The research, led by Min Jeong Son, Kwang-Hee Bae and Won Kon Kim of the Korea Research Institute of Bioscience and Biotechnology together with colleagues in Korea, began with a simple observation. When mice were exposed to cold or given a beta-adrenergic drug that mimics cold stimulation, DUSP4 levels rose sharply in their inguinal white adipose tissue and brown fat. Conversely, when the animals were kept at thermoneutrality, a warm 30 degrees Celsius at which thermogenic genes are largely switched off, DUSP4 expression fell. The enzyme was also strongly induced in cultured adipocytes treated with norepinephrine, the chemical messenger that sympathetic nerves release to activate heat production. Notably, the induction was confined to the adipocytes themselves rather than to immune or stromal cells in the tissue, and it was most pronounced in the inducible beige fat depot, while the epididymal fat pad, which is metabolically inert with respect to thermogenesis, showed little response.

To find out whether DUSP4 actually matters for metabolism, the team studied mice in which the Dusp4 gene had been deleted. On a standard chow diet the knockout animals looked essentially normal, with only a modestly higher body weight and slightly reduced thermogenic gene expression. But when the researchers challenged them with a high-fat diet, the picture changed dramatically. The DUSP4-deficient mice gained significantly more weight than controls, accumulated enlarged fat cells across multiple depots, and showed the molecular hallmarks of a fat tissue that had shifted from burning to storing: thermogenic genes such as Ucp1 and Pgc1alpha were substantially suppressed, while lipogenic genes including Srebp1c, Fas and Acc were elevated. The animals also developed markedly worse glucose tolerance and insulin resistance than their wild-type littermates on the same diet.

The most striking evidence came from cold-exposure experiments. When placed at 6 degrees Celsius for a week, mice lacking DUSP4 could not defend their body temperature as effectively as normal mice. Their rectal temperatures dropped further and stayed lower throughout the challenge, and infrared thermal imaging confirmed that their body surfaces were cooler. Indirect calorimetry revealed the underlying cause: the knockout mice consumed less oxygen, produced less carbon dioxide and expended significantly less energy than controls during cold exposure, even though they ate the same amount of food. Their inguinal fat also failed to undergo the normal mitochondrial remodeling that cold triggers, showing blunted increases in mitochondrial DNA copy number, reduced levels of respiratory chain proteins and an attenuated induction of mitochondrial biogenesis markers such as Tfam, Erralpha and Nrf1.

These defects pointed to a problem in the signaling machinery that switches on thermogenic genes. The researchers focused on CRTC3, a transcriptional coactivator that is abundant in adipocytes and known to control energy metabolism through its location within the cell. CRTC3 is regulated by phosphorylation: when serine residues on the protein are decorated with phosphate groups, it is retained in the cytoplasm and unable to assist gene transcription; when those phosphates are removed, CRTC3 moves into the nucleus, where it can boost the activity of genes such as Ucp1. What had remained unclear was which phosphatase, if any, performs this dephosphorylation in fat cells.

The answer, according to the new study, is DUSP4. Co-immunoprecipitation experiments showed that DUSP4 physically associates with CRTC3, an interaction that was strengthened by forskolin, a compound that mimics the cAMP signaling triggered by cold. When the team overexpressed DUSP4 in mature inguinal adipocytes, the serine phosphorylation of CRTC3 dropped; when they expressed a catalytically dead version of the enzyme, in which a single cysteine at position 284 was replaced by serine, CRTC3 phosphorylation actually increased. Crucially, DUSP4’s effect was selective: phosphorylation of AMPK and CREB, upstream signaling proteins, was left intact, indicating that DUSP4 acts at a specific phosphatase layer rather than globally dampening kinase pathways. In the fat tissue of DUSP4-knockout mice, CRTC3 phosphorylation was elevated, consistent with the enzyme’s role in living animals.

The functional consequences followed directly. With DUSP4 present, CRTC3 accumulated in the nucleus, as shown by both immunofluorescence imaging and biochemical fractionation of cells into nuclear and cytosolic compartments. Nuclear-localized CRTC3 enhanced the activity of reporter constructs driven by the Ucp1 and Pgc1alpha promoters, and this boost was abolished by the catalytically inactive mutant. In epistasis experiments using stable DUSP4-overexpressing adipocytes, silencing CRTC3 with shRNA blunted the UCP1 induction that DUSP4 would otherwise drive, while adding extra CRTC3 amplified it. Oxygen consumption measurements tracked the same pattern: CRTC3 knockdown reduced mitochondrial respiration in DUSP4-overexpressing cells, whereas CRTC3 overexpression raised it. Together these results establish CRTC3 as the functional mediator through which DUSP4 activates the thermogenic program.

