Autophagy-linked transcriptome screen in pulp stem cells names CASP3, DDIT3, ATF6 and ERN1 as hub genes; rapamycin raised VEGF in a mouse pulpitis model

What was tested

Kailun Wu, Xiaolan Guo (co-first authors) and colleagues at Nanfang Hospital and Southern Medical University in Guangzhou, China, published in the International Dental Journal on 7 October 2026 (open access under CC BY-NC-ND; the article is assigned to the December 2026 issue, volume 76, issue 6). The study asks whether autophagy-related RNA networks regulate the angiogenic differentiation of dental pulp stem cells (DPSCs) in pulpitis, where early inflammation often comes with local microcirculation disorders.

The in vitro arm ran RNA sequencing on DPSCs before and after angiogenic differentiation induction, screened differentially expressed (DE) messenger RNAs, long non-coding RNAs (lncRNAs) and microRNAs (miRNAs), intersected the lists with an autophagy gene set, and ranked candidates through a protein-protein interaction network using the MCODE and MCC algorithms. A competitive endogenous RNA (ceRNA) network, in which lncRNAs sponge miRNAs to modulate mRNAs, was then built. The in vivo arm used a mouse pulpitis model to test the autophagy activator rapamycin, reading inflammation and angiogenesis markers.

What they found

The sequencing arm identified 2,013 DE lncRNAs, 5,089 DE mRNAs and 148 DE miRNAs. Enrichment analysis placed the differentially expressed genes in autophagy-related pathways, named in the abstract as “Autophagy-animal” and “FoxO signalling pathway”. Four hub genes came out of the intersection: CASP3, DDIT3, ATF6 and ERN1. The constructed ceRNA network contains 15 lncRNAs, 4 miRNAs and 4 mRNAs.

In the mouse pulpitis model, the rapamycin group showed reduced expression of IL-18, mTOR and p62, and significantly increased expression of VEGF and ET.

One observation is worth stating plainly: the four named hub genes are not canonical autophagy machinery. CASP3 is the executioner caspase of apoptosis, while DDIT3 (CHOP), ATF6 and ERN1 (IRE1 alpha) are endoplasmic reticulum stress and unfolded-protein-response genes. The “autophagy-related” label comes from the gene-set intersection, and the network’s named center of gravity sits in ER stress and apoptosis biology adjacent to autophagy. On the animal markers, the directions are consistent with increased autophagic flux (lower mTOR, lower p62) and reduced inflammasome signaling (lower IL-18), but note that the abstract does not expand “ET”: if it is endothelin, a potent vasoconstrictor and pain mediator in pulp, then its rise alongside VEGF is not a clean pro-angiogenic signature, and the abstract does not resolve which.

What it does not show

The structured abstract reports no sample sizes, donor counts, statistics or effect sizes, so none are quoted here. The ceRNA network is a computational prediction: ceRNA relationships require experimental validation such as luciferase reporter assays, and the abstract names none. The mouse pulpitis model is not human disease, rapamycin is a blunt pharmacological probe rather than a pulpitis therapeutic, and no delivery, dosing or safety work appears. Angiogenic differentiation of DPSCs in dishes is a proxy for the clinical problem, which is revascularization of a necrotic tooth; no tooth-level regeneration outcome is reported.

Where it sits in the field

For the pulp and dentin repair program, blood vessel regrowth is the bottleneck in regenerative endodontics, and this paper works one level upstream of the clinical readouts: mechanism-level RNA network in human pulp stem cells, plus a pharmacological probe in mice. It is not the first autophagy-linked ceRNA network in pulpitis; Wang and colleagues built and experimentally validated one from the GSE92681 human pulpitis dataset in BMC Genomics in 2023 (DOI 10.1186/s12864-023-09363-9). The incremental contribution here is the angiogenic-differentiation angle and the rapamycin animal arm. This is mechanism work on the existing dentition, not a third-dentition route, and it moves no program tier; the current field assessment stands at /field/.

Where we differ from the coverage

We found no press or popular coverage of this paper to differ from. Against the paper itself, we bound two framings in the abstract. “Offering a preliminary theoretical basis for targeted intervention” describes a computational network plus a single probe in one mouse model, not evidence that any targeted intervention works. And “autophagy-related hub genes” overstates the identity of the four named hubs, which are apoptosis and ER-stress genes. We also flag the unresolved “ET” marker above rather than reading it as angiogenesis.

Provenance: grounded in the publisher’s complete structured abstract for Wu et al., International Dental Journal, published online 7 October 2026, DOI 10.1016/j.identj.2026.111199, read in full via the ScienceDirect record, with bibliographic metadata cross-checked against Crossref and OpenAlex. The full text was not retrievable from our server because the publisher’s site blocks automated access, so sample sizes, statistics and figure-level values are not quoted and every claim above is bounded to what the abstract states. Method and sourcing standard at /method/.