Mutations in the CLN3 gene cause Juvenile Neuronal Ceroid Lipofuscinosis (JNCL), a form of Batten disease that is grouped within the broad class of lysosomal storage diseases. JNCL displays a primary central nervous system phenotype characterized by rapid onset blindness, wide spread brain atrophy and reversal of learned abilities with death occurring 10-20 years after symptom onset. The mechanisms underlying these phenotypes are not known. CLN3 encodes CLN3, a protein with no known molecular function. CLN3 is expressed at very low levels natively in most cells, and is highly hydrophobic. Similar to other lysosomal storage diseases, it is difficult to ascertain the primary versus the secondary defects when the protein functions along the endosomal-lysosomal pathway. In JNCL one common finding among several labs, in various cellular systems, is a fluid-phase endocytotic defect. I took this commonality as a key to CLN3 function, and pursued cell biological pathways required for fluid-phase endocytosis. Fluid-phase endocytosis is regulated by cycling of the small GTPase Cdc42 and I discovered increased Cdc42-GTP in CLN3-null mouse brain endothelial cells. In mouse brain endothelial cells enhanced Cdc42-GTP increased Cdc42 dependent signaling, filopodial formation, and retarded cell migration. I also found reduced plasma membrane association of ARHGAP21, a known negative regulator of Cdc42. My data supports a model where loss of CLN3 reduces ARHGAP21 plasma membrane recruitment, and causes aberrant Cdc42 activation. Thus irregular Cdc42 activation underlies the commonly reported fluid-phase endocytic defects in JNCL. Therapeutic development for JNCL has been hampered in part from the varying phenotypes ascribed to CLN3 deficiency. My discovery that the fluid-phase endocytic defects result from Cdc42 pathway aberrations, which in turn contributed to multiple downstream phenotypes, opened the door to novel JNCL therapeutics. Here I present work showing that a Cdc42 inhibitor corrects the Cdc42 dependent defects in vitro and multiple defects in a JNCL mouse model.