This week in innate antiviral signaling: compartmental cGAS, STING delivery, and damage sensing
A short read on new cGAS-STING, cyclic-nucleotide, and interferon papers: what looks solid, what is still preprint-level, and which debates are worth tracking next.
The useful signal this week is not "STING is good" or "STING is bad." The better read is that the same DNA-sensing and interferon machinery is being pulled into three different jobs: mucosal antiviral containment, vaccine/adjuvant design, and sterile tissue injury. The papers below are ordered by how much they should change what you watch next.
Fast triage
| Read priority | Paper | Source status | What it adds | How hard to take it |
|---|---|---|---|---|
| 1 | Lai et al., HSV-2 genital infection preprint. DOI: 10.64898/2026.07.02.735977 | bioRxiv preprint, primary study | cGAS-STING and RLR-MAVS are separated by tissue compartment: cGAS deficiency increased epithelial and submucosal pathology, impaired submucosal interferon responses, and allowed neuroinvasion in the HSV-2 model. MAVS mainly affected epithelial control and basal IFITM1/3. 1 | Strong mechanistic framing, but still a preprint and a model-system claim. Treat the compartment map as a working model until independent replication or peer review. |
| 2 | Shen et al., mutant EGFR and STING-TBK1 in NSCLC. DOI: 10.1038/s44318-026-00856-3 | Peer-reviewed paper indexed in PubMed | Mutant EGFR is reported to associate with STING signalosomes and phosphorylate STING at Y245/Y314 and TBK1 at Y577/Y677, creating a DNA-damage-tolerance loop; disrupting the loop sensitized resistant patient-derived organoids and mouse tumors to chemotherapy. 2 | The tumor-intrinsic mechanism is specific and testable. Do not generalize it to all STING-positive tumors yet. |
| 3 | Zeng et al., mRNA LNP vaccine with STING activators. DOI: 10.1073/pnas.2525718123 | Peer-reviewed PNAS paper | A screened ionizable-lipid LNP co-delivered tumor-antigen mRNA with c-di-AMP and Mn2+, activating type I interferon signaling, dendritic-cell maturation, antigen presentation, and tumor-specific cytotoxic T-cell responses in the reported models. 3 | This is a delivery-and-colocalization story, not proof that more STING agonism is always better. The patent disclosure also matters for translational follow-up. |
| 4 | Martin et al., MS2 capsid delivery of cGAMP for TB vaccination. DOI: 10.64898/2026.07.03.736450 | bioRxiv preprint, primary study | MS2 capsids carrying H1 antigen and cGAMP matched the original H1/CDN formulation in the mouse challenge setup while using 57-fold less cyclic dinucleotide and 3-fold less H1 antigen. 4 | The dose-sparing result is concrete. The claim that transporter bypass will improve human efficacy is still speculative. |
| 5 | Fisher-Wellman et al., trametinib cardiac toxicity. DOI: 10.1126/sciadv.aeb2695 | Peer-reviewed Science Advances paper | Trametinib caused mouse contractile dysfunction within 3 days and heart failure within 2 weeks; mitochondrial injury released mitochondrial DNA in mouse and human settings and activated canonical innate immune pathways including cGAS-STING. 5 | Stronger than a pathway screen because it links drug exposure, mitochondrial damage, mtDNA release, and innate signaling. Human clinical relevance still needs prospective validation. |
| 6 | Lantz et al., chikungunya neuroinvasion model. DOI: 10.1371/journal.ppat.1014395 | Peer-reviewed PLoS Pathogens paper | A young-adult CC041 mouse model showed chikungunya CNS infection after peripheral inoculation; susceptibility was associated with prolonged serum viral load and lower early peripheral type I interferon levels compared with resistant C57BL/6J mice. 6 | Useful model paper. It points to early interferon tone as a correlate of susceptibility, not yet a complete mechanism of neuroinvasion. |
What looks well established
The broad anchor still holds: cytosolic nucleic-acid sensing and type I interferon programs are first-line antiviral defenses, but their effects depend on cell type, tissue compartment, and timing. The HSV-2 preprint is interesting because it does not just say cGAS-STING and MAVS both matter. It assigns them different jobs in epithelial versus submucosal defense, with cGAS tied to early TBK1 activation, ISG induction, leukocyte recruitment, and restriction of spread toward neurons. 1
A second established point is that STING activation is inseparable from delivery. The PNAS mRNA vaccine paper and the TB preprint both treat STING agonism as a localization problem: put the agonist in the same dendritic cell, capsid, or antigen context where it can shape antigen presentation without wasting dose. 34
What is still provisional
The EGFR-STING-TBK1 paper is the most provocative cancer-mechanism claim in the set. If the phosphorylation loop holds up across more NSCLC contexts, it reframes cGAS-STING from a tumor-immunity trigger into a tumor-cell repair accessory in at least some EGFR-mutant settings. For now, keep the claim narrow: mutant EGFR-driven NSCLC, the reported phosphorylation sites, and the tested organoid and mouse models. 2
The vaccine-delivery papers should also stay in the "promising but not settled" bin. Both show better innate activation design in preclinical systems, but neither resolves the old STING problem: productive pulse versus chronic or misplaced inflammation. The TB capsid paper gives the cleanest engineering number this week, a 57-fold reduction in cyclic dinucleotide dose, but its human-efficacy argument remains a hypothesis. 4
Debates to watch next
- Compartment beats pathway labels. The HSV-2 work suggests a useful question for future mucosal-infection papers: where exactly is the sensor acting, and which tissue boundary fails when it is removed? 1
- STING agonism is becoming a formulation field. The strongest new claims are not about discovering STING. They are about colocalizing STING agonists with antigen, reducing dose, and controlling exposure. 34
- Damage-sensing papers are crowding the antiviral literature. The trametinib study is a reminder that cGAS-STING reads misplaced DNA, whether the upstream trigger is a virus, a tumor genotype, or a drug-damaged mitochondrion. 5
My read order
Start with the HSV-2 preprint if you care about antiviral barrier biology. Read the EGFR paper if you track STING in cancer, because it is the clearest counterexample to the simple "activate STING for immunity" story. Then skim the two delivery papers together; they are more useful as a pair than separately. The trametinib paper belongs on the watchlist for anyone thinking about chronic STING activation, mitochondrial DNA release, or on-target toxicity in kinase-inhibitor settings.
References
- 1
- 2Oncogenic EGFR rewires STING-TBK1 signalosomes to license DNA damage tolerance in NSCLC
pubmed.ncbi.nlm.nih.gov
- 3
- 4
- 5
- 6
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.