Knocking down FAM110A suppresses colon adenocarcinoma progression by inhibiting the Nrf2/HO-1 axis to induce ferroptosis Authors: Qilong Li, Zhiqiang Li, Yuke Lin, Qian Xu, Xianrong Lu, Juan Zhou, Jie Yi, Jun Chen, Chunkang Yang Published: 2026 ABSTRACT BackgroundThe family with sequence similarity 110 member A (FAM110A) is implicated in cell cycle regulation, but its role in colon adenocarcinoma (COAD) remains unknown. Through integrative bioinformatics analysis, clinical tissue sample validation, and experimental validation, we investigated the role of FAM110A in COAD.MethodsWe employed a comprehensive bioinformatics approach, utilizing The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) datasets, alongside various R packages for data acquisition and analysis. The clinical COAD tissue samples and their corresponding paracancerous normal colonic mucosa samples were collected for immunohistochemical (IHC) staining to verify the expression of FAM110A in clinical specimens. After knocking down FAM110A, the cell viability, migration, invasion, cell cycle, apoptosis, reactive oxygen species (ROS), malondialdehyde (MDA), 4-hydroxynonenoic acid (4-HNE), Fe2+, and glutathione (GSH) levels, and FAM110A, glutathione peroxidase 4 (GPX4), solute carrier family 7 member 11 (SLC7A11), nuclear transcription factor erythroid 2-related factor 2 (Nrf2), and heme oxygenase 1 (HO-1) protein expression levels of HCT-116 and SW620 cells were determined. Furthermore, rescue experiments were performed by overexpressing Nrf2 in FAM110A-knockdown HCT-116 cells to verify the regulatory relationship between Nrf2 and FAM110A in cell growth and ferroptosis.ResultsMultiple differentially expressed genes in COAD in the TCGA-COAD and GSE20916 datasets were identified. The elevated expression of FAM110A in tumor samples compared to normal tissues suggested its potential diagnostic role in COAD. Multivariate Cox regression models incorporating clinical variables indicated that FAM110A was one of the independent prognostic factors for COAD. The drug sensitivity analysis revealed correlations between FAM110A and the efficacy of several anticancer drugs. Clinical tissue IHC analysis further confirmed the significantly higher expression of FAM110A in COAD tumor tissues compared with paired paracancerous normal colonic mucosa tissues. The immune infiltration analysis revealed that FAM110A could serve as a biomarker for predicting response to immunotherapies for COAD. FAM110A was enriched in multiple pathways. Knocking down FAM110A reduced cell viability, migration, invasion, and GSH levels, as well as GPX4, SLC7A11, Nrf2, and HO-1 protein expression levels, arrested the cell cycle, and elevated apoptosis, ROS, MDA, 4-HNE, and Fe2+ levels in HCT-116 and SW620 cells. Overexpression of Nrf2 reversed the effects of FAM110A knockdown on HCT-116 cell growth and ferroptosis.ConclusionsOur findings underscore the diagnostic and prognostic significance of FAM110A in COAD. Knocking down FAM110A alleviated COAD via Nrf2/HO-1 axis inhibition to promote ferroptosis, and Nrf2 overexpression could rescue the ferroptosis and growth inhibition induced by FAM110A deficiency, providing a theoretical basis for developing drugs targeting FAM110A for treating COAD. Introduction Colon adenocarcinoma (COAD), a common and globally prevalent cancer, is the main type of colorectal cancer (CRC), originating from adenomatous lesions [1]. According to the latest survey, the morbidity of COAD ranks third, and its mortality ranks second among all cancers [2]. With the global economic growth and the continuous improvement of cancer screening methods, it was discovered that the morbidity of COAD among young people was increasing year by year [3]. In some provinces with higher income in China, the mortality rate of COAD has increased as well [4]. However, the causes of COAD are still unclear at present. Some studies have shown that high-fat diet, chronic intestinal inflammation, and genetics are risk factors for the occurrence and development of COAD [5, 6]. Despite advances in screening and treatment modalities, the five-year survival rate of COAD has only marginally improved, particularly in cases diagnosed at later stages [7]. Local recurrence and distant metastasis are key factors in COAD treatment failure. Although the current therapeutic methods, such as surgery and chemoradiotherapy, have made significant progress in treating COAD, tumor metastasis or recurrence occurs in 60% of patients after surgical resection or chemoradiotherapy [8]. Therefore, identifying novel biomarkers and developing therapeutic targets are urgently needed to enhance diagnostic accuracy, predict therapeutic responses, and improve patient outcomes of COAD. The family with sequence similarity 110 member A (FAM110A), one of the FAM110 family, is located on the short arm of chromosome 20 [9]. The three members of the FAM110 family, including FAM110A, FAM110B, and FAM110C, are located in the centrosome, assisting in the cell cycle progression and the formation of microtubules [10]. The FAM110A protein was first identified when using the yeast two-hybrid experiments to search for interaction proteins of centrosome and spindle pole-associated protein involved in cell cycle regulation [10]. It was indicated that the expression of FAM110A protein was regulated by the cell cycle, namely, the FAM110A expression increased during the