The study, published in Nature Communications, focuses on a protein called USP28, a deubiquitinating enzyme that removes ubiquitin tags from other proteins, thereby protecting them from degradation. The researchers found that USP28 is elevated in patients with sepsis and in a mouse model of the condition, and that this elevation triggers a cascade of events that ultimately kills endothelial cells through a violent form of programmed cell death known as necroptosis.

At the center of this cascade sits another protein: non-muscle myosin heavy chain IIA, or NMMHC IIA. The team demonstrated that USP28 directly removes K48-linked ubiquitin chains from NMMHC IIA at a specific amino acid position, lysine 1413, stabilizing the protein and preventing its destruction by the cellular waste-disposal system. With NMMHC IIA stabilized, it accumulates in the cell and begins promoting the trafficking of interferon-alpha/beta receptor 1 (IFNAR1) to the plasma membrane, where it becomes more active.

Once IFNAR1 reaches the cell surface, it amplifies signaling through STAT1, a transcription factor that drives inflammatory gene expression. This enhanced IFNAR1/STAT1 signaling pushes endothelial cells toward necroptosis — a form of cell death that is particularly damaging in the context of sepsis because it triggers further inflammation as cells rupture and release their contents. The researchers found that when they deleted USP28 specifically from endothelial cells in mice, or blocked it pharmacologically, sepsis-induced multi-organ injury was significantly reduced.

Crucially, the team uncovered a self-reinforcing loop: STAT1, once activated, transcriptionally increases USP28 expression, which in turn stabilizes more NMMHC IIA, which drives more IFNAR1 to the membrane, which activates more STAT1. This positive feedback loop explains how a relatively modest initial inflammatory signal can spiral into widespread vascular destruction during sepsis.

The researchers tested their hypothesis from multiple angles. When they overexpressed USP28 in endothelial cells, sepsis-induced organ damage worsened. Conversely, when they overexpressed NMMHC IIA or IFNAR1 in mice lacking endothelial USP28, the protective effect of USP28 deletion was largely abolished, confirming that these proteins act downstream of USP28 in the same pathway.

The study was conducted using a polymicrobial sepsis model in male mice, which involves puncturing the cecum to induce a bacterial infection that mimics human sepsis. The researchers also analyzed blood samples from septic patients, obtained through Zhongda Hospital at Southeast University, to confirm that USP28 elevation is relevant to human disease and not merely a rodent phenomenon.

The work was led by Jiahui Tang, Dingji Hu, and Mengchen Xu, with corresponding authors Yuanyuan Zhang, Songqiao Liu, and Junping Kou, spanning China Pharmaceutical University, Southeast University’s Zhongda Hospital, and The First People’s Hospital of Lianyungang. The research was funded by the National Natural Science Foundation of China and the Graduate Research and Innovation Projects of Jiangsu Province.

While the findings are compelling, several questions remain before this pathway can be targeted in clinical practice. The study was conducted exclusively in male mice, and sex differences in sepsis biology are well documented. Additionally, the precise pharmacological profile of USP28 inhibitors — their selectivity, toxicity, and pharmacokinetics — will need careful evaluation. Nevertheless, the identification of a specific, druggable enzymatic node in the endothelial damage cascade represents a meaningful advance in understanding how sepsis destroys blood vessels, and offers a concrete molecular target around which new therapies could eventually be designed.

The ubiquitin system, which the USP28 discovery exploits, is one of the most versatile regulatory networks in mammalian cells. Ubiquitin is a small 76-amino-acid protein that can be attached to target proteins singly or in chains of varying topology, and the shape of those chains dictates fate: K48-linked chains classically flag proteins for destruction by the proteasome, while K63-linked chains tend to scaffold signaling complexes. Deubiquitinating enzymes such as USP28 reverse this process by cleaving ubiquitin from substrates, and the human genome encodes roughly a hundred of them. Several members of this family, including USP4, USP15, and A20, have already been implicated in inflammatory and vascular disease, which lends biological plausibility to the new findings and suggests that endothelial cells rely heavily on ubiquitin editing to calibrate their response to inflammatory stress.

Necroptosis itself has emerged over the past decade as a distinctive cell death program that blends features of both regulated death and inflammatory rupture. Unlike apoptosis, which packages cellular contents into membrane-bound vesicles that are quietly consumed by phagocytes, necroptosis ends with plasma membrane rupture and the release of damage-associated molecular patterns such as HMGB1 and interleukin-33, molecules that further inflame neighboring tissue. The pathway is orchestrated by receptor-interacting protein kinases RIPK1 and RIPK3 together with the executioner protein MLKL, and it is normally held in check by caspase-8, which cleaves and disables the necroptotic machinery. When caspase-8 activity is compromised, as can occur during heavy cytokine exposure, necroptosis becomes a dominant fallback. Endothelial cells appear particularly susceptible to this form of death during systemic infection, and loss of the endothelial barrier is widely regarded as a central event in the vascular leak, coagulation disruption, and organ failure that define severe sepsis.

