Search

Why do neurodegenerative diseases tend totarget specific brain regions? To begin toaddress this phenomenon known as selectivevulnerability we combined two powerful tools,an allelic series of Prnp knock-in mouse lineswith the high precision of mouse adapted scrapie strains. The first mouse line modeled fatalfamilial insomnia (FFI), and in old age spontaneously developed neuronal loss and reactivegliosis in the thalamus. The second line modeled Creutzfeldt-Jakob disease (CJD), spontaneously developing PrPres aggregates, spongiosis,and reactive gliosis in the hippocampus. In bothmodels the cerebellum was also affected: atrophied in the FFI model and loaded with PrPresaggregates in the CJD model. A third mouseline expressed a mutated N-terminal polybasicregion (PBR), but was disease free. We therefore challenged our allelic series with two scrapie strains with similar incubation periods(RML and 22L), one of which was reported to aggressively target the cerebellum in vivo(22L). Based on incubation periods, the FFImice were highly resistant to both scrapiestrains. In contrast, CJD mice were resistant to22L but very sensitive to RML. The oppositewas true for PBR mice which were highly resistant to RML but much more prone to 22L thanFFI and CJD mice. Interestingly, multiplemarkers of neurodegeneration revealed a divergence of responses to the two strains in variousbrain regions. In particular, various markers ofgliosis appeared in different brain regionsdepending on the mouse line:scrapie strain combination, suggesting the existence of a diversityof glial phenotypes in these brains.

One of the most remarkable features of human prion diseasesis that different prion protein (PrP) mutants lead to very different familial disease forms, with different shapes of PrP aggregates, different degrees of spongiform degeneration, and differentlocations of neuronal loss and reactive gliosis. In recent years wehave developed an allelic series of PrP knock-in mice to studyhow mutations confer toxic properties on PrP. We have recentlycompared and contrasted mice modeling fatal familial insomnia(FFI) and familial Creutzfeldt-Jakob disease (CJD). Both mutations caused neurodegenerative processes but the diseases werequite different. The FFI model had neuronal loss, reactive gliosis,and proteinase K sensitive PrP aggregates, predominantly in thethalamus, while the CJD model had spongiosis and proteinase Kresistant PrP aggregates, most prominently in the hippocampus.Based on neuropathological and behavioral changes, we judgedthe FFI mutation to be more toxic. We hypothesized that the FFImutant’s higher toxicity was due to its proximity to one of the cysteines comprising PrP’s sole disulfide bridge, thereby interferingwith bridge formation and resulting in a destabilized and toxicspecies. In support of this conjecture, the steady state level of theFFI mutant PrP is reduced ~ 70%, consistent with most moleculesnot folding properly and being triaged for degradation soon aftersynthesis. We further hypothesized that complete elimination ofthe bridge via mutagenesis would result in a disease resemblingthat of the FFI model, but more severe. We tested this hypothesisby developing knock-in mice expressing PrP with both cysteinesreplaced with serines, which are structurally similar. These micedeveloped disease spontaneously, but surprisingly, it was morelike that of the CJD model, with wide-spread spongiosis. In aneffort to create a version with more severe disease we resorted tothe commonly used random integration strategy to over-express the same PrP mutant. Surprisingly, despite increased expressionlevels, the random integration mice were normal in all regards,notably lacking the spongiosis present in the knock-in line.Moreover, the random integration line was surprisingly effectiveat delaying scrapie infection, dominantly. These results indicatethat the mechanisms of toxicity of mutant PrPs are difficult topredict and can be altered when PrP is expressed ectopically.