In sepsis, the injured kidney ships mitochondrial DNA to the lungs — and the lungs are injured by the delivery, in mice and human data

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

Mitochondrial DNA is a danger signal wherever it appears outside mitochondria — the immune system reads it as damage. This paper shows an organ deliberately packaging that signal into vesicles and mailing it to another organ, with lethal effect. It explains a piece of how sepsis becomes multi-organ failure, and it names a step — vesicle formation in the kidney — that could in principle be interrupted.

How does the kidney-to-lung signal work?

In septic mice and in cultured renal tubular epithelial cells, sepsis upregulated sorting nexin 9, a protein involved in forming mitochondrial-derived vesicles. The kidney released more mitochondria-rich extracellular vesicles into the circulation. Those mitoEVs reached the lung and were taken up by pulmonary endothelial cells — the lining of the lung's blood vessels. Inside those cells, the vesicles' mitochondrial DNA activated the cGAS-STING innate immune pathway, with the DNA sensor ZBP1 identified as a potential cooperative partner. The result was endothelial barrier disruption, immune-cell recruitment and respiratory failure: acute lung injury manufactured at a distance by another organ's mitochondria.

What shows the pathway is causal, not just present?

The interventional half of the paper: blocking mitoEV formation in kidney tubules reduced the vesicles' release and protected septic mice against lung injury. That moves the finding beyond correlation inside the animal model — interrupt the shipping step and the downstream organ fares better. The human half of the paper is deliberately weaker and labeled as such by its design: in patients with sepsis, higher levels of circulating mitoEVs were associated with lung injury and with death. Association in patients can reflect severity rather than cause, which is exactly why the mouse intervention matters and exactly why it does not yet prove the same lever exists in human sepsis.

What could this mean for sepsis care?

Two possibilities, both unproven in people. As a biomarker, circulating mitoEV levels tracked the outcomes clinicians most fear in sepsis, so a validated mitoEV measure could one day help identify which patients are heading toward lung failure. As a target, the pathway has named, druggable-looking steps — sorting nexin 9-driven vesicle formation in the kidney, vesicle uptake in the lung, and cGAS-STING signaling itself, a pathway already under drug development for other inflammatory diseases. Sepsis has humbled decades of single-target therapies, so the honest label is a mechanistically complete, animal-proven axis with a human association — a map, not a treatment.

Where the evidence stands

• Established: Sepsis commonly progresses to multiple organ failure including acute lung injury; mitochondrial DNA outside mitochondria acts as a danger signal that can activate innate immune pathways including cGAS-STING; and cells release mitochondria-containing extracellular vesicles.
• Preliminary: In septic mice and cultured kidney cells (Communications Biology, Oct 30, 2026; DOI 10.1038/s42003-026-11217-5), sorting nexin 9-driven mitoEV release from injured kidney carried mtDNA to lung endothelium, activating cGAS-STING and causing barrier failure, and blocking vesicle formation protected the lungs; in sepsis patients, higher circulating mitoEVs were associated with lung injury and death.
• Absent: Proof that the same pathway drives human lung injury, any therapy targeting it, and any validated clinical mitoEV test.