S and G2/M phases of cell division [11]. Recently, one study revealed that FAM110A acted as a prognostic biomarker for breast cancer after analyzing bulk and single-cell sequencing [12]. In addition, another previous study indicated the oncogenic role of FAM110A in pancreatic cancer via combining bioinformatic analysis and in vivo and in vitro validation [13]. Although the previous studies mentioned above demonstrated the role of FAM110A in different kinds of cancer, the role of FAM110A in COAD is still unclear. Ferroptosis, as a unique mode of cell death, is driven by iron-dependent lipid peroxidation and regulated by various cellular events, including redox equilibrium, mitochondrial activity, and lipids and iron metabolism [14]. Since the proposal of ferroptosis, an increasing number of studies have continuously confirmed its close association with cancers, especially for COAD. It was demonstrated that inducing ferroptosis by increasing reactive oxygen species (ROS) and Fe2+ levels or down-regulating glutathione (GSH) levels and glutathione peroxidase 4 (GPX4) expression contributes to the clinical treatment of COAD [15]. Meanwhile, inhibiting ferroptosis promotes the progression and drug resistance of colon cancer [16]. More and more studies have affirmed that the nuclear transcription factor erythroid 2-related factor 2 (Nrf2)/heme oxygenase 1 (HO-1) signaling pathway plays an essential role in oxidative damage, ferroptosis, pyroptosis, and autophagy [17–19]. After activating Nrf2, it binds to antioxidant response elements via nuclear translocation to initiate transcriptional expression of antioxidant genes, including HO-1 [20]. The activated HO-1 plays a protective role in ferroptosis by clearing harmful free radicals and reducing iron accumulation [21]. Some previous studies indicated that inhibiting the Nrf2/HO-1 pathway contributed to promoting the ferroptosis of colorectal cancer [22, 23]. However, it is still unclear whether FAM110A is associated with the ferroptosis of COAD, especially through regulating the Nrf2/HO-1 pathway. Against the above backgrounds, our study aims to shed light on the pivotal role of FAM110A in COAD and its utility as a biomarker for diagnosis and prognosis through bioinformatics approaches. Meanwhile, based on bioinformatics results, our study also investigated whether knocking down FAM110A alleviated COAD by inhibiting the Nrf2/HO-1 axis to promote ferroptosis through in vitro validation studies, and further explored whether overexpressing Nrf2 could rescue the effects of FAM110A knockdown on cell growth and ferroptosis via rescue experiments, providing the framework and foundation for future research into the targeted treatment of COAD. Materials and methods Single-cell sequencing analysis The single-cell sequencing analysis was performed based on the HRA000201 dataset downloaded from the Genome Sequence Archive (GSA, https://ngdc.cncb.ac.cn/gsa-human) database, which includes 21 CRC patients (8 males and 13 females) and 6 cancer-free elderly normal individuals (3 males and 3 females). After transferring the merged matrix to Rstudio (version 4.3.1), the Seurat software package (v.4.2.0) was further adopted to annotate cell subpopulations. The ggplot2 software package was used to visualize cell subpopulations on the UMAP plots and analyze the expression of FAM110A in various cell subpopulations. In addition, the cell subpopulations were also mapped onto the UMAP plots based on the disease staging and types. Data collection After downloading the dataset (TCGA-COAD) from the Cancer Genome Atlas database (https://portal.gdc.can-cer.gov/) and excluding data lacking clinical information, a total of 477 COAD samples and 41 normal samples were collected. Meanwhile, after being standardized to the Fragments Per Kilobaseper Million format, the clinical data of each sample was obtained through the UCSC Xena database (https://xena.ucsc.edu/), listed in Table 1 [24]. In addition, the GSE20916 (COAD) dataset was downloaded from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/), whose specific information is shown in Table 2. A total of 101 COAD samples and 44 normal samples in the GSE20916 dataset were included in this study. Table 1The baseline characteristics of TCGA-COAD patientsCharacteristicsOveralln477Age (%) <= 60141 (29.6) > 60336 (70.4)Gender (%) Female226 (47.4) Male251 (52.6)M (%) M0348 (74.0) M166 (14.0) MX56 (11.9)N (%) N0283 (59.3) N1108 (22.6) N286 (18.0)T (%) T1&T295 (19.9) T3322 (67.5) T460 (12.6)TCGA The Cancer Genome Atlas, COAD colon adenocarcinoma Table 1The baseline characteristics of TCGA-COAD patientsCharacteristicsOveralln477Age (%) <= 60141 (29.6) > 60336 (70.4)Gender (%) Female226 (47.4) Male251 (52.6)M (%) M0348 (74.0) M166 (14.0) MX56 (11.9)N (%) N0283 (59.3) N1108 (22.6) N286 (18.0)T (%) T1&T295 (19.9) T3322 (67.5) T460 (12.6)TCGA The Cancer Genome Atlas, COAD colon adenocarcinoma The baseline characteristics of TCGA-COAD patients TCGA The Cancer Genome Atlas, COAD colon adenocarcinoma Table 2The GEO microarray chip informationGSE20916PlatformGPL570TypeArraySpeciesHomo sapiensTissueColonSamples in the COAD group101Samples in the normal group44ReferencePMID: 20,957,034GEO Gene Expression Omnibus Table 2The GEO microarray chip informationGSE20916PlatformGPL570TypeArraySpeciesHomo sapiensTissueColonSamples in the COAD group101Samples in the normal group44ReferencePMID: 20,957,034GEO