Abstract
When people ask what is in an exosome, they usually expect a short answer. A growth factor. A peptide. Something nameable. The honest answer is less tidy and far more interesting: an exosome derived from pluripotent stem cells carries tens of thousands of distinct molecular species.
We know this because we had it measured. The component inventory discussed in this article comes from an independent compositional analysis of our own exosome preparations, performed by a university analytical center rather than in-house. We did not select what appeared on that list, and we did not curate it afterward. What follows is a reading of what the analysis found — including the entries that are less convenient to explain.
The named components represent a small fraction of the total inventory, and that fraction is still remarkable. Reading through it, what stands out is not the presence of any one celebrated molecule. It is the pattern. The components are not a random assortment of useful ingredients. They sort into coherent functional families, and those families map, one after another, onto the recognized hallmarks of biological aging. Genomic instability. Telomere attrition. Epigenetic drift. Loss of proteostasis. Deregulated nutrient sensing. Mitochondrial dysfunction. Cellular senescence. Stem cell exhaustion. Altered intercellular communication.
A pluripotent stem cell did not assemble that cargo by consulting a list. It assembled it because that is what a cell does when it is doing the work of building and maintaining living tissue. What follows is a guided walk through those families — what is there, what each group does, and why the composition as a whole behaves differently from the sum of its parts.
One: The Genomic Repair Toolkit
Every cell in the body sustains DNA damage continuously — tens of thousands of lesions per cell per day from metabolic byproducts, radiation, and ordinary replication error. Aging is, in large part, the story of repair capacity failing to keep pace with damage accumulation.
Representative components: BRCA1, PARP1, MSH2, PRKDC, TP53BP1, XRCC5, RPA1, FEN1, TDP1, ZSWIM7, DDB1, MCM3, KPNA2, RPS3, UBA1, PSMD2, HIST4H4
What is notable here is coverage rather than any single entry. These components span all of the major repair pathways: homologous recombination for accurate double-strand break repair, non-homologous end joining for rapid double-strand break resolution, mismatch repair for replication errors, and single-strand and base excision repair for the routine daily damage. PARP1 acts as the immediate damage sensor. TP53BP1 and BRCA1 arbitrate which double-strand break pathway a cell will use. PRKDC executes end joining and also patrols chromosome ends.
A pharmaceutical approach to DNA repair supplies one enzyme. This is a complete toolkit arriving together, in the proportions a pluripotent cell maintains for its own genome — the most stringently protected genome in the body.
Two: Telomere Maintenance
Telomeres are the protective sequences capping each chromosome. They shorten with each division, and once critically short they trigger senescence or death. Telomere attrition is one of the most direct and measurable links between cellular behavior and organismal aging.
Representative components: TERT, TERC, DKC1, RTEL1, ZSCAN4, TCP1, CCT4, CCT7, CCT8, SIRT6, XRCC5, PRKDC, HIST4H4
TERT and TERC are the catalytic protein and RNA template of telomerase itself. DKC1 stabilizes the telomerase RNA. RTEL1 resolves the structural knots that make telomeres fragile. The T-complex chaperone subunits regulate telomerase assembly and function.
ZSCAN4 deserves separate mention. It lengthens telomeres by a recombination-based mechanism that operates independently of telomerase, and it does so alongside a genome-stabilizing effect. Pluripotent cells use both routes. That redundancy is characteristic of how living systems protect what matters most: not one mechanism, but a primary system and a structurally different backup.
Three: Pluripotency and Epigenetic Information
This family is the one that most sharply distinguishes pluripotent stem cell exosomes from mesenchymal stem cell exosomes. MSC-derived vesicles do not carry this cargo. They cannot — an MSC is a committed adult cell and does not express these factors.
Representative components: POU5F1 (OCT4), MYC, KLF4, NANOG, LIN28A, SALL4, ZFP42 (REX1), ESRRB, PRDM14, DPPA4, GLIS1, TET1, HELLS, AHCY
Aging is now understood to be substantially an epigenetic phenomenon. The DNA sequence in an eighty-year-old cell is largely the same sequence it had at twenty; what has changed is the pattern of marks that tell the cell which genes to read. That pattern degrades over decades. The information is not lost so much as it becomes noisy.
