Recognition sequence design for peptidyl modulators of beta-amyloid aggregation and toxicity.

beta-Amyloid (Abeta), the primary protein component of Alzheimer's plaques, is neurotoxic when aggregated into fibrils. We have devised a modular strategy for generating compounds that inhibit Abeta toxicity, based on linking a recognition element for Abeta to a disrupting element designed to interfere with Abeta aggregation. One such compound, with the 15-25 sequence of Abeta as the recognition element and a lysine hexamer as the disrupting element, altered Abeta aggregation kinetics and protected cells from Abeta toxicity [Ghanta et al. (1996) J. Biol. Chem. 271, 29525]. To optimize the recognition element, peptides of 4-8 residues composed of overlapping sequences within the 15-25 domain were synthesized, along with hybrid compounds containing those recognition sequences coupled to a lysine hexamer. None of the recognition peptides altered Abeta aggregation kinetics and only two, KLVFF and KLVF, had any protective effect against Abeta toxicity. The hybrid peptide KLVFF-KKKKKK dramatically altered Abeta aggregation kinetics and aggregate morphology and provided significantly improved protection against Abeta toxicity compared to the recognition peptide alone. In contrast, FAEDVG-KKKKKK possessed only modest inhibitory activity and had no marked effect on Abeta aggregation. The scrambled sequence VLFKF was nearly as effective a recognition domain as KLVFF, suggesting the hydrophobic characteristics of the recognition sequence are critical. None of the cytoprotective peptides prevented Abeta aggregation; rather, they increased aggregate size and altered aggregate morphology. These results suggest that coupling recognition with disrupting elements is an effective generalizable strategy for the creation of Abeta inhibitors. Significantly, prevention of Abeta aggregation may not be required for prevention of toxicity.

Treatment of spontaneously arising retinoblastoma tumors in transgenic mice with an attenuated herpes simplex virus mutant.

The use of viruses to treat tumors has received renewed interest with the availability of genetically defined attenuated mutants. Herpes simplex virus (HSV) type 1 in particular has been shown to be effective for tumors of neuronal origin. However, the model systems used for these studies rely on the use of explanted tumor cells in immunodeficient animals. We have used a recently developed transgenic mouse model, wherein mice spontaneously develop retinoblastomas, to determine if a mutant HSV has a therapeutic effect against an endogenously arising tumor in an immunocompetent host. The injection of 1 x 10(6) PFU of the neuroattenuated HSV-1/HSV-2 recombinant RE6 into the vitreous of transgenic mice resulted in a significant inhibition of tumor growth compared to injection of medium alone (P = 0.0063). Immunohistochemical analysis of viral antigen showed that viral replication was restricted to focal areas of the tumors and the retinal pigment epithelium. Viral growth was not significantly different in the eyes of transgene-positive and transgene-negative mice, suggesting that enhanced replication in tumor cells may not explain the effects. Tumor cells in the treated eyes were significantly less differentiated than those in the untreated eyes (P = 0.04), suggesting that the virus may replicate better in certain cell types in the tumors. Although the injection of RE6 resulted in a difference in tumor size, the treatment did not result in the elimination of tumors in any of the mice improvements in the efficacy of tumor control are needed if this therapy is to be of use.