Abstract
Background:
Lung adenocarcinoma (LUAD) is the most common histological subtype and a leading cause of cancer mortality worldwide. Cellular Retinol Binding Protein-1 (CRBP1) regulates intracellular retinol bioavailability and several transcriptional activities in normal and neoplastic cells. High CRBP1 expression in LUAD has been previously shown to correlate with worse tumor grade and prognosis. We therefore investigated the putative mechanisms through which CRBP1 expression influences LUAD progression.
Methods:
We analyzed genomic copy number variations, miRNA and gene expression, as well as protein levels and ubiquitination status in CRBP1-transfected A549 cells and in CRBP1-expressing LUAD tissues.
Results:
Bioinformatics analysis revealed that high expression of CRBP1 in both transfected A549 cells and LUAD tissues is associated with miR-1915-3p downregulation. In silico analysis of the interactions between deregulated miR-1915-3p and gene expression identified EGFR and RARα as potential targets. Consistently, western blot analysis confirmed EGFR and RARα upregulation in CRBP1-transfected A549 cells. However, CRBP1-induced EGFR upregulation occurred independently of proteasomal degradation, as no changes in EGFR ubiquitination or protein stability were observed after MG132 treatment. Finally, pathway enrichment analysis revealed that CRBP1 expression is related to the modulation of several genes belonging to cancer pathways.
Conclusions:
Our findings identify a CRBP1-driven molecular landscape characterized by miR-1915-3p suppression and EGFR/RARα upregulation, providing a new hypothetical framework for understanding potential mechanisms of LUAD progression, which requires further validation in larger cohorts.
Background
Non-Small-Cell-Lung Cancer (NSCLC) accounts for more than 80% of all lung cancers and includes three main histological subtypes: adenocarcinoma, squamous cell carcinoma, and large cell carcinoma (). Lung Adenocarcinoma (LUAD) is the most common histological type of NSCLC (, ). LUAD appears morphologically heterogeneous and, besides genetic alterations and environmental influences, epigenetic factors seem to be involved in its pathogenesis (). Accordingly, molecular testing has been introduced for personalized management of oncogene-addicted NSCLC (). More than 60% of LUADs display molecular alterations amenable to targeted therapy (). Among these, the Epidermal Growth Factor Receptor (EGFR) represents one of the most significant therapeutic targets (). EGFR activation leads to a conformational change of its intracellular phosphorylation components, influencing tumor cell cycle progression, survival, proliferation and angiogenesis (, ). While its mutational frequency varies across different ethnicities, EGFR signaling is central to the pathogenesis and clinical management of LUAD (, ). Retinoids are a class of chemical compounds derived from vitamin A implicated in cell development, differentiation, growth, and death (). Retinoids also influence several intracellular signaling pathways involved in cancer initiation and progression (). In the last few decades, retinoids received considerable attention as anti-carcinogens alone or in combination with existing therapies (). It has been described that all-trans retinoic acid reduced drug resistance in LUAD-derived cell cultures (). Biological activity of retinoids is normally mediated through the binding to specific nuclear receptors, namely Retinoic Acid Receptors (RARα, β, and γ) and Retinoid X Receptors (RXRα, β, and γ) (). In addition, cellular retinol-binding proteins (CRBPs) and cellular retinoic acid-binding proteins (CRABPs) mediate intracellular trafficking of retinoids ().
Cellular retinol-binding protein-1 (CRBP1) is a 15 kDa cytosolic binding protein crucial for the uptake and subsequent intracellular esterification of retinol, so regulating its bioavailability (). CRBP1 localizes within mitochondria-associated membranes and colocalizes and surrounds growing lipid droplets during retinol ester biosynthesis (). Aberrant expression of retinoid receptors influences the biological behavior of several human cancers (, –). CRBP1 expression in LUAD is associated with the transcriptional upregulation of EGFR/Akt/Erk pathway (, –). Moreover, LUAD tissues with high cell expression of CRBP1 (CRBP1High) were previously found to display higher tumor grade and worse patients’ prognosis in a larger clinical cohort ().
MiRNAs are small, non-coding RNAs, approximately 18–23 nucleotides in length, implicated in post-transcriptional gene regulation (). MiRNAs are considered major epigenetic factors playing a key role in several processes, including differentiation, apoptosis and development of normal and neoplastic cells (). Aberrant expression of miRNAs plays a relevant role in lung cancer progression (). MiRNAs are located in common copy number variation (CNV) regions () or, alternatively, are underrepresented in polymorphic regions, likely due to their relevant biological role (, ). Recent studies demonstrated that retinoids also influence miRNA expression (). Therefore, in this pilot study, we used a highly selected subgroup of these historical samples to investigate whether CRBP-1 expression in LUAD cells is associated with specific miRNA dysregulation and its related pathways.
Materials and methods
Cell transfection
Human A549 LUAD cells (Sigma-Aldrich, St. Louis, USA) maintained in RPMI 1640 (Lonza BioPharma AG, Switzerland) in 10% FBS (Fetal Bovine Serum, Sigma-Aldrich) were transfected by using a vector pTargeT Mammalian expression system (Promega, Italy) carrying the whole sequence of CRBP-1 gene (NM_002899) and the gene for the resistance to G418 (CRBP1+, Promega, Italy), or the G418-resistance gene alone (CRBP1-), as reported ().
