Background/Objectives: Natural killer (NK) cells play a crucial role in immune surveillance against melanoma, yet they frequently exhibit dysfunction in the tumor microenvironment. This study aims to establish an NK cell activation-related prognostic signature and identify potential druggable targets to overcome NK cell dysfunction. Methods: A prognostic signature was developed using the TCGA-SKCM cohort and validated across independent datasets. NK cell activation and cytotoxicity were evaluated in melanoma-NK-92MI co-culture systems via flow cytometry. Mechanistic studies employed Western blotting, co-immunoprecipitation, ELISA, and qRT-PCR. Single-cell RNA-seq data were used to analyze cell–cell communication. Results: A four-gene NK cell activation signature was identified and validated for prognostic significance across five independent melanoma datasets. Among the identified genes, cyclin B1 (CCNB1) emerged as a novel therapeutic target for overcoming NK cell resistance. In vivo, pharmacological inhibition of the CCNB1/Cyclin-dependent kinase 1 (CDK1) complex with RO-3306 significantly suppressed melanoma growth by enhancing NK cell infiltration and IFN-γ production. In vitro, CCNB1 knockdown in melanoma cells augmented NK-92MI activation, as evidenced by increased expression of CD69, CD107a, IFN-γ, and NKG2D, thereby improving NK cell-mediated cytotoxicity. Mechanistically, in melanoma cells, the CCNB1/CDK1 complex phosphorylates STAT3, activating the IL-6/STAT3 positive feedback loop, which upregulates PD-L1 and enables resistance to NK cell-mediated cytotoxicity. Beyond its role in immune evasion, CCNB1 also promoted melanoma invasiveness by inducing epithelial–mesenchymal transition (EMT) through the TGF-β-SMAD2/3 signaling. Conclusions: This study establishes CCNB1/CDK1 as a novel immunotherapeutic target and uncovers a new role for CDK1 inhibitors in enhancing NK cell function and suppressing melanoma progression.

Immune checkpoint inhibitors (ICIs) improve cancer outcomes but can lead to immune-mediated hepatotoxicity (IMH). Real-world data on severe IMH remain limited. This study aimed to identify risk factors and evaluate clinical outcomes, including the safety of ICIs rechallenge. This retrospective dual-center study included patients who developed IMH after ICIs therapy at two high-volume hospitals in China. Univariable and multivariable logistic regression were used to identify independent risk factors. The timing of onset was assessed by treatment cycles, and outcomes of patients who underwent rechallenge were analyzed. Among 408 patients with IMH, 175 had severe disease. Multivariable analysis identified hepatitis B virus (HBV) infection (OR 1.99, 95% CI 1.16-3.40), liver malignancy (OR 1.89, 95% CI 1.13-3.18), cholelithiasis (OR 2.28, 95% CI 1.41-3.71), and elevated lactate dehydrogenase (LDH) (OR 1.87, 95% CI 1.04-3.39) as independent risk factors. IMH occurred mainly within the first four ICIs cycles, with significantly earlier onset in HBV-positive patients (median 2 vs. 3 cycles, P = 0.038). Among 130 patients with grade 3 IMH, 94.6% recovered. ICIs rechallenge was performed in 65 patients after adequate recovery of liver function; 61/65 (93.8%) were able to continue ICIs, and recurrent hepatotoxicity was generally mild. Severe IMH was independently associated with HBV infection, liver malignancy, cholelithiasis, and elevated LDH. Grade 3 IMH was largely reversible and, in selected patients with adequate liver function recovery and without severe systemic irAEs, could allow cautious and individualized ICIs rechallenge.