The reprogramming factors present in this cargo — OCT4, KLF4, MYC, together with NANOG, LIN28A, SALL4 and the pluripotency network — are the same class of factors that laboratory work has shown, in animal models, can partially restore youthful epigenetic patterns without erasing cell identity. TET1 and HELLS govern DNA methylation and demethylation directly. AHCY sits at the center of cellular methylation capacity.
We are careful with this section. Partial epigenetic reprogramming is an active research field, most of the compelling evidence is preclinical, and we do not claim that an exosome infusion reproduces a laboratory reprogramming protocol. What we do observe is that these factors are present in the cargo, at physiological levels, in the company of everything else described in this article — and that context, as the next section explains, changes what they do.
Four: Metabolic Regulation and the Longevity Pathways
This family contains the molecules most often discussed in longevity research, and it contains an unusually complete set of them.
Representative components: SIRT1, SIRT3, SIRT4, SIRT5, SIRT6, SIRT7, FOXO3, FOXO4, MTOR, TFEB, NADSYN1, FGF23, GDF11, EPRS, VCP, PTEN
All seven sirtuins are represented. These NAD-dependent enzymes govern mitochondrial function, metabolic efficiency, genomic stability, and stress adaptation, and their decline is among the better-characterized features of aging tissue. NADSYN1 supplies the NAD they require to work at all — the cargo brings both the enzyme family and the cofactor pathway that fuels it.
FOXO3 and FOXO4 are transcription factors central to stress resistance, cellular repair, and autophagy. FOXO3 variants are the most reproducibly replicated genetic association with exceptional human longevity, identified independently in Japanese, European, and other centenarian populations.
TFEB is the master regulator of lysosomal biogenesis and autophagy — the cellular recycling machinery that clears damaged proteins and organelles, and which slows markedly with age. MTOR and PTEN complete the nutrient-sensing axis.
Five: Oxidative Stress Defense
Oxidative damage is continuous and cumulative. The response to it is not a single antioxidant but a layered defense system with sensing, signaling, and enzymatic components.
Representative components: NFE2L2 (NRF2), CAT, PRDX6, SIRT3, FKBP4, GSKIP, NPM1, GOT1, FIS1
NRF2 is the master transcriptional switch for the cell’s antioxidant program — activating it upregulates dozens of protective genes at once. Catalase and peroxiredoxin-6 are the enzymatic effectors that neutralize peroxides directly. SIRT3 governs reactive oxygen species production at the mitochondrial source rather than mopping up downstream. FIS1 participates in mitochondrial fission and quality control, the process by which damaged mitochondria are segregated and cleared.
Again the pattern: a regulator, its effectors, and the upstream source control, arriving as a set.
Six: Neurotrophic Support
The nervous system has limited regenerative capacity and depends heavily on trophic support to maintain existing neurons, synapses, and glia.
Representative components: BDNF, GDNF, CNTF, NTF3, NTF4, NGF, PAX6, GOT1, SIRT1, VEGFA
BDNF supports synaptic plasticity and neuronal survival and is among the most consistently studied molecules in cognitive resilience. GDNF is notably supportive of dopaminergic neurons. CNTF, NTF3, NTF4, and NGF cover complementary neuronal populations. GOT1 scavenges excess glutamate, the excitatory neurotransmitter whose accumulation drives excitotoxic injury.
One technical point worth stating precisely, because it is frequently confused. VEGF-A appears in this cargo, and vascular endothelial growth factor supports the perfusion on which all neural tissue depends. But VEGF-A delivered as a standalone gene therapy is a strictly local intervention — its effect extends only a few centimeters from the injection site and it does not act systemically. Where the goal is perfusion support within the central nervous system, the pluripotent stem cell-derived exosome is the appropriate vehicle: it crosses the blood-brain barrier and carries VEGF-A as part of its native cargo. The vesicle is what makes the destination reachable.
Seven: Vascular and Angiogenic Signaling
No tissue repairs itself without blood supply. Microvascular rarefaction — the progressive loss of small vessels — is an early and underappreciated feature of tissue aging, and it constrains everything else.
Representative components: VEGFA, KDR (VEGFR2), FLT4 (VEGFR3), PIGF, PROK1, FGF2, HGF, hsa-mir-126, ICAM1
The cargo carries both the angiogenic ligands and their receptors, which matters: aged endothelium often responds poorly to growth factor signals because receptor expression has declined. Supplying ligand alone addresses half the problem. FLT4 extends this to lymphatic vessels, which govern tissue drainage and immune trafficking and are rarely addressed at all. The microRNA miR-126 is among the most endothelium-specific regulators known and reinforces the same program at the post-transcriptional level.
