Language Selection

Get healthy now with MedBeds!
Click here to book your session

Protect your whole family with Orgo-Life® Quantum MedBed Energy Technology® devices.

Advertising by Adpathway

         

 Advertising by Adpathway

Ten Innate Lymphoid Cell Genes Predict Cervical Cancer Survival in Computational Study

2 hours ago 3

PROTECT YOUR DNA WITH QUANTUM TECHNOLOGY

Orgo-Life the new way to the future

  Advertising by Adpathway

Researchers investigating the cervical cancer tumor microenvironment have identified ten prognostic genes linked to innate lymphoid cells through combined single-cell transcriptomic analyses, machine learning modeling, and laboratory validation. In the study published in Infectious Agents and Cancer, the team isolated key gene expression modules associated with tissue-resident innate lymphoid cells and developed a predictive risk-scoring model to evaluate patient survival. Cervical cancer ranks globally as the fourth most commonly diagnosed malignancy among women, with high-risk human papillomavirus variants acting as the primary etiological drivers. While early screening and vaccination initiatives have reduced disease burden, advanced stages continue to present unfavorable survival outcomes, prompting investigators to search for cellular markers that can clarify tumor biology and patient risk.

Innate lymphoid cells represent a class of tissue-resident immune effectors that maintain barrier integrity, support immunological homeostasis, and execute rapid local responses when prompted by environmental tissue stress signals, alarmins, and cytokines. Derived from common lymphoid progenitors through an established differentiation hierarchy, these cells mature into natural killer cells, lymphoid tissue-inducing cells, and helper subsets. Despite their recognized presence within the endometrium and uterine stroma, the operational states and clinical significance of innate lymphoid cell populations within cervical cancer tissue have remained insufficiently defined. To address this knowledge gap, the investigators analyzed single-cell sequencing profiles alongside bulk transcriptomic datasets to map how these specific immune subsets correlate with patient prognosis.

The experimental workflow began with single-cell RNA sequencing data obtained from Gene Expression Omnibus accession GSE208653, comprising cervical specimens across normal tissue, human papillomavirus infection, precancerous cervical lesions, and cervical cancer. Quality control procedures and doublet filtering retained 53,885 high-quality singlet cells for downstream computational processing. Unsupervised clustering initially resolved fourteen distinct cell clusters, categorizing ten general lineages including T cells, neutrophils, epithelial cells, myeloid cells, fibroblasts, plasma cells, mast cells, B cells, endothelial cells, and innate lymphoid cells. Secondary clustering focused specifically on innate lymphoid cells resolved fifteen subclusters, which were categorized into three principal subsets: natural killer and type 1 innate lymphoid cells, type 2 innate lymphoid cells, and type 3 innate lymphoid cells based on canonical marker profiles.

Transcriptional profiling showed that the natural killer and type 1 subset predominated across all evaluated tissue specimens, defined by the expression of NKG7, TBX21, and IKZF3 without CD3, GATA3, or RORC. Pseudotime trajectory reconstruction established that natural killer and type 1 cells were represented throughout the entirety of differentiation, whereas type 2 cells appeared primarily during intermediate and late phases. Cell-cell communication analyses further mapped signaling networks connecting these immune populations to neighboring cells in the microenvironment, detecting altered pathways such as CD99, collagen-CD44 signaling from fibroblasts, and human leukocyte antigen interactions between neutrophils and lymphoid subsets across the cervical tissue landscape.

To isolate gene networks specifically coupled to innate lymphoid cells, the authors applied high-dimensional weighted gene co-expression network analysis to aggregated metacell profiles derived from normal and cancer specimens. Network topology analysis selected a soft threshold power of seven to achieve scale-free distribution, grouping expressed transcripts into seven co-expression modules. Evaluating correlation metrics between module eigengenes and cellular populations revealed that the turquoise module exhibited the strongest positive correlation with innate lymphoid cells, reaching a Pearson coefficient of 0.34. Extracting the transcripts comprising this turquoise module yielded 1,282 key module genes, which were subsequently analyzed for functional pathway enrichment.

Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment assessments indicated that these module genes were heavily implicated in leukocyte cell-cell adhesion, vacuolar membrane structures, GTPase regulatory activity, T cell receptor signaling, and Th17 cell differentiation pathways. Concurrently, differential expression testing of bulk transcriptomic data from merged cancer cohorts identified 9,184 differentially expressed genes, comprising 3,798 downregulated and 5,386 upregulated transcripts. Univariate Cox proportional hazards modeling filtered these candidate transcripts down to 363 survival-associated genes. Intersecting the 363 survival-associated genes with the 1,282 innate lymphoid cell module genes produced an overlapping set of 21 candidate prognostic genes.

To select the most robust prognostic architecture from these intersecting candidates, the investigators evaluated 101 machine learning algorithm combinations spanning random survival forests, elastic networks, CoxBoost, and support vector machine approaches. A combined Least Absolute Shrinkage and Selection Operator and Stepwise Cox regression framework emerged as the optimal formulation by maximizing the concordance index under five-fold cross-validation. This procedure isolated ten prognostic genes: AGPAT4, UCP2, NDUFB7, SELL, TNF, WDR45, TFRC, FCRL3, TBX21, and AKNA. An integrated prognostic risk score was formulated by multiplying each gene’s expression value by its corresponding regression coefficient.

Model performance evaluation in the primary training cohort showed an area under the receiver operating characteristic curve of 0.750 for three-year overall survival, accompanied by a concordance index of 0.726. Stratification of patients into high- and low-risk cohorts using cross-validated cutoff thresholds demonstrated significantly shortened overall survival among high-risk individuals on Kaplan-Meier analysis. Independent external validation in cohort GSE30759 yielded a three-year area under the curve of 0.646 and a concordance index of 0.651, while validation in cohort GSE44001 showed a three-year area under the curve of 0.714, verifying consistent risk differentiation across disparate patient series without parameter re-estimation.

Functional pathway assessments across stratified risk tiers revealed marked immune alterations between groups. Gene set enrichment and variation analyses highlighted heightened activity in transforming growth factor beta signaling, Wnt signaling, and extracellular matrix receptor interactions within high-risk profiles, contrasted against suppressed oxidative phosphorylation and immunodeficiency cascades. Single-sample enrichment of 28 immune cell subsets showed significant compositional shifts in seventeen populations, including activated B cells and effector memory CD8 T cells. Furthermore, correlation testing demonstrated associations between elevated risk scores and heightened tumor immune exclusion alongside diminished calculated drug sensitivity across multiple therapeutic compounds.

Experimental laboratory testing evaluated expression levels of the candidate markers in biopsy specimens collected from five cervical squamous cell carcinoma patients alongside matched adjacent normal tissues. Quantitative reverse transcription polymerase chain reaction verified significant upregulation of FCRL3, NDUFB7, TNF, TBX21, UCP2, and TFRC in tumor tissues, while AGPAT4 and AKNA demonstrated significant downregulation. Western blot assays subsequently confirmed elevated FCRL3 protein abundance in tumor samples relative to controls. The authors noted that while these computational and preliminary laboratory findings delineate candidate prognostic markers and immune signatures, prospective validation in larger multi-center cohorts and detailed functional experiments remain necessary before clinical translation.

In mapping the expression profiles of the ten prognostic markers across innate lymphoid subsets, the researchers observed distinct lineage-dependent distributions. Five signature genes, namely AGPAT4, NDUFB7, TNF, WDR45, and TBX21, were predominantly enriched within natural killer and type 1 innate lymphoid cells. In contrast, FCRL3 and AKNA displayed preferential expression within type 2 innate lymphoid cells, whereas UCP2, SELL, and TFRC localized primarily to type 3 innate lymphoid cells. Within the broader single-cell landscape, detailed surface and transcriptional criteria distinguished these cell types: natural killer cell maturation depended on both EOMES and TBX21, while helper type 1 cells relied exclusively on TBX21. The protein-protein interaction network constructed from key co-expression module transcripts further highlighted 313 nodes and 344 edges, identifying specific interactive partnerships including SUN2 with the SYNE family and ADRB2 with ARRB2 and SLC9A3R1.

