> 매일 아침 research-watcher가 27개 query family × PubMed/GEO를 스캔해서, Brown Biotech 파이프라인 관점에서 decision-ready 인사이트로 정리합니다. 매일 06:00 KST 자동 생성. **Category focus:** AI Drug Discovery **Published:** 2026-10-27 06:06 KST **Entry ID:** 138 **Tags:** #refractory cancer #esophageal squamous cell carcinoma #ESCC #cancer-associated fibroblasts #CAF #extracellular vesicles #ANP32A #ANXA2 #STAT3 #SLC29A4 #glutathione #GSH --- ## 🔬 Today's Top Findings ### 1. [Refractory-Cancer Lane + AI Drug Discovery + Stromal EV-ANP32A Glutathione Axis + Cisplatin Resistance + Tetrahedral DNA Nanostructure siRNA Delivery] GSE351727 'Cancer-associated fibroblast-derived ANP32A drives cisplatin resistance in esophageal squamous cell carcinoma' (Homo sapiens, score 6, n=6, pdat 2026/10/25 = FRESHEST NOVEL accession of the day, raw_file_availability yes, license NCBI public repository, query 'tumor metabolism single-cell') is the FIRST NOVEL stromal-EV-ANP32A cisplatin-resistance dataset of the day ### 2. [Longevity & Senolytics Lane + Brown Adipose Thermogenesis + YBX3 RNA-Binding Protein + BCAA Flux + PPARgamma-Dependent Mitochondrial OXPHOS] GSE277858 'Effects of knockdown of Ybx3 on gene expression during brown adipogenic differentiation of C3H1' (Mus musculus, score 7, n=8, pdat 2026/09/30, raw_file_availability yes, license NCBI public repository, query 'aging muscle atlas') is the FIRST NOVEL BAT-RBP thermogenesis dataset of the day ### 3. [AI Drug Discovery Lane + Synthetic RNA Circuit + LAG-3 Isoform + FOXO1 / IL-12 Payloads + Exhausted-CD8-T-Cell-Selective Expression] GSE317552 'Harnessing T-cell exhaustion-associated RNA splicing to enhance cancer immunotherapy' (Mus musculus ### 4. Homo sapiens, score 6, n=10, pdat 2026/10/22, raw_file_availability yes, license NCBI public repository, query 'protein design') is the FIRST NOVEL exhaustion-state-gated RNA-circuit dataset of the day ## 📋 Synthesis Loop 001 daily digest for 2026-10-27, generated from research-watcher output/latest/hits.jsonl (114 hits across 27 queries, scan regenerated 2026-10-27 06:03 KST). The score-9 top band was entirely RECYCLED against the retained id:133-id:137 window (GSE328275 TNBC-CD45 immune atlas, GSE328422 sequential lymph-node metastasis spatial, GSE326553 Oprm1-A118G opioid CosMx, GSE281463/64/65 chromatin-GWAS bundle), so the actionable harvest again comes from the score-7 and score-6 ladder after accession-level dedupe against the page-level GSE retention set (75 NOVEL accessions remain of 114). Signal 1 is GSE351727 (Homo sapiens, n=6, pdat 2026/10/25, the freshest NOVEL accession in today's scan): a stromal cell-cancer cell communication axis in ESCC where cancer-associated fibroblasts raise intracellular glutathione in tumor cells through extracellular-vesicle-delivered ANP32A, with a defined ANXA2-STAT3-SLC29A4 downstream chain and a tetrahedral DNA-nanostructure siRNA delivery proof of concept that reduced GSH, increased lipid peroxidation and suppressed ESCC xenograft growth. Signal 2 is GSE277858 (Mus musculus, n=8, score 7): the mechanistic dataset behind YBX3, a brown-adipose-enriched RNA-binding protein that is cold-induced and cAMP-driven, required for brown adipocyte differentiation, and that stabilizes Slc3a2 mRNA to license branched-chain-amino-acid catabolism in a PPARgamma-dependent manner; Ybx3 loss damps thermogenesis and worsens diet-induced obesity while overexpression protects against metabolic dysfunction. Signal 3 is GSE317552 (Mus musculus; Homo sapiens, n=10, pdat 2026/10/22): discovery of a previously unannotated RNA isoform of LAG-3 expressed specifically in intratumoral exhausted CD8+ T cells, plus a synthetic RNA splicing switch that conditionally expresses FOXO1 and IL-12 only in the exhausted state, improving TCR-T anti-tumor activity versus constitutive expression. All three are accession-level NOVEL versus id:137 (liver-aging Epi Multiome, preinvasive breast spatial pair, MASH liver spatial) and versus the wider id:133-id:137 window, so the novelty gate is met at 3/3 against a required 2/2, with three distinct BB lanes covered (refractory cancer / AI-DD, longevity-metabolism, AI-DD immune engineering). --- ## 🎯 Highlights ### 1. [Signal 1 - GSE351727 CAF-derived ANP32A drives cisplatin resistance in ESCC, Homo sapiens, score 6, n=6, pdat 2026/10/25 = FRESHEST NOVEL accession of the day]: the strongest new refractory-cancer signal in today's scan because it identifies the CARRIER of the resistant phenotype rather than the phenotype itself. CAFs elevate glutathione abundance in ESCC cells via extracellular-vesicle-mediated communication; proteomic profiling of CAF EVs plus prognostic analysis nominated ANP32A as the effector; ANP32A depletion in CAFs abolished the EV capacity to raise GSH, suppress ROS and lipid peroxidation, and protect ESCC cells from cisplatin cytotoxicity. Mechanistically ANP32A binds ANXA2 and STAT3, and that interaction is required for efficient STAT3 activation; ChIP-based analysis then identified SLC29A4 (PMAT) as the downstream STAT3 target essential for ANP32A-dependent GSH homeostasis and cisplatin resistance. Deletion mapping pinned amino acids 123-161 of ANP32A as the region required for ANXA2 association and the full drug-resistant phenotype - which converts a correlative biomarker into a druggable interface and a candidate peptide-mimic / small-molecule angle. Critically, tumor-targeted tetrahedral DNA nanostructures delivering siANP32A lowered GSH, raised lipid peroxidation and suppressed ESCC xenograft growth, so the entry point is not only diagnostic. Why it matters for BB: this is exactly the shape of the refractory-cancer brief Brown Biotech sells - a stromal-immune-metabolic escape mechanism with a named target, a defined interaction domain, and an in-vivo delivery proof of concept - and it is human material, so the reasoning layer can be validated against clinical relevance immediately. It also reinforces id:135 and id:136's carrier-and-vehicle thesis: here the carrier is the CAF-derived EV, the vehicle is the DNA nanostructure, and the target sits in the stromal cell rather than in the tumor cell that actually fails to respond. ### 2. [Signal 2 - GSE277858 YBX3 as a brown-adipose RNA-binding protein controlling BCAA flux and thermogenesis, Mus musculus, score 7, n=8, pdat 2026/09/30]: the highest-scoring NOVEL accession in today's scan and the first NOVEL brown-adipose post-transcriptional-regulation dataset to enter the retention window. The dataset is the transcriptomic backbone of a mechanistic claim: YBX3 is a BAT-enriched RNA-binding protein induced by cold stimulation and beta3-adrenergic/cAMP signaling and suppressed by warm stimulation; in vitro loss- and gain-of-function show it is essential for brown adipocyte differentiation and thermogenesis; BAT-specific Ybx3 loss damps thermogenesis and exacerbates diet-induced obesity while Ybx3 overexpression promotes thermogenesis and confers protection against diet-induced metabolic dysfunction. Transcriptome analysis plus mitochondrial stress test show Ybx3 deficiency compromises mitochondrial oxidative phosphorylation, and mechanistically YBX3 stabilizes Slc3a2 mRNA to facilitate branched-chain-amino-acid flux and promote BCAA catabolic rate-limiting enzyme expression in a PPARgamma-dependent manner. Why it matters for BB: Brown Biotech's HBAT brown-adipose and senescence workstream has so far been run at the phenotype and flux level; this is the missing post-transcriptional CONTROL POINT - a single RNA-binding protein whose loss-of-function is sufficient to cause thermogenic failure and whose gain-of-function is protective, which is the ideal shape for a target-nomination brief and for a peptide-adjacent or RNA-modulating intervention thesis. It also links the longevity lane to a genuinely druggable node (BCAA catabolism is upstream of the mTOR and insulin axes, so the reader can trace a therapeutic chain without leaving the dataset. ### 3. [Signal 3 - GSE317552 exhaustion-state-gated synthetic RNA circuit via a LAG-3 isoform, Mus musculus; Homo sapiens, score 6, n=10, pdat 2026/10/22]: the most infrastructure-forward AI drug discovery signal of the day, and the first NOVEL RNA-processing-circuit dataset in the retention window. The authors discovered and characterized RNA isoforms specifically expressed in human intratumoral exhausted CD8+ T cells relative to peripheral CD8+ subsets, including a previously unannotated isoform encoding the immune checkpoint LAG-3 that governs LAG-3 protein expression in exhausted CD8+ T cells; they then harnessed the physiologic regulation of LAG-3 splicing