HLA-G functions as a tumor-intrinsic driver of growth and survival in renal cell carcinoma

ABSTRACT

HLA-G is a non-classical MHC class I molecule with potent immunoregulatory functions that is aberrantly expressed in multiple malignancies, yet its tumor-intrinsic role remains poorly defined. To characterize this potential oncogenic role of HLA-G in clear cell renal cell carcinoma (ccRCC), we utilized integrated transcriptomic, in vitro, and in vivo approaches. Analysis of the Cancer Genome Atlas (TCGA) ccRCC cohort revealed that elevated HLA-G expression was associated with immunosuppressive programs and cell populations. Interrogation of a publicly available ccRCC single-cell RNA sequencing dataset revealed that HLA-G expression within tumor epithelial clusters is associated with hypoxia-driven, metabolic transcriptional programs. Multiplex immunofluorescence of human ccRCC specimens confirmed the presence of tumor cell–intrinsic HLA-G expression in advanced disease. Functional studies using RCC7 cells expressing the full-length canonical HLA-G isoform (RCC7/HLA-G1) demonstrated increased proliferation, migration, clonogenicity, cell-cycle progression, and resistance to apoptosis compared with HLA-G-negative RCC7wt cells. RCC7/HLA-G1 xenografts exhibited accelerated tumor growth accompanied by the activation of proliferative, stemness-related, and metabolic programs. Multi-omics analyses further revealed enhanced mitochondrial activity and redox metabolic adaptation in HLA-G-expressing tumors. Mechanistically, HLA-G expression was associated with increased VEGF-C expression and enhanced VEGFR3 signaling, suggesting the activation of a VEGF-C/VEGFR3-associated pro-survival pathway. In three-dimensional tumor spheroid immune cell co-culture models, HLA-G expression reduced CD8⁺ T-cell-mediated cytotoxicity while promoting regulatory T-cell expansion and macrophage polarization toward an immunosuppressive M2-like phenotype. Collectively, these findings establish HLA-G as a key contributor to tumor progression and immune suppression in ccRCC and support HLA-G as a promising therapeutic target.

Introduction

Renal cell carcinoma (RCC) is the sixth most common cancer in men and ninth in women, with an estimated 81,800 new cases diagnosed annually in the United States. ¹ The three predominant histologic subtypes are clear cell renal cell carcinoma (ccRCC, 75%–80%), papillary RCC (10%–15%), and chromophobe RCC (~5%). ² RCC is considered one of the most highly immune-infiltrated solid tumors, containing a diverse mixture of immune cells, including T cells, macrophages, and myeloid-derived suppressor cells (MDSCs), within its tumor microenvironment (TME). This immune richness underlies RCC’s robust responsiveness to immune-checkpoint blockade (ICB) therapy. ³ Clinical trials employing monoclonal antibodies targeting the PD-1/PD-L1 and CTLA-4 pathways (nivolumab, pembrolizumab, and ipilimumab) have demonstrated durable clinical responses and significantly improved survival in both metastatic and locally advanced RCC. ^{4 , 5} Consequently, combinatorial regimens integrating PD-1 blockade with VEGF-targeted tyrosine kinase inhibitors (TKIs) have emerged as frontline standards of care for RCC. ^{6 , 7} Despite these advances, however, a substantial proportion of RCC patients ultimately develop acquired resistance to ICB, highlighting the urgent need to identify complementary immune-evasion pathways that can be therapeutically targeted. ^{8 , 9} Within this context, the non-classical MHC class I molecule human leukocyte antigen G (HLA-G) has gained significant attention as a potential mediator of immune escape in RCC. Originally discovered for its pivotal role in maintaining feto-maternal tolerance during pregnancy, ^{10 , 11} HLA-G has since been recognized as a key factor in promoting long-term allograft acceptance in solid-organ transplantation and is involved in a variety of other immunological processes, including viral infections and autoimmune disorders. ^12-16 Through interactions with the inhibitory receptors LILRB1 (ILT2), LILRB2 (ILT4), and KIR2DL4, HLA-G suppresses natural killer (NK) cell cytotoxicity, T-cell proliferation, and dendritic cell maturation, thereby establishing local immunosuppression or tolerance. ^{12 , 17} More recently, elevated HLA-G expression has been detected in numerous malignancies, including RCC, melanoma, and hepatocellular carcinoma (HCC). ^{18 , 19} Consequently, HLA-G represents a promising yet under-explored target, whose expression and tumor-intrinsic functions in RCC warrant detailed investigation.

In this study, we sought to dissect the tumor-intrinsic role of HLA-G in RCC using integrated transcriptomic, in vitro, and in vivo approaches. Analysis of The Cancer Genome Atlas (TCGA) datasets revealed that HLA-G is robustly overexpressed across multiple human malignancies, including clear-cell and papillary RCC, and its expression positively correlates with immunosuppressive cellular signatures within the tumor microenvironment. To determine whether HLA-G exerts direct oncogenic effects independent of immune modulation, we utilized RCC7 cell models stably expressing HLA-G1 (the full-length membrane-boundcanonical isoform of HLA-G) ²⁰ and compared their phenotypic and transcriptional profiles with those of their wild-type counterparts lacking HLA-G. Functional assays demonstrated that HLA-G expression markedly enhances tumor cell proliferation, migration, and cell cycle progression while suppressing apoptotic signaling. Xenograft studies in immunodeficient mice further indicated that HLA-G promotes accelerated tumor growth and activates transcriptional programs linked to stemness and metabolic fitness. Mechanistically, RNA sequencing and pathway analyses revealed upregulation of PI3K–AKT–mTOR, MAPK, and RAS signaling cascades, accompanied by the repression of interferon-responsive and immune-activation pathways. Beyond tumor-intrinsic functions, in three-dimensional tumor-spheroid immune-based co-culture models, HLA-G expression conferred resistance to immune-mediated cytotoxicity and promoted an immunosuppressive microenvironment characterized by reduced CD8⁺ T-cell activity and the expansion of regulatory T cells and M2-like macrophages. Collectively, our findings identify HLA-G as a dual-function molecule in RCC, possessing both extrinsic immunosuppressive and intrinsic oncogenic properties. By directly promoting tumor proliferation, survival, and stem-like transcriptional and metabolic reprogramming, as well as shaping an immunosuppressive TME, HLA-G emerges as a multifaceted contributor to RCC progression. These results expand the current paradigm of HLA-G biology beyond immune tolerance, suggesting that targeting HLA-G may provide therapeutic benefit not only through immune restoration but also by disrupting tumor-intrinsic survival signaling.

