BACKGROUND: Major depressive disorder (MDD) during adolescence significantly jeopardizes both mental and physical health. However, the etiology underlying MDD in adolescents remains unclear. METHODS: A total of 74 adolescents with MDD and 40 health controls (HCs) who underwent comprehensive clinical and cognitive assessments were enrolled. Differential expression analysis was conducted on plasma extracellular vesicles (EVs) carrying long non-coding RNAs (lncRNAs) and messenger RNAs (mRNAs) by microarray analysis. Two possible lncRNA-miR-mRNA networks were established and candidate regulatory axes were generated using the StarBase, miRDB, and TargetScan bioinformatics databases. Quantitative real-time polymerase chain reaction (qRT-PCR) was used to validate the candidate molecules and signaling axes in a clinical cohort. RESULTS: A total of 3752 dysregulated lncRNAs and 1789 dysfunctional mRNAs were identified. Two candidate regulatory axes (AC156455.1/miR-126-5p/AAK1 and CCDC18-AS1/miR-6835-5p/CCND2) with potential connections with MDD were selected. The candidate molecules exhibit differential expression patterns among adolescents with MDD and HCs, as well as before and after treatment with sertraline in adolescents with MDD. Furthermore, AAK1, CCDC18-AS1, and miR-6835-5p expressions exhibited significant differences between the response and non-response groups. Baseline expression of CCDC18-AS1, miR-6835-5p, and CCND2 could predict the therapeutic effect of sertraline, which may be associated with reducing suicidal ideation and improving cognitive function. CONCLUSION: Our study may provide insights into the understanding of the underlying pathological mechanisms in adolescents with MDD.