to generate a synthetic RNA for conditional expression of therapeutic payloads exclusively in exhausted CD8+ T cells. That switch conditionally expressed the transcription factor FOXO1 and IL-12 and promoted anti-tumor activity of TCR T cells versus cells with constitutive expression of the same proteins. Why it matters for BB: this is a state-detection-plus-payload-delivery architecture rather than a single-target story, and the discriminator is the exhausted state itself - which is the same logic BB's AI-DD lane applies when it argues for carrying a payload on a state marker rather than on a constitutively expressed receptor. It is also the clearest recent example of the modality we keep asking clients about: RNA isoform-level engineering as a programmable-delivery layer, directly usable as a hypothesis-generation frame for the refractory-cancer service, where checkpoint resistance is the standard failure mode. The dataset spans mouse and human, so the cross-species translation step is already paid for. ### 4. [Cross-signal synthesis - the new observation of the day is that all three NOVEL signals move the control point from the cell to the STATE, and each one supplies a delivery vehicle matched to that state]: Signal 1 moves it from the tumor cell to the CAF (the stromal cell sets GSH, and the vehicle is a tetrahedral DNA nanostructure carrying siRNA into the stroma); Signal 2 moves it from the pathway to the post-transcriptional layer (the state is the beta3-adrenergic/cAMP-driven thermogenic program, and the lever is the stability of a single mRNA, Slc3a2, gating BCAA flux); Signal 3 moves it from the receptor to the splicing state (the state is intratumoral exhaustion, and the vehicle is a synthetic RNA switch that only fires there). Read together they compose with id:135's amplifying-layer thesis and id:136's carrier-and-vehicle thesis into a three-digest arc with a single proposition: THE PRODUCTIVE TARGET IS THE STATE-CARRIER, AND EVERY USABLE INTERVENTION CARRIES ITS OWN STATE DETECTOR. The Brown Biotech implication is concrete - a named 'state-detection-and-delivery' method note for the AI-DD and refractory-cancer lanes, because the same reasoning move (find the state, then route the payload to it) is what the clients in these lanes are actually paying for, and today's three datasets are three independent demonstrations of it. --- ## 💼 Next Steps - **[Open GSE351727 in GEO (CAF-ANP32A glutathione axis, ESCC cisplatin resistance, n=6)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE351727)** - **[Open GSE277858 in GEO (YBX3 brown-adipose RBP / BCAA flux, n=8)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE277858)** - **[Open GSE317552 in GEO (LAG-3 isoform / synthetic RNA circuit, n=10)](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE317552)** - **[Read ANP32A cisplatin resistance literature](https://pubmed.ncbi.nlm.nih.gov/?term=ANP32A+cancer+associated+fibroblast+cisplatin+resistance)** - **[Read YBX3 brown adipose thermogenesis literature](https://pubmed.ncbi.nlm.nih.gov/?term=YBX3+brown+adipose+thermogenesis)** - **[Read T-cell exhaustion RNA splicing literature](https://pubmed.ncbi.nlm.nih.gov/?term=T+cell+exhaustion+RNA+splicing+LAG-3+isoform)** - **[Request refractory-cancer brief on stromal EV resistance mechanisms](https://brownbio.tech/services/refractory-cancer#brief)** - **[Request AI drug discovery brief on state-gated delivery](https://brownbio.tech/services/ai-drug-discovery#brief)** - **[Request longevity brief on brown-adipose target nomination](https://brownbio.tech/services/longevity#brief)** - **[View AI drug discovery service](https://brownbio.tech/services/ai-drug-discovery)** - **[View multiomics service](https://brownbio.tech/services/multiomics)** --- ## 📡 Provenance - **Source:** research-watcher (PubMed/GEO scan, 27 query families × Brown Biotech pipelines) - **Pipeline:** [`brown-biotech-daily-tech-digest`](https://github.com/ohbryt/brown-biotech-platform) — 06:00 KST cron - **Generated:** 2026-10-27 06:06 KST - **Repo:** `ohbryt/brown-biotech-platform` --- _Auto-published by `brown_biotech_research_digest_publisher.py` · [brownbio.tech](https://brownbio.tech) · Decision-ready research, daily._
Brown Biotech Research Digest — 2026-10-27
PubMed/GEO scan · research-watcher · 06:00 KST