The team, working with young and aged C57BL/6J mice, began by confirming what epidemiological studies had long hinted at. Auditory brainstem response testing showed that 85 percent of the aged animals failed to respond to sounds below 70 decibels, with click-evoked and frequency-specific thresholds markedly elevated across the 4 to 32 kilohertz range. Distortion product otoacoustic emissions were also impaired at high frequencies, indicating that the cochlea’s mechanical amplifier had faltered alongside its neural signaling. In short, these mice were a faithful model of human presbycusis.

To find out what was happening at the cellular level, the investigators isolated pericytes from the stria vascularis, the metabolically ravenous, ribbon-like structure that powers the cochlea and houses the blood-labyrinth barrier. Using fluorescence-activated cell sorting to purify pericytes marked by PDGFR-beta, they performed transcriptome sequencing on cells from young and aged mice. The results were striking: both Igf1 and Slc25a33, a gene encoding a mitochondrial nucleotide carrier, were significantly downregulated in the aged pericytes. Pathway enrichment analyses flagged the TGF-beta signaling pathway as one of the most differentially activated routes in the aging cells, alongside terms linked to cell junctions and the cytoskeleton.

Immunofluorescence staining of cochlear sections confirmed the story in tissue. IGF-1 and SLC25A33 proteins, which normally co-localize within the stria vascularis, showed a pronounced age-dependent decline, and the change was most dramatic in pericytes themselves. Critically, the levels of the IGF-1 receptor did not differ between young and aged pericytes, suggesting that the defect lies in the supply of the hormone rather than in the cell’s ability to sense it. When the researchers blocked IGF-1 signaling in cultured human brain vascular pericytes with the inhibitor NVP-AEW541, SLC25A33 messenger RNA and protein both fell; adding recombinant IGF-1 pushed them back up. Endothelial cells, by contrast, did not respond in the same way, marking pericytes as the specific cellular target of this regulatory axis.

Why would losing SLC25A33 matter so much? The protein is a mitochondrial carrier that helps sustain oxidative phosphorylation and redox balance. When the team silenced Slc25a33 with shRNA, flow cytometry using MitoSOX Red revealed a surge in mitochondrial reactive oxygen species leaking from the electron transport chain. The pericytes then began to change character. Alpha-smooth muscle actin, the signature of a contractile, myofibroblast-like state, rose sharply at both the messenger RNA and protein levels, while E-cadherin, a junctional protein associated with quiescent, barrier-stabilizing pericytes, dropped. The cells were undergoing a phenotypic switch from quiet vascular support cells into activated, fibrosis-prone contractile cells.

The mechanistic thread connecting these events proved to be TGF-beta signaling. In SLC25A33-deficient pericytes, TGF-beta messenger RNA and protein were elevated, and phosphorylated Smad2, the activated form of the transcriptional mediator, increased robustly. Total Smad2, Smad3, and Smad4 levels were unchanged, and phosphorylated Smad3 did not budge, indicating a selective engagement of the canonical Smad2 arm. When the researchers treated the deficient cells with MitoTEMPO, a mitochondrial antioxidant, the excess reactive oxygen species fell and TGF-beta, phospho-Smad2, and alpha-SMA all declined, placing mitochondrial oxidative stress upstream of the signaling cascade. Conversely, blocking the TGF-beta receptor with LY2109761 restored E-cadherin and suppressed the contractile markers, confirming that the pathway mediates the differentiation program.

The decisive test came in living animals. Delivering recombinant IGF-1 directly into the cochlea of aged mice through the round window membrane boosted SLC25A33 expression in the stria vascularis, while the IGF-1 inhibitor deepened its decline. Using adeno-associated virus vectors tailored for inner ear transduction, the team then overexpressed or knocked down Slc25a33 in the aged cochlea. Overexpression reduced alpha-SMA and TGF-beta signaling and raised E-cadherin; knockdown did the opposite. Transmission electron microscopy laid bare the structural consequences: in IGF-1-suppressed mice, pericytes showed abnormal morphology, loose associations with endothelial cells, and widened tight junctions, hallmarks of a leaky blood-labyrinth barrier. Exogenous IGF-1 preserved those intimate pericyte-endothelial contacts and restored tight junction integrity.

Most importantly for patients, the intervention worked functionally. IGF-1 supplementation significantly lowered click-evoked auditory brainstem response thresholds in aged mice, and at 12 kilohertz the treated animals heard markedly better than both untreated and IGF-1-suppressed controls. In mice whose Slc25a33 had been silenced by viral knockdown, treatment with the TGF-beta receptor inhibitor LY2109761 raised levels of the tight junction proteins Claudin-5 and Occludin in the stria vascularis, easing barrier disruption. The findings were further validated in an independent mouse strain, CBA/CaJ, where aged animals likewise showed reduced IGF-1 and SLC25A33 protein in the stria vascularis, indicating that the mechanism generalizes across genetic backgrounds.