Materials and methods

Cell line

The RCC7 cell line, derived from a ccRCC patient, was previously described. ²¹ RCC7wt lacks HLA-G expression, whereas RCC7/HLA-G1 was generated by lentiviral transduction with HLA-G1 cDNA provided by Dr. Carosella. RCC7wt, RCC7/HLA-G1, and RCC10RGB (RCB1151; RIKEN Cell Bank) cells were cultured under standard conditions/reagents as described in Supplementary Table 1 and passaged at 75%–80% confluence.

Xenograft studies

NOD scid γ (NSG; NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ; JAX #005557) mice were obtained from The Jackson Laboratory and used at 6–8 weeks of age. Age- and sex-matched mice were housed under pathogen-free conditions at Augusta University. All procedures were approved by the Augusta University IACUC (#2008-0051), conducted in accordance with institutional guidelines, ARRIVE 2.0 recommendations, and AVMA guidelines. Cell line-derived xenografts were established by subcutaneous injection of 1 × 10⁶ RCC7wt or RCC7/HLA-G1 cells in 100 µL PBS into the left flank (n = 4/group). The procedures were performed under 2%–3% isoflurane anesthesia. Tumor growth was measured every 2 d for 20 d using digital calipers, and the volume was calculated as V = L × W² × (π/6). Measurements were performed by a blinded investigator. Mice were euthanized by CO₂ inhalation followed by cervical dislocation at the study endpoint or upon reaching humane tumor size limits.

Fluorescent multiplexed immunohistochemistry assay, imaging, and quantitation

Human FFPE ccRCC tissues (Grades I–IV) were obtained from US Biomax (HuCAT389, HuCAT390, HuCAT396) and the Georgia Cancer Center Biorepository (RRID: SCR_02727) (Supplementary Table 2); unmatched normal kidney tissue (KD242) was included as a control. All specimens were de-identified and collected with informed consent in accordance with the Declaration of Helsinki. Multiplex immunofluorescence was performed using the Opal 7-Color Manual IHC Kit (Akoya Biosciences) according to the manufacturer's instructions. Following deparaffinization, antigen retrieval, and blocking, the sections were stained sequentially with anti-HLA-G (Abcam) and anti-CAIX (Novus Biologicals) antibodies. Signal detection was performed using Opal HRP polymer and tyramide signal amplification, with DAPI nuclear counterstaining. Images were acquired using the Vectra Automated Quantitative Pathology Imaging System and analyzed by spectral unmixing and single-cell quantification using InForm v2.3.0.

TCGA and single-cell RNA-seq analyses

TCGA-KIRC transcriptomic and clinical data were obtained from UCSC Xena. Log₂-transformed TPM matrices were analyzed individually. Differential HLA-G expression between tumor and adjacent normal tissues, across stages, immune infiltration (TIMER2.0), and differential gene expression between HLA-G-high and HLA-G-low tumors were evaluated using limma (adjusted p < 0.05), with pathway enrichment performed using clusterProfiler in R. Single-cell RNA-seq data from ccRCC patients (GEO: GSE242299) were analyzed using Seurat v5 following quality control, log-normalization, PCA, UMAP, and unsupervised clustering. The cell populations were annotated using lineage markers, and the tumor epithelial cells were stratified into HLA-G⁺ and HLA-G⁻ populations. Differential expression and Hallmark gene set enrichment analyses (MSigDB) were performed, with adjusted p < 0.05 considered significant.

Flow cytometry, apoptosis, and cell cycle analysis

RCC7wt and RCC7/HLA-G1 cells (1 × 10⁶) were stained with fluorochrome-conjugated antibodies (BioLegend) against HLA-G receptors (ILT2, ILT3, and ILT4), VEGFR2, VEGFR3, and immune markers (FOXP3, CD210, CD45, CD4, CD8, Granzyme B, CD11b, CD163, CD14, and CD73) (Supplementary Table 3). HLA-G expression in RCC10RGB cells was confirmed using an anti-HLA-G antibody and an isotype control. The cells were blocked with Human TruStain FcX (BioLegend), stained in FACS buffer for 45 min at 4 °C, and incubated with Zombie Aqua viability dye. Intracellular staining was performed using the Cyto-Fast Fix/Perm Buffer Set (BioLegend). For apoptosis analysis, the cells were resuspended in Annexin V Binding Buffer (1 × 10⁶ cells/mL) and stained with APC Annexin V and 7-AAD for 15  min at room temperature. For cell-cycle analysis, cells were fixed in 70% ethanol, treated with RNase A (100  µg/mL; Qiagen), and stained with propidium iodide (50 µg/mL; BioLegend) for 15  min at 4 °C. Flow cytometry data were acquired on Attune NxT (Thermo Fisher Scientific) or FACSCanto (BD Biosciences) instruments and analyzed using FlowJo v10.

CCK-8 and WST cell viability assays

Cells (5 × 10³/well) were seeded in 96-well plates and cultured under standard conditions. For proliferation assays, RCC7wt and RCC7/HLA-G1 cells were monitored for 0–72 h using the Cell Counting Kit-8 (CCK-8; Tocris Bioscience). For the inhibitor studies, RCC7/HLA-G1 cells were treated with increasing concentrations of 87G, JNJ-78306358, or SAR131675 for the indicated time periods to determine dose- and time-dependent effects and to identify optimized concentrations for subsequent experiments. RCC7wt and RCC10RGB cells were treated with selected concentrations of 87G or JNJ-78306358 as indicated. At each time point, CCK-8 or WST reagent was added according to the manufacturer’s instructions, incubated for 1 h at 37 °C, and the absorbance was measured at 450 nm using a Synergy HTX microplate reader (BioTek). The cell viability was normalized to that of the corresponding untreated control.

Migration and clonogenicity assays

For the wound-healing assays, the cells were seeded in 6-well plates and grown to ~80% confluence. Wounds were generated using a sterile 1000 µL pipette tip, washed with PBS, and cultured in DMEM containing 2% FBS and 1% penicillin/streptomycin at 37 °C with 5% CO₂. Wound closure was monitored using a BZ-X fluorescence microscope (Keyence) and quantified using an ImageJ plug-in. ²² For the clonogenicity assays, cells were seeded at 500 cells/well in 6-well plates (n = 3). After 14 d, the colonies were washed with PBS, fixed with 4% paraformaldehyde for 5 min, stained with 0.05% crystal violet in PBS for 30 min, and washed with PBS before analysis.