The catalytic activity of DUSP4 proved essential at every level. Cells expressing wild-type DUSP4 responded to norepinephrine with robust UCP1 production, increased mitochondrial mass visible with MitoTracker staining, higher mitochondrial DNA copy number, and a coordinated induction of respiratory chain genes such as Ndufa9, Sdha, Uqcrc2, Cox4 and Atp5a1. Cells carrying the inactive mutant showed none of these responses, and the mutant also failed to dephosphorylate p38 MAPK, a known DUSP4 substrate, confirming that its enzymatic deadness was genuine. In other words, it is not merely the presence of DUSP4 but its phosphatase activity that turns on the fat-burning machinery.

Perhaps the most clinically suggestive result came from a rescue experiment. The researchers injected a lentivirus carrying the Dusp4 gene directly into the inguinal fat pads of knockout mice, then exposed the animals to cold. Restoring DUSP4 in this single depot reduced lipid droplet accumulation, raised UCP1 and PGC1alpha expression, decreased CRTC3 phosphorylation, enhanced CRTC3 nuclear localization, and brought UCP1 levels back to near-normal. This depot-targeted reconstitution demonstrates that DUSP4 acts in an adipocyte-autonomous fashion, although the authors caution that whether such localized rescue can improve whole-body metabolic parameters would require adipose-wide delivery strategies and dedicated systemic phenotyping.

The findings arrive amid intense interest in thermogenic fat as a therapeutic target. Adult humans do possess metabolically active brown fat that can be activated by cold, its presence correlates with better cardiometabolic health in large cohort studies, and the beta3-adrenergic drug mirabegron has been shown to increase brown fat activity and improve insulin sensitivity in people. Because DUSP4 is an enzyme, it is considered pharmacologically tractable: small molecules that enhance its catalytic activity or stabilize its interaction with CRTC3 could in principle augment energy expenditure. The authors acknowledge important limitations, including the use of a global knockout model, relatively modest replicate numbers for some assays, and the fact that the specific DUSP4-sensitive phosphorylation sites on CRTC3 have not yet been mapped. Validation in human adipose tissue will also be essential. Still, the identification of a DUSP4-CRTC3-UCP1 axis adds a genuinely new node to the thermogenic signaling network, and it suggests that the road to obesity treatment may one day run through a phosphatase that teaches fat cells to burn rather than store.

Subject of Research: Regulation of adipocyte thermogenesis by the phosphatase DUSP4 via CRTC3 dephosphorylation and UCP1 expression

Article Title: DUSP4 activates adipocyte thermogenesis via CRTC3 dephosphorylation-dependent UCP1 expression

Article References: Son, M. J., Kook, H. U., Jung, J., Park, S.-J., Jeon, Y. J., Oh, K.-J., Lee, E.-W., Han, B. S., Choi, J. H., Bae, K.-H., & Kim, W. K. (2026). DUSP4 activates adipocyte thermogenesis via CRTC3 dephosphorylation-dependent UCP1 expression. Experimental & Molecular Medicine. https://doi.org/10.1038/s12276-026-01850-9

Image Credits: AI Generated

DOI: 10.1038/s12276-026-01850-9

Keywords: DUSP4, CRTC3, UCP1, thermogenesis, obesity, adipocytes, brown fat, beige fat, phosphatase, mitochondria, insulin resistance, energy expenditure

The study, conducted by a team at Tokyo University of Science and Nippon Medical School and led by Professor Tadayoshi Hayata, focused on a gene called CTD nuclear envelope phosphatase 1, or CTDNEP1. This gene encodes a phosphatase, an enzyme class that removes phosphate groups from target molecules, placing it at the crossroads of intracellular signaling and lipid metabolism. Phosphatases of this kind help regulate how cells respond to their environment, how membranes are built and maintained, and how growth signals are transmitted. Because these are precisely the processes that go awry in cancer, abnormalities in CTDNEP1 have attracted attention before. Earlier work has suggested links between the gene and medulloblastoma, a malignant brain tumor that primarily affects children, and the same research group had previously reported that low CTDNEP1 expression was associated with poor prognosis in pancreatic ductal adenocarcinoma.

