Abstract Background: Centrosome amplification (CA, the presence of >2 centrosomes) is a hallmark of several tumors. CA perturbs mitosis by generating nonphysiological pulling forces, leading to chromosome missegregation and aneuploidy. This condition may be advantageous for tumor outgrowth providing it gets corrected by clustering centrosomes into a pseudo-bipolar conformation for successful cell division. Molecular determinants of CA and therapeutic opportunities in breast cancer (BC) are still poorly understood. Preliminary data from our group show that somatic NF1 loss of function (LOF), common upon metastatic progression, is associated in vitro and in patients with selective sensitivity to maytansinoids such as T-DM1. Here, we explored the molecular basis of this increased sensitivity. Methods: Multiple CRISPR/Cas9-generated NF1KO or WT clones of HER2+ BC cell lines (BT474, SKBR3 and HCC1954) were assessed for sensitivity to T-DM1 or DM1 in BrdU-based and clonogenic assays, with RNAseq being performed. BT474 cells were furthered engineered with the FUCCI(Ca) reporter for live cell cycle imaging. Ploidy was assessed through flow cytometry. Aneuploidy in the GENIE cohort (restricted to BC patients analyzed with NF1-covering panels) was assessed by generating a segmentation score (sum of the absolute ploidy scores in segments covered by NGS panels). Centrosomes, spindle conformation and chromosome abnormalities were studied by confocal microscopy in cells synchronized both with RO3306 or thymidine block. Results: All lines showed increased sensitivity to T-DM1 in KO vs WT in BrdU-based and clonogenic assays; similar results were obtained with DM1 only, suggesting independence from HER2 targeting. RNAseq differential analysis showed significant enrichment for gene sets involved in mitotic spindle in KO but not WT cells upon T-DM1 treatment. FUCCI analysis showed significantly longer permanence in G2/M in NF1KO compared to WT cells (24.8 vs 17.9% in G2/M, p=1.81E-07). In vehicle-treated synchronized cells, significantly more KO cells showed >2 centrosomes (21.6 vs 4.7%, p<0.00001) and multiple pseudo-bipolar mitotic figures with narrow intercentriolar distances, indicative of efficient clustering. This was associated with more frequent chromosome misalignment (26.7 vs 6.7%, p=0.038). In the GENIE cohort, segmentation score was higher for patients with NF1 LOF mutations vs NF1 WT (median 46.7 vs 41.1, p=0.0023), indicative of more common aneuploidy. Upon T-DM1, KO cells exhibited significantly more non-bipolar spindles with massively wider intercentriolar distances. Conclusions: Somatic NF1 loss causes aneuploidy due to CA. This likely favors metastatization and can be exploited therapeutically. CA associated with other common oncogenic events (RAS and BRCA1/2 mutations, PTEN loss) may represent a general biomarker for drugs inhibiting centrosome clustering in BC. Citation Format: Bruno Achutti Duso, Elena Gavilan Dorronzoro, Giulia Tini, Maria de Filippo, Marica Ippolito, Chiara Soriani, Simona Rodighiero, Stefano Santaguida, Pier Giuseppe Pelicci, Luca Mazzarella. Somatic NF1 loss in breast cancer leads to centrosome amplification, aneuploidy and increased sensitivity to T-DM1 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1164.