Gene Expression Omnibus The GEO microarray chip information GEO Gene Expression Omnibus Differential expression genes (DEGs) analysis The R package limma was adopted to analyze the DEGs between the COAD and normal samples in the TCGA-COAD and GSE20916 datasets. In addition, after dividing the COAD samples in the TCGA-COAD dataset into FAM110A-high and FAM110A-low expression groups based on the median expression level of FAM110A, the R package DESeq2 was adopted to analyze the DEGs between the FAM110A-high expression and FAM110A-low expression groups. Setting |logFC| > 0.5 and adj.p < 0.05 as well as |logFC| > 1 and adj.p < 0.05 as the thresholds of DEGs between the COAD and normal samples as well as between the FAM110A-high expression and FAM110A-low expression groups, respectively. The results of the DEGs analysis were presented in a volcano map using the R package ggplot2, and the DEGs were displayed in a Venn diagram. Co-expressed genes analysis To screen the co-expressed genes, the pairwise correlation analysis between FAM110A and DEGs generated from the FAM110A-high and FAM110A-low expression groups of COAD samples of the TCGA-COAD and dataset was performed. The top 20 genes related to FAM110A were ultimately selected as co-expressed genes. Moreover, to verify the expression of co-expressed genes in the GSE20916 dataset, the pairwise correlation analysis between FAM110A and other genes was performed again after dividing the COAD samples in the GSE20916 dataset into FAM110A-high and FAM110A-low expression groups. The results of the co-expressed genes analysis were presented as correlation and co-expression heatmaps using the R package ggplot2. Diagnostic analysis Before the diagnostic analysis, the FAM110A expression level between the COAD and normal groups in the TCGA-COAD and GSE20916 were compared. To evaluate the diagnostic effect of FAM110A in COAD, the R package pROC was adopted to draw the receiver operating characteristic (ROC) curve of FAM110A to calculate the area under the curve (AUC) value. The larger the AUC value, the better the diagnostic effect. Prognostic analysis To investigate the clinical prognostic value of FAM110A in COAD, the FAM110A expression, age, gender, and clinical stages (M stage, N stage, and T stage) were employed to construct univariate/multivariate Cox regression models, displayed in the forest plots. After that, the R-package timeROC was used to analyze the time-dependent ROC curves of FAM110A and the diagnostic model in the TCGA-COAD dataset within 1, 2, and 3 years, and the AUC value was calculated. Subsequently, after dividing the COAD samples in the TCGA-COAD dataset into FAM110A-high and FAM110A-low expression groups based on the median expression level of FAM110A, a risk factor chart between FAM110A high and low expression grouping and multivariate Cox regression analysis risk score was drew. Finally, based on the results of multivariate Cox regression analysis, a nomogram was established to predict the 1-, 2-, and 3-year survival probabilities of COAD patients, and a calibration curve was established to evaluate the accuracy and resolution of the nomogram. The nomogram and calibration curve were established using the R package RMS. Drug sensitivity analysis In this study, the correlations between FAM110A and drug sensitivity were evaluated by comprehensively analyzing three large-scale pharmacogenomics databases, including the Genomics of Drug Sensitivity in Cancer (GDSC, https://www.cancerrxgene.org/), Cancer Cell Line Encyclopedia (CCLE, https://sites.broadinstitute.org/ccle/), and CellMiner (https://discover.nci.nih.gov/cellminer/) databases. Immunohistochemical (IHC) analysis Based on the Human Protein Atlas (HPA) database (www.proteinatlas.org/), the IHC analysis results were presented to compare the FAM110A expression between normal colon tissues and colon tumor tissues. Regulatory networks construction To analyze the relationship between transcription factors (TF) and FAM110A, as well as between miRNAs and FAM110A, the ChIPBase database (http://rna.sysu.edu.cn/chipbase/) and TarBase database (http://www.microrna.gr/tarbase) were adopted to obtain FAM110A-related TF and miRNAs, respectively. Next, the Cytoscape software was employed to visualize the FAM110A-TF and FAM110A-miRNA regulatory networks. Gene ontology (GO) and kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis In this study, the R package clusterProfiler was used to perform the GO and KEGG enrichment analysis on the FAM110A and co-expressed genes. P value < 0.05 and FDR value (q value) < 0.25 were considered statistically significant. Gene set enrichment analysis (GSEA) After all genes in the FAM110A-high and FAM110A-low groups of COAD samples in the TCGA-COAD dataset were sorted according to their logFC values and downloading the c2.cp.all.v2022.1.Hs.symbols.gmt dataset from the Molecular Signatures Database (MSigDB, https://www.gsea-msigdb.org/gsea/msigdb) database, the R-package clusterProfiler was used for GSEA on all genes. The screening criteria for GSEA are adj.p < 0.05 and FDR value (q value) < 0.25. Copy number variations (CNV) analysis After selecting the “Masked Copy Number Segment” data as the CNV data, the GISTIC2.0 analysis on the downloaded and processed CNV fragments was performed to analyze the CNV of COAD samples in the TCGA-COAD dataset. Immune infiltration analysis After labeling various