CLN6, spanning the endoplasmic reticulum membrane, is a protein of unknown function. Mutations in the CLN6 gene are linked to an autosomal recessively inherited disorder termed CLN6 disease, classified as a form of the neuronal ceroid lipofuscinoses (NCL). The pathogenesis of CLN6 disease remains poorly understood due to a lack of information about physiological roles CLN6 plays. We previously demonstrated that CLN6 has the ability to prevent protein aggregate formation, and thus hypothesized that the abrogation of CLN6's anti-aggregate activity underlies the development of CLN6 disease. To test this hypothesis, we narrowed down the region vital for CLN6's anti-aggregate activity, and subsequently investigated if pathogenic mutations within the region attenuate CLN6's anti-aggregate activity toward four aggregation-prone αB-crystallin (αBC) mutants. None of the four αBC mutants was prevented from aggregating by the Arg106ProfsX truncated CLN6 mutant, the human counterpart of the nclf mutant identified in a naturally occurring mouse model of late infantile-onset CLN6 disease. In contrast, the Arg149Cys and the Arg149His CLN6 mutants, both associated with adult-onset CLN6 disease, blocked aggregation of two out of and all of the four αBC mutants, respectively, indicating that CLN6's anti-aggregate activity is differentially modulated according to the substitution pattern at the same amino acid position. Collectively, we here propose that the graded reduction in CLN6's anti-aggregate activity governs the clinical course of late infantile- and adult-onset NCL.