Small Molecules Alone Can Rapidly Reprogram Cells to Pluripotency — Here's How
Scientists uncover how RNA splicing and histone modification interact to supercharge chemical reprogramming, opening new doors for regenerative medicine.
Summary
Researchers can now turn ordinary mouse skin-type cells into stem cells using only a cocktail of small molecules, in about 7 to 12 days. How this works has been unclear. This study looked at RNA splicing, the process that lets one gene produce several protein versions. The team found that a splicing regulator called PTBP3 acts as a brake on reprogramming. Switching it down roughly doubled the number of stem cell colonies. PTBP3 changes how another gene, SCMH1, is spliced. SCMH1 helps place a gene-silencing tag, H2AK119 ubiquitination, on inflammatory genes. Lowering inflammation signaling helped reprogramming, and the drugs amlexanox and BMS-345541 improved efficiency even more. The work was done in mouse cells, but it points to ways of making cell rejuvenation and regeneration safer and more efficient.
Detailed Summary
Turning mature cells back into pluripotent stem cells is a cornerstone of regenerative medicine and rejuvenation research. Most methods rely on gene-delivered transcription factors. This team previously built a fast chemical reprogramming (FCR) system that makes induced pluripotent stem cells (iPSCs) within 7 to 12 days using small molecules alone, but the underlying biology was poorly understood. This study asks how post-transcriptional control, specifically alternative splicing (AS), shapes that process.
The authors first compared RNA-binding protein (RBP) expression and splicing in mouse embryonic fibroblasts (MEFs), chemically induced iPSCs (fiPSCs) and embryonic stem cells. Spliceosome genes were the most enriched RBP category. Splicing profiles of fiPSCs clustered with embryonic stem cells and apart from MEFs. About 1,000 skipped exons differed between MEFs and fiPSCs. MEF-preferred exons were shorter (median 86 vs 104.5 bp) and had longer flanking introns (4,358 vs 3,409 bp). So small molecules alone rewired the splicing landscape.
A shRNA screen of spliceosome-related RBPs identified PTBP3 as a negative regulator. Knocking it down roughly doubled fiPSC colony numbers without changing cell numbers. A second shRNA reproduced this, and rescue with an shRNA-resistant Ptbp3 reversed it. Overexpression suppressed colonies and Oct4 and Dppa5a induction. Knockdown of the related Ptbp1 had minimal effect. Ptbp3 knockdown also helped adult mouse lung fibroblasts but not transcription-factor reprogramming. The TGF-β inhibitor 616452 induced Ptbp3, and Smad2 knockdown raised Ptbp3 expression. Passage-12 knockdown fiPSCs expressed OCT4, SOX2 and NANOG and formed all three germ layers.
Bulk RNA-seq at day 8 found 421 upregulated and 630 downregulated protein-coding genes after knockdown, and 319 of these overlapped published PTBP3 binding data. Inflammatory response and NF-kB activity terms fell. The small nucleolar RNA Rian, linked to iPSC developmental potential, rose, and the negative reprogramming regulator MMVL30-int fell. Single-cell RNA-seq of 21,995 cells showed more intermediates and nascent iPSCs after knockdown, upregulated lysosome, autophagy, mTOR, proteasome and longevity-regulating pathways, and lower variability, with the Fano factor falling in over 95% of genes.
Splicing analysis found about 600 cassette-exon, 800 intron-retention, 200 Alt5 and 200 Alt3 events changed. Cassette exons near PTBP3 binding motifs implicated polycomb pathways, and 93 of 211 cassette-exon genes had direct binding evidence. Scmh1, a PRC1 component, shifted its isoform balance. Per the abstract, the PTBP3-SCMH1 axis regulates H2AK119ub deposition at inflammatory genes, and the inhibitors amlexanox and BMS-345541 further improved FCR efficiency.
Implications are mainly mechanistic. Chemical reprogramming activates its own brakes, and inflammation and splicing control can be tuned. Caveats: all work used mouse cells and shRNA, with no human cells or in vivo rejuvenation. Some later sections of the paper were not available for this summary.
Key Findings
- Ptbp3 knockdown increased fiPSC colony numbers about 2-fold without changing overall cell numbers, and a second shRNA and rescue experiments confirmed the effect
- About 1,000 skipped exons differed between MEFs and fiPSCs, and fiPSC splicing profiles clustered with embryonic stem cells rather than MEFs
- Deep RNA-seq after Ptbp3 knockdown found 421 upregulated and 630 downregulated genes, 319 of which overlapped published PTBP3 binding data
- Single-cell RNA-seq of 21,995 cells showed more intermediates and nascent iPSCs after knockdown, with the Fano factor falling in over 95% of genes
- Ptbp3 knockdown shifted about 600 cassette-exon and 800 intron-retention events, and 93 of 211 cassette-exon genes showed direct PTBP3 binding, including the PRC1 component Scmh1
- Removing the TGF-β inhibitor 616452 from the cocktail sharply cut Ptbp3 expression, and Smad2 knockdown raised Ptbp3
- The anti-inflammatory compounds amlexanox and BMS-345541 further improved reprogramming efficiency (per the abstract)
Methodology
The study used mouse embryonic fibroblasts and mouse lung fibroblasts in a 12-small-molecule chemical reprogramming protocol, comparing fiPSCs with embryonic stem cells. Methods included an shRNA screen scored by colony formation, rescue and overexpression experiments, qPCR, bulk RNA-seq at day 8, 10x scRNA-seq at day 12 (21,995 cells), vast-tools splicing analysis, and H2AK119ub profiling across reprogramming. Colony data came from three pooled independent experiments (n = 3 each) analyzed with Welch's unpaired t tests, while the RNA-seq of knockdown samples used n = 2 independent experiments and the scRNA-seq used Wilcoxon rank-sum tests.
Study Limitations
All experiments used mouse cells in culture and relied on shRNA knockdown, so human relevance, in vivo effects and long-term safety are untested. Effects were specific to chemical reprogramming and minimal in transcription-factor reprogramming, and some RNA-seq comparisons used only two replicates. The full text was truncated for this analysis, so the authors' stated limitations and any conflict-of-interest disclosures were not available.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: