https://www.ijstemcell.com/journal/view.html?uid=703&vmd=Full
chatGPT(5.5paid)
The paper, “Rejuvenation Potential of Developmental Genes Downregulated in Aging” by Min et al. (2026), proposes an interesting alternative to Yamanaka-factor partial reprogramming: rather than partially pushing old cells toward pluripotency, identify developmental maintenance programs that are progressively lost with age and restore those directly.
1. Summary
Central hypothesis
The authors start from the observation that development and ageing show partly opposing transcriptional trajectories. They hypothesize that genes which are:
upregulated during development but downregulated during ageing
may constitute a set of genes capable of restoring youthful cellular function without inducing pluripotency. They call the eventual candidates rejuvenation-inducing factors (RIFs).
This is intended to avoid an important problem with OSKM partial reprogramming: c-Myc is oncogenic, OSK can itself produce teratomas under some circumstances, and OSKM is difficult to package into AAV.
How they found the genes
The authors assembled ageing-associated genes from five sources, including Tabula Muris Senis, mouse plasma proteomics, HAGR/GenAge, GTEx-derived transcriptional ageing data and a large blood-expression meta-analysis.
They then compared these with single-cell RNA-seq from 811 mouse preimplantation embryonic cells, covering development from zygote through blastocyst.
A striking result is that genes that decline with ageing tend to increase during embryonic development. In the detectable dataset, the authors found progressive developmental upregulation of ageing-down genes, whereas ageing-up genes were relatively stable.
They classified genes into four categories:
| Ageing | Development | Genes | Main enrichment |
|---|---|---|---|
| Up | Up | 78 | Translation, autophagy |
| Up | Down | 113 | Transcription/development |
| Down | Up | 290 | Development, ribosome biogenesis |
| Down | Down | 180 | Cancer-related pathways |
The third category—ADDU: Aging Down–Development Up—is the important one.
Identification of the four RIFs
The authors then added a clever third filter.
They asked which ADDU genes also increase during the intermediate phase of OSKM reprogramming, before full pluripotency has been established.
That produced 31 candidate genes. After excluding genes associated with cancer/senescence and selecting factors with suitable developmental/regulatory functions, they chose four:
DDX21 + PHGDH + SERBP1 + TOMM70A
Functionally, these represent quite different aspects of cellular maintenance:
- DDX21 — ribosome biogenesis/rRNA processing
- PHGDH — serine biosynthesis and one-carbon metabolism
- SERBP1 — RNA stability/translational regulation
- TOMM70A/Tom70 — mitochondrial protein import and mitochondrial biogenesis
The paper therefore ends up identifying a rather interesting combination spanning mitochondria → metabolism → RNA → ribosomes/protein synthesis, rather than classical pluripotency transcription factors.
Experimental result
The authors overexpressed all four genes together in primary mouse tail-tip fibroblasts.
Five days later they found:
- a very large reduction in SA-β-gal-positive cells;
- reduced Cdkn1a/p21;
- reduced Cdkn2a/p16;
- increased Lmnb1.
Importantly, the fibroblasts retained fibroblast morphology. This differs from OSKM intermediate cells, which show morphological evidence of cell-fate change. The authors interpret this as rejuvenation without reprogramming/dedifferentiation.
2. What is novel?
I think there are three levels of novelty, with the first being the most important.
1. Rejuvenation-factor discovery by intersecting ageing, development and partial reprogramming
Rather than asking “which reprogramming factors make cells younger?”, the authors effectively ask:
Which genes are turned on as a young organism is constructed, subsequently decline with ageing, and are transiently restored during successful reprogramming?
That three-way intersection is an elegant way of searching for candidate rejuvenation mechanisms.
Conceptually:
Development ↑
∩ Ageing ↓
∩ Partial reprogramming ↑
→ candidate rejuvenation genes
This is more interesting than simply finding genes negatively correlated with chronological age.
2. Attempting rejuvenation without pluripotency
The paper separates two processes that are often conflated:
dedifferentiation/reprogramming ≠ rejuvenation
The authors’ proposal is that the rejuvenating part of OSKM may be reproducible by restoring particular homeostatic/developmental systems without activating the pluripotency network.
Their cells retain fibroblast morphology, supporting this possibility, although—as discussed below—it doesn’t yet prove it.
3. The identity of the four factors is mechanistically interesting
The resulting factors are not simply four alternative transcription factors.
They affect fundamental cellular infrastructure:
TOMM70A
↓
mitochondrial protein import/biogenesis
PHGDH
↓
serine / one-carbon metabolism
SERBP1
↓
mRNA stability / translation
DDX21
↓
rRNA processing / ribosome biogenesis
This suggests that the computational pipeline may be picking up a general decline in biosynthetic and information-processing capacity, rather than a specific “youth transcription programme.”
That may ultimately be more important than the particular four-gene cocktail.
3. Critique
The paper is conceptually strong but experimentally very preliminary. In my view the language occasionally gets ahead of what the experiments establish.
The biggest problem: they have demonstrated anti-senescence markers, not rejuvenation
Their experimental evidence essentially consists of:
SA-β-gal ↓ + p16 ↓ + p21 ↓ + Lamin B1 ↑
after five days of four-gene overexpression.
