Mitochondria conduct cancer's spread — and a second organelle may be confounding the drugs aimed at them, a review argues

Edited by Samael A. Tejada, Founder & Editor-in-Chief

Reviews earn their keep when they change what the next experiment must measure. This one's demand is concrete: until researchers measure both organelles at once — paired mitochondrial and peroxisomal morphometry, organelle-resolved redox, genetic rescue separating the two — a claim that a drug works 'through mitochondria' in metastasis should be read as unproven, however clean the adaptor-level selectivity looks.

How do mitochondria power invasion, on this account?

Metastatic cells frequently show higher mitochondrial membrane potential, greater oxidative phosphorylation capacity and more dynamic fission-fusion cycling than non-metastatic counterparts, with the strongest evidence in breast cancer, melanoma and hepatocellular carcinoma models. Mitochondria physically relocate to lamellipodia, invadopodia and the leading edge of migrating cells, where local ATP production drives actin polymerization and secretion of matrix-degrading proteases such as MMP-2 and MT1-MMP; restricting that trafficking reduces matrix invasion in experimental systems, though glycolysis can also supply cortical ATP depending on tumor lineage and nutrient state. Mitochondrial ROS sit in a narrow pro-invasive window — enough to oxidize PTEN and phosphatases and prolong PI3K-AKT signaling and activate NF-kB, AP-1 and HIF-1alpha, but not so much as to kill the cell — and VDAC1, the outer membrane's most abundant channel, is elevated in multiple solid tumors and couples glycolysis to mitochondria through hexokinase II.

Why do peroxisomes complicate the drug story?

Mitochondria and peroxisomes divide using shared machinery: FIS1 and MFF adaptors serve both organelles, while MIEF1 and MIEF2 are mitochondria-only. A compound like P110, which blocks the DRP1-FIS1 interaction, is therefore adaptor-selective but organelle-agnostic — its phenotype could reflect altered peroxisomal division as easily as mitochondrial division — and Mdivi-1 additionally inhibits mitochondrial Complex I independently of DRP1. Peroxisomes also set the conditions mitochondria work under: they shorten very-long-chain fatty acids and pass medium-chain products back for complete oxidation, they generate and detoxify hydrogen peroxide, and their ether-lipid synthesis shapes membrane behavior and ferroptosis susceptibility. The same peroxisomal pathway can therefore support invasion and create a ferroptotic liability, making the direction of any intervention impossible to predict from peroxisome abundance alone.

What about mitochondria moving between cells — and the limits?

The review also covers intercellular mitochondrial transfer through tunneling nanotubes, extracellular vesicles and gap junctions: adipose stromal cells transferring mitochondria to breast cancer cells and conferring multidrug resistance, and, in an immune-evasion mechanism, cancer-cell mitochondria carrying mtDNA mutations entering tumor-infiltrating T cells, resisting mitophagy via co-transferred USP30 and progressively replacing the T cells' own mitochondria, leaving them senescent and less responsive to PD-1 blockade. Against all of this the authors set their own limit plainly: oxidant transfer at mitochondria-peroxisome contacts was characterized outside metastatic cancer, several contact architectures come from yeast or non-malignant systems, and the decisive experiment — perturbing the contact while preserving both organelles, then measuring invasion or colonization — has not been done.

Where the evidence stands

• Established: Mitochondria relocate to the invasive leading edge of migrating cancer cells and supply local ATP for actin dynamics and protease secretion in experimental models; mitochondria and peroxisomes share the fission adaptors FIS1 and MFF, and peroxisomes process very-long-chain fatty acids and hydrogen peroxide.
• Preliminary: The mechanometabolic-conductor framework and the mitochondria-peroxisome interface as a pharmacological confound (Frontiers in Pharmacology review, Oct 27, 2026) synthesize model evidence across breast cancer, melanoma and hepatocellular carcinoma, with lineage-dependent effects (for example PGC-1alpha promoting metastasis in breast models yet suppressing invasion in melanoma).
• Contested: Whether mitochondria-peroxisome contacts causally drive invasion, transit or colonization — the review itself states this has not been demonstrated and proposes the contact-perturbation experiment that would test it.
• Absent: Clinical evidence that targeting the mitochondria-peroxisome interface changes metastasis or survival in patients.