infiltrating immune cell types, the enrichment scores calculated by the single-sample gene-set enrichment analysis were used to represent the relative abundance of immune cell infiltration in each sample, and the immune cell infiltration matrix of COAD samples in the TCGA-COAD dataset was obtained. Subsequently, the expression differences of immune cells between the FAM110A-high expression and FAM110A-low expression groups in COAD samples from the TCGA-COAD dataset were presented using the R package ggplot2. The immune cells with significant differences between the two groups were selected for subsequent analysis. The Spearman algorithm was used to calculate the correlation among immune cells and between FAM110A expression and immune cells. The results of the above correlation analysis were presented as a heatmap and bubble plot using the R package pheatmap and the R package ggplot2, respectively. Tumor mutation burden (TMB), microsatellite instability (MSI), immune checkpoint genes (ICG), and human leukocyte antigen (HLA) analysis After downloading the TMB and MSI data from the cBioPortal database (https://www.cbioportal.org/), the Mann-Whitney U test was adopted to calculate the differences in TMB and MSI between FAM110A-high expression and FAM110A-low groups in the COAD samples of the TCGA-COAD dataset. Based on the previous study, 47 ICG were obtained, and the expression differences of these ICG between the FAM110A-high expression and the FAM110A-low expression groups in the COAD sample of the TCGA-COAD dataset were analyzed [25]. After regarding “Human Leukocyte Antigen” as the keyword to search in the GeneCards database (https://www.genecards.org/), a total of 34 HLA family genes were screened, and their expression differences between the FAM110A-high expression and FAM110A-low expression groups of the COAD sample in the TCGA-COAD were analyzed.After regarding “Human Leukocyte Antigen” as the keyword to search in the GeneCards database (https://www.genecards.org/), a total of 34 HLA family genes were screened, and their expression differences between the FAM110A-high expression and FAM110A-low expression groups of the COAD sample in the TCGA-COAD were analyzed. Clinical tissue sample collection and IHC staining A total of 15 pairs of formalin-fixed paraffin-embedded (FFPE) COAD tumor tissue samples and their corresponding paracancerous normal colonic mucosa samples (distance from tumor edge ≥ 5 cm, confirmed to be non-tumor tissue by pathological examination) were collected from the Fujian Cancer Hospital between 2023 and 2024. All samples were obtained with written informed consent from patients, and the study was approved by the Ethical Review Committee of the Second Hospital of Nanping City (Approval No.: [2023] No.28). FFPE COAD tumor tissue was prepared into 4 µm-thick sections for IHC staining. The sections were deparaffinized in xylene, rehydrated through a graded ethanol series, and subjected to antigen retrieval by boiling in citrate buffer for 15 min. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 10 min, and non-specific binding was blocked with 5% bovine serum albumin for 30 min at room temperature. The sections were then incubated with anti-FAM110A (1:200, Lot: PA5-67037, Invitrogen, Carlsbad, USA) at 4°C overnight, followed by incubation with a horseradish peroxidase-conjugated secondary antibody (1:500, Proteintech, Wuhan, China) for 1 h at room temperature. The staining was visualized using 3,3’-diaminobenzidine chromogen, and the sections were counterstained with hematoxylin. The staining results were observed and recorded using an IX83 microscope (Olympus, Tokyo, Japan), and the percentage of positive cells was calculated. Cell culture The human intestinal epithelial cell line (HIEC-6 cells) and human colon cancer cell lines (DLD-1, HCT-116, SW480, and SW620 cells) without mycoplasma contamination were bought from the National Collection of Authenticated Cell Cultures (Beijing, China), all of which were authenticated by Yiping Jiang, a professor at Fujian Medical University with expertise in cell biology. Apart from HIEC-6 cells were cultured in a DMEM medium (Lot: 11965092, Gibco, Grand Island, USA), other cells were all cultured in an RPMI 1640 medium (Lot: 12633020, Gibco, Grand Island, USA). All culture media were supplemented with 10% fetal bovine serum (Lot: FBSSA500-S, Aus GeneX, Sydney, Australia), penicillin (100 U/mL), and streptomycin (100 µg/mL). All cells were cultured in the surroundings of 37℃ and 5% CO2. The culture medium was exchanged every two days. Cell experiment protocol and cell transfection To select the appropriate cell lines for subsequent studies, all cell lines were divided into five groups, namely, HIEC-6, DLD-1, HCT-116, SW480, and SW620 groups. All cells were normally cultured. To select the optimal transfectant, HCT-116 and SW620 cells were randomly divided into four groups, namely, the control, sh-NC, sh-FAM110A-1, and sh-FAM110A-2 groups. Apart from the control group, HCT-116 and SW620 cells in the other groups were transfected with corresponding transfectants. To investigate the effects and related mechanisms of knocking down FAM110A in colon cancer cells, HCT-116 and SW620 cells were stochastically divided into two groups, including the sh-NC and sh-FAM110A groups. For rescue experiments, HCT-116 cells were randomly divided into sh-NC, sh-FAM110A, sh-FAM110A + oe-NC, and