High FIBP expression is associated with a ferroptosis-resistant transcriptional phenotype and adverse survival in acute myeloid leukemia: An integrative TCGA-LAML analysis
TL;DR
High FIBP expression is associated with an iron-buffering and antioxidant transcriptional phenotype and adverse survival in AML, but the current data do not demonstrate direct FIBP control of ferroptosis or independent prognostic value of the proposed sensitivity score.
Abstract
Background: FIBP has been linked to adverse acute myeloid leukemia (AML) features, but whether its expression identifies a ferroptosis-related biological state is unknown. We evaluated the association of FIBP with iron handling, lipid-peroxidation defense, immune signatures, and survival while explicitly separating association from causal mechanism. Methods: The KEGG ferroptosis pathway was compared with the STRING first-shell FIBP network. Legacy TCGA-LAML RNA-sequencing data were analyzed in 173 samples after log2(RSEM+1) transformation. Eleven prespecified genes covering iron homeostasis, antioxidant defense, lipid remodeling, and mitochondrial function were compared across median-defined FIBP groups and correlated with continuous FIBP expression. Marker-derived regulatory T-cell (Treg) and M2-like macrophage scores were examined exploratorily. A ferroptosis sensitivity score was defined as mean Z(TFRC, SLC11A2, ACSL4, LPCAT3) minus mean Z(FTH1, GPX4, AIFM2, DHODH). Overall survival was analyzed in 149 matched patients using Kaplan-Meier estimates, log-rank tests, 3-year restricted mean survival time (RMST), continuous-variable Cox models with proportional-hazards diagnostics, bootstrap-corrected C-index, and time-dependent area under the curve (AUC). Results: The strict STRING and KEGG sets had no direct gene overlap. Four target genes were higher in FIBP-high AML after transcriptome-wide false-discovery correction: FTH1, GPX4, AIFM2/FSP1, and LPCAT3. FIBP correlated positively with GPX4 (rho=0.619), FTH1 (rho=0.512), LPCAT3 (rho=0.502), and AIFM2 (rho=0.285), and negatively with SLC11A2 (rho=-0.220) and ACSL4 (rho=-0.196). The global FIBP-M2 association was nominal (rho=0.157; P=0.039) but did not remain significant after correction across the two immune scores (q=0.077); the Treg association was not reproduced. Survival differed among low-FIBP, high-FIBP/high-score, and high-FIBP/low-score groups (log-rank P<0.001). Three-year RMST was 729, 492, and 448 days, respectively; the two high-FIBP groups did not differ (RMST difference -44 days; bootstrap 95% CI, -236 to 137). Continuous FIBP remained associated with mortality after adjustment for the score and age (hazard ratio per SD, 1.420; 95% CI, 1.119-1.803; P=0.004), whereas the score was not independently associated (P=0.657). Adding the score did not improve the FIBP-only C-index (0.637 vs 0.636; delta 0.0008, bootstrap 95% CI, -0.0470 to 0.0014). Conclusions: High FIBP expression is associated with an iron-buffering and antioxidant transcriptional phenotype and adverse survival in AML, but the current data do not demonstrate direct FIBP control of ferroptosis or independent prognostic value of the proposed sensitivity score. GPX4-, FSP1-, and ferritin-centered defenses are testable biological hypotheses rather than established therapeutic indications.