Eight: Tissue Repair, Matrix Remodeling, and the Anti-Fibrotic Cluster
Fibrosis is what happens when repair goes wrong: functional tissue is replaced by scar. It is the terminal common pathway in a large fraction of chronic organ disease — pulmonary, hepatic, renal, cardiac, and cutaneous alike.
Representative components: TGFB3, MXRA5, RCN3, SCGB3A2, TIMP1, TIMP2, LUM, FN1, TGFBI, SOX9, WNT4, BMP4, BMP5, BMP7, hsa-miR-29a-3p, hsa-miR-17-5p, HGF
TGF-beta 3 is worth singling out. Where TGF-beta 1 drives scarring, the beta-3 isoform is associated with scarless healing — it is the isoform that predominates in fetal wound repair, which resolves without a scar. Its presence alongside the anti-fibrotic set is one of the more telling details in the entire composition.
MXRA5 and RCN3 limit excessive collagen deposition. SCGB3A2 has specifically documented anti-fibrotic activity in lung tissue. The microRNA miR-29a-3p is a direct post-transcriptional brake on collagen synthesis, and miR-17-5p opposes pulmonary fibrosis by targeting thrombospondin-2. TIMP1 and TIMP2 regulate the matrix metalloproteinases that determine whether matrix is remodeled constructively or degraded destructively.
This is not a group of molecules that shuts repair down. It is a group that steers repair toward regeneration rather than scar.
Nine: Immune Modulation — Including the Parts That Look Wrong
Here we should be candid, because a careful reader will notice it anyway.
Representative components: IL10, IL1RN, TNFRSF1B, TGFB1, hsa-mir-146a, hsa-mir-21 — alongside TNF, IL6, IL1A, IFNG, CCL2, CCL5, CCL11, hsa-mir-155
The composition contains anti-inflammatory components and pro-inflammatory components. Interleukin-10 and the interleukin-1 receptor antagonist are among the body’s most powerful endogenous brakes on inflammation, and miR-146a is a well-characterized negative regulator of inflammatory signaling. But tumor necrosis factor, interleukin-6, interleukin-1 alpha and interferon gamma are also present, and eotaxin (CCL11) is a molecule specifically associated in the literature with aged circulation.
It would be easy to omit those entries and present a cleaner list. We think that would be both dishonest and a misreading of what the cargo is.
Inflammation is not a defect. It is the opening phase of every repair process the body performs. A wound that does not inflame does not heal. What distinguishes healthy repair from chronic degeneration is not the absence of inflammatory signaling but its resolution — the orderly transition from an inflammatory phase to a proliferative one to a remodeling one. A cargo containing only anti-inflammatory signals would not be a repair signal at all. It would be an immunosuppressant.
What the pluripotent cell supplies is both arms of the system in the proportions it maintains for itself, so that the receiving tissue can set its own balance. This is modulation, not suppression, and the distinction is the whole point.
Ten: Muscle, Body Composition, and Growth Signaling
Representative components: FST (Follistatin), GDF11, IGF1, IGFBP1, IGFBP2, IGFBP3, IGFBP4, IGFBP6, GH1, KITLG, KIT
Follistatin is the natural antagonist of myostatin, the brake on muscle growth. It is associated with preserved lean mass, reduced fat accumulation, and improved metabolic profile, and it has become one of the most sought-after targets in longevity gene therapy for exactly those reasons.
The insulin-like growth factor system appears in unusual completeness — the ligand together with six binding proteins. That matters, because IGF-1’s relationship to longevity is genuinely bidirectional: necessary for tissue anabolism, but chronically elevated signaling is associated with reduced lifespan across species. The binding proteins are precisely what determines how much free IGF-1 is available and where. A cargo containing IGF-1 without its binding proteins would be a cruder instrument than what is actually present here.
The Question of Balance: Why the Combination Is the Therapy
This brings us to the objection most often raised about pluripotent stem cell biology, and to what we consider the most important idea in this article.
