Genome-wide association and fine-mapping studies have identified over 40 susceptibility regions for type 1 diabetes (T1D), a common autoimmune disease; however, most of the disease-associated variants are noncoding, and it remains a challenge to understand their biological contributions to T1D pathogenesis. One identified T1D risk locus is located at chromosome 21q22.3 where the most likely candidate gene is UBASH3A, a negative regulator of NF-κB signaling. Various noncoding variants in UBASH3A have been shown to be associated with T1D or other autoimmune diseases. Here we investigated four such SNPs-rs11203202, rs80054410, rs11203203, and rs1893592. We discovered a novel role for rs1893592 in T1D and showed that its minor allele protects against T1D. Our haplotype analysis identified three T1D-associated UBASH3A haplotypes, and revealed that risk for T1D is affected by additive effects of these four UBASH3A variants. In human primary CD4+ T cells, upon T-cell receptor stimulation, the minor allele of rs1893592 was associated with both a significant reduction in the overall mRNA levels of UBASH3A, and an increase in the proportion of a normally occurring, but low-abundant, UBASH3A transcript that retains intron-9 sequences and cannot produce full-length UBASH3A protein. This reduction in UBASH3A, as a consequence of the minor allele at rs1893592, resulted in increased secretion of IL-2, a key cytokine that is required for T-cell activation and function but is deficient in some T1D subjects. Our study provides new mechanistic insights into how rs1893592 affects T1D and autoimmunity, and how interactions between multiple T1D-associated, noncoding variants influence the disease risk.