatment. “The really cool thing from this paper…is they only had to increase LDLR expression twofold [to see effects on Aβ],” said Mary Jo LaDu of the University of Illinois at Chicago. “That predicts a really nice potential for a treatment strategy.” Another favorable feature is that the other major LDLR ligand, ApoB, is too big to cross the blood-brain barrier. Thus, increasing LDLR would likely only affect ApoE and not ApoB.

However, figuring out how to target LDLR over other related proteins may prove challenging. “It comes down to specificity. This is a clean genetic approach, but you can’t do that in humans,” LaDu said, noting that four members of the LDLR family are expressed in the brain.

Yadong Huang, who studies ApoE at the Gladstone Institute of Neurological Disease, San Francisco, found the study “quite exciting” on a broader level. Last week, his lab published a study in Cell Stem Cell showing that ApoE4 operates via GABAergic signaling pathways to impair neurogenesis in mice (Li et al., 2009 and ARF related news story). Huang said he thinks ApoE is still understudied, and that the field underestimates its role in AD pathogenesis. “I hope increased interest in ApoE research is the trend,” he said, noting that 60 to 80 percent of AD patients have at least one copy of the high-risk E4 allele.—Esther Landhuis.

Reference:
Kim J, Castellano JM, Jiang H, Basak JM, Parsadanian M, Pham V, Mason SM, Paul SM, Holtzman DM. Overexpression of Low-Density Lipoprotein Receptor in the Brain Markedly Inhibits Amyloid Deposition and Increases Extracellular Abeta Clearance. Neuron. 2009 Dec 10;64:632-44. Abstract