sh-FAM110A + oe-Nrf2 groups. For cell transfection, after corresponding transfectants (including sh-NC (5’-CCGGCAACAAGATGAAGAGCACCAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTTG-3’), sh-FAM110A-1 (sense: 5’-GCAAGGACTGACTCTCCAA-3’ and antisense: 5’-TTGGAGAGTCAGTCCTTGC-3’), sh-FAM110A-2 (sense: 5’-GCGTGATCAAGTGGTTGTA-3’ and antisense: 5’TACAACCACTTGATCACGC-3’), oe-NC (sense: 5’-TTCTCCGAACGTGTCACGT-3’ and antisense: 5’-ACGTGACACGTTCGGAGAA-3’), and oe-Nrf2 (sense: 5’-ATGATGGACTTGGAGTTGCCAC-3’ and antisense: 5’-TCACTCGTCTTCTTGTAAATCAG-3’)) synthesized by Sangon Biotech (Shanghai, China) were mixed with the Lipofectamine 3000 transfection reagent (Lot: L3000075, Thermo Fisher Scientific, Massachusetts, USA), they were respectively transfected to colon cancer cells for 96 h. Cell viability assay During transfection, the CCK-8 agent (Lot: C0038, Beyotime, Shanghai, China) was incubated with HCT-116 and SW620 cells at 37℃ for 2 h every 24 h. After incubation, their cell viability was determined based on the absorbance at 450 nm using a Varioskan LUX microplate reader (Thermo Fisher Scientific, Massachusetts, USA). All cell viability experiments were independently repeated three times, and the absorbance values were expressed as the mean ± standard deviation. Statistical significance of the differences in cell viability between the sh-NC and sh-FAM110A groups was analyzed using Student’s t-test, with a P value < 0.05 considered statistically significant. Wound healing experiment After finishing the transfection and seeding HCT-116 and SW620 cells onto the 6-well plate, two parallel vertical lines were delimited. Then, they were incubated with serum-free culture medium for 24 h. An IX83 microscope (Olympus, Tokyo, Japan) was adopted for visualization at 0 h and 24 h, respectively. The migration rate was determined using the ImageJ software. Transwell assay Transwell assays were performed to detect the invasion ability of HCT-116 and SW620 cells and the migration ability of HCT-116 cells. For invasion detection, 200 µL of serum-free medium containing transfected HCT-116 or SW620 cells was seeded into the upper chamber of Matrigel-coated Transwell inserts, and 600 µL of complete medium was added to the lower chamber. For migration detection, the same seeding procedure was applied using uncoated Transwell inserts with HCT-116 cells only. After 24 h of incubation, cells on the upper surface of the membrane were removed, and cells that had migrated or invaded to the lower surface were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet (Lot: C0121, Beyotime, Shanghai, China) for 15 min. The number of stained cells was counted under a microscope to evaluate cell migration and invasion abilities. Flow cytometry This study adopted flow cytometry to determine the ROS levels, cell cycle, and apoptosis. To determine the ROS level, after all transfected HCT-116 and SW620 cells were stained with 1 mL DCFH-DA (10 µM, Lot: S0033S, Beyotime, Shanghai, China) for 20 min at 37℃ in the dark, they were washed with the serum-free RPMI 1640 medium. Then, all cells were collected for subsequent detection. To determine the cell cycle, all transfected HCT-116 and SW620 cells were collected and fixed with a 70% ethanol solution for 24 h. Then, they were stained with 0.5mL propidium iodide (PI) staining solution in the dark for 30 min. To determine the cell apoptosis, after gathering all transfected HCT-116 and SW620 cells, they were stained by a mixed solution containing 500 µL binding buffer, 5 µL AnnexinV-FITC, and 5 µL PI in the dark for 10 min. The ROS levels, cell cycle, and cell apoptosis were determined using a NovoCyte D2040R flow cytometer (Aisen Gene, Shenzhen, China). Determination of malondialdehyde (MDA), 4-hydroxynonenoic acid (4-HNE), Fe2+, and glutathione (GSH) levels According to the producer’s instructions, the MDA assay kit (Cat., ADS-W-YH002, Adbio, Yancheng, China), Fe2+ assay kit (Cat., ADS-W-QT027, Adbio), and reduced GSH assay kit (Cat., G0206W, Grace Biotechnology, Suzhou, China) were used to determine the MDA, Fe2+, and GSH levels in HCT-116 and SW620 cells, and the human 4-HNE ELISA kit (Cat., SP11758, Saipei Biotechnology, Wuhan, China) was used to determine the 4-HNE level in the supernatant of HCT-116 and SW620 cells. QRT-PCR After the total RNAs in HCT-116 and SW620 cells were extracted by using the Trizol agent (Lot: T1908, Takara, Kyoto, Japan), the PrimeScript™ RT reagent Kit with gDNA Eraser (Lot: RR047A, Takara, Kyoto, Japan) was employed to reverse transcribe total RNAs into cDNA. Then, the cDNA was amplified by adopting the SYBR Green™ Premix Ex Taq™ II (Lot: RR820A, Takara, Kyoto, Japan). The reverse transcription and amplification processes were performed in a qRT-PCR instrument (Rotor-GeneQ, QIAGEN, Hilden, Germany). The 2−ΔΔCt approach was applied to quantify the FAM110A mRNA expression level. The sequences of upstream and downstream primers are listed in Table 3, among which GAPDH is the internal reference. Table 3The primer sequences for qRT-PCRPrimersSequence (5’→3’)SKA1ForwardGTCCCTGGCTACCTGCTACReverseCTGTCACACAAGTCGATGAGGGAPDHForwardGTCTCCTCTGACTTCAACAGCGReverseACCACCCTGTTGCTGTAGCCAA Table 3The primer sequences for qRT-PCRPrimersSequence (5’→3’)SKA1ForwardGTCCCTGGCTACCTGCTACReverseCTGTCACACAAGTCGATGAGGGAPDHForwardGTCTCCTCTGACTTCAACAGCGReverseACCACCCTGTTGCTGTAGCCAA The primer sequences for qRT-PCR Western blot After the