The involvement of NMMHC IIA in this pathway adds an intriguing dimension because this motor protein is better known for its structural roles. Non-muscle myosin IIA, encoded by the MYH9 gene, powers actin-based contractility, cell migration, and junctional integrity, and mutations in MYH9 cause a spectrum of human disorders featuring thrombocytopenia, nephropathy, and hearing loss. The finding that a cytoskeletal motor also functions as a trafficking factor for a cytokine receptor illustrates how proteins frequently moonlight in signaling roles distinct from their mechanical duties. Whether USP28-mediated stabilization of NMMHC IIA also affects its canonical cytoskeletal functions in endothelial cells, such as barrier maintenance or leukocyte transmigration, remains an open question that could broaden the physiological significance of the axis.

The IFNAR1/STAT1 arm of the pathway connects the study to a large body of work on type I interferons in sepsis. Type I interferons, best known for antiviral defense, have been shown in multiple models to worsen bacterial sepsis outcomes when excessively activated, partly through promoting inflammatory cell death and tissue damage. STAT1, the canonical downstream transcription factor, drives expression of hundreds of interferon-stimulated genes, some of which sensitize cells to necroptosis by upregulating components of the death machinery. The positive feedback loop described in the study, in which STAT1 transcriptionally upregulates USP28, fits a recurring theme in inflammatory biology in which interferon signaling amplifies its own inputs. Similar self-sustaining circuits have been described in lupus, psoriasis, and antiviral responses, and their tendency to become self-perpetuating may explain why inflammation in sepsis can continue to escalate even after the inciting pathogen has been cleared.

From a translational standpoint, the pharmacological arm of the study deserves attention. USP28 inhibitors have been under development in oncology, where the enzyme supports the stability of proteins such as c-Myc and Notch, and several small-molecule candidates have entered preclinical characterization. Repurposing such compounds for sepsis would face distinct challenges: sepsis is an acute, time-critical condition in which drugs must act rapidly, and the same ubiquitin-editing pathways that promote pathology in endothelial cells may serve protective roles elsewhere. The observation that endothelial-specific deletion of USP28 attenuated injury while global overexpression worsened it suggests that cell-type selectivity matters, and systemic inhibition could carry off-target consequences in tissues where USP28 is beneficial, including proliferating compartments of the intestine and bone marrow.

The restriction of the animal work to male mice is a notable limitation given the epidemiology of sepsis, which shows sex-dependent differences in incidence, immune response, and outcome. Estrogen-linked differences in interferon signaling are particularly well documented, and since the pathway described here converges on STAT1, a hormone-sensitive transcriptional regulator, the findings may not translate identically to female biology. Human validation, meanwhile, rests on circulating USP28 measurements from a single clinical cohort, and whether endothelial USP28 levels in patients track with disease severity, organ failure scores, or mortality will require prospective study.

Methodologically, the polymicrobial cecal ligation and puncture model used in the study is the most widely adopted experimental surrogate for human sepsis because it reproduces the mixed bacterial burden and sustained inflammation characteristic of abdominal infections, though it imperfectly captures the diversity of clinical sepsis arising from pneumonia, urinary sources, or hospital-acquired organisms. The use of endothelial-specific genetic tools, rather than global knockout, strengthens the causal interpretation by isolating the vascular compartment, and the rescue experiments, in which NMMHC IIA or IFNAR1 overexpression reversed the protection seen with USP28 loss, follow the logic of epistasis testing that is considered the gold standard for ordering genes within a pathway.

Looking forward, the most immediate questions are whether USP28 activity can be measured noninvasively in patients to identify those with active endothelial necroptosis, whether existing proteasome or interferon-signaling modulators could be repositioned to disrupt the loop, and whether the K1413 ubiquitination site on NMMHC IIA offers a more selective point of intervention than the catalytic cleft of USP28 itself. Answering these questions will determine whether the USP28–NMMHC IIA axis moves from mechanistic insight to therapeutic reality in a disease that still lacks a single approved drug targeting its core pathology.

Subject of Research: The role of the deubiquitinating enzyme USP28 in promoting endothelial necroptosis and multi-organ injury during sepsis

Article Title: USP28 potentiates endothelial necroptosis via deubiquitinating NMMHC IIA in septic male mice

Article References: Tang, J., Hu, D., Xu, M., Chen, H., Gong, S., Li, F., Yu, B., Zhang, Y., Liu, S., & Kou, J. (2026). USP28 potentiates endothelial necroptosis via deubiquitinating NMMHC IIA in septic male mice. Nature Communications. https://doi.org/10.1038/s41467-026-78666-8

Image Credits: AI Generated

DOI: 10.1038/s41467-026-78666-8

Keywords: sepsis, USP28, endothelial dysfunction, necroptosis, NMMHC IIA, IFNAR1, STAT1, ubiquitination, Nature Communications, potentiates, endothelial, deubiquitinating