Human apoptosis PCR arrays and quantitative real-time PCR

Total RNA from RCC7wt and RCC7/HLA-G1 cells was extracted using TRIzol (Thermo Fisher Scientific), purified with the RNeasy Mini Kit (Qiagen), and reverse-transcribed (1 µg RNA) using the RT² First Strand Synthesis Kit (Qiagen). Apoptosis-related genes were profiled using the RT² Profiler Human Apoptosis PCR Array (Qiagen) according to the manufacturer’s instructions. Relative expression was calculated by the ΔΔCt method and normalized to ACTB, GAPDH, and HPRT1. RT-qPCR validation was performed using gene-specific primers (Supplementary Table 4) on an Applied Biosystems StepOnePlus system, normalized to GAPDH, and expressed as the fold change in RCC7/HLA-G1 cells relative to RCC7wt cells.

Bulk RNA-sequencing and data analysis

Xenograft tumors from RCC7wt and RCC7/HLA-G1 mice were harvested 20 d post-implantation, dissociated using the gentleMACS Octo Dissociator and Human Tumor Dissociation Kit (Miltenyi Biotec), and processed for RNA extraction using the RNeasy Mini Kit (Qiagen). The RNA quality (RIN ≥ 8) was verified on an Agilent 2100 Bioanalyzer. Library preparation and paired-end sequencing were performed at the Augusta University Integrated Genomics Core (RRID: SCR_026483) using an Illumina NovaSeq 6000. Reads were aligned to GRCh38 with HISAT2, quantified using featureCounts, and analyzed with limma-voom in R. DEGs were defined as |log₂FC| ≥ 1 and adjusted p < 0.05 (Benjamini–Hochberg). Functional enrichment and GSEA (KEGG, GO, Reactome) were performed using clusterProfiler.

Untargeted metabolomic analysis

Untargeted metabolomic profiling of RCC7wt and RCC7/HLA-G1 xenograft tumors was performed at the Georgia Cancer Center Proteomics Core (RRID: SCR_027673) using UPLC–high-resolution Orbitrap mass spectrometry. Metabolites were extracted from frozen tissue with 80% methanol, homogenized, centrifuged, dried, reconstituted, and analyzed on an Orbitrap Fusion Tribrid mass spectrometer coupled to an Ultimate 3000 UPLC system (Thermo Fisher Scientific). Metabolites were separated on a C18 column using a water/acetonitrile gradient containing 0.1% formic acid. Data were acquired in positive and negative electrospray ionization modes using data-dependent acquisition and processed with Compound Discoverer v3.3 for feature detection, alignment, normalization, annotation, and differential metabolite analysis.

3D spheroid formation and drug treatment

RCC7wt and RCC7/HLA-G1 cells were cultured in RCCM medium described previously. ²³ For spheroid formation, 1 × 10⁴ cells were seeded in 200 µL RCCM per well in ultra-low attachment 96-well U-bottom plates (Corning) and centrifuged at 500 × g for 1 min. After 3 d, uniformly formed spheroids were selected for further experiment. Optimized drug doses determined from WST assays were used for spheroid experiments. The spheroids were treated with 87G, JNJ-78306358, or SAR131675 and imaged every 24 h using a Keyence BZ-X700 microscope (10×). After 8 d of treatment, the drug-containing medium was replaced with drug-free RCCM, and the spheroids were monitored until Day 16. The spheroid area was quantified using FIJI/ImageJ and normalized to the corresponding Day 0 area. Representative images are shown at Day 0, Day 8, and Day 16 following drug withdrawal/rescue. Day 8 images were selected to demonstrate visible treatment-associated morphological changes.

Spheroid infiltration assays with allogeneic PBMCs and macrophages

Peripheral blood from de-identified healthy donors was obtained from the Shepeard Community Blood Center (Augusta, GA) with written informed consent. In accordance with institutional guidelines, this study was determined not to involve human subjects research. The selected PBMC samples were HLA-typed to RCC7 cells to confirm minimal allogenicity. The spheroid co-culture was established as described. ²³ PBMCs were cultured in RPMI containing 10% FBS and penicillin–streptomycin and activated with plate-bound anti-CD3 and anti-CD28 antibodies (2  µg/mL each; Thermo Fisher Scientific) for 48 h. Activated PBMCs (1 × 10⁵) were added to RCC spheroids at a 10:1 effector-to-target ratio and co-cultured for 3 d in the presence of propidium iodide (Molecular Probes). PBMC infiltration and tumor cell death were monitored using a Keyence BZ-X700 microscope. Spheroids were dissociated with Accutase (Innovative Cell Technologies), and T-cell phenotypes were analyzed by flow cytometry using FMO controls for gating. For the macrophage assays, monocytes isolated using the EasySep Human Monocyte Isolation Kit (STEMCELL Technologies) were cultured in RPMI containing human AB serum (Sigma) and penicillin–streptomycin and differentiated with M-CSF (100 ng/mL; PeproTech) for 96 h. The macrophages (3 × 10⁴) were labeled with Hoechst dye (BD Pharmingen) and co-cultured with RCC spheroids for 3 d. Images were acquired on Day 3 using a Nikon AX R NSPARC confocal microscope (10×, 405 nm). For intracellular cytokine staining, 1 × Brefeldin A (BioLegend) was added during the final 4–6 h before harvest.

Statistics

Statistical analyses were performed using GraphPad Prism 9. Data are presented as mean ± SD unless otherwise stated. Normality was assessed before analysis. Comparisons between two groups were performed using unpaired two-tailed Student’s t-tests. Tumor growth, time-course, and dose‒response experiments were analyzed using two-way ANOVA with appropriate post hoc multiple-comparison correction. P < 0.05 was considered statistically significant.

