FTH1 Loss Exposes Liver-Specific PDAC Metastasis Vulnerability

Key Takeaways

• In preclinical pancreatic ductal adenocarcinoma (PDAC) liver metastasis models, ferritin heavy chain (FTH1) emerged as a liver-specific dependency, and its loss suppressed liver metastatic growth while sparing the other tested growth settings.
• The control-versus-sgFth1 tumor-burden gap was approximately eight-fold in immunocompetent hosts versus approximately two-fold in NOD scid gamma (NSG) hosts, and lung colonization was not impaired, supporting both liver-context and immune-mediated components.
• Ferric ammonium citrate (FAC)-modeled iron overload was associated with proliferation suppression, glycolytic rewiring, reduced oxygen consumption rate (OCR), and mitochondrial injury in FTH1-deficient PDAC cells.
• Under the same high-iron conditions, FTH1-deficient cells were also associated with lower Na+/K+-ATPase activity, reduced ATP1A1, cytosolic calcium accumulation, selective IL36G induction, and renewed liver metastatic growth after Il36g deletion.

Pancreatic ductal adenocarcinoma commonly disseminates to the liver, where tumor cells face nutrient, oxygen, and host-immune conditions that differ from those in the pancreas. Because the liver is central to systemic iron handling, metastatic growth there may depend on iron-buffering programs that standard culture systems and primary-tumor models do not fully reveal. Investigators tested whether that liver-specific metabolic pressure could expose a selective vulnerability in pancreatic ductal adenocarcinoma.

An in vivo metabolism-focused CRISPR/Cas9 screen in a KPC-derived pancreatic ductal adenocarcinoma liver-metastasis model used the HY19636 cell line, a hemisplenectomy platform, and a metabolism-targeted library covering approximately 2,900 genes, with significant depleted genes defined at false discovery rate (FDR) < 0.2. Dependencies were compared with prior screens in 2D culture, 3D culture, and subcutaneous tumors, then followed through validation with two independent sgRNAs against ferritin heavy chain (FTH1), a second KPC-derived line, direct liver injection, lung metastasis testing, immunocompetent and NOD scid gamma (NSG) hosts, ferric ammonium citrate (FAC) loading, metabolomics, electron microscopy, sodium and calcium assays, RNA sequencing, RNA-stability studies, and later orthotopic pancreatic and human-cell-line confirmation; liver-metastasis validation used sgLuc n = 10, sgFth1#1 n = 10, and sgFth1#2 n = 12, and the NSG comparison used sgLuc n = 8 and sgFth1 n = 8.

FTH1 was the most representative liver-specific vulnerability from the screen, and knockout reduced liver metastatic burden without a comparable phenotype in 2D culture, 3D culture, subcutaneous tumors, orthotopic pancreatic tumors, or the reported lung-metastasis model. Direct liver injection also reduced tumor formation, and the phenotype remained present but weaker in NSG hosts, supporting both cell-autonomous and non-cell-autonomous components.

Under modeled high-iron conditions, FTH1-deficient cells exposed to FAC at 100 uM, commonly for 48 hours, showed dose-dependent proliferation suppression without observed cell death or rescue by Ferrostatin, along with glycolytic rewiring, pyruvate and lactate accumulation, reduced oxygen consumption rate (OCR), lower succinate dehydrogenase complex iron sulfur subunit B (SDHB), altered pyruvate dehydrogenase phosphorylation, and mitochondrial cristae disruption. In the mechanistic analysis of FTH1 loss, LKB1 suppression, and IL36G signaling, the same state was associated with reduced Na+/K+-ATPase activity, lower ATP1A1 abundance, altered sodium handling, cytosolic calcium accumulation, selective IL36G induction only in FAC-treated FTH1-deficient cells, and lower IL36G after ATP1A1, Fth1, or Lkb1 rescue; Il36g deletion restored liver metastatic growth and reduced CD8a-positive T-cell infiltration, while structural and RNA-stability experiments supported an ion-sensitive LKB1 5′ untranslated region RNA G-quadruplex with downstream disruption of nonsense-mediated decay (NMD).

These findings are preclinical and come from mouse models and murine and human cell lines rather than patient-outcome data. The proposed LKB1 5′ untranslated region G-quadruplex and the LKB1-SMG1-NMD connection are the authors’ supported mechanistic model, not settled clinical biology, and the downstream immune program remains incompletely defined because the observed drop in CD8a-positive infiltration does not establish whether those cells were the direct antitumor effectors or a broader recruitment signal. The authors also cautioned that direct FTH1 targeting is unlikely to be a viable therapeutic strategy and that the vulnerability appeared liver-context dependent in the reported models.

The authors concluded that iron handling can shape pancreatic ductal adenocarcinoma liver colonization through a cascade linking iron overload to metabolic and ionic stress, reduced LKB1 signaling, IL36G induction, and immune-mediated tumor control.

Clinician Questions

How did investigators show that FTH1 loss was a liver-specific vulnerability rather than a general PDAC growth defect?

Investigators mapped the phenotype across multiple settings and found it only when tumor cells were growing in the liver. The absence of a comparable effect in 2D culture, 3D culture, subcutaneous tumors, orthotopic pancreatic tumors, and the lung-metastasis model, together with reduced growth after direct liver injection, argues that FTH1 loss impaired liver outgrowth in the reported systems rather than causing a general pancreatic ductal adenocarcinoma growth defect.

What evidence linked reduced LKB1 to altered translation rather than loss of Lkb1 mRNA or increased protein turnover in iron-overloaded PDAC cells?

In the stressed cells, Lkb1 messenger RNA was not reported to fall when LKB1 protein decreased, and chloroquine plus bortezomib did not restore LKB1, which argued against lysosomal or proteasomal turnover as the primary explanation. Exogenous Lkb1 lacking the 5′ untranslated region was also resistant to FAC-associated suppression, and reverse-transcriptase footprinting with structural assays supported an ion-sensitive RNA G-quadruplex in that 5′ untranslated region, which is why the authors interpreted the effect as translational.

What remains unresolved about IL36G-driven immunity in PDAC liver metastasis?

Il36g deletion restored tumorigenic potential and reduced CD8a-positive T-cell infiltration, but the source says it remains unclear whether those CD8a-positive cells were direct antitumor effectors or a marker of broader immune recruitment. The reported immune response downstream of Il36g therefore remains incompletely defined and would require deeper immune-subtype analysis to clarify which cells mediate tumor control.