Lung adenocarcinoma tissues and immunohistochemistry
We investigated a small group of LUAD tissues from a previous cohort of patients, which had been originally published and described by Doldo et al. (). These historical samples were originally collected as routine diagnostic specimens and analyzed anonymously in the present study as a secondary data analysis. In accordance with the Declaration of Helsinki and local institutional frameworks governing the secondary use of fully de-identified and untraceable archival material, specific prospective institutional review board approval and active patient informed consent were waived by the local Independent Ethics Committee.
Tumor classification, subtyping, grading, and staging were in accordance with WHO 2021 criteria () and the Tumor-Node-Metastasis (TNM) staging system (TNM) (, ). We randomly selected two groups of LUAD tissues (): the first group with high tumor tissue expressing CRBP1 (CRBP1High; n=4), while the second group included those with low tumor expression of CRBP1 (CRBP1Low; n=4) (see Supplementary Table 1). To investigate a possible immunological phenotype modulation in CRBP1High versus CRBP1Low LUAD, an immunostaining was performed on Tissue Microarray (TMA) LUAD tissue sections using a rabbit polyclonal anti-CRBP-1 (clone FL-135, Santa Cruz Biotechnology, Heidelberg, Germany), as already reported (), and a rabbit anti-PD-L1 antibody (sp263 clone, Ventana, Roche Diagnostics, Basel, Switzerland) using Ventana BenchMarck Ultra stainer. PD-L1 expression was quantitatively evaluated by two independent pathologists using the standardized Tumor Proportion Score (TPS), defined as the percentage of viable tumor cells showing partial or complete membrane staining. Samples were classified according to established clinical thresholds: negative (TPS < 1%), low expression (TPS 1–49%), and high expression (TPS ≥ 50%), in accordance with the VENTANA SP263 assay interpretation guidelines for non-small cell lung cancer (, ).
Nucleic acid extraction
MiRNAs and DNA were extracted from LUAD representative paraffin-embedded tissue blocks maintaining patients’ anonymity, as reported (). Briefly, Hematoxylin & Eosin (H&E)-stained sections were reviewed by light microscopy and tumor areas carefully selected and marked by a pathologist. Nucleic acid extraction was performed using RecoverAll™ Total Nucleic Acid Isolation Kit (Thermo Fisher Scientific, Waltham, MA, USA), according to manufacturer’s instruction.
For in vitro studies, total RNA and miRNAs were isolated from CRBP1-transfected and empty-transfected A549 using a mirVana miRNA Isolation Kit (TermoFisher Scientific) according to the manufacturer’s protocol. Total RNA concentration was determined using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific).
MiRNA expression analysis
To assess differences in miRNA expression pattern between CRBP1High and CRBP1Low LUAD tissues and between CRBP1-transfected and empty-transfected A549 cells, we analyzed 830 human miRNAs using a nCounter® Human v3 miRNA Expression Assay (NanoString Technologies, Seattle, WA, USA) according to the manufacturer’s instructions. Hybridized miRNA-CodeSet complexes were purified using the fully automated nCounter Prep Station (NanoString Technologies), and reporters were counted using the nCounter Digital Analyzer (NanoString Technologies). Data analysis was performed using nSolver Software version 4.0 (NanoString Technologies). Analysis of miRNA expression from A549 cells was performed from Pharmadiagen (CBM scrl, Trieste, Italy) in triplicate. Given the exploratory nature of this initial screening on 830 microRNAs, a strict and biologically relevant descriptive fold-regulation threshold (± 2.5) was applied to select candidates for downstream validation, rather than relying on formal multi-test corrected p-values. This approach strictly follows the manufacturer’s data analysis guidelines for exploratory pilot series ().
Gene expression analysis
For gene expression analysis, total RNA from CRBP1-transfected and empty-transfected A549 cells was analyzed using a nCounter XT Customer Gene Expression Assay kit (NanoString Technologies) as reported (). Briefly, RNA samples were hybridized in the presence of target-specific capture and purified onto the nCounter Prep Station (NanoString Technologies), according to the manufacturer’s recommendation. The nCounter Digital Analyzer (NanoString Technologies) was used for imaging and data analysis by using the nSolver Software version 4.0. Each sample was normalized on three optimal housekeeping genes with the most stable expression from the full custom panel, automatically selected by the software: C13orf23 (NM_025138.2), CYR61 (NM_001554.3) and HIST1H1D (NM_005320.2). Fold regulation was considered significant for descriptive values ± 2.5. This threshold served as an exploratory screening filter to identify candidate nodes for downstream validation, following official manufacturer recommendations ().
Blot analysis
Western blot analysis was performed as reported (). Briefly, proteins from CRBP1-transfected and empty-transfected A549 cells were electroblotted and incubated with polyclonal rabbit anti-RARα (1:500; Santa Cruz Biotechnology, Dallas, TX, USA) and anti-EGFR antibodies (1:500; Cell Signaling Technology, Danvers, MA, USA), as well as polyclonal mouse anti-ubiquitin (1:500; Santa Cruz Biotechnology), followed by horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (1:5000; Sigma Aldrich, St. Louis, MO, USA) and goat anti-rabbit IgG (1:2000; Sigma Aldrich). Densitometric analysis was performed for three independent experiments using a polyclonal mouse α-tubulin (1:1000; Sigma-Aldrich) to normalize. For EGFR accumulation, CRBP1-transfected and empty-transfected A549 cells were treated with 10μM of MG132 proteasome inhibitor (Sigma-Aldrich) for 6 and 10 hours. This drug regimen was selected based on established protocols in lung adenocarcinoma models, where 10 μM MG132 for 6 hours is extensively validated to achieve a complete proteasomal blockade for EGFR stability assays in A549 cells (). All experiments were performed in triplicate.
