For the first time, it is reasonable to speculate on the interacting pathways contributing to HDL2 pathogenesis. Data that need to be accounted for include the phenotypic and genetic similarity to HD and other polyglutamine disorders (rather than to DM1), the presence of protein inclusions staining with 1C2 antibodies, and the mild phenotype of JP3 knockout mice. We propose the tentative model shown in Fig. 8, recognizing that some aspects of the model are almost certainly wrong. The left-most pathway depicts the possible role of poly-amino acid tracts, either polyalanine or polyleucine encoded from JPH3, or polyglutamine cryptically encoded from the reverse strand. The dotted lines indicate that this pathway is not well supported by the available data. However, this pathway is the most obvious explanation for the protein aggregates observed in HDL2. Perhaps a small amount of protein with an expanded polyglutamine tract, undetectable by conventional assays, is sufficient to seed aggregates which are ultimately composed of other proteins. Toxicity would stem from one or more of the mechanisms postulated to contribute to other polyglutamine diseases.

The righthand pathway depicts the possible role of loss of JPH3 expression. Assays of JPH3 transcript and protein in HDL2 brain provide modest support for loss of expression. Decreased expression could arise from a direct effect of the repeat on transcription or splicing, or by sequestration of the transcript into RNA foci. Loss of expression might lead to incomplete construction of junctional complexes, with disruptions in calcium flux and increased vulnerability to factors such as excitotoxity and metabolic stress. Loss of JPH3, a membrane protein, might contribute to the formation of acanthocytes, but perhaps only in predisposed individuals.

The largest arrows depict a pathway of RNA toxicity. The CUG repeat leads to RNA foci formation. Whether these foci themselves contribute to toxicity is unclear, but the net result is a decrease in functional MBNL1, with subsequent effects on splicing of many genes, destabilizing neurons. CUG expansions may induce toxicity through mechanisms other than MBNL1 and splicing dysregulation. We predict that cells subject to CUG repeat toxicity may aggregate protein, partially accounting for HDL2 protein aggregates.

This model is speculative, but provides a basis for further exploration of HDL2 pathogenesis.

Acknowledgements The authors would like to thank Drs. Susan E. Holmes, Christopher A. Ross, Ruth H. Walker, Amanda Krause, Charles Thornton, Olga Pletnikova, Juan Troncoso, Adam Rosenblatt, Nancy Sachs, and Elizabeth O'Hearn for their insights into HDL2. We also thank the HDL2 families who have so willing cooperated with our investigations. This work was supported by the Hereditary Disease Foundation and NIH NS16375.

0 0

Post a comment