The study thereby assembles a complete causal chain: aging lowers IGF-1, which suppresses the mitochondrial carrier SLC25A33 in strial pericytes; mitochondrial reactive oxygen species accumulate; the TGF-beta/Smad2 pathway activates; pericytes transform into contractile, pro-fibrotic cells; the blood-labyrinth barrier breaks down; and hearing deteriorates. This mirrors the pericyte-to-myofibroblast transitions documented at the blood-brain barrier, where elevated alpha-SMA similarly undermines vascular stability, and it echoes the ROS-driven epithelial-to-mesenchymal transitions long observed in other tissues. The cochlea, with its extraordinarily high metabolic demand and dense mitochondrial population, appears uniquely vulnerable to this failure mode.

Questions remain before the discovery can be translated into therapy. Whether IGF-1 acts directly on pericytes through its receptor or partly through paracrine networks will require cell-specific knockout models, and IGF-1 almost certainly regulates additional mitochondrial and cytoprotective genes that contribute to barrier stability. Species differences between human and mouse pericytes also warrant caution. Still, the identification of SLC25A33 as a biomarker and potential drug target, and the demonstration that both hormone supplementation and TGF-beta blockade can partially rescue barrier structure and auditory function in aged animals, opens a genuinely new frontier. If future work confirms the axis in humans, protecting the inner ear’s microscopic vasculature could become a strategy for preserving hearing in a rapidly aging world population.

Subject of Research: The IGF-1/SLC25A33 regulation of blood-labyrinth barrier pericytes in age-related hearing loss

Article Title: IGF‐1/SLC25A33 Maintains Blood‐Labyrinth Barrier Integrity by Suppressing TGF‐β/Smad2‐Mediated Pericyte Activation in Age‐Related Hearing Loss

Article References: Xu, R., Tang, J., Lin, H., Li, Q., Feng, Y., Zhang, X., Hu, X., Yang, J., Li, Y., & Bai, Y. (2026). IGF ‐1/ SLC25A33 Maintains Blood‐Labyrinth Barrier Integrity by Suppressing TGF ‐β/Smad2‐Mediated Pericyte Activation in Age‐Related Hearing Loss. Aging Cell, 25(10), Article e70712. https://doi.org/10.1111/acel.70712

Image Credits: AI Generated

DOI: 10.1111/acel.70712

Keywords: age-related hearing loss, IGF-1, SLC25A33, pericytes, blood-labyrinth barrier, TGF-beta, Smad2, mitochondrial ROS, stria vascularis, presbycusis, cochlea, Aging Cell

The research team set out to decode the regulatory architecture of subcutaneous fat deposition by comparing two dramatically divergent duck lines: the Pekin duck, a fast-growing commercial breed prized for its thick layer of subcutaneous fat, and the Liancheng duck, a leaner indigenous Chinese breed. Rather than relying on a single type of genomic data, the investigators assembled an unusually rich multi-omics portrait of the animals’ subcutaneous adipose tissue, integrating epigenomic maps of open chromatin generated with ATAC-seq, histone modification profiles produced by CUT&Tag-seq targeting H3K27ac, three-dimensional genome organization captured through chromatin interaction mapping, and transcriptomic profiles of gene expression across the two breeds.

This layered approach allowed the researchers to move beyond the classic limitation of association studies, which can flag a genomic region as relevant but rarely identify the actual causal variant and its target gene. The team catalogued differential open chromatin regions between the fat and lean lines, annotated them as candidate cis-regulatory elements such as enhancers and super-enhancers, and then linked these regulatory elements to distant target genes using chromatin loops and topologically associating domains, the three-dimensional compartments within which enhancers typically operate. The result was a comprehensive catalogue of variant-to-gene interactions, termed IMVGI, that connected noncoding variants to the genes they plausibly regulate.

To separate true regulators from statistical noise, the researchers developed what they describe as a mixed-strategy gene prioritization framework. This combined weighted gene co-expression network analysis, or WGCNA, which groups genes into modules whose coordinated expression tracks the trait of interest, with a Bayesian model that integrates multiple independent lines of evidence and assigns each candidate gene a posterior probability of being genuinely involved in fat deposition. The Bayesian machinery, validated through robustness checks across different prior distributions, distilled the field of candidates down to 112 high-confidence genes—a manageable set from which the team could hunt for the master switches controlling avian adiposity.