Results

HLA-G is strongly expressed in RCC

To investigate the role of HLA-G across cancers, we analyzed TCGA RNA-seq data from tumor and adjacent normal tissues. HLA-G was significantly overexpressed in multiple malignancies, including cholangiocarcinoma (CHOL), esophageal carcinoma (ESCA), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), uterine corpus endometrial carcinoma (UCEC), and thyroid carcinoma (THCA) (Figure 1A). Within RCC, both KIRC and KIRP exhibited significantly higher HLA-G expression than matched normal tissues (Figure 1B). In KIRC samples, HLA-G expression was positively correlated with M-MDSC and Treg signatures, suggesting that HLA-G-high tumors are characterized by an immunosuppressive TME (Figure 1C). To dissect this, TCGA-KIRC tumors were stratified into HLA-G-high and low groups and analyzed by ssGSEA. Among the top 20 enriched pathways in HLA-G-high tumors (FDR < 0.05), immune tolerance-associated pathways predominated, including peripheral tolerance induction and positive regulation of T-cell tolerance induction (Figure 1D). To characterize the cellular context of HLA-G expression, we analyzed the ccRCC single-cell RNA-seq dataset GSE242299 (n = 8 tumors). ²⁴ Cluster identities were defined based on DEG analysis using the Seurat FindAllMarkers function and annotated according to established lineage-specific marker genes. Unsupervised clustering identified tumor epithelial clusters (6 and 12), within which HLA-G expression was predominantly localized in Cluster 12 (Figure 1E, F). Notably, HLA-G⁺ tumor epithelial cells co-expressed established carcinoma markers, including KRT8, ²⁵ KRT18, ²⁶ CA9, ^{27 , 28} PAX8, ^{29 , 30} NDUFA4L2, ³¹ and EPCAM, ³² as demonstrated by UMAP feature visualization (Figure 1F) and DotPlot analysis (Figure 1G). Differential expression analysis comparing HLA-G⁺ and HLA-G⁻ tumor epithelial cells revealed distinct transcriptional programs. GSEA demonstrated significant enrichment of hypoxia and glycolysis pathways in HLA-G⁺ cells (Figure 1H). Specifically, the expression of canonical hypoxia-responsive genes, including SLC2A1, ³³ NDUFA4L2, ³¹ BNIP3, ³⁴ VEGFA, ³⁵ and LOX ³⁶ was upregulated in HLA-G⁺ cells, alongside increased expression of glycolytic enzymes such as HK2, ³⁷ LDHA, ³⁸ ALDOC, ³⁹ and PGK1. In contrast, KRAS signaling and inflammation-associated pathways were relatively enriched in HLA-G⁻ tumor cells. Heatmap visualization confirmed the coordinated activation of hypoxia- and metabolism-associated genes in HLA-G⁺ tumor epithelial cells (Figure 1I). Immunofluorescence analysis of human ccRCC specimens further validated the elevated HLA-G protein expression across Grade I, II, III, and IV tumors compared with that in normal kidney (Figure 1J). Across 3 different Grade III and Grade IV ccRCC samples, HLA-G expression co-localized with carbonic anhydrase IX (CAIX)-positive carcinoma cells, whereas in Grade I and II tumors, HLA-G displayed a more diffused stromal expression within the TME (Supplementary Figure 1). This spatial redistribution suggests that HLA-G expression may be dynamically regulated during tumor progression, with early-stage enrichment within the TME followed by acquisition of tumor cell-intrinsic expression in advanced disease.

Figure 1.

HLA-G is upregulated in RCC and linked to immune suppression and translational activity. (A) Pan-cancer TCGA analysis showing elevated HLA-G expression across multiple tumor types. (B) HLA-G is significantly higher in clear cell (KIRC) and papillary (KIRP) RCC compared with the normal kidney. (C) TIMER analysis reveals positive correlations between HLA-G and immunosuppressive cells (Tregs, MDSCs) in KIRC. (D) ssGSEA of GO biological process (GO:BP) terms identifies enrichment of various immune-tolerogenic pathways in HLA-G-high tumors. (E) UMAP visualization of integrated scRNA sequencing data (GSE242299) from eight ccRCC tumors showing major cellular populations identified by unsupervised clustering. (F) UMAP feature expression and (G) dot plot demonstrating the localization of HLA-G within tumor epithelial clusters (clusters 6 and 12) and co-expression with canonical epithelial markers. (H) GSEA comparing HLA-G⁺ and HLA-G⁻ tumor epithelial cells reveals enrichment of hypoxia and glycolysis pathways in HLA-G⁺ cells. (I) The representative heatmap shows coordinated upregulation of hypoxia-responsive and glycolytic genes with HLA-G expression. (J) Fluorescent multiplex IHC images of normal kidney and clear cell RCC (ccRCC) Grade I, II, and IV tissue samples. Scale bar: 20 μm. CAIX (red), DAPI (blue), and HLA-G (green). Quantification of overall HLA-G expression as the MFI in different zones of tumor and control tissues (n = 4 zones per tissue section). Data represent mean ± SEM: significance was determined by two-tailed t-test (**p < 0.01, ****p < 0.0001).

HLA-G promotes cell growth and migration while inhibiting apoptotic signaling in RCC7

To define the functional consequences of HLA-G expression in ccRCC, we compared RCC7/HLA-G1 cells with parental RCC7wt cells lacking HLA-G1. Flow cytometry and RT-qPCR confirmed robust HLA-G expression in RCC7/HLA-G1 cells (Figure 2A, B). CCK-8 assays demonstrated significantly increased proliferation in RCC7/HLA-G1 cells (Figure 2C), and clonogenic assays revealed enhanced colony formation (Figure 2D). Wound healing analysis showed increased migratory capacity (Figure 2E) while cell cycle analysis revealed a reduced G₁-phase fraction with a corresponding increase in G₂/M phase cells in RCC7/HLA-G1 cells (Figure 2F). Annexin V/7-AAD staining showed a significant reduction in late apoptotic cells (Figure 2G), indicating enhanced survival. RT² Human Apoptosis Profiler Array analysis identified upregulation of key anti-apoptotic genes and the downregulation of pro-apoptotic genes in RCC7/HLA-G1 cells (Figure 2H). RT-qPCR validation confirmed increased expression of BIRC2, BCL2, and BCL10 and reduced expression of CASP2 and CASP3 (Figure 2I). To evaluate the therapeutic potential of HLA-G inhibition, RCC7/HLA-G1 cells were treated with the anti-HLA-G antibody 87G or the HLA-G-targeting therapeutic agent JNJ-78306358 ^{40 , 41} , which disrupts HLA-G receptor interactions. Both treatments induced significant dose- and time-dependent reductions in cell viability (Supplementary Figures 2A, B). To validate these findings in an independent model, we selected RCC10RGB cells based on reported HLA-G expression in the Human Protein Atlas and confirmed low but detectable endogenous HLA-G expression by flow cytometry (Supplementary Figure 2C). Consistent with RCC7/HLA-G1 cells, pharmacological HLA-G inhibition significantly reduced RCC10RGB cell viability (Supplementary Figure 2D). Collectively, these findings support a critical role for HLA-G in RCC cell survival and suggest that HLA-G promotes tumor progression by enhancing proliferation and survival while suppressing apoptotic pathways.