Copy number variation analysis
Copy Number Variation (CNV) analysis was performed using nCounter® v2 Cancer CN Assay (NanoString Technologies) according to standard protocol. Briefly, after extraction and quantification, DNA was fragmented and denatured at 95 °C for 5 minutes to produce single strands. Denatured DNA was then hybridized with capture and reporter probes and hybridized DNA-CodeSet complexes purified using the fully automated nCounter Prep Station (NanoString Technologies) and reporters counted using the nCounter Digital Analyzer (NanoString Technologies). Data analysis was performed using the nSolver Software version 4.0 (NanoString Technologies). CNV was determined by averaging over three probes per region and normalized on the average count value for the 10 INV probes (invariant regions of 10 autosomes, present in two chromosomal copies), according to the manufacturer’s protocol. If the average CNV was below 1.4, the gene was considered to have one copy; if it was between 1.6 and 2.4, it was considered to have two copies; and if it was between 2.6 and 3.4, it was considered to have three copies. To fully demonstrate and account for inter-tumor heterogeneity within this pilot clinical cohort, individual profiling data were comprehensively evaluated alongside collective group averages.
Bioinformatic and statistical analysis
For the assessment of miRNA regulatory role in gene pathways and the identification of target genes in LUAD, we performed a functional characterization using miRanda () and DIANA-miRPath 3.0 (). Gene pathway enrichment analysis was performed using the DAVID Functional Annotation Tool (). To find miRNAs located in CNV regions, we searched for miRNAs within DECIPHER, a database of genomic variation with associated phenotypes that facilitates the identification and interpretation of specific pathogenic genetic variation (). To assess common deregulated miRNAs in CRBP1High/Low LUAD tissues and CRBP1-transfected A549 cells, we intersected the deregulated miRNAs using the InteractiVenn tool (). Results were expressed as the mean ± standard error mean (SEM). Differences were analyzed by Student t-test and considered statistically significant for p < 0.05.
Results
CRBP1High LUAD and CRBP1-transfected A549 cells show modulation of specific miRNAs
To evaluate if CRBP1 expression associates with a specific miRNA modulation, we analyzed a panel of 830 human miRNAs. Interestingly, CRBP1 expression associated with the deregulation of specific miRNAs in CRBP1High LUAD tissues and CRBP1-transfected A549 cells. In particular, in CRBP1High LUAD tissues (Figure 1A), one miRNAs resulted down-regulated (hsa-miR-1915-3p) and 6 up-regulated (hsa-miR-29b, hsa-miR-1260b, hsa-miR-4454, hsa-miR-7975, hsa-miR-548v and hsa-miR-874-5p). CRBP1-transfected A549 cells (Figure 1B) showed eight down-regulated miRNAs (hsa-miR-4488, hsa-miR-4516, hsa-miR-494, hsa-miR-1915-3p, hsa-miR-6720, hsa-miR-601, hsa-miR-4286, hsa-miR-221-3p) and fifteen up-regulated miRNAs (hsa-miR-30b-5p, hsa-let-7c, hsa-miR-30e-5p, hsa-miR-186-5p, hsa-miR-222-3p, hsa-miR-95, hsa-miR-1246, hsa-miR-32-5p, hsa-miR-99a-5p, hsa-miR-361-3p, hsa-miR-30a-5p, hsa-miR-26b-5p, hsa-miR-342-3p, hsa-miR-574-3p and hsa-miR-30d-5).
Figure 1
We intersected deregulated miRNAs in CRBP1High LUAD tissues and CRBP1-transfected A549 cells by using the InteractiVenn tool (http://www.interactivenn.net/). Interestingly, miR-1915-3p resulted downregulated in both CRBP1High LUAD tissues and CRBP1-transfected A549 cells (Figure 2).
Figure 2
We investigated by bioinformatics analysis those gene pathways associate with deregulation of miR-1915-3p. Heatmap analysis revealed that miR-1915-3p is mainly involved in Fatty Acid Biosynthesis and Metabolism, Cancer Development, Steroid Biosynthesis, Prion disease, TP53 Signaling Pathway and Cell cycle (Figure 3).
Figure 3
CRBP1 expression associates with cancer-related gene pathways
Moreover, we analyzed by gene expression analysis a panel of cancer-associated genes implicated in tumor growth, metastasis and chemoresistance in CRBP1-trasfected A549 cells. Our in vitro results showed that seventeen genes were at least 2.5-fold down-regulated (KLF10, EDN1, CHGB, TYMS, CDH1, ICAM1, HIF1A, TNFAIP3, STAT6, NDUFV3, PSMC4, PTGS2, EXOC2, TRAF4, HDAC1, BTBD15 and TP53) and 4 up-regulated (C11orf58, NTS, DNAJB9, UBE2B) in CRBP1-transfected compared to empty-transfected A549 cells (Figure 4A). Pathway enrichment analysis confirmed that six of those deregulated genes belong to cancer pathways (Figure 4B).
