Accelerating Treatment Development for Psychosis in AD: MODEL AD+P Psychotic symptoms occur in ~ 40-60% of individuals with Alzheimer Disease (AD with psychosis, AD+P). Numerous studies have found that the AD+P phenotype is associated with more rapid cognitive decline than AD subjects without psychosis (AD-P). Current, empirically developed treatments for psychosis in AD have limited efficacy, do not alter the more rapid disease progression, and are associated with substantial toxicity, including excess mortality. The goal of this renewal is to close several critical gaps in knowledge that currently inhibit effective treatment development for psychosis in AD: delineation of the neurobiologic signature distinguishing AD+P from AD-P; identification of drugs that reverse this signature, and; development of construct and face valid models of AD+P in which to evaluate these drugs. We look to achieve this goal by building on a series of accomplishments from the current funding period, in which we: 1) Identified a genome-wide significant locus for AD+P risk in SUMF1 and found reduced SUMF1 mRNA expression in excitatory neurons in AD+P; 2) Confirmed associations of AD+P with excess phosphotau burden and newly demonstrated that AD+P is associated with excess loss of dorsolateral prefrontal cortex glutamatergic neurons relative to AD-P; 3) Showed that AD+P (relative to AD-P) is further characterized by reduced postsynaptic density (PSD) yield and profound loss of PSD proteins exceeding the loss of glutamate neurons; 4) Characterized the PSD protein signature of AD+P relative to AD-P, finding reduced levels of a network of protein kinases and regulators of the actin cytoskeleton, and; 5) Generated computational predictions of drugs that may reverse the PSD protein signature of AD+P. We will now conduct a set of integrated experiments that will: Aim 1) more finely resolve signaling alterations in AD+P by characterizing the PSD phosphoproteome and kinome signature of AD+P; Aim 2) test computationally identified drugs for effects on the PSD protein and phosophoprotein signature in mouse models of AD pathology, and; Aim 3) reduce SUMF1 expression in mouse models of AD pathology to evaluate if this manipulation recapitulates the neurobiology of psychosis in AD and could thus serve as a construct and face valid model to evaluate for predictive validity in future pharmacological studies. If successful, we will have more finely characterized the PSD biology of AD+P, validated several available drugs as potential novel treatment candidates, and established the first construct and face valid model of AD+P for drug testing. We will also share with the larger research community extensive data on the PSD proteome and phosphoproteome of: AD+/-P, normal elderly, multiple mouse models of AD, and drug treatments. These outcomes will help accelerate development of experimental therapeutics for AD+P.