total proteins in HCT-116 and SW620 cells were extracted by RIPA lysis solution, the total proteins were denatured in boiling water. Then, the denatured proteins were loaded on an SDS-PAGE gel for separation. Next, after the separated protein was transferred to the PVDF membranes, the membranes were blocked by skim milk. The membranes were then immersed in primary antibodies at 4℃ for one night, including anti-FAM110A (1: 1000, Lot: PA5-67037), anti-GPX4 (1: 1000, Lot: 30388-1-AP), anti-solute carrier family 7 member 11 (SLC7A11, 1: 1000, Lot: 26861-1-AP), anti-Nrf2 (1: 2000, Lot: 20733), anti-HO-1 (1: 1000, Lot: 66743-1), and anti-β-actin (1: 50000, Lot: 20536-1-AP). Apart from the anti-FAM110A and anti-Nrf2 were respectively bought from Invitrogen (Carlsbad, USA) and CST (Danvers, USA), other antibodies were obtained from Proteintech (Wuhan, China). Finally, after the membranes were immersed in secondary antibodies for 30 min and stained by ECL chemiluminescent substrate, the grayscale value of each blot was analyzed using ImageJ software (Version 1.8.0), and protein quantification was performed using a relative quantification method with β-actin as the internal reference protein. The relative expression level of the target protein was calculated as the ratio of the grayscale value of the target protein band to the grayscale value of the corresponding β-actin band. All Western blot experiments were independently repeated three times, and the relative expression levels are expressed as the mean ± standard deviation. Statistical analysis All data produced from independently triplicate experiments were exhibited as mean ± standard deviation. Statistical significance was determined using Student’s t-test or one-way ANOVA followed by Tukey’s post-hoc test. After using the Benjamini-Hochberg method to adjust the p-value for multiple comparisons, a p-value of less than 0.05 was considered statistically significant. Results Single-cell sequencing analysis validated that FAM110A was highly expressed in CRC patient cells After performing the single-cell sequencing analysis on samples of 21 CRC patients (8 males and 13 females) and 6 cancer free elderly normal individuals (3 males and 3 females), the Seurat software package was adopted to annotate cells into 21 subpopulations, mainly including normal epithelial, cancer cells, endothelial, fibroblast, T cells, B cells, mast cells, NK cells, and other myeloid (Fig. 1A). In addition, after visualizing the gene exprssion on a UMAP plot, it was found that FAM110A was highly expressed in CRC patients (the red parts, Fig. 1B). The FAM110A expression in the above nine dominant cell subpopulations was further analyzed, indicating that the FAM110A was highly expressed in the mast cells (Fig. 1C). Moreover, after mapping the cell subpopulations onto the UMAP plots based on the disease staging and types, it was discovered that FAM110A was highly expressed in the cells of patients with IIA stage (Fig. 1D) and primary CRC (Fig. 1E). Fig. 1Single cell clustering and FAM110A expression in CRC tissues. A The UMAP plot of the cell subpopulations in the HRA000201 dataset. B The plot of the FAM110A expression of the cell subpopulations in the HRA000201 dataset. C The violin plot of FAM110A expression in the dominant cell subpopulations in the HRA000201 dataset. The UMAP plot of the cell subpopulations in the HRA000201 dataset based on the disease staging (D) and types (E) Single cell clustering and FAM110A expression in CRC tissues. A The UMAP plot of the cell subpopulations in the HRA000201 dataset. B The plot of the FAM110A expression of the cell subpopulations in the HRA000201 dataset. C The violin plot of FAM110A expression in the dominant cell subpopulations in the HRA000201 dataset. The UMAP plot of the cell subpopulations in the HRA000201 dataset based on the disease staging (D) and types (E) The DEGs analysis and co-expressed genes analysis in the FAM110A-high and FAM110A-low expression groups of the TCGA-COAD and GSE20916 datasets After obtaining the TCGA-COAD and GSE20916 datasets, the R package limma was used to standardize the GSE20916 dataset, and the differences before and after standardization are shown in Fig. 2A and B. Further analysis revealed 21,851 DEGs between the COAD and normal samples in the TCGA-COAD dataset, among which 15,212 were up-regulated, and 6639 were down-regulated (Fig. 2C). In addition, there were 5012 DEGs between the COAD and normal samples in the GSE20916 dataset, among which 2628 were up-regulated, and 2384 were down-regulated (Fig. 2D). Moreover, based on the above results, the Venn diagram indicated that there were 1683 co-upregulated genes (Fig. 2E) and 1525 co-downregulated genes (Fig. 2F) between the TCGA-COAD and GSE20916 datasets. After dividing the COAD samples in the TCGA-COAD dataset into FAM110A-high and FAM110A-low expression groups, the DEGs between the two groups were presented in a volcano plot (Fig. 2G). Subsequently, after establishing the correlation analysis between FAM110A and all DEGs, the top 20 genes with the highest absolute value of correlation coefficient were identified as co-expressed genes, including SRXN1, ANKRD36B, ARHGAP5, CEP85L, KCNAB1, SHPRH, NOXO1, NBEAL1, PDE10A, ANKRD36, SLC16A7, ZBTB20, IGLJ3, ARHGEF10, CCDC88A, NPTN-IT1, FOXP1-IT1, SCN9A, MICU3, and ENTPD1-AS1. After that, the correlation between FAM110A and those co-expressed genes