Figure 4
The analysis of Copy Number Variations (CNVs) in CRBP1High LUAD tissue revealed significant genomic instability compared with CRBP1Low ones. As reported in Table 1, the analysis showed a shared copy number gain (3 copies) in key oncogenic and regulatory genes and a copy number loss (1 copy) in several tumor suppressor loci and regulatory genes. While Table 1 summarizes the consensus copy number states for the cohort, the complete raw resolution of inter-tumor heterogeneity, showing the individual normalized counts for each single patient sample and specific gene probe, is provided in Supplementary Table 2.
Table 1
| Chromosome | Gene | CNV region | Localization | # of copies |
|---|---|---|---|---|
| chr01 | TP73 | 1p36.32 | chr1:3,569,129-3,652,765 | 1 |
| MYCL1 | 1p34.2 | chr1:40,361,096-40,367,687 | 1 | |
| CDKN2C | 1p32.3 | chr1:51,435,642-51,440,309 | 3 | |
| REG4 | 1p12 | chr1:120,336,641-120,354,203 | 1 | |
| MCL1 | 1q21.3 | chr1:150,547,027-150,552,214 | 3 | |
| chr03 | FHIT | 3p14.2 | chr3:59,735,036-61,237,133 | 1 |
| chr05 | TERT | 5p15.33 | chr5:1,253,287-1,295,162 | 1 |
| chr07 | SHH | 7q36.3 | chr7:155,595,558-155,604,967 | 1 |
| chr08 | CSMD1 | 8p23.2 | chr8:2,792,875-4,852,328 | 1 |
| chr09 | MELK | 9p13.2 | chr9:36,572,859-36,677,680 | 3 |
| chr12 | MDM2 | 12q15 | chr12:69,201,971-69,239,320 | 3 |
| chr13 | GPC5 | 13q31.3 | chr13:92,050,935-93,519,487 | 1 |
| chr17 | ITGB4 | 17q25.1 | chr17:73,717,516-73,753,899 | 1 |
| chr18 | DCC | 18q21.2 | chr18:49,866,542-51,062,273 | 1 |
| chr20 | EEF1A2 | 20q13.33 | chr20:62,119,365-62,130,668 | 1 |
Gene copy number variation in CRBP1High compared to CRBP1Low lung adenocarcinoma tissues.
Gene copy number estimates (1, 2, or 3 copies) were derived from nCounter automated multi-probe normalization protocols using strict software-driven thresholds relative to invariant regions. Loci with normal copy number states (2 copies) across the cohort are omitted.
CRBP1-transfected A549 cells show upregulation of miR-1915-3p putative target genes RARα and EGFR
Bioinformatics analysis identified RARα and EGFR as potential targets of miR-1915-3p (Table 2).
Table 2
| miRNA | Target gene | Alignment | mirSVR score* |
|---|---|---|---|
hsa-miR-1915![]() | RARα ![]() | ![]() | -0.029 |
| EGFR ↑ | ![]() | -0.056 |
Interaction between miR-1915 and its putative target genes.
*mirSVR score is an estimate of the miRNA effect on the mRNA expression level. The more negative the score, the greater the effect.
↓, down-regulated; ↑, up-regulated.
Based on this prediction, we performed Western blot analysis to verify the protein expression of these candidate genes. As expected, RARα and EGFR expression was significantly higher in CRBP1-transfected A549 cells compared to empty vector-transfected ones (Figures 5A, B). These findings are consistent with our previously reported study, in which we demonstrated that CRBP1High LUAD tissues show higher EGFR protein expression compared with CRBP1Low ones ().
Figure 5
Although CRBP1 transfection induced an overall increase in protein ubiquitination (p<0.05; Figures 6A, B), consistent with the upregulation of the ubiquitin conjugator UBE2B (see Figure 4A), the accumulation of EGFR in these cells appears to be independent of the ubiquitin-proteasome system. Indeed, treatment with the proteasome inhibitor MG132 (10 μM) failed to further increased EGFR level (Figures 6C, D), suggesting that its upregulation is not due to impaired proteasomal degradation, supporting a potential mechanism of miRNA-mediated epigenetic regulation.
Figure 6
CRBP-1 expression does not correlate with PD-L1 immune check point expression in LUAD tissues
In order to verify a potential link between CRBP1 and PD-L1 expression, we performed PD-L1 immunostaining of TMA sections in LUAD tissues. Quantitative immunohistochemical scoring revealed that all specimens in both the CRBP1High (n=4) and CRBP1Low (n=4) subgroups were completely negative, showing a TPS of 0% (TPS < 1%). Consequently, no quantitative or clinical differences in PD-L1 expression were observed between the two backgrounds (see Supplementary Figure 1).
