Growth Hormone‐Related Gene Signatures Reveal Diagnostic Biomarkers, Immune Landscapes, and Potential Therapeutic Targets in Sepsis Through WGCNA and Machine Learning Analyses
Growth-Hormon-assoziierte Gensignaturen zeigen diagnostische Biomarker, Immunlandschaften und potenzielle therapeutische Ziele bei Sepsis durch WGCNA- und Machine-Learning-Analysen
Immunity, Inflammation and Disease · 2026-10 · 14 (10)
DOI: 10.1002/iid3.70566 ↗Auf einen Blick
JAK1 und GATA3 wurden als wachstumshormonbezogene Kandidaten-Biomarker bei Sepsis identifiziert. Der kombinierte Marker zeigte in Trainings- und Validierungskohorten die beste diagnostische Trennschärfe.
Fragestellung
Welche growth-hormone-related genes unterscheiden Sepsis und spiegeln Unterschiede im Immunstatus wider?
Methode
Aus zwei GEO-Mikroarray-Kohorten kombiniert; GSE26378 und GSE28750 zur Validierung. Differentialanalyse, WGCNA, Schnittmenge mit 1205 GHRGs, PPI und vier Machine-Learning-Verfahren; ROC, Nomogramm, Kalibrierung und Decision-Curve-Analyse.
Ergebnisse
Die integrierte Kohorte ergab 2552 DEGs und 1873 Gene aus sepsisassoziierten WGCNA-Modulen; die Schnittmenge mit 1205 GHRGs umfasste 40 Gene. PPI reduzierte auf 23 Hub-Kandidaten. JAK1 und GATA3 wurden konsistent beibehalten und waren in Sepsis herunterreguliert.
Schlussfolgerung der Autor:innen
JAK1 und GATA3 erscheinen als vielversprechende wachstumshormonbezogene diagnostische Biomarker bei Sepsis und sind mit immunologischer Heterogenität verbunden. Die kombinierte Signatur könnte für die molekulare Stratifizierung nützlich sein und braucht weitere Validierung.
Kernaussagen
- JAK1 und GATA3 wurden als stabile diagnostische Kandidaten identifiziert.
- Die kombinierte Signatur trennte Sepsis besser als jedes Gen allein.
- Immunprofile unterschieden sich zwischen zwei auf JAK1/GATA3 basierenden Subtypen.
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Fachartikel in der Zeitschrift Immunity, Inflammation and Disease, online erschienen am 28.10.2026. Verfasst von Xiuhua Li und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Emergency Medicine,, Affiliated Jinhua Hospital, Zhejiang University School of Medicine, Jinhua, Zhejiang, China. Hauptthemen laut MEDLINE: Sepsis, Machine Learning, Growth Hormone, Transcriptome. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Background
Sepsis reshapes inflammatory, immune, and metabolic activity, yet the behavior of genes linked to growth hormone biology within this response has not been well defined. We therefore searched for growth hormone‐related genes (GHRGs) that could distinguish sepsis and asked whether the resulting markers also reflected variation in host immune state.
Methods
For model development, two Gene Expression Omnibus (GEO) microarray cohorts were combined, while GSE26378 and GSE28750 were kept separate for validation. Sepsis‐associated expression signals were first defined by differential analysis and weighted gene co‐expression network analysis (WGCNA), then intersected with 1205 GHRGs. A protein–protein interaction step and four machine‐learning procedures were used to reduce the candidate set. Classification was examined with receiver operating characteristic (ROC) curves and a two‐gene nomogram, with calibration and decision‐curve analyses used as complementary checks. We also examined pathway enrichment, immune‐cell patterns, predicted regulatory interactions, candidate compounds, and expression‐based molecular subgroups.
Results
The integrated cohort yielded 2552 DEGs and 1873 genes from sepsis‐associated WGCNA modules; their overlap with the GHRG collection contained 40 genes. PPI module analysis reduced this set to 23 hub candidates. Machine‐learning screening consistently retained JAK1 and GATA3, both of which were downregulated in sepsis and showed stable diagnostic performance across the training and external validation cohorts. Their combined model provided stronger discrimination than either gene alone. Functional analyses linked the two genes to T‐cell signaling, antigen processing, immune regulation, mitochondrial metabolism, and translational processes. Immune profiling revealed reduced lymphocyte‐associated features together with enhanced myeloid signatures in sepsis. When septic samples were grouped by JAK1/GATA3 expression, two subtypes with different immune and transcriptional profiles emerged.
Conclusions
JAK1 and GATA3 emerged as promising growth hormone‐related candidate diagnostic biomarkers associated with immune heterogeneity in sepsis. Their combined signature may provide a useful framework for molecular stratification and warrants further experimental and clinical validation.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Department of Emergency Medicine,, Affiliated Jinhua Hospital, Zhejiang University School of Medicine, Jinhua, Zhejiang, China.
ORCID-Profile: Yingwei Ding, Bingwen Zhang
Zitieren
APA
Li, X., Liu, A., Ding, Y., Chen, G., & Zhang, B. (2026). Growth Hormone‐Related Gene Signatures Reveal Diagnostic Biomarkers, Immune Landscapes, and Potential Therapeutic Targets in Sepsis Through WGCNA and Machine Learning Analyses. Immunity, Inflammation and Disease, 14(10). https://doi.org/10.1002/iid3.70566
Vancouver
Li X, Liu A, Ding Y, Chen G, Zhang B. Growth Hormone‐Related Gene Signatures Reveal Diagnostic Biomarkers, Immune Landscapes, and Potential Therapeutic Targets in Sepsis Through WGCNA and Machine Learning Analyses. Immunity, Inflammation and Disease. 2026;14(10). doi:10.1002/iid3.70566
BibTeX
@article{Li2026,
author = {Li, Xiuhua and Liu, Ang and Ding, Yingwei and Chen, Gang and Zhang, Bingwen},
title = {Growth Hormone‐Related Gene Signatures Reveal Diagnostic Biomarkers, Immune Landscapes, and Potential Therapeutic Targets in Sepsis Through WGCNA and Machine Learning Analyses},
journal = {Immunity, Inflammation and Disease},
year = {2026},
volume = {14},
number = {10},
doi = {10.1002/iid3.70566}
}
EUVIMED