Figure 2.

HLA-G promotes cell growth and migration while inhibiting apoptotic signaling in RCC7 cells. (A) Flow-cytometric and (B) RT-qPCR validation confirming marked upregulation of HLA-G1 expression in RCC7/HLA-G1 cells compared with RCC7wt. (C) CCK-8 proliferation assay showing significantly increased cell proliferation in RCC7/HLA-G1 cells relative to RCC7wt. (D) The colony formation assay demonstrates enhanced clonogenicity of RCC7/HLA-G1 cells compared to RCC7wt. (E) Wound healing assay demonstrating enhanced migratory capacity in RCC7/HLA-G1 cells. (F) Cell-cycle analysis by propidium iodide staining showing reduced G₁-phase and increased G₂/M-phase fractions in RCC7/HLA-G1 cells, consistent with accelerated G₁/S transition. (G) Annexin-V/7-AAD staining indicates a significant reduction in late-apoptotic cells in RCC7/HLA-G1 versus RCC7wt. (H) RT² Human Apoptosis Profiler Array illustrating the upregulation of anti-apoptotic genes and the downregulation of pro-apoptotic genes in RCC7/HLA-G1cells in comparison to RCC7wt. (I) Custom RT-qPCR validation confirming increased expression of pro-apoptotic BIRC2, BCL2, and BCL10 genes, and decreased expression of anti-apoptotic CASP2 and CASP3 expression in RCC7/HLA-G1 cells. Data represent mean ± SEM of four independent experiments: significance was determined by two-tailed t-test or two-way ANOVA as appropriate (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

HLA-G promotes tumor growth and stem-like transcriptional reprogramming in RCC7 xenografts in vivo

Our previous data indicated the role of HLA-G in promoting the proliferation, migration, and cell survival of RCC7 cells. To evaluate the tumorigenic effects of HLA-G in vivo, RCC7/HLA-G1 and RCC7wt cells were implanted into NSG mice. RCC7/HLA-G1 tumors exhibited significantly accelerated growth over 20 d compared with RCC7wt controls (Figure 3A). Bulk RNA sequencing of xenograft tumors revealed distinct clustering between the groups based on the top differentially expressed genes (Figure 3B). RCC7/HLA-G1 tumors demonstrated upregulation of genes associated with proliferation (MAPK1 and MAPK3), angiogenesis (VEGFC and VEGFR3), survival (BIRC3), and PI3K–AKT–mTOR signaling (AKT1, MTOR, and EGFR), while IFN-γ-responsive chemokines (CXCL9, CXCL10, and CXCL11) and pro-apoptotic genes (CASP2 and CASP3) were downregulated compared with RCC7wt (Figure 3C, 3D). GSEA demonstrated activation of the PI3K-AKT, RAP1, and RAS signaling pathways (Figure 4A) and strong enrichment of cell cycle-associated programs, including DNA replication and mitotic progression (Figure 4B), in RCC7/HLA-G1 tumors. GO molecular function analysis further identified enrichment of extracellular matrix-related and transporter activities (Figure 4C), whereas Reactome analysis indicated suppression of keratinization-associated pathways (Figure 4D). Filtering GO:BP (biological processes) for stem cell-related terms revealed significant enrichment of stem cell proliferation and maintenance programs in RCC7/HLA-G1 tumors (Figure 4E). RT-qPCR confirmed increased expression of the stemness-associated factors OCT4 and NANOG (Figure 4F). Collectively, these results indicate that HLA-G expression in renal carcinoma cells enhances proliferative, survival, and stemness-associated pathways while repressing differentiation and immune-interaction programs, highlighting its dual role in promoting tumor growth and establishing an immunosuppressive tumor phenotype.

Figure 3.

HLA-G enhances tumor growth and induces transcriptional reprogramming in RCC xenografts. (A) Schematic of the xenograft model showing subcutaneous implantation of 3 × 10⁶ RCC7/HLA-G1 or RCC7wt cells into immunodeficient NSG mice. Tumor growth kinetics over a 20-day period showing significantly accelerated growth in RCC7/HLA-G1 xenografts compared with RCC7wt (****p < 0.0001). Data represent mean ± SEM of n = 5 mice per group; significance was determined by two-way ANOVA. (B) Representative tumor images at Day 20 showing a larger tumor mass in RCC7/HLA-G1 mice. (C) The hierarchical clustering heatmap depicts the top 500 DEGs between RCC7/HLA-G1 and RCC7wt tumors, demonstrating distinct global transcriptional profiles. (D) Volcano plot highlighting key upregulated genes in red (AKT1, MTOR, RAF1, and EGFR) and downregulated genes in green (CASP2, CASP3, IL6, and CXCL9-11) of RCC7/HLA-G1 vs RCC7wt xenografts.

Figure 4.

HLA-G drives the activation of proliferative and stemness-associated pathways in RCC. (A) KEGG pathway enrichment of upregulated genes in RCC7/HLA-G1 tumors showing activation of PI3K–Akt, RAP1, and RAS signaling, and ECM–receptor interaction. (B) GSEA enrichment of GO:BP terms related to cell proliferation in RCC7/HLA-G1 tumors. (C) GO molecular function analysis highlighting enrichment of extracellular matrix structural and adhesion-related terms. (D) Reactome analysis showing the downregulation of keratinization and cornified envelope formation in RCC7/HLA-G1 tumors. (E) Stem cell-related GO analysis identifies enrichment of terms linked to stem cell differentiation, proliferation, and maintenance in RCC7/HLA-G1 tumors. (F) RT-qPCR analysis of key stem cell factors in RCC7/HLA-G1 and RCC7wt tumors. Data represent mean ± SEM of n = 3 technical replicates. Significance was determined by a two-tailed t-test. ( **p < 0.01, ***p < 0.001).