Discussion
The present study highlights a complex CRBP-1-associated regulatory network in LUAD. While we previously observed that CRBP1High LUAD tissues exhibit a more aggressive phenotype (), our current pilot screening suggests that higher expression of CRBP1 is accompanied by a specific genomic and transcriptomic signature. This signature appears to be characterized by chromosomal instability and the deregulation of key microRNAs in both CRBP1High LUAD tissues and CRBP1-transfected A549 cells. Among these, miR-1915-3p downregulation was the only alteration shared between both models. Our in vivo and in vitro data align with previous studies concerning the role of miR-1915-3p in LUAD tumorigenesis and progression, as well as with scientific evidence from other cancer types. In fact, it has been reported that miR-1915-3p downregulation in non-small cell lung cancer (NSCLC) tissues and cell lines inversely associates with clinical TNM stage and overall survival (). Moreover, miR-1915-3p significantly suppresses migration, invasion, and epithelial-mesenchymal transition (EMT) in NSCLC cells. These findings support a tumor-suppressive role and anti-metastatic therapeutic potential for miR-1915-3p in lung cancer treatment (). These findings support a tumor-suppressive role and anti-metastatic therapeutic potential for miR-1915-3p in lung cancer treatment, acting through a regulatory network downstream of epigenetic modifications like the METTL3/YTHDF2 axis (). Concurrently, experimental exosomal delivery of miR-1915-3p in vitro was documented to decrease EMT marker expression, increase oxaliplatin sensitivity, and induce a parallel E-cadherin upregulation in drug-resistant colorectal cancer cell lines (). The observed copy number gains of oncogenic drivers such as MDM2 and MCL1, alongside the loss of tumor suppressors like FHIT and DCC, provide a possible structural basis for the aggressive phenotype previously described in CRBP1High compared with CRBP1Low LUAD tissues (). In this context, we specifically focused our genomic investigation on Copy Number Variations (CNV) rather than mutational status. This choice was driven by the need to determine whether the high EGFR protein level was sustained by structural gene amplification or, conversely, by broader chromosomal instability and post-transcriptional shifts associated with the CRBP1High background. It is well established that EGFR protein overexpression does not always overlap with gene mutations (, ). While mutations primarily affect protein activity and drug sensitivity, high protein levels can be independently sustained by gene copy number gains or, as our data suggest, by the loss of microRNA-mediated repression. Interestingly, the downregulation of miR-1915-3p, observed consistently in both CRBP1High LUAD tissue and CRBP1-transfected A549 cells, emerges as a potential functional link. Bioinformatics analysis and subsequent protein validation identified EGFR and RARα as putative targets of miR-1915-3p. The simultaneous upregulation of these proteins in CRBP1-transfected A549 cells, sustained by the absence of a direct gene amplification (CNV gain) for EGFR and RARα in CRBP1High LUAD tissues, strongly suggests a mechanism of post-transcriptional de-repression. This implies that the loss of miR-1915-3p might be a primary driver for EGFR and RARα accumulation, potentially contributing to the aggressive phenotype even in the absence of traditional gene gains or specific activating mutations. According to the literature, the elevation of RARα further reinforces EGFR levels by promoting its transcription, creating a powerful oncogenic loop (). Previous studies have indicated that RARα protein interacts with PI3K on the cell membrane and activates the PI3K-Akt/mTOR pathway (, ). Data from the TCGA PanCancer Atlas (accessed via cBioPortal) demonstrate that high EGFR-RARα co-expression correlates with a poorer prognosis in LUAD patients. Notably, our findings extend this observation, demonstrating that this aggressive phenotype is specifically rooted in a CRBP1-high background, where CRBP1 levels directly correlate with the upregulation of both RARα and EGFR. Furthermore, our investigation into the ubiquitin-proteasome system provides additional layers of complexity. While CRBP1 expression appears to modulate the ubiquitination landscape, specifically through the deregulation of UBE2B, the accumulation of EGFR appears to be independent of proteasomal degradation. The failure of MG132 to further stabilize EGFR levels indicates that CRBP1-mediated EGFR upregulation may involve a mechanism of miRNA-mediated epigenetic regulation, likely through miR-1915-3p regulatory activity. Finally, the absolute lack of baseline PD-L1 expression (TPS of 0%) across all examined tissues suggests that the pro-tumorigenic role of CRBP1 in LUAD is primarily driven by cell-intrinsic pathways, such as fatty acid metabolism and cell cycle regulation, rather than through the active modulation of this specific immune checkpoint. Nevertheless, this study presents some limitations. First, as an exploratory pilot screening with a limited sample size (four CRBP1High vs. four CRBP1Low human specimens), it lacks independent statistical power; thus, clinical and prognostic correlations are strictly inferred from our previously published large-scale cohort. Furthermore, due to the small sample size and discovery-driven nature of this profiling, formal statistical testing and multiple testing corrections were omitted for both the initial screenings (Figures 1, 4A) and the bioinformatic pathway analysis (Figure 3). Reporting uncorrected nominal p-values across these multi-target panels was intentionally avoided to prevent methodologically misleading false-positive results. Additionally, direct RT-qPCR validation was not performed for hsa-miR-1915-3p or the wider transcriptomic panel. Consequently, these multi-gene profiles remain descriptive, while validation efforts were targeted at downstream effector proteins (EGFR and RARα). Additionally, a direct physical interaction between miR-1915-3p and the 3’UTR of EGFR/RARα was not verified via dual-luciferase reporter assays, nor were functional rescue experiments performed. It must also be noted that the bioinformatic mirSVR scores for these predicted targets are relatively weak, suggesting that the direct post-transcriptional repressive loop might be modest or part of a more complex indirect network. Second, although the MG132 drug regimen used is extensively validated, direct experimental verification of proteasomal block via positive controls was not available within our current setup. Taking these constraints into account, the regulatory loop illustrated in Figure 7 remains a speculative model based on bioinformatic predictions and correlative protein changes that requires future mechanistic verification. Collectively, however, our findings point toward an intriguing hypothetical framework where CRBP1 drives LUAD progression through a synergistic combination of structural genomic alterations and microRNA-mediated regulatory shifts, identifying miR-1915-3p and its downstream targets as potential focal points for prognostic purposes and therapeutic intervention.