Those are certainly compatible with reduced cellular senescence, but they are not sufficient to establish that the cells have become biologically younger.
The authors themselves describe the four genes as producing “anti-senescence and rejuvenation-related phenotypes.”
The stronger conclusion that these factors can “reverse aging” is therefore premature.
There are no measurements of, for example:
- epigenetic age;
- DNA methylation clocks;
- chromatin accessibility;
- transcriptome-wide age reversal after treatment;
- mitochondrial respiration;
- proteostasis;
- DNA damage;
- telomere dysfunction;
- replicative lifespan;
- functional cellular performance.
A convincing rejuvenation paper would ideally show movement across multiple independent ageing dimensions.
No individual-factor experiments
Another major weakness is that the four genes are transfected together.
Consequently, the experiment does not establish whether:
all four are necessary,
one is sufficient, or
there is genuine synergy between them.
A straightforward factorial experiment—four individual genes, six pairs, four triples and the four-factor combination—would be extremely informative.
It might turn out, for example, that TOMM70A alone accounts for much of the phenotype.
Potential proliferation confound
This is particularly important.
DDX21 and PHGDH can support biosynthesis and proliferation. A reduction in the fraction of SA-β-gal-positive cells could therefore arise partly because healthier/proliferating cells expand faster than senescent cells.
The methods say that they counted the percentage of β-gal-positive cells in ten microscopic locations.
That doesn’t by itself distinguish:
senescent cells becoming non-senescent
from
non-senescent cells preferentially proliferating.
Lineage tracing or longitudinal single-cell tracking would address this.
Weak statistical depth
Figure 4B explicitly says statistical comparisons were not performed because biological replicate numbers were limited.
The principal molecular validation is also only three independent qPCR experiments.
That’s reasonable for an exploratory study, but not strong enough for claims about a robust rejuvenation intervention.
“Safer than OSKM” is a hypothesis, not a demonstrated result
The authors argue that none of the four RIFs is classified as an oncogene and that fibroblast morphology remains unchanged.
That is encouraging but does not establish safety.
Chronic enhancement of ribosome biogenesis, serine metabolism and mitochondrial biogenesis could potentially facilitate proliferation in premalignant cells even if the factors themselves are not canonical oncogenes.
Long-term transformation assays and, ultimately, in-vivo tumorigenicity studies would be required.
Development is not necessarily the opposite of ageing
This is the deeper conceptual issue.
The paper frequently treats developmental upregulation + ageing downregulation as evidence of a youthful programme.
But an embryonic cell is not simply an extremely young adult somatic cell.
Preimplantation embryos have extraordinarily unusual:
- metabolism;
- proliferation;
- chromatin architecture;
- transcription;
- DNA methylation;
- mitochondrial biology;
- cell-cycle regulation.
Thus:
embryonic ↑ + old ↓
does not necessarily imply
adult-young ↑.
Interestingly, the authors acknowledge that adding postnatal maturation stages could improve the approach.
I think that is actually crucial. Comparing young adult → old adult with postnatal maturation → adult homeostasis might produce a much more physiologically relevant RIF set than preimplantation embryos.
4. An especially interesting interpretation
There is a potentially deeper result hidden in this paper.
The authors’ analysis finds that ageing-down/development-up genes are enriched in ribosome biogenesis, while their final four factors involve mitochondrial import, serine metabolism, RNA regulation and ribosome production.
So the result can be interpreted in two ways.
The authors’ interpretation is:
developmental programmes are lost with ageing, and restoring them rejuvenates cells.
But another interpretation is:
ageing progressively reduces the cell’s capacity to maintain mitochondrial, transcriptional/translational and biosynthetic infrastructure; restoring that capacity suppresses the senescent phenotype.
The latter interpretation does not require ageing to be a literal reversal of development.
Overall assessment
I would rate this as a highly interesting hypothesis-generating paper with a clever computational discovery strategy, but relatively weak experimental validation.
Its strongest contribution is not yet demonstrating a four-gene anti-ageing therapy. Rather, it provides a useful conceptual framework:
Ageing datasets
↓
identify genes lost with age
Developmental datasets
↓
select those involved in building/maintaining youthful cellular capacity
Partial-reprogramming datasets
↓
select those restored during rejuvenation but before pluripotency
DDX21 + PHGDH + SERBP1 + TOMM70A
↓
candidate non-pluripotency rejuvenation programme
The next decisive experiment would be RNA-seq plus epigenetic-age/chromatin measurements before and after the four-RIF treatment, alongside individual-factor and proliferation controls. If those showed that the intervention moves the whole cellular state toward that of young fibroblasts rather than merely reducing p16/p21 and SA-β-gal, the claim of genuine rejuvenation would become much stronger.
There is also a particularly interesting connection here to your mitochondrial → citrate → nuclear acetyl-CoA → histone acetylation model: TOMM70A and PHGDH provide plausible routes by which the RIF cocktail could alter mitochondrial/metabolic support for chromatin and transcription, while DDX21 and SERBP1 act farther downstream on RNA/ribosome biology. The paper does not test that mechanism, but it would be quite interesting to map the four RIFs onto that pathway.