That history made the new result all the more striking. When the team examined CTDNEP1 in renal cell carcinoma, they found that higher expression of the gene was associated with poorer survival outcomes in clear cell renal cell carcinoma, the most prevalent subtype of the disease. This is the opposite direction of the association the group had documented in pancreatic cancer, where low expression signaled worse outcomes. The same molecule, in other words, can carry a completely different clinical meaning depending on the cancer type and even the subtype within a cancer type. Professor Hayata noted that in the earlier pancreatic cancer work, low CTDNEP1 expression was associated with poor prognosis, while in clear cell renal cell carcinoma the team observed the opposite direction, underscoring that context determines the gene’s clinical significance.

To understand the scale of the analysis, it helps to appreciate the landscape of kidney cancer itself. Renal cell carcinoma accounts for roughly ninety percent of all kidney cancers and arises when cells lining the kidney’s small tubules begin to grow uncontrollably. Within this category, clear cell renal cell carcinoma represents approximately eighty percent of cases and is named for the clear-looking appearance of its cancer cells under the microscope. Papillary renal cell carcinoma, the second most common subtype at roughly ten to fifteen percent of cases, instead forms distinctive finger-like structures called papillae. The two subtypes differ not only in appearance but in their cellular origins, genetic makeup, and the tumor microenvironments in which they develop. Yet reliable molecular markers for assessing prognosis in these subtypes have remained limited, which is precisely the gap the Japanese team set out to address.

The researchers’ first clue came from a broader observation: CTDNEP1 expression was elevated in tumor tissue compared with normal tissue across several cancer types, including renal cell carcinoma. Encouraged by this pattern, they turned to the TCGA Pan-Cancer Atlas, a comprehensive and widely used public database of genomic and clinical data, to examine the gene’s expression in 512 patients with clear cell renal cell carcinoma and 283 patients with papillary renal cell carcinoma. The team divided patients into groups based on their CTDNEP1 expression levels and compared survival outcomes between the groups, carefully accounting for confounding variables including age, sex, and pathological stage. This kind of database-driven approach, sometimes called an in silico study, allows researchers to detect associations across large patient cohorts that no single hospital could assemble on its own.

The results in clear cell renal cell carcinoma were clear-cut. Patients whose tumors showed higher levels of CTDNEP1 had significantly poorer survival outcomes, and the association was particularly pronounced in Stage III disease. Across the overall clear cell cohort, the hazard ratio for overall survival in the high-expression group was 2.16 compared with the low-expression group, meaning that high expression roughly doubled the risk of death over the study period after adjustment for the measured covariates. When the researchers drilled down to Stage III clear cell renal cell carcinoma specifically, high CTDNEP1 expression was significantly associated with poorer overall survival, poorer disease-specific survival, and shorter progression-free interval, although no significant association emerged for disease-free interval. These stage-specific findings suggest the gene’s expression may be most informative in the intermediate stage of disease, where prognostic refinement could have the greatest practical impact on treatment decisions.

In papillary renal cell carcinoma, by contrast, the picture was far murkier. The association between CTDNEP1 expression and prognosis was limited and inconsistent, with no consistent relationship observed for overall survival, disease-specific survival, or progression-free interval. Even the findings for disease-free interval shifted depending on which statistical model the researchers applied. This subtype-dependent divergence is not merely a statistical curiosity. The authors suggest that the genetic and biological differences between the two subtypes may explain it: clear cell and papillary renal cell carcinoma may arise from different types of kidney cells, carry different genetic backgrounds, and develop within different tumor environments, any of which could influence the biological processes connected to CTDNEP1 expression.

Supporting that interpretation, the team identified different biological pathways linked to CTDNEP1 expression in each subtype. In clear cell renal cell carcinoma, higher expression of the gene was associated with immune- and inflammation-related pathways, hinting that CTDNEP1 may intersect with the tumor’s inflammatory milieu in ways that worsen outcomes. In papillary renal cell carcinoma, the gene’s expression was instead associated with immune responses and metabolic processes, including pathways involved in energy production. These distinct pathway signatures reinforce the idea that the same gene can be wired into different biological circuits in different tumor contexts, a phenomenon increasingly recognized across cancer genomics and one that complicates any attempt to read a biomarker’s meaning from one cancer type and apply it to another.