Figure 7
Conclusion
In conclusion, our preliminary findings suggest that high CRBP1 expression in lung adenocarcinoma is closely associated with a distinct landscape of genomic instability. While further investigations are required to definitively map the hierarchical interactions within this network, the integration of miRNA analysis and CNV data supports a model where the aggressiveness of CRBP1High LUADs results from a synergistic effect: the direct gain of anti-apoptotic genes (such as MDM2 and MCL1) and the potential de-repression of key oncogenic pathways (like EGFR) due to the loss of regulatory effectors like miR-1915-3p. These findings highlight the potential of CRBP1 as a robust prognostic biomarker and provide an intriguing rationale for exploring retinol-focused targeted therapeutic strategies aimed at counteracting the specific dysregulated gene pathways identified in this subtype of LUAD patients.
Statements
Data availability statement
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. All relevant data supporting the findings of this study are included within the article and its Supplementary Information Files.
Ethics statement
The requirement for prospective ethical approval was waived because this study represents a retrospective, secondary data analysis using fully de-identified and anonymous archival tissue material. These historical samples were originally collected as routine diagnostic specimens, and prior to the initiation of this study, they were completely anonymized and rendered untraceable to living participants, in accordance with the Declaration of Helsinki and local institutional guidelines. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements. The human samples used in this study were acquired from Historical archival blocks of lung adenocarcinoma tissues originally collected as routine diagnostic specimens from a past cohort, which were analyzed anonymously as a secondary data analysis of a previously published cohort, and were fully de-identified prior to the initiation of this study.
Author contributions
TC: Writing – original draft, Investigation. AO: Supervision, Writing – review & editing, Resources. MR: Writing – original draft, Investigation. MS: Writing – review & editing, Supervision. MQ: Methodology, Investigation, Writing – original draft. AF: Writing – review & editing, Investigation, Formal analysis, Conceptualization. ED: Methodology, Investigation, Writing – original draft, Project administration. FC: Methodology, Data curation, Conceptualization, Software, Writing – original draft.
Funding
The author(s) declared that financial support was not received for this work and/or its publication.
Acknowledgments
The authors thank Sabrina Cappelli for her technical support.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc.2026.1983972/full#supplementary-material
References
1
NicholsonAGTsaoMSBeasleyMBBorczukACBrambillaECooperWAet al. The 2021 WHO classification of lung tumors: Impact of advances since 2015. J Thorac Oncol. (2021) 16:1832–51. doi: 10.1016/j.jtho.2021.11.003
2
InamuraK. Lung cancer: Understanding its molecular pathology and the 2015 WHO classification. Front Oncol. (2017) 7:193. doi: 10.3389/fonc.2017.00193
3
TravisWD. Pathology of lung cancer. Clin Chest Med. (2002) 23:65–81. doi: 10.1016/s0272-5231(03)00061-3
4
KimDLeeYSKimDHBaeSC. Lung cancer staging and associated genetic and epigenetic events. Mol Cells. (2020) 43:1–9. doi: 10.14348/molcells.2020.2246
5
HendriksLEKerrKMMenisJMokTSNestleUPasaroAet al. Oncogene-addicted metastatic non-small-cell lung cancer: ESMO clinical practice guideline for diagnosis, treatment and follow-up. Ann Oncol. (2023) 34:339–57. doi: 10.1016/j.annonc.2022.12.009
6
LiamCKMallawathantriSFongKM. Is tissue still the issue in detecting molecular alterations in lung cancer? Respirology. (2020) 25:917–8. doi: 10.1111/resp.13823
7
PenroseHHellerSCableCMakboulRChadalawadaGChenYet al. Epidermal growth factor receptor mediated proliferation depends on increased lipid droplet density regulated via a negative regulatory loop with FOXO3/Sirtuin6. Biochem Biophys Res Commun. (2016) 469:370–6. doi: 10.1016/j.bbrc.2015.11.119
8
ScottASalgiaRM. Biomarkers in lung cancer: From early detection to novel therapeutics and decision making. Biomarkers Med. (2008) 2:577–86. doi: 10.2217/17520363.2.6.577
9
NapoliJL. Retinol metabolism in LLC-PK1 cells: Characterization of retinoic acid synthesis by an established mammalian cell line. J Biol Chem. (1986) 261:13592–7. doi: 10.1016/s0021-9258(18)67061-9
10
BushueNWanYJ. Retinoid pathway and cancer therapeutics. Adv Drug Delivery Rev. (2010) 62:1285–98. doi: 10.1016/j.addr.2010.07.003
11
TripathiSKPandeyKPandaMSpinellaMJRengasamyKRBiswalBK. The potential of retinoids for combination therapy of lung cancer: Updates and future directions. Pharmacol Res. (2019) 147:104331. doi: 10.1016/j.phrs.2019.104331
12
YaoWGuoQSunYChenJFuWQiuZ. All-trans retinoic acid reduces cancer stem cell-like cell properties in EGFR-TKI-resistant non-small cell lung cancer cells. BMC Cancer. (2020) 20:319. doi: 10.1186/s12885-020-06818-0
13
MarquezHACardosoWV. Vitamin A-retinoid signaling in pulmonary development and disease. Mol Cell Pediatr. (2016) 3:28. doi: 10.1186/s40348-016-0054-6
14
DoldoECostanzaGAgostinelliSTarquiniCFerlosioAArcuriGet al. Vitamin A, cancer treatment and prevention: The new role of cellular retinol binding proteins. BioMed Res Int. (2015) 2015:624627. doi: 10.1155/2015/624627
15
NapoliJL. Biosynthesis and metabolism of retinoic acid: Roles of CRBP and CRABP in retinoic acid homeostasis. J Nutr. (1993) 123:362–6. doi: 10.1093/jn/123.suppl_2.362
16
JiangWNapoliJL. Reorganization of cellular retinol-binding protein type 1 and lecithin:retinol acyltransferase during retinyl ester biosynthesis. Biochim Biophys Acta Gen Subj. (2012) 1820:859–69. doi: 10.1016/j.bbagen.2012.03.016
17
DoldoECostanzaGFerlosioAPasseriDBernardiniSScioliMGet al. CRBP-1 expression in ovarian cancer: A potential therapeutic target. Anticancer Res. (2014) 34:3303–12.