was presented in a heat map (Fig. 2H). It showed that, apart from SRXN1, NOXO1, and IGL3, FAM110A was negatively related to 17 other genes. Meanwhile, based on the GSE20916 dataset, the correlation between FAM110A and those co-expressed genes was presented in another heat map (Fig. 2I). It was presented that FAM110A was positively related to SRXN1, KCNAB1, NOXO1, NBEAL1, ANKRD36, IGLJ3, and SCN9A, and negatively related to ANKRD36B, ARHGAP5, CEP85L, SHPRH, ZBTB20, ARHGEF10, NPTN-IT1, FOXP1-IT1, and ENTPD1-AS1. In addition, the expression levels of the 20 co-expressed genes between FAM110A-high and FAM-110 A-low expression groups of COAD samples in the TCGA-COAD and GSE20916 datasets were presented in Fig. 2J and K. All co-expressed genes showed significant statistical differences between FAM110A-high and FAM-110 A-low expression groups of COAD samples in the TCGA-COAD dataset. Apart from CCDC88A, MICU3, NBEAL1, PDE10A, and SLC16A7, other co-expressed genes showed significant statistical differences between FAM110A-high and FAM-110 A-low expression groups of COAD samples in the GSE20916 dataset. Finally, the co-expression of FAM110A and those co-expressed genes in FAM110A-high and FAM110A-low expression groups of COAD samples in the TCGA-COAD and GSE20916 datasets was presented in a heat map (Fig. 2L and M). Fig. 2The DEGs analysis and co-expressed genes analysis in the FAM110A-high and FAM110A-low expression groups of the TCGA-COAD and GSE20916 datasets. The boxplots of the GSE20916 dataset before (A) and after (B) normalization. The volcano plots of DEGs between the COAD and normal samples in the TCGA-COAD (C) and GSE20916 (D) datasets. The Venn diagrams of the up-regulated (E) and down-regulated (F) genes in the TCGA-COAD and GSE20916 datasets. (G) The volcano plots of DEGs between the FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD dataset. The correlation heat map between FAM110A and co-expressed genes in FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD (H) and GSE20916 (I) datasets. The differences in the expression levels of co-expressed genes in FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD (J) and GSE20916 (K) datasets. The co-expression heat map of FAM110A and co-expressed genes in FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD (L) and GSE20916 (M) datasets. ***/**/*P < 0.001/0.01/0.05 The DEGs analysis and co-expressed genes analysis in the FAM110A-high and FAM110A-low expression groups of the TCGA-COAD and GSE20916 datasets. The boxplots of the GSE20916 dataset before (A) and after (B) normalization. The volcano plots of DEGs between the COAD and normal samples in the TCGA-COAD (C) and GSE20916 (D) datasets. The Venn diagrams of the up-regulated (E) and down-regulated (F) genes in the TCGA-COAD and GSE20916 datasets. (G) The volcano plots of DEGs between the FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD dataset. The correlation heat map between FAM110A and co-expressed genes in FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD (H) and GSE20916 (I) datasets. The differences in the expression levels of co-expressed genes in FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD (J) and GSE20916 (K) datasets. The co-expression heat map of FAM110A and co-expressed genes in FAM110A-high and FAM110A-low expression groups of the COAD samples in the TCGA-COAD (L) and GSE20916 (M) datasets. ***/**/*P < 0.001/0.01/0.05 The diagnostic value of FAM110A on COAD As the above results indicated that FAM110A was highly expressed in CRC patients, we subsequently validated the expression differences of FAM110A between the COAD and normal samples in the TCGA-COAD and GSE20916 datasets. As presented in Fig. 3A and B, FAM110A was highly expressed in the COAD sample of the TCGA-COAD and GSE20916 datasets. Meanwhile, the ROC curve was established to evaluate the diagnostic value of FAM110A on COAD. As illustrated in Fig. 3C and D, the AUC values of FAM110A in the TCGA-COAD and GSE20916 datasets were both larger than 0.7, indicating that FAM110A possesses good diagnostic value for COAD. We subsequently conducted the IHC analysis on the FAM110A expression in COAD tissue and normal colon tissue based on the HPA database. It showed that FAM110A was highly expressed in COAD tissue (Fig. 3E) compared with the normal colon tissue (Fig. 3F). Fig. 3The diagnostic value of FAM110A on COAD. The comparison of FAM110A expression differences between the COAD and normal samples in the TCGA-COAD (A) and GSE20916 (B) datasets. The ROC curves of FAM110A in the TCGA-COAD (C) and GSE20916 (D) datasets. The IHC analysis of FAM110A in COAD tissue (E) and normal colon tissue (F). ***P < 0.001 The diagnostic value of FAM110A on COAD. The comparison of FAM110A expression differences between the COAD and normal samples in the TCGA-COAD (A) and GSE20916 (B) datasets. The ROC curves of FAM110A in the TCGA-COAD (C) and GSE20916 (D) datasets. The IHC analysis of FAM110A in COAD tissue (E) and normal colon tissue (F). ***P < 0.001 The prognostic value of FAM110A on COAD To investigate the impact of FAM110A on the prognosis of COAD patients, the correlation between FAM110A expression, age, gender, and clinical stages, as well as clinical prognosis, was explored in the TCGA-COAD dataset using univariate and multivariate Cox regression analysis. As illustrated in Fig. 4A and B, and Table 4, the three clinical stages (T stage, N