Abstract Oncogenic mutations in the KRAS gene are well-established drivers of cancer. Promising preclinical strategies including RNA interference (RNAi) have been developed to target oncogenic KRAS function, yet a clinically effective therapy to directly target KRAS remains to be achieved. While genetic knockdown of mutant KRAS (mKRAS) with RNAi is one promising approach, current methods are not mutant-selective, and thus they also decrease normal KRAS levels, raising concerns about normal cell toxicity. Here, we describe a custom short interfering RNA (siRNA) oligonucleotide (D1) designed to preferentially bind the most commonly occurring RAS missense mutations in codons 12 and 13. Our approach designed the siRNA to have three mismatches binding wild-type mRNA, but only two mismatches for the oncogenic mRNAs, thus having binding affinities more favorable for the mutant-specific sequences. Using NIH/3T3 mouse fibroblasts stably expressing human KRAS WT or mutant cDNA sequences and KRAS mutant lung cancer cell lines, we determined that D1 siRNA preferentially reduced mKRAS mRNAs versus the wild-type sequence at the transcriptional and protein levels. Phenotypically, D1 siRNA significantly reversed mKRAS-induced morphologic and growth transformation of NIH/3T3 cells. Furthermore, D1 siRNA significantly impaired the proliferation of several mKRAS lung cancer cells in 2D as well as 3D spheroids embedded in an extracellular matrix. To overcome clinical limitations of delivering the siRNA to lung tumors, we evaluated lipid nanoparticles (LNPs) as carriers of D1 siRNAs in orthotopic mKRAS lung cancer models and observed that Cy5.5-labeled LNPs are highly enriched in lung tumors and metastases. Compared with control siRNA-LNP treated mice, 344SQ (KRAS G12D) tumor-bearing mice injected once with D1 siRNA-LNPs had drastically decreased KRAS protein levels in tumors 72 h later (73% decrease, p=0.001). Using an orthotopically injected luciferase-labeled A549 (KRAS G12S) human lung cancer model, D1 siRNA-LNP treated mice had pronounced reductions in luciferase signal (56% reduction, p=0.045), chest wall metastases (67% reduction, p<0.001) and contralateral lung metastases (73% reduction, p<0.01). Interestingly, D1 siRNA-LNP did not impair tumor growth when applied in the 344SQ mKRAS orthotopic mouse lung model, consistent with the mKRAS-independent growth of this line. This result supports our conclusion that the effects of D1 siRNA-LNP in the A549 model are KRAS specific and highlight the importance of determining RAS-dependency a priori for KRAS-targeting therapeutics. Taken together, our data indicate a novel use of RNAi to specifically target oncogenic KRAS-driven cancers. Citation Format: Bjoern Papke, Amanda E.D. Van Swearingen, Anne Y. Feng, Salma H. Azam, Emily B. Harrison, Runying Yang, Adrienne D. Cox, Channing J. Der, Chad V. Pecot. Silencing of oncogenic KRAS by a mutant-favoring short interfering RNA [abstract]. In: Proceedings of the AACR Special Conference on Targeting RAS-Driven Cancers; 2018 Dec 9-12; San Diego, CA. Philadelphia (PA): AACR; Mol Cancer Res 2020;18(5_Suppl):Abstract nr B32.