18
EvertsHBAkuailouEN. Retinoids in cutaneous squamous cell carcinoma. Nutrients. (2021) 13:153. doi: 10.3390/nu13010153
19
NagaiYAmbinderAJ. The promise of retinoids in the treatment of cancer: Neither burnt out nor fading away. Cancers (Basel). (2023) 15:3535. doi: 10.3390/cancers15143535
20
DoldoECostanzaGFerlosioAPompeoEAgostinelliSBellezzaGet al. High expression of cellular retinol binding protein-1 in lung adenocarcinoma is associated with poor prognosis. Genes Cancer. (2015) 6:490–502. doi: 10.18632/genesandcancer.89
21
Quintero BarceinasRSGarcia-RegaladoAArechaga-OcampoEVillegas-SepulvedaNGonzalez-De la RosaCH. All-trans retinoic acid induces proliferation, survival, and migration in A549 lung cancer cells by activating the ERK signaling pathway through a transcription-independent mechanism. BioMed Res Int. (2015) 2015:404368. doi: 10.1155/2015/404368
22
ZitoGNaselliFSaievaLRaimondoSCalabreseGGuzzardoCet al. Retinoic acid affects lung adenocarcinoma growth by inducing differentiation via GATA6 activation and EGFR and Wnt inhibition. Sci Rep. (2017) 7:4770. doi: 10.1038/s41598-017-05047-z
23
HammondSM. An overview of microRNAs. Adv Drug Delivery Rev. (2015) 87:3–14. doi: 10.1016/j.addr.2015.05.001
24
O'BrienJHayderHZayedYPengC. Overview of MicroRNA biogenesis, mechanisms of actions, and circulation. Front Endocrinol (Lausanne). (2018) 9:402. doi: 10.3389/fendo.2018.00402
25
Ghafouri-FardSShooreiHBranickiWTaheriM. Non-coding RNA profile in lung cancer. Exp Mol Pathol. (2020) 114:104411. doi: 10.1016/j.yexmp.2020.104411
26
WongKKDeLeeuwRJDosanjhNSKimmLRChengZHorsmanDEet al. A comprehensive analysis of common copy-number variations in the human genome. Am J Hum Genet. (2007) 80:91–104. doi: 10.1086/510560
27
FelekkisKVoskaridesKDweepHStichtCGretzNDeltasC. Increased number of microRNA target sites in genes encoded in CNV regions. Evidence for an evolutionary genomic interaction. Mol Biol Evol. (2011) 28:2421–4. doi: 10.1093/molbev/msr078
28
MarcinkowskaMSzymańskiMKrzyżosiakWJKozłowskiP. Copy number variation of microRNA genes in the human genome. BMC Genomics. (2011) 12:183. doi: 10.1186/1471-2164-12-183
29
Gholikhani-DarbroudR. MicroRNA and retinoic acid. Clin Chim Acta. (2020) 502:15–24. doi: 10.1016/j.cca.2019.12.003
30
GroomePABolejackVCrowleyJJKennedyCKrasnikMGirardNet al. The IASLC lung cancer staging project: Validation of the proposals for revision of the T, N, and M descriptors and consequent stage groupings in the forthcoming (seventh) edition of the TNM classification of malignant tumours. J Thorac Oncol. (2007) 2:694–705. doi: 10.1097/jto.0b013e31812d05d5
31
YoshizawaAMotoiNRielyGJSimaCSGeraldWLKrisMGet al. Impact of proposed IASLC/ATS/ERS classification of lung adenocarcinoma: Prognostic subgroups and implications for further revision of staging based on analysis of 514 stage I cases. Mod Pathol. (2011) 24:653–64. doi: 10.1038/modpathol.2010.232
32
RebelattoMCMidhaAMistryASabalosBRobbinsPLiXet al. Development of a programmed cell death ligand-1 immunohistochemical assay using antibody SP263 for non-small cell lung cancer. Mod Pathol. (2016) 29:1337–44. doi: 10.1038/modpathol.2016.136
33
Ventana Medical Systems. VENTANA PD-L1 (SP263) Assay Interpretation Guide for Non-Small Cell Lung Cancer. Tucson, AZ: Roche Diagnostics (2019).