Among these candidates, one regulatory element stood out. The team identified a key selected SNP—a single nucleotide change that bears the signature of natural or artificial selection, as assessed by cross-population extended haplotype homozygosity and fixation index analyses—that acts as a remote control for the expression of the ADM gene, which encodes adrenomedullin, a peptide signaling molecule. Critically, this variant does not sit near the ADM promoter. Instead, it lies within a distal enhancer, and the study’s chromatin interaction maps show that this enhancer physically loops across a long genomic distance to contact the ADM locus, delivering regulatory input from afar. Comparisons of chromatin accessibility, enhancer activity and ADM expression between Pekin and Liancheng ducks consistently supported this long-distance regulatory relationship.

The mechanistic detail is where the study becomes particularly striking. Through transcription factor motif scanning of the sequence flanking the variant, the researchers found that the two alleles differ in their ability to recruit SMAD2, a transcription factor best known for its role in the TGF-beta signaling pathway. Allele-specific binding was supported by experiments including electrophoretic mobility shift assays with supershift validation using purified SMAD2 protein, alongside in silico mutagenesis predictions generated with the AlphaGenome model trained on human data and cross-species comparisons in the homologous chicken region. In effect, the selected SNP rewires the enhancer’s protein-binding landscape, changing how strongly SMAD2 can dock onto the DNA and thereby tuning the volume of ADM expression in the fat tissue.

What does ADM actually do once its expression is dialed up? The study presents converging evidence that adrenomedullin functions as a cellular signal driving fat accumulation by promoting the proliferation and differentiation of preadipocytes, the precursor cells that populate subcutaneous adipose tissue and mature into fat-storing adipocytes. Histological staining of fat tissue from the two breeds revealed differences in cell size consistent with divergent adipogenic activity, and ligand-receptor interaction analyses showed that ADM’s receptor components, including CALCRL and RAMP2, are expressed in the relevant tissue compartments, with single-cell resolution data from chicken preadipocytes corroborating the signaling axis. Human phenome-wide association data for the ADM locus further linked the gene to metabolic traits, hinting that the regulatory logic uncovered in ducks may echo in mammalian biology.

The broader significance of the work lies in its demonstration that functional dark matter—the noncoding majority of the genome—can be systematically interrogated when the right biological system and the right analytical toolkit are brought together. The duck’s extreme lipogenic efficiency acted as a magnifying glass, amplifying the phenotypic consequences of regulatory variation so that the causal architecture became visible. The team’s integrative pipeline, spanning epigenomics, 3D genomics, co-expression networks and Bayesian statistics, offers a template that could be applied to other livestock species and, potentially, to the interpretation of human GWAS loci that have long resisted functional annotation. In an independent segregating duck population, variants linked through the IMVGI framework explained a meaningful portion of the variance in subcutaneous fat phenotypes, underscoring the predictive power of the approach.

There are also immediate practical implications for agriculture. Subcutaneous fat percentage, thickness and weight are central economic traits in duck production, influencing carcass value, feed efficiency and consumer preference in markets where duck fat is a prized ingredient. The high-resolution epigenomic map and the prioritized candidate gene list produced by this study constitute what the authors describe as a precision blueprint for genetic selection, offering breeders molecular markers that track the fat-associated alleles identified through selection signature analyses. Marker-assisted or genomic selection programs built on such variants could reshape fat deposition in duck lines with far greater precision than traditional phenotypic selection alone.

For the biomedical community, the study adds a compelling chapter to the growing recognition that distal enhancers, rather than protein-coding mutations, frequently hold the keys to complex metabolic traits. The finding that a single selected SNP can reorchestrate transcription factor recruitment at a remote enhancer and, through that molecular switch, reshape an entire tissue’s fat-storing capacity illustrates the elegance and economy of regulatory evolution. As the authors note, fat deposition represents a global health threat whose genetic architecture remains unresolved; by illuminating how a duck enhancer commands ADM signaling to build fat, the research opens a window onto regulatory mechanisms that may well operate, in variant forms, across the vertebrate lineage—including in humans grappling with the genetics of obesity.

Subject of Research: Regulatory genetics of fat deposition in ducks

Article Title: A selected distal enhancer orchestrates avian fat deposition in coordination with ADM signaling

Article References: Liu, H., Zhang, H., Tang, H., Liu, S., Liu, D., Mou, Q., Wang, Z., Xiao, Y., Zhang, L., Zhang, Y., Yuan, L., Hou, S., & Zhou, Z. (2026). A selected distal enhancer orchestrates avian fat deposition in coordination with ADM signaling. BMC Biology. https://doi.org/10.1186/s12915-026-02748-8

Image Credits: AI Generated

DOI: 10.1186/s12915-026-02748-8

Keywords: fat deposition, duck, enhancer, ADM, SMAD2, epigenomics, ATAC-seq, 3D genome, WGCNA, Bayesian gene prioritization, preadipocyte, GWAS