HLA-G expression induces metabolic reprogramming in RCC7 tumors

Bulk RNA sequencing revealed broad transcriptional reprogramming in RCC7/HLA-G1 tumors across multiple signaling pathways. To determine whether these transcriptional changes were accompanied by corresponding functional alterations, we performed untargeted metabolic profiling of RCC7/HLA-G1 and RCC7wt tumors using UPLC-high-resolution Orbitrap mass spectrometry. As illustrated in the Venn diagram (Figure 5A), RCC7/HLA-G1 tumors contained 86 unique metabolites distinct from RCC7wt. Using the MetaboAnalyst platform, we carried out an overrepresentation analysis (ORA) based on the Small Molecule Pathway Database (SMPDB). Enrichment analysis revealed that RCC7/HLA-G1 tumors exhibited significantly elevated levels of metabolites associated with the mitochondrial electron transport chain, catecholamine biosynthesis, the Warburg effect, and glutathione metabolism (Figure 5B). KEGG pathway analysis further identified increased tyrosine metabolism, TCA cycle activity, and glycerophospholipid metabolism (Figure 5C). Consistent with the metabolomic findings, RCC7/HLA-G1 tumors showed increased expression of nuclear-encoded oxidative phosphorylation genes across all 5 electron transport chain complexes-NDUFS1, SDHA, UQCRB, COX5A, and ATP5F1A, relative to RCC7wt tumors (Figure 5D). Live RCC7/HLA-G1 cells also exhibited significantly elevated intracellular ROS levels, as assessed by CellROX, providing a dynamic measure of oxidative stress (Supplementary Figure 3). Supportingly, re-analysis of the GSE242299 dataset demonstrated enrichment of glycolysis, oxidative phosphorylation, reactive oxygen species (ROS), and fatty acid metabolism signatures in HLA-G⁺ tumor cells compared with those in HLA-G⁻ (Figure 5E). Collectively, these findings indicate that HLA-G expression promotes a metabolic state characterized by enhanced mitochondrial activity and redox balance, consistent with increased bioenergetic and antioxidant capacity.

Figure 5.

Untargeted metabolomic profiling and pathway enrichment analysis of RCC7/HLA-G1 versus RCC7wt tumors. (A) Venn diagram illustrating the distribution of detected metabolites between RCC7wt and RCC7/HLA-G1 tumors. (B) Over-representation analysis (ORA) performed using the small molecule pathway database (SMPDB) identified significant enrichment of pathways in RCC7/HLA-G1 tumors. (C) KEGG pathway analysis of the enriched metabolites demonstrated upregulation of specific pathways in RCC7/HLA-G1 tumors compared with RCC7wt tumors. (D) RT-qPCR analysis of the nuclear-encoded mitochondrial oxidative phosphorylation genes NDUFS1, SDHA, UQCRB, COX5A, and ATP5F1A in RCC7/HLA-G1 tumors relative to RCC7wt tumors. Data are presented as mean ± SEM from three independent experiments. **q < 0.01, ***q < 0.001, multiple unpaired t-tests with false discovery rate (FDR) correction using the two-stage Benjamini‒Hochberg method. (E) Violin plots depicting Hallmark glycolysis, oxidative phosphorylation, reactive oxygen species, and fatty acid metabolism module scores in HLA-G⁺ and HLA-G⁻ tumor epithelial cells derived from re-analysis of the ccRCC single-cell RNA sequencing dataset GSE242299.

HLA-G expression correlates with increased VEGFR3 signaling in RCC7

To explore receptor-mediated mechanisms downstream of HLA-G, we assessed the expression of the canonical HLA-G receptors ILT2, ILT3, and ILT4. Flow cytometry revealed minimal expression of these receptors in RCC7wt and RCC7/HLA-G1 tumors (Figure 6A), indicating that HLA-G-mediated tumor cell effects may occur independently of classical ILT2/ILT3 receptor signaling within tumor cells. Previous studies reported increased VEGF-C expression in RCC7/HLA-G1 cells, ⁴² which was also observed in RCC7/HLA-G1 tumors (Supplementary Figure 4). Flow cytometric analysis revealed comparable VEGFR2 expression between groups, whereas VEGFR3, the primary VEGF-C receptor, was significantly elevated in RCC7/HLA-G1 cells (Figure 6B). Despite minimal expression of canonical HLA-G receptors on tumor cells, HLA-G blockade significantly reduced cell viability, supporting a tumor-intrinsic, non-canonical role for HLA-G in RCC growth. To further investigate the functional contribution of HLA-G and VEGFR3 signaling, RCC7/HLA-G1 spheroids were treated with the anti-HLA-G antibody 87 G, JNJ-78306358, or the VEGFR3 inhibitor SAR131675. ⁴³ Treatment from Day 0 to Day 8 significantly reduced spheroid growth, with direct HLA-G blockade producing a stronger anti-proliferative effect than VEGFR3 inhibition (Figure 6C, D). These findings suggest that the VEGF-C/VEGFR3 axis contributes to, but does not fully account for, the HLA-G-associated phenotype. Replacement of inhibitor-containing media with fresh media partially restored spheroid growth, indicating that HLA-G-dependent signaling supports sustained proliferative capacity, as quantified by spheroid area measurements (Figure 6D). Consistent with the absence of HLA-G expression, 87G, JNJ-78306358, or SAR131675 did not significantly affect RCC7wt cell viability (Supplementary Figure 5). Collectively, these data support a central role for HLA-G in RCC cell survival and growth. Although the underlying mechanisms remain to be fully defined, HLA-G may induce the VGF-C/VEGFR2 axis via various pathways, including the PI3K/AKT, MAPK/ERK, STAT3, and NF-κB, which regulate proliferation, survival, and immune evasion. Further studies are required to determine the relative contribution of these pathways to HLA-G-mediated RCC progression.

Figure 6.

HLA-G1 promotes VEGFR3-associated signaling and spheroid growth in RCC7 cells. (A) Flow cytometric analysis of the HLA-G receptors ILT2, ILT3, and ILT4 in RCC7wt and RCC7/HLA-G1 cells. The representative histograms show minimal expression of all three receptors in both cell lines. Isotype controls are shown in gray. (B) Flow cytometric assessment of VEGFR2 and VEGFR3 expression in RCC7wt and RCC7/HLA-G1 cells. Representative histograms and quantification of VEGFR2⁺ and VEGFR3⁺ cell populations are shown. HLA-G1 expression was associated with a significant increase in VEGFR3⁺ cells. Data are presented as mean ± SD; p < 0.05. (C) Representative brightfield images of RCC7/HLA-G1 spheroids treated with the anti-HLA-G antibody 87G, the HLA-G-targeting agent JNJ-78306358, or the VEGFR3 inhibitor SAR131675. The spheroids were treated for 8 d, followed by an 8-day recovery period in drug-free medium. Images were acquired at Day 0, Day 8 (treatment phase), and Day 16 (rescue phase following drug rescue/withdrawal). Scale bars = 100 μm. Images are representative of three independent experiments. (D) Quantification of spheroid growth based on brightfield image analysis. The spheroid area was measured using FIJI/ImageJ, normalized to the corresponding Day 0 spheroid area, and expressed as the fold change relative to Day 0. Data are presented as mean ± SD (n = 4 spheroids per treatment group). Statistical significance was determined relative to the untreated control at the corresponding time point. ns, not significant (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