The researchers are careful about what their findings do and do not establish. The study demonstrates an association rather than a direct cause-and-effect relationship, and further experimental work is needed to determine whether CTDNEP1 directly contributes to renal cell carcinoma progression. The stage-specific analyses were exploratory and based on limited patient numbers, and the threshold used to separate high- and low-expression groups was derived from the same cohort being analyzed, which means the results require validation in independent patient cohorts before they can inform clinical practice. Even so, the study positions CTDNEP1 as a candidate prognostic biomarker, particularly for clear cell renal cell carcinoma, and Professor Hayata emphasized that the findings suggest the gene may serve as a candidate prognostic marker in renal cell carcinoma while calling for further experimental studies to clarify whether it directly contributes to disease progression. The research, published in Volume 46, Issue 10 of the journal Anticancer Research, also stands as an example of cross-institutional collaboration, with first author Mayuka Nii, a Ph.D. candidate at Tokyo University of Science, working alongside Kazutaka Ota, then a third-year medical student at Nippon Medical School who participated through the school’s Project Semester III research placement course. For a disease in which subtype-specific prognostic tools remain scarce, a gene whose meaning flips between cancer types is a vivid reminder that in modern oncology, context is everything.

Subject of Research: Subtype-dependent association of CTDNEP1 gene expression with survival outcomes in renal cell carcinoma

Article Title: Gene expression linked to survival differs by kidney cancer subtype

Article References: Gene expression linked to survival differs by kidney cancer subtype. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: renal cell carcinoma, CTDNEP1, clear cell renal cell carcinoma, papillary renal cell carcinoma, prognostic biomarker, gene expression, TCGA Pan-Cancer Atlas, survival analysis, kidney cancer, phosphatase, tumor microenvironment, Anticancer Research

The investigation began with a large-scale survey of human gene expression. The team mined the Genotype-Tissue Expression database, known as GTEx, which catalogues RNA sequencing data from hundreds of donors. Comparing individuals aged 20 to 49 with those aged 60 and above across 25 tissues, they found that EYA4 expression was significantly elevated in 12 aging tissues, while remaining unchanged in the other 13. This pattern suggested that EYA4 is not a passive bystander in aging but a gene whose expression is actively remodeled as organisms grow older, echoing a growing appreciation that senescence-associated genes can feed back to regulate the aging trajectory itself.

To connect that tissue-level observation to a cellular mechanism, the researchers turned to primary human diploid fibroblasts, the classic workhorse of senescence research. They generated replicatively senescent HFF-1 and BJ fibroblasts by culturing them beyond 50 population doublings, roughly double the lifespan of young control cells at about 25 doublings. They also induced senescence chemically, treating young fibroblasts with bleomycin, a DNA-damaging agent, for three days. In both models, senescence was confirmed by elevated senescence-associated beta-galactosidase activity and increased p21, a cyclin-dependent kinase inhibitor that serves as a canonical senescence marker. Crucially, EYA4 rose in parallel, at both the messenger RNA and protein levels, in every senescent culture examined.

The functional test followed. Using small interfering RNAs to knock down EYA4 in fibroblasts approaching replicative senescence, the team observed a striking rescue: levels of p21 and senescence-associated beta-galactosidase dropped, and EdU incorporation assays showed that many cells retained their capacity to synthesize DNA and divide. The same held true in bleomycin-induced senescence, where EYA4 depletion again reduced senescence markers and preserved proliferative capacity. Together, these experiments demonstrated that EYA4 is not merely correlated with senescence but actively promotes it, and that removing it can delay both the replicative and stress-induced forms of the process.

What made this result mechanistically intriguing is that EYA4 is an unusual protein with two distinct biochemical faces. Its C-terminal Eya domain carries tyrosine phosphatase activity and mediates protein interactions, while its N-terminal domain harbors threonine phosphatase activity and transcriptional co-activation functions. To determine which activity drives p21 upregulation, the researchers engineered phosphatase-deficient mutants targeting each domain separately. Remarkably, both mutants promoted p21 expression just as strongly as wild-type EYA4. The senescence-promoting function of EYA4 therefore depends not on its enzymatic phosphatase activity but on its transcriptional activation capacity, a distinction that matters because it narrows the search for the relevant molecular partners.

