Why some people struggle to shed residual fat even after meaningful weight loss may owe less to willpower or metabolism broadly defined, and more to a molecular memory etched into the immune cells residing in fat tissue. This finding reframes obesity not merely as a metabolic state but as one that reprograms cellular machinery in ways that actively resist reversal.
The research, published in Science Translational Medicine, identifies a specific RNA-processing breakdown in adipose tissue macrophages (ATMs) that persists well beyond weight loss itself. Using multiomics profiling and gene-targeting tools in mouse models, the investigators found that roughly 52% of obesity-induced changes in messenger RNA splicing within ATMs remained altered after weight loss — a molecular fingerprint of prior obesity. A key driver was dysfunction in CWC22, a component of the exon junction complex that governs how pre-mRNA is stitched together. In CWC22-impaired macrophages, the gene Scarb1 undergoes aberrant exon skipping, shifting production toward the SR-BII isoform of scavenger receptor class B. The resulting SR-BI/SR-BII heterodimers are flagged for endoplasmic reticulum-associated degradation, depleting surface SR-BI and crippling efferocytosis — the macrophage-mediated clearance of dead cells. This in turn suppresses inosine release from apoptotic cells, a signal that normally drives lipolysis in white adipose tissue. Crucially, antisense oligonucleotide correction of Scarb1 splicing restored the entire cascade — SR-BI surface expression, efferocytosis, inosine availability, and fat loss — in CWC22-deficient mice. Human adipose biopsies corroborated the mechanism: CWC22 nuclear localization, essential for proper splicing, was markedly reduced in ATMs from obese individuals compared to lean controls.
This work is potentially paradigm-shifting in its mechanistic specificity. The field has long recognized epigenetic memory in macrophages after obesity, but attributing persistent dysfunction to a splicing complex — and linking it causally to a druggable isoform imbalance — opens a genuinely new therapeutic avenue. Antisense oligonucleotides are already clinically validated in other diseases, lending near-term translational credibility. Key limitations include the predominantly murine mechanistic data and the relatively small human histological dataset. Whether CWC22 nuclear exclusion in human ATMs is functionally causal, or merely correlative, awaits interventional human study.