The reprogramming factors described in family three include MYC and KLF4. MYC in particular is a well-known proto-oncogene. In laboratory reprogramming work, forced overexpression of these factors in isolation — sustained, unregulated, at supraphysiological levels, delivered by integrating vectors — has been associated with tumor formation. This is a real finding and we do not dispute it.
But it is a finding about isolation, not about the factors themselves.
Consider what accompanies those reprogramming factors in this cargo. KLF4 itself is documented as a tumor suppressor in most tissue contexts. PTEN is one of the most important tumor suppressors known, a direct brake on the PI3K-AKT proliferative axis. PADI4 has documented tumor-suppressive function. ERRFI1 (MIG6) restrains epidermal growth factor receptor signaling and carries antitumor activity. The microRNA miR-34a is a potent tumor suppressor and inhibitor of cancer stem cells. IGFBP4 inhibits multiple cancer cell types in vivo and in vitro; IGFBP6 promotes apoptosis in transformed cells and inhibits angiogenesis, acting as a tumor suppressor. TIMP2 suppresses the proliferative response to angiogenic factors. Interferon gamma exerts antiproliferative effects on transformed cells. FOXO3 and FOXO4 themselves induce apoptosis in cells that have accumulated irreparable damage.
And beyond the named entries on any composition list, the pluripotent microenvironment carries factors whose antitumor activity has been characterized in its own right. Lefty, a member of the TGF-beta superfamily secreted abundantly by pluripotent stem cells, is the best-studied of these: it is a principal mediator of the long-observed capacity of the embryonic microenvironment to reprogram aggressive tumor cells toward a less malignant phenotype. The embryonic environment is not permissive to tumor growth. It is actively hostile to it.
So the picture is not “reprogramming factors, which cause cancer.” The picture is reprogramming factors delivered at physiological levels, embedded within a secretome that simultaneously carries a dense and overlapping set of tumor-suppressive, pro-apoptotic, anti-angiogenic and differentiation-promoting signals. The cargo contains its own checks. That is not a coincidence and it is not our editorial gloss on the data — it is the arrangement a pluripotent cell maintains in order to build an entire organism from a single cell without producing a tumor in the process.
The same logic applies elsewhere in the cargo, and it is worth seeing that the principle generalizes. FOXO3 is a longevity factor, but forced overexpression of FOXO3 in skeletal muscle drives atrophy, because FOXO3 also transcribes the muscle-wasting genes. MYC is a growth driver and a proto-oncogene. IGF-1 is anabolic and, in excess, life-shortening. VEGF-A is essential for perfusion and, unopposed, pathological. Every one of these molecules is beneficial within a range and harmful outside it. Every one of them arrives here accompanied by its natural counterweight.
This is the fundamental difference between supplying a molecule and supplying a signal. Gene therapy that overexpresses a single factor must get the dose right, and must hold it right, in every tissue, indefinitely, without feedback. A native secretome delivers each factor at the concentration a living cell selected, in the company of its regulators, to a recipient cell that retains its own capacity to respond, ignore, or modulate.
The Intelligence of the Living
We should be plain about the limits of our understanding. Nobody has mapped the interactions among sixty thousand molecular species. Nobody can write the equation for how this cargo behaves in a given tissue on a given day. We can name a fraction of the components and describe what each does in isolation, which is a long way from understanding the system.
But we are not required to understand a system completely in order to recognize that it is coherent — or to observe what it does.
The composition described in this article is the working secretome of a pluripotent stem cell: the signal a cell sends when its task is to build tissue, maintain a genome across an enormous number of divisions, and construct an organism. The balance it holds — between growth and restraint, between inflammation and its resolution, between proliferation and apoptosis — is not something we added. It is what the cell already knew.
What we do is far more modest than the language around this field sometimes suggests. We do not engineer the signal. We do not improve it. We collect it, under conditions that keep the cells healthy and pluripotent, and we deliver it intact to tissue that has lost the capacity to generate it for itself.
This is also, we think, the honest explanation for something we observe clinically and could not have predicted from a component list: the breadth of response. A cargo assembled around a single mechanism should produce a narrow effect. What is reported instead — across tissue types, across organ systems, in conditions that share no obvious common pathway — is what one would expect from a signal that addresses many hallmarks of aging simultaneously because it was never built to address just one.
We understand the parts we can name. The results suggest the whole is doing more than the parts we can name would account for. That is not a gap we intend to paper over with confident language. It is the reason our work is worth doing, carefully.
