Amyloid plaques are still the main attraction, and tangles have certainly not left the stage, but there is still a lot of interest at this meeting (off-Broadway, so to speak) in other potential players in AD pathology. An example is ApoE, widely recognized as the most common risk factor for the disease. A panel of experts provided their views on the potential role of this protein in AD pathology. Although there was a disproportionate amount of previously published work presented by most speakers, there were enough new results (especially from Weisgraber) to maintain the interest of the relatively small audience. Ever since the genetic association between the ApoE4 isoform and AD risk was discovered by Allen Roses and colleagues, a number of laboratories have been developing and testing hypotheses as to what this lipid-transport protein might be doing. As pointed out by Judes Poirier, one of the presenters, these hypotheses fall into two broad classes: ApoE (especially E3) is good or ApoE (especially E4) is bad. The majority of hypotheses are based on the notion that E4 is simply not as good as E3 in performing some vital function.
Whether or not E4 is bad or just ineffective, Weisgraber described strategies for studying its effects in a transgenic mouse model and for making it more like E3. This group has previously found that "domain interactions" between Arg-61 and Glu-255 are important in making E4 behave differently from E2 or E3. The novel mouse is one in which mouse ApoE has been "humanized" by replacing the normal threonine at position 61 with arginine. (Of several species examined, human ApoE is the only one with arginine at this position.) The ApoE from this humanized mouse shows preferential association with VLDL, as does human ApoE4. The humanized gene is under the control of the natural mouse enhancers and tissue-specific elements and there is no endogenous mouse ApoE expressed. These mice will be studied as potential models for biological effects of ApoE4. Weisgraber also described a drug-screening strategy to find compounds that will specifically interfere with the domain interaction present in E4. A molecular modeling approach (the DOCK program) was used to assess how well putative ligands fit into the site that mediates interaction between the critical residues in E4. 65 compounds showed inhibition of the VLDL-binding preference of E4 when tested in an emulsion assay. Thirteen of these compounds reduced binding activity of E4 to that normally obtained with E3 but had no effect on E3 binding. The goal is to identify compounds that make E4 behave more like E3 as a means of treating the ApoE-related pathology of AD.
David Holtzman followed with a review of mostly published studies involving several transgenic mice in which APP is overexpressed in the presence or absence of different ApoE backgrounds. The absence of enodgenous ApoE results in less Aß deposition in brain, with almost no fibrillar Aß and no dystrophic neurites at 12 and 15 months of age. In the presence of ApoE, however, abundant thioflavin S-positive plaques with dystrophic neurites are present. Lack of ApoE also reduces vascular amyloid deposition. When human ApoE is expressed in APP mice (without mouse ApoE), there is initial suppression of amyloid deposits. However, E4 mice show 10-fold greater Aß deposits and neuritic plaques than E3 mice at 15 months of age. This isoform-specific effect is pronounced at 21 months when the E4 mice show significant neuritic dystrophy. The hypothesis is that ApoE plays a critical role in conversion of Aß to fibrillar (toxic) forms and that E4 may be doing this more than E3. (Surprisingly, Holtzman did not mention the possibility that ApoE may itself be a source of neurotoxic activity.) He also presented more recent data on the levels of Aß in CSF collected from the mouse cisterna magna (a heroic task!). There is a decline in CSF Aß in the E4 mouse over time, perhaps due to deposition or clearance and Aß levels are correlated in CSF and plasma, suggesting that clearance into plasma may be occurring.
Poirier provided an overview of studies based on the assumption that ApoE normally plays a positive role that needs to be enhanced in order to modify the pathology of the disease. He noted that others have shown that plasma ApoE is lowest in E4 homozygotes. Poirier's group has shown the same result in AD patients with similar trends for hippocampal ApoE levels (lowest in 4/4 samples). Low throughput screening was used to identify compounds that will increase ApoE production in rat astrocytes on the assumption that boosting ApoE will be of benefit in treating AD. A wide variety of agents were identified, one of which, Probucol, was found to be especially potent in vitro and was also found to cause increased ApoE in mouse brain in vivo. An initial clinical trial has been carried out with Probucol in AD patients in which ApoE levels in lumbar CSF showed an average increase of 20% but the increase was greatest in subjects lacking the E4 allele. Interestingly, the cognitive status of these patients appeared to stabilize and there was a strong positive correlation between cognitive performance and CSF ApoE.
Bales returned to transgenic mice, reviewing mostly published data demonstrating that the absence of ApoE results in no thioflavin S plaques and no neuritic dystrophy in transgenic amyloid mice. Also, there is a different pattern of amyloid staining in the absence of ApoE. When ApoE is present, it appears in all the thioflavin S-positive plaques (as it does in AD plaques). Replacing mouse ApoE with human apoE demonstrates that E4 mice develop thioflavin S-positive plaques, but there are none in E2 or E3 mice. She also reported on APP mice that were fed a fat-enriched diet for 12 weeks. This resulted in more Aß staining and more thioflavin S-positive plaques.
Mucke also covered transgenic territory in the context of understanding gender- and age-dependent effects of ApoE on disease risk. Women are more affected than men by the presence of E4, although this effect is less prominent at early and later ages. In APP mice, dystrophic neurites are associated with plaques and there is an age-dependent decline in synaptophysin staining [SYN-IR], which negatively correlates with Aß42 levels. Hippocampal slices from these mice exhibit deficits in basal synaptic transmission but these effects are independent of plaque load. One possible explanation is that plaque formation is separate from degenerative effects of Aß. When combined with ApoE transgenic mice (under control of GFAP or NSE), behavioral deficits emerge in the E4 mice but not the E3 mice. In fact, E4 may exert a dominant negative effect under some conditions, perhaps through effects on Aß, but there also may be differences in the ApoE isoforms on neuronal protection. He also reported on gene chip analyses to look for changes in hippocampal gene expression in these mice. 618 of 13,033 screened genes showed changes. It remains to be determined which ones will be informative.
Smith rounded out the symposium with a return to the gender gap as well as evaluating the contribution of trauma (head injury is a risk factor for AD). Mice in which ApoE expression is driven by the GFAP promoter were studied. Estrogen results in induction of ApoE in the brain, but with regional variations. Ovariectomy results in increased mortality in APP mice, an effect that is apparently not due to increased Aß. Using a fluid percussion head injury model, induction of ApoE and GFAP were found. He also examined the contribution of environmental or genetic variations to phenotypic results obtained with transgenic lines. C57BL/6 and FVB/N mice were studied with and without endogenous ApoE. The FVB/N mice are good breeders but harbor the rd mutation that causes blindness. Using an olfactory-cued eight-arm radial maze (does not require vision), all six-week-old mice showed learning over five days. However, at six months, only the C57 ApoE-/- mice show deficits, which are not due to increased corticosterone. Thus, both age-dependent and genetic strain-dependent effects on cogniitve behavior exist and need to be considered when evaluating transgenic models.
If there was a common message in this symposium, it was hard to decipher. ApoE clearly has an effect on amyloid deposition and there are interesting isoform-specific effects of ApoE on transgenic mouse phenotypes. However, whether ApoE ever gets the chance to share the limelight with amyloid or tau is unknown and will likely depend on a unifying hypothesis that still remains to be established for ApoE's role.—Keith Crutcher
No Available References
No Available Further Reading