I suspect that the shift in senescent cell identity—specifically the transition from ‘regenerative’ senescent fibroblasts found in young models to the ‘stiffness-locked’ populations seen in aged tissue—is the primary molecular switch controlling whether adipose fibrosis can be reversed. In obesity, fibrosis acts as an adaptive response to mechanical stress. In aging, however, I believe the loss of morphogenetic gene programs leaves senescent fibroblasts unable to deploy latent matrix-remodeling enzymes, even when we hit the AMPK pathway with pharmacological tools. This essentially leaves the ECM irreversibly cross-linked.

While recent work highlights how semaglutide restores proteolytic balance through AMPK activation [https://pmc.ncbi.nlm.nih.gov/articles/PMC12898281/], I suspect this effect depends entirely on the epigenetic state of the local fibroblast population. Young senescent fibroblasts maintain a developmental program that still allows for matrix degradation [https://doi.org/10.1101/2025.06.08.658533]. In aging, the accumulation of senescent markers and the depletion of these programs create a ‘mechanosensing dead-end.’

My model suggests three stages:

1. The Stiffness-Senescence Threshold: As adipocytes hypertrophy and stiffen the ECM [https://pmc.ncbi.nlm.nih.gov/articles/PMC7503256/], local fibroblasts shift into a senescent state.
2. Qualitative Loss of Remodeling: In the young, these cells transiently provide remodeling factors. In the aged, they’re trapped in a pro-fibrotic phenotype by the buildup of COL6 endotrophin fragments [https://pmc.ncbi.nlm.nih.gov/articles/PMC7503256/], which block the transcriptional activation of MMPs.
3. Refractory Response: This likely explains why fibrosis in older patients often ignores metabolic therapies; the fibroblasts have lost the circuitry needed to translate AMPK-mediated signals into actual tissue remodeling.

We can test this by running single-cell RNA-seq and ATAC-seq on adipose fibroblasts from young versus aged humans who are undergoing caloric restriction or GLP-1 receptor agonist treatment.

• Prediction: Fibroblasts from aged, fibrotic adipose tissue will exhibit a closed chromatin state at the loci encoding MMP-1, MMP-9, and ADAMTS proteins compared to young fibrotic controls.
• Falsification: If aged fibroblasts show the same transcriptional plasticity as young ones when treated with AMPK activators, then the idea that qualitative aging of senescent cells limits fibrosis reversal is incorrect.

This implies that we may need to combine senolytics or epigenetic rejuvenation with metabolic stabilizers to unlock ECM remodeling in aged patients, as metabolic intervention alone can’t bypass the cell-intrinsic loss of morphogenetic capacity.