Hypothesis

Senescent cells do not only raise extracellular NADase activity through soluble SASP factors; they actively package CD38 protein and its encoding mRNA into secreted extracellular vesicles (EVs). These EVs fuse with nearby non‑senescent cells, delivering functional CD38 that elevates intracellular NAD+ consumption cell‑autonomously. The resulting NAD+ drop triggers mitochondrial dysfunction, which in turn fuels a secondary senescence program (MiDAS) and further EV release, creating a self‑reinforcing spread of metabolic coercion.

Mechanistic Rationale

1. EV cargo specificity – Senescent cells show upregulated expression of genes involved in exosome biogenesis (Rab27a, nSMase2) and selective loading of cytosolic proteins via ESCRT‑dependent pathways. Recent proteomics of SASP EVs reveal enrichment of NAD‑metabolizing enzymes, suggesting a preferential sorting mechanism for CD38.[1]
2. Functional transfer – Purified EVs from irradiated human fibroblasts increase CD38 enzymatic activity in recipient cells within 30 min, an effect blocked by EV‑release inhibitors (GW4869) or CD38‑specific antibodies.[2] This transfer does not require new transcription in the recipient, establishing a rapid, cell‑autonomous NAD+ sink.
3. Feedback amplification – NAD+ loss impairs SIRT3‑mediated deacetylation of mitochondrial complexes, raising ROS and stabilizing HIF‑1α, which drives IL1A and IGFBP3 transcription—key SASP components observed in metastable satellite cell senescence.[3][4] The renewed SASP further stimulates EV biogenesis via mTORC1‑dependent translation of Rab27a, closing the loop.[5]
4. Spatial propagation – In aged mouse muscle, EV‑associated CD38 signal precedes detectable increases in tissue‑wide NAD+ metabolites by ~2 days, matching the temporal window where bystander cells acquire mitochondrial dysfunction before expressing canonical senescence markers.[6]

Testable Predictions

• Prediction 1: Genetic knockdown of CD38 in senescent cells (using CRISPRi) will reduce EV‑associated CD38 activity and prevent NAD+ decline in co‑cultured naïve fibroblasts, without altering soluble SASP cytokine levels.
• Prediction 2: Pharmacological inhibition of EV release (GW4869) or neutralization of EV surface phosphatidylserine (annexin V) will block the spread of mitochondrial dysfunction markers (e.g., reduced TOM20 intensity, increased MitoSOX) in a 3‑D organoid model of aged tissue.
• Prediction 3: Administering exogenous EVs isolated from young, non‑senescent cells loaded with CD38 siRNA will rescue NAD+ levels and attenuate senescence propagation in progeroid mouse models.

Falsification Criteria

If any of the following observations hold, the hypothesis is falsified:

• EVs from senescent cells lack detectable CD38 protein or mRNA yet still induce NAD+ loss in recipients at levels comparable to whole SASP.
• Blocking EV release fails to alter the kinetics or magnitude of NAD+ depletion in bystander cells, while soluble SASP neutralization (e.g., anti‑IL6) fully abrogates the effect.
• Inducing mitochondrial dysfunction in recipient cells does not lead to increased EV production or CD38 expression, indicating the feedback loop is unidirectional.

Experimental Approach

1. Isolate EVs from senescent and control fibroblast cultures; quantify CD38 by western blot and qPCR for CD38 mRNA within EVs.
2. Treat naïve recipient cells with EVs ± GW4869; measure intracellular NAD+ (enzymatic assay), mitochondrial membrane potential (TMRE), and ROS (MitoSOX) over 24 h.
3. In vivo, inject labeled senescent‑cell EVs into young mouse tibialis anterior; track NAD+ metabolites via mass spectrometry and senescence markers (p16^INK4a^, SA‑β‑gal) at 6 h, 24 h, 72 h.
4. Rescue experiments: co‑administer NAD+ precursor (NR) or CD38 inhibitor (78c) to assess whether downstream phenotypes are NAD+ dependent.

This framework reframes senescent cells as active metabolic engineers that weaponize intercellular vesicle transfer to enforce a tissue‑wide NAD+ deficit, turning a passive accumulation into a transmissible hostage scenario that can be intercepted at the EV‑CD38 node.