Introduction: Filamin A (FLNA) is a multifunctional actin-binding protein involved in cytoskeletal remodeling, cell migration, cell signaling, and signal transduction (Zhou, 2021). While its role has been extensively studied in solid tumors, its function in acute myeloid leukemia (AML) remains poorly defined. Here, we provide a comprehensive multi-omic characterization of FLNA in AML, integrating transcriptomic, clinical, and pharmacological datasets, complemented by functional assays.

Methods: We analyzed FLNA expression across AML cohorts (TCGA, HOVON, BeatAML) and normal hematopoietic populations. Gene set enrichment analysis (GSEA), immune deconvolution, and drug response correlations were performed using bulk RNA-seq, and pharmacotranscriptomic data. Dependency scores from genome-wide CRISPR screens (DepMap portal) were analyzed to assess the essentiality of FLNA. In vitro FLNA knockdown was conducted in MLL-rearranged AML cell lines (MV4-11, MOLM-13) followed by assessment of drug sensitivity to Venetoclax, Cytarabine and Daunorubicin using cell viability assays. Survival analyses included univariate and multivariate Cox regression to determine prognostic value.

Results: High FLNA expression was associated with adverse clinical and cytogenetic features, including complex karyotype, del(7q), trisomy 8, leukocytosis, and elevated peripheral blast counts. In univariate Cox regression, FLNA-high patients had a 3.81-fold increased risk of death (p = 0.003) and a 3.38-fold increased risk of relapse or death (p = 0.008). These associations remained significant in multivariate models adjusted for age, white blood cell count, cytogenetic risk, and peripheral blasts, with FLNA independently predicting poor outcomes (HR = 3.78 for OS, p = 0.003; HR = 3.36 for DFS, p = 0.008). Transcriptomic analysis from the TCGA cohort revealed that FLNA-high AML samples exhibit activation of inflammatory (IL-6/JAK/STAT3, NF-κB), lysosomal, glycolytic, and cytoskeletal pathways, alongside suppression of MYC and E2F targets. This expression profile is suggestive of a quiescent and stress-adapted phenotype, reminiscent of leukemic stem cell–enriched states (TAKAO et al., 2025). Combined GSEA across TCGA, HOVON, and BeatAML cohorts (using C2 and C5 gene sets) revealed enrichment in transcriptional programs linked to metabolic rewiring and vesicle trafficking, and notably, MLL-rearranged AML signatures. This prompted further investigation, which showed that FLNA expression tracked closely with MLL-rearranged AML cell lines (MV4-11, MOLM-13), confirmed at transcript and protein levels. External datasets, such as BloodSpot, supported this pattern, while the limited MLL+ representation in TCGA (n=5) may have masked this relationship in direct comparisons. Besides, FLNA was most highly expressed in FAB M5 (monocytic) AML, consistent with its enrichment in normal monocytes and the monocytic immune profile seen in FLNA-high cases. Immune deconvolution suggested an immunosuppressive microenvironment in FLNA-high AML, with enrichment of monocytic and macrophage-like signatures and relative depletion of lymphocytic components. This profile may reflect a tolerogenic niche permissive to leukemic growth, aligning with the inflammatory and cytoskeletal programs enriched in these cases. Pharmacotranscriptomic data from the BeatAML cohort revealed that FLNA expression correlates with ex vivo resistance to Venetoclax, Azacitidine, and Cytarabine. However, in vitro FLNA knockdown paradoxically increased Venetoclax IC50 in MLL-rearranged cell lines, highlighting context-dependent effects. Genome-wide CRISPR data did not indicate essentiality for basal viability, suggesting FLNA is relevant under therapeutic or metabolic stress.

Conclusions: FLNA emerges as a multifaceted modulator of AML biology, associated with immune evasion, metabolic rewiring, chemoresistance, and poor prognosis. These findings position FLNA as a candidate biomarker and potential therapeutic target, warranting further investigation into its context-specific roles in AML pathogenesis and treatment resistance.

This content is only available as a PDF.
Sign in via your Institution