That search led to SIX2, a transcription factor from the SIX family, whose members are famous for partnering with EYA proteins during development. Because EYA proteins lack any intrinsic DNA-binding ability, they depend on SIX partners to be escorted into the nucleus and tethered to specific promoters. Mining GTEx data through the GEPIA platform, the team found that SIX1, SIX2 and SIX4 expression correlated positively with EYA4 across tissues, and that SIX2 was the most abundantly expressed of the three in the 12 tissues where EYA4 rises with age, as well as in HeLa cells used for mechanistic assays. Co-immunoprecipitation experiments then confirmed that EYA4 and SIX2 physically interact, with the Eya domain of EYA4 serving as the primary binding interface.

The partnership proved functionally essential. Immunofluorescence microscopy showed that SIX2 overexpression drove EYA4 into the nucleus, while SIX2 knockdown diminished EYA4’s nuclear localization. Overexpressing SIX2 alone increased p21, but silencing SIX2 abolished EYA4’s ability to boost p21, and conversely, silencing EYA4 negated SIX2’s effect. Chromatin immunoprecipitation followed by quantitative PCR revealed that SIX2 binds to the P21 promoter within roughly one kilobase upstream of the transcription start site, and dual-luciferase reporter assays confirmed that this binding region is sufficient to activate P21 transcription. The model that emerges is one of mutual dependence: SIX2 recruits EYA4 to DNA, and EYA4 supplies the transcriptional activation power that neither protein can provide alone.

The p53 connection added a final layer of complexity. p53, the famed tumor suppressor, binds the P21 promoter about 2.4 kilobases upstream of the transcription start site and is classically regarded as p21’s master regulator. When the researchers generated p53-knockout HeLa cells using CRISPR, p21 expression fell to undetectable levels, and neither EYA4 nor SIX2 overexpression could restore it. Rescuing p53 in those cells brought p21 back to normal, and EYA4 knockdown then reduced p21 and senescence-associated secretory phenotype factors as before. Importantly, further experiments showed that p53 does not disrupt the EYA4-SIX2 interaction itself, nor does it prevent SIX2 from binding the P21 promoter, and the promoter remained responsive to p53-independent activation by an AKT inhibitor. This suggests p53 acts as a permissive gatekeeper for transcriptional initiation while the EYA4-SIX2 complex fine-tunes expression levels, though the authors caution that the precise mechanism requires further study.

The study’s implications extend in several directions. Because p21 is persistently activated in replicative and DNA damage-induced senescence, and elevated p21 has been linked to metabolic and cardiovascular diseases, the EYA4-SIX2 axis offers a candidate target for interventions aimed at delaying age-related pathology. The researchers also note that EYA4 plays dual roles in cancer, promoting some tumors through its phosphatase activity while suppressing others, so any therapeutic strategy would need to account for context. Open questions remain, including whether EYA4 partners with other SIX family members in different tissues, whether the axis also governs developmentally programmed senescence, and whether the complex preferentially regulates early-stage rather than late-stage senescence. The authors acknowledge the absence of endogenous co-immunoprecipitation data, owing to antibody limitations, but point to structural studies of related EYA-SIX complexes as strong support for a direct interaction. As senescence research moves from correlation toward mechanism, EYA4’s emergence as a p53-dependent, phosphatase-independent driver of p21 transcription offers a fresh and testable entry point into the biology of growing old.

Subject of Research: The role of EYA4 and SIX2 in regulating p21 transcription and cellular senescence during aging

Article Title: EYA4 promotes cellular senescence by enhancing P21 transcription through interaction with SIX2

Article References: Li, X., Mao, P., Chen, D., Li, L., Fang, H., Huang, J., & Liu, H. (2026). EYA4 promotes cellular senescence by enhancing P21 transcription through interaction with SIX2. Advanced Biotechnology, 4(2), Article 17. https://doi.org/10.1007/s44307-026-00109-8

Image Credits: AI Generated

DOI: 10.1007/s44307-026-00109-8

Keywords: EYA4, SIX2, p21, cellular senescence, p53, aging, GTEx, transcriptional regulation, phosphatase, fibroblasts, gene expression, senescence markers