HLA-G promotes an immunosuppressive microenvironment in RCC

Our data demonstrated a tumor-intrinsic role of HLA-G in RCC. However, HLA-G is also well recognized for its immunomodulatory properties in several pathological contexts, including transplantation and autoimmune disorders. We therefore investigated whether HLA-G could also modulate immune responses within the RCC TME. To address this, we established an in vitro 3D spheroid co-culture system in which RCC7wt or RCC7/HLA-G1 tumor spheroids were co-cultured with CD3/CD28-activated PBMCs derived from healthy donors (Supplementary Figure 6). In RCC7wt spheroids, the addition of activated PBMCs induced prominent tumor cell apoptosis. In contrast, this cytotoxic effect was significantly attenuated in RCC7/HLA-G1 spheroids (Figure 7A). Flow cytometric analysis of dissociated spheroids revealed a significant reduction in the frequency of CD8⁺ Granzyme B⁺ cytotoxic T cells in RCC7/HLA-G1 spheroids compared with RCC7wt spheroids (Figure 7B). Concurrently, RCC7/HLA-G1 spheroids promoted the expansion of CD4⁺CD25⁺Foxp3⁺ regulatory T cells (Tregs) (Figure 7C) and increased the expression of the immunosuppressive cytokine IL-10 within CD4⁺CD25⁺ T cells (Figure 7D). To further examine potential myeloid interactions, we re-analyzed the RCC scRNA-seq dataset (GSE242299). This analysis revealed that the HLA-G receptors LILRB1 (ILT2) and LILRB2 (ILT4) were predominantly expressed in tumor-associated macrophages (TAMs), together with the classical macrophage markers CD68 and CD163 (Figure 7E). CellChat analysis predicted extensive intercellular communication across tumor and immune populations, with HLA-G⁺ tumor cells (cluster 12) showing strong predicted interactions with LILRB2⁺ TAMs (cluster 5) (Figure 7F). These observations suggested that HLA-G-expressing tumor cells might influence TAM polarization within the TME. To test this hypothesis, CD14⁺ monocytes were differentiated into macrophages using M-CSF and co-cultured with RCC7wt or RCC7/HLA-G1 spheroids (Supplementary Figure 7). Flow cytometric analysis revealed comparable levels of macrophage infiltration across both groups (Figure 7G). However, RCC7/HLA-G1 spheroids induced a significant expansion of CD11b⁺CD14⁺CD163⁺ anti-inflammatory M2-like macrophages (Figure 7H) and increased the expression of the immunosuppressive ectoenzyme CD73 across macrophage populations (Figure 7I). Notably, we also observed a significant expansion of ILT4⁺ macrophages in RCC7/HLA-G1 spheroids (Figure 7J), suggesting a positive feedback loop. Taken together, these findings indicate that HLA-G expression contributes to the establishment of an immunosuppressive TME in RCC, characterized by reduced CD8⁺ cytotoxic activity, the expansion of Treg cells, IL-10 production, and the polarization of macrophages toward an M2-like immunosuppressive phenotype.

Figure 7.

HLA-G promotes an immunosuppressive tumor microenvironment in RCC. (A) RCC7wt or RCC7/HLA-G1 tumor spheroids co-cultured with CD3/CD28-activated PBMCs. Tumor cell apoptosis was significantly reduced in RCC7/HLA-G1 spheroids following PBMC-mediated cytotoxicity. (B) Flow-cytometric analysis of dissociated spheroids showing reduced frequencies of CD8⁺ Granzyme B⁺ cytotoxic T cells, (C) increased CD4⁺CD25⁺FOXP3⁺ regulatory T cells (Tregs), and (D) expanded CD4⁺CD25⁺IL-10⁺ T cells in RCC7/HLA-G1 spheroids. (E) Re-analysis of the RCC scRNA-seq dataset (GSE242299) demonstrating the expression of HLA-G receptors LILRB1 and LILRB2 in TAMs expressing CD68 and CD163. (F) CellChat analysis showing predicted intercellular communication between HLA-G⁺ tumor cells (cluster 12) and LILRB2⁺ TAMs (cluster 5). (G) Flow-cytometric quantification of macrophage infiltration following co-culture of M-CSF-differentiated CD14⁺ macrophages with RCC7wt or RCC7/HLA-G1 spheroids. (H) Flow-cytometric analysis reveals expansion of CD11b⁺CD14⁺CD163⁺ M2-like macrophages. (I) Increased expression of CD73 and (J) elevated ILT4⁺ macrophage populations in RCC7/HLA-G1 spheroids. Data represent mean ± SEM of n = 5 technical replicates. Significance was determined by a two-tailed t-test. (NS - not significant, *p < 0.05,**p < 0.01, ***p < 0.001, ****p < 0.0001).

Discussion

Fetal development and tumorigenesis share notable biological parallels, including rapid cellular proliferation, elevated anti-apoptotic signaling, and increased telomerase activity. ⁴⁴ Both the placental environment and the TME promote immune tolerance through the expansion of regulatory T cells and suppressive macrophage populations. ⁴⁵ Within these contexts, HLA-G functions as a key tolerogenic immune checkpoint that exhibits overlapping expression patterns between placental tissues and malignant tumors. Physiologically, HLA-G is expressed on extravillous trophoblasts that invade the maternal decidua – an invasive process analogous to tumor infiltration and metastasis. In tumors, HLA-G provides a selective advantage by enabling immune evasion through the suppression of T and B cell activation, the inhibition of NK and CD8⁺ cytotoxic responses, and the impairment of neutrophil and dendritic cell functions. Consistent with these roles, elevated HLA-G expression has been documented across multiple malignancies, including HCC, ⁴⁶ RCC, ^{47 , 48} chronic lymphocytic leukemia, ^{49 , 50} and non-Hodgkin’s lymphoma. ^{51 , 52} Here, we demonstrated that HLA-G co-expression with CAIX- an established marker for RCC- across Grade III and IV ccRCC patient samples, suggesting that tumor cell-intrinsic HLA-G expression is associated with higher-grade disease. These histological findings require validation in a larger annotated ccRCC cohort to establish their broader clinical significance.

