Human genetics link Alzheimer's disease pathogenesis to excessive accumulation of amyloid-beta (Abeta) in brain, but the symptoms do not correlate with senile plaque burden. Since soluble Abeta aggregates can cause synaptic dysfunctions and memory deficits, these species could contribute to neuronal dysfunction and dementia. Here we explored selective targeting of large soluble aggregates, Abeta protofibrils, as a new immunotherapeutic strategy. The highly protofibril-selective monoclonal antibody mAb158 inhibited in vitro fibril formation and protected cells from Abeta protofibril-induced toxicity. When the mAb158 antibody was administered for 4 months to plaque-bearing transgenic mice with both the Arctic and Swedish mutations (tg-ArcSwe), Abeta protofibril levels were lowered while measures of insoluble Abeta were unaffected. In contrast, when treatment began before the appearance of senile plaques, amyloid deposition was prevented and Abeta protofibril levels diminished. Therapeutic intervention with mAb158 was however not proven functionally beneficial, since place learning depended neither on treatment nor transgenicity. Our findings suggest that Abeta protofibrils can be selectively cleared with immunotherapy in an animal model that display highly insoluble Abeta deposits, similar to those of Alzheimer's disease brain.