34
NanoString Technologies. nCounter Expression Data Analysis Guide (MAN-C0011-04). Seattle, WA: NanoString Technologies (2022).
35
D'AmicoFDoldoEPisanoCScioliMGCentofantiFProiettiGet al. Specific miRNA and gene deregulation characterize the increased angiogenic remodeling of thoracic aneurysmatic aortopathy in Marfan syndrome. Int J Mol Sci. (2020) 21:6886. doi: 10.3390/ijms21186886
36
FerlosioAArcuriGDoldoEScioliMGDe FalcoSSpagnoliLGet al. Age-related increase of stem marker expression influences vascular smooth muscle cell properties. Atherosclerosis. (2012) 224:51–7. doi: 10.1016/j.atherosclerosis.2012.07.016
37
DingKJiangXWangZZouLCuiJLiXet al. JAC4 inhibits EGFR-driven lung adenocarcinoma growth and metastasis through CTBP1-mediated JWA/AMPK/NEDD4L/EGFR axis. Int J Mol Sci. (2023) 24:8794. doi: 10.3390/ijms24108794
38
BetelDKoppalAAgiusPSanderCLeslieC. Comprehensive modeling of microRNA targets predicts functional non-conserved and non-canonical sites. Genome Biol. (2010) 11:R90. doi: 10.1186/gb-2010-11-8-r90
39
VlachosISZagganasKParaskevopoulouMDGeorgakilasGKKaragkouniDVergoulisTet al. DIANA-miRPath v3.0: Deciphering microRNA function with experimental support. Nucleic Acids Res. (2015) 43:W460–6. doi: 10.1093/nar/gkv403
40
HuangDWShermanBTLempickiRA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. (2009) 4:44–57. doi: 10.1038/nprot.2008.211
41
BraginEChatzimichaliEAWrightCFHurlesMEFirthHVBevanAPet al. DECIPHER: Web-based resource for clinical interpretation of genomic variants. Nucleic Acids Res. (2014) 42:D993–9. doi: 10.1093/nar/gkt937
42
HeberleHMeirellesGVda SilvaFRTellesGPMinghimR. InteractiVenn: A web-based tool for the analysis of sets through Venn diagrams. BMC Bioinf. (2015) 16:169. doi: 10.1186/s12859-015-0611-3
43
WangMLiXZhangYHanJXieYYangJet al. MicroRNA-1915-3p inhibits cell migration and invasion by targeting SET in non-small-cell lung cancer. BMC Cancer. (2021) 21:1218. doi: 10.1186/s12885-021-08961-8
44
XiaoZLiuYLiQLiuQLiuYLuoYet al. EVs delivery of miR-1915-3p improves the chemotherapeutic efficacy of oxaliplatin in colorectal cancer. Cancer Chemother Pharmacol. (2021) 88:1021–31. doi: 10.1007/s00280-021-04348-5
45
PinterFPapayJAlmasiASapiZSzaboEKrenacsTet al. Correlation of EGFR mutations with chromosomal alterations and expression of EGFR, HER2, and VEGF in non-small cell lung cancer. Lung Cancer. (2008) 60:235–45. doi: 10.1016/j.lungcan.2008.03.011
46
BethuneGBethuneDRidgwayNXuZ. Epidermal growth factor receptor (EGFR) in lung cancer: An overview and update. J Thorac Dis. (2010) 2:32–51. doi: 10.1016/j.jep.2016.12.035
47
GuXYJiangYLiMQHanPLiuYLCuiBB. Over-expression of EGFR regulated by RARA contributes to 5-FU resistance in colon cancer. Aging (Albany NY). (2020) 12:156–77. doi: 10.18632/aging.102607
48
MasiaSAlvarezSde LeraARBarettinoD. Retinoic acid activates the phosphoinositide 3-kinase/Akt pathway in neuronal cells in a retinoid acid receptor-dependent manner. Mol Endocrinol. (2007) 21:2391–402.
49
MasiaSAlvarezSde LeraARBarettinoD. Rapid, nongenomic actions of retinoic acid: RARalpha signaling at the plasma membrane. Mol Cancer. (2018) 17:61. doi: 10.1210/me.2007-0062
Summary
Keywords
LUAD, CRBP1, miR-1915-3p, EGFR, RARα
Citation
Ferlosio A, Centofanti F, Doldo E, Cenci T, Qordja M, Rapanotti MC, Scioli MG and Orlandi A (2026) MiRNA-1915-3p dysregulation characterizes aggressive CRBP1-expressing lung adenocarcinoma cells in vitro and in patient-derived tissues. Front. Oncol. 16:1983972. doi: 10.3389/fonc.2026.1983972
Received
31 August 2026
Revised
01 October 2026
Accepted
10 October 2026
Published
09 November 2026
Volume
16 - 2026
Edited by
Chiara Nicolazzo, Università Link Campus, Italy
Updates
Copyright
© 2026 Ferlosio, Centofanti, Doldo, Cenci, Qordja, Rapanotti, Scioli and Orlandi.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Maria Giovanna Scioli, scioli@med.uniroma2.it
†These authors have contributed equally to this work and share first authorship
‡Deceased
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.