To investigate the intracellular signaling programs mediated by HLA-G in RCC, we utilized RCC7 cell lines derived from a ccRCC patient, in which the parental RCC7wt cells naturally lack HLA-G expression, while the lentivirally transduced RCC7/HLA-G1 variant robustly expresses the canonical HLA-G1 isoform. Through comparative in vitro and xenograft analyses, we systematically defined the signaling, metabolic, and immune consequences of HLA-G expression in RCC. Our findings demonstrate that HLA-G confers strong oncogenic properties in RCC cells. RCC7/HLA-G1 cells displayed enhanced proliferation, clonogenic growth, migration, and survival compared with RCC7wt cells. These effects were accompanied by accelerated G₁/S transition and suppression of apoptotic signaling, as evidenced by the upregulation of BIRC2, BCL2, and BCL10 and the downregulation of CASP2 and CASP3, consistent with HLA-G-driven apoptotic resistance reported in hepatocellular carcinoma and gliomas. ⁵³ Transcriptomic analysis of RCC7/HLA-G1 xenograft tumors confirmed the upregulation of MAPK1, MAPK3, AKT1, MTOR, and EGFR with enrichment of PI3K–AKT, RAP1, and RAS signaling – central regulators of proliferation, metabolic adaptation, and therapy resistance. RCC7/HLA-G1 tumors additionally exhibited enrichment of stemness-associated gene signatures linked to therapeutic resistance and recurrence in RCC, suggesting that HLA-G promotes tumor plasticity through coordinated activation of proliferative and stem-like transcriptional programs. Metabolomic profiling further revealed enhanced mitochondrial activity, glutathione metabolism, and alterations in tyrosine and glycerophospholipid pathways, indicative of a metabolically active, redox-balanced state supporting rapid proliferation. Mechanistically, VEGF-C/VEGFR3 signaling has emerged as a downstream component of HLA-G-mediated tumor remodeling, with RCC7/HLA-G1 tumors displaying elevated VEGF-C expression and VEGFR3⁺ cell expansion. However, VEGFR3 inhibition produced only modest spheroid growth suppression relative to direct HLA-G blockade, indicating that VEGF-C/VEGFR3 represents one node within a broader signaling network. Mechanistically, our transcriptomic data revealed concurrent upregulation of both VEGFC and VEGFR3, together with enrichment of multiple signaling pathways known to regulate their expression, including the PI3K-AKT-mTOR, RAS, and RAP1 signaling pathways in RCC7/HLA-G1 tumors (Figures 3 and 4). These findings and others ⁴² suggest a potential model in which HLA-G expression promotes oncogenic signaling networks that subsequently enhance VEGF-C/VEGFR3 signaling; however, additional mechanistic studies are needed to dissect the exact mechanism. The significant antiproliferative effect of HLA-G blockade despite minimal ILT2/ILT3/ILT4 tumor cell expression suggests the involvement of a non-canonical receptor or binding partner, the identity of which warrants future investigation.

Beyond tumor-intrinsic effects, HLA-G contributed to immunosuppressive TME remodeling. In 3D spheroid co-culture models, RCC7/HLA-G1 spheroids resisted PBMC-mediated cytotoxicity, accompanied by reduced CD8⁺ Granzyme B⁺ T cells, expanded CD4⁺CD25⁺FOXP3⁺ regulatory T cells, increased IL-10 expression, and polarization of macrophages toward a CD11b⁺CD14⁺CD163⁺ M2-like phenotype with elevated CD73. Analysis of ccRCC single-cell transcriptomic datasets further supported these observations, demonstrating the expression of HLA-G receptors LILRB1 and LILRB2 in TAMs and predicted interactions between HLA-G-expressing tumor cells and LILRB2⁺ macrophage populations. While spheroid co-culture systems do not fully recapitulate the complex human RCC TME, the consistency between our co-culture findings with patient-derived transcriptomic data supports a key role for HLA-G in coordinating lymphoid and myeloid immune suppression in the TME. Future studies employing models that incorporate autologous immune components and patient-specific TME will be necessary to extend these findings.

In summary, this study identifies HLA-G as a multifaceted regulator of RCC progression. HLA-G simultaneously promotes tumor-intrinsic oncogenic signaling, including activation of PI3K–AKT–mTOR, MAPK, and RAP1/RAS pathways, while driving metabolic adaptation and establishing an immunosuppressive TME. Through coordinated effects on tumor proliferation, immune evasion, and macrophage polarization, HLA-G emerges as a central regulator of RCC tumor biology. Targeting this pathway may therefore offer new opportunities to overcome immune resistance and improve therapeutic responses in RCC and other solid tumors.

Supplementary Material

Supplementary Material

3 Ajith_Supplementary Information.docx

Funding Statement

The author(s) declare that financial support was received for the research and/or publication of this article. This study was supported by research funding through the National Institutes of Health/National Cancer Institute Grant No. RO1CA172230.

Acknowledgments

We acknowledge the support and contribution of the Integrated Genomics Core Shared Resources at the Georgia Cancer Center, Augusta University (RRID: SCR_026483). The authors would like to acknowledge the Augusta University Proteomics and Mass Spectrometry Core (RRID: SCR_027673) for assistance with the metabolomic analysis. We acknowledge the support and contribution of the Georgia Cancer Center Biorepository, Augusta University, for providing human samples used in this research study (RRID: SCR_027271). We thank Marina Daouya for the experimental support. We thank Rhea-Beth Markowitz for critically reading the manuscript. We also acknowledge the Georgia Cancer Center at Augusta University community for their insightful feedback and support.

Disclosure of potential conflicts of interest

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: ArrayExpress database at EMBL-EBI (www.ebi.ac.uk/arrayexpress) under accession number E-MTAB-16395.

Ethics statement

The study was conducted in accordance with the Declaration of Helsinki. This non-interventional study utilized de-identified, publicly available datasets (e.g., TCGA) and either commercially procured human specimens (e.g., Biomax) or samples obtained from the Georgia Cancer Center Biorepository, which contains no identifiable private information. No interaction with human subjects was involved. In accordance with the U.S. Department of Health and Human Services regulations (45 CFR 46.104(d)(4)), such secondary research is classified as exempt; therefore, Institutional Review Board (IRB) approval was not required.

Prior presentation of the work

This work has not been previously presented at any scientific meeting.

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/2162402X.2026.2717680.