The computational architecture also reflected deliberate methodological controls designed to handle disparate clinical data sources. The primary training cohort combined 308 cervical squamous cell carcinoma cases from the Cancer Genome Atlas, including 283 samples with complete overall survival records, alongside ten normal cervical tissue specimens from the Genotype-Tissue Expression repository. To prevent introducing artificial artifacts between distinct sequencing platforms, the investigators avoided mandatory batch correction during preliminary differential expression exploration, relying instead on two external transcriptomic cohorts and clinical tissue assays to confirm signature stability and eliminate platform-derived bias.

The study placed particular emphasis on AGPAT4 and TBX21, highlighting their previously unrecognized prognostic roles in cervical cancer. AGPAT4 encodes an outer mitochondrial membrane enzyme that converts lysophosphatidic acid into phosphatidic acid; its reduced expression in cervical carcinoma mirrors patterns documented in thyroid cancer, where low levels alter microenvironmental lipid signaling. Meanwhile, TBX21 serves as a master regulator of type 1 helper T cell development and exhibited a strong positive correlation of 0.74 with activated CD8 T cell infiltration in the tumor microenvironment. Conversely, TFRC demonstrated an inverse correlation of minus 0.39 with activated CD8 T cells. TFRC functions in cellular iron absorption via transferrin-mediated endocytosis, an essential metabolic pathway whose overexpression has been linked to advanced clinical staging, lymph node metastasis, and altered kinase signaling cascades.

Complementing the immunological characterization, drug response simulations across 198 pharmacologic agents revealed significant IC50 disparities for 114 compounds between risk cohorts. Four candidate therapeutics, specifically PF4708671, dactolisib, leflunomide, and AZD8055_1059, exhibited significantly higher IC50 values in high-risk patients, indicating diminished therapeutic sensitivity. The authors also discussed potential innate immune checkpoint mechanisms, noting that regulatory networks such as CD47-SIRPalpha signaling can constrain macrophage phagocytosis and innate lymphoid cytotoxicity, promoting immune evasion. Although the machine learning signature reliably stratified overall survival across public databases and verified consistent transcript and protein shifts for markers like FCRL3, the authors emphasized substantial limitations. The biological mechanisms linking these genes to tumor progression remain uncharacterized, requiring rigorous prospective human studies, downstream pathway dissection, and in vivo animal models before clinical implementation.

Subject of Research: Identification and validation of innate lymphoid cell-related prognostic genes in cervical cancer

Article Title: Identification of prognostic genes associated with innate lymphoid cells in cervical cancer through bioinformatics analyses and experimental assays

Article References: Sun, J., Xu, Y., Guo, K., & Feng, Q. (2026). Identification of prognostic genes associated with innate lymphoid cells in cervical cancer through bioinformatics analyses and experimental assays. Infectious Agents and Cancer, 21(1), Article 75. https://doi.org/10.1186/s13027-026-00788-x

Image Credits: AI Generated

DOI: 10.1186/s13027-026-00788-x

Keywords: cervical cancer, innate lymphoid cells, prognostic genes, machine learning, transcriptomics, single-cell RNA sequencing, hdWGCNA, FCRL3, TBX21, TFRC, UCP2, tumor microenvironment, bioinformatics, biomarkers, overall survival, gene expression, risk stratification

Read Entire Article

         

        

Start the new Vibrations with a Medbed Franchise today!  

Protect your whole family with Quantum Orgo-Life® devices

  Advertising by Adpathway