stage, and M stage), age, and FAM110A expression may be independent prognostic factors for COAD. Subsequently, the time-dependent ROC curves at 1, 2, and 3 years were established based on the FAM110A expression and the Cox regression prognostic model risk score, respectively. It showed that the predicted accuracy of FAM110A on the prognosis of COAD was low, and its prediction performance was the best in the first year (AUC = 0.632, Fig. 4C). Meanwhile, the predicted accuracy of the Cox regression prognostic model risk score on the prognosis of COAD was moderate, and its prediction performance was the best in the second year (AUC = 0.770, Fig. 4D). In addition, the risk factors plot and nomogram based on the multivariate Cox regression model revealed that the efficacy of FAM110A expression and age in the prognostic risk model of COAD were the highest and lowest, respectively (Fig. 4E and F). Finally, the calibration curves of the multivariate Cox regression model declared that the prediction performance of the model in the first year was better than that in the second and third years (Fig. 4G-I). Fig. 4The prognostic value of FAM110A on COAD. The forest plots of univariate (A) and multivariate Cox regression analysis (B) based on the FAM110A expression, gender, age, and clinical stages. The time-dependent ROC curves of the FAM110A expression (C) and Cox regression prognostic model risk score D. The risk factor plot (E) and nomogram (F) of the multivariate Cox regression model. The calibration curves of the multivariate Cox regression model at 1-year (G), 2-year (H), and 3-year (I) The prognostic value of FAM110A on COAD. The forest plots of univariate (A) and multivariate Cox regression analysis (B) based on the FAM110A expression, gender, age, and clinical stages. The time-dependent ROC curves of the FAM110A expression (C) and Cox regression prognostic model risk score D. The risk factor plot (E) and nomogram (F) of the multivariate Cox regression model. The calibration curves of the multivariate Cox regression model at 1-year (G), 2-year (H), and 3-year (I) Table 4Results of univariable and multivariable Cox analysis of the TCGA-COAD datasetCharacteristicsTotal(N)Univariate analysisMultivariate analysisHazard ratio (95% CI)P valueHazard ratio (95% CI)P valueAge470 > 60333ReferenceReference <= 601370.655 (0.408–1.052)0.0800.461 (0.282–0.754)0.002Gender470 FEMALE223Reference MALE2471.131 (0.762–1.680)0.540M Stage470 M0348ReferenceReference MX561.611 (0.880–2.950)0.1231.413 (0.758–2.636)0.277 M1664.212 (2.699–6.575)< 0.0012.321 (1.365–3.949)0.002N Stage470 N0279ReferenceReference N11071.645 (0.989–2.735)0.0551.182 (0.675–2.068)0.559 N2844.041 (2.571–6.349)< 0.0012.530 (1.479–4.328)< 0.001T Stage470 T1&T293ReferenceReference T4597.891 (3.178–19.591)< 0.0014.660 (1.770–12.266)0.002 T33182.962 (1.284–6.833)0.0112.024 (0.853–4.801)0.110FAM110A4701.604 (1.164–2.210)0.0041.553 (1.129–2.135)0.007 Table 4Results of univariable and multivariable Cox analysis of the TCGA-COAD datasetCharacteristicsTotal(N)Univariate analysisMultivariate analysisHazard ratio (95% CI)P valueHazard ratio (95% CI)P valueAge470 > 60333ReferenceReference <= 601370.655 (0.408–1.052)0.0800.461 (0.282–0.754)0.002Gender470 FEMALE223Reference MALE2471.131 (0.762–1.680)0.540M Stage470 M0348ReferenceReference MX561.611 (0.880–2.950)0.1231.413 (0.758–2.636)0.277 M1664.212 (2.699–6.575)< 0.0012.321 (1.365–3.949)0.002N Stage470 N0279ReferenceReference N11071.645 (0.989–2.735)0.0551.182 (0.675–2.068)0.559 N2844.041 (2.571–6.349)< 0.0012.530 (1.479–4.328)< 0.001T Stage470 T1&T293ReferenceReference T4597.891 (3.178–19.591)< 0.0014.660 (1.770–12.266)0.002 T33182.962 (1.284–6.833)0.0112.024 (0.853–4.801)0.110FAM110A4701.604 (1.164–2.210)0.0041.553 (1.129–2.135)0.007 Results of univariable and multivariable Cox analysis of the TCGA-COAD dataset The drug sensitivity and regulatory network analysis of FAM110A After confirming the diagnostic and prognostic values of FAM110A on COAD, we subsequently predicted the sensitivity of FAM110A to common anti-cancer drugs through IC50 values based on the GDSC, CCLE, and CellMiner databases. Based on the GDSC dataset, six drugs that were negatively related to FAM110A have been found, including trametinib, selumetinib, refametinib, PD0325901, AICA ribonucleotide, and afatinib (Fig. 5A). Based on the CCLE dataset, five drugs that were negatively related to FAM110A have been found, including AZD0530, lapatinib, paclitaxel, panobinostat, and PD-0332991 (Fig. 5B). Based on the CellMiner dataset, five drugs that were positively related to FAM110A have been found, including n-(3, 5-dimethoxybenzyl)-4-methoxyaniline, diethyldithiocarbamate, cupric, 9-phenoxy, and acridine (Fig. 5C). After obtaining the TFs combined with FAM110A from the ChIPBase database, the FAM110A-TF regulatory network was constructed. As presented in Fig. 5D, there were 32 TFs related to FAM110A, including CDK9, ETS1, LDB1, RFX1, E2F1, ETV1, MAZ, RUNX1, E2T4, ETV5, NFKB1, SPI1, EBF1, FLI1, NFYA, TBP, EGR1, GABPA, NFYB, TCF12, ELF1, HDAC1, NRF1, TFAP2A, EP300, IRF4, POLR2A, TFAP2C, ERG, KMT2B, RELA, and ZEB1. Meanwhile, after obtaining the miRNAs combined with FAM110A from the TarBase database, the FAM110A-miRNA regulatory network was constructed. As presented in Fig. 5E, there were 8 miRNAs related to FAM110A, including miR-522-5p, miR-107, miR-126-3p, miR-195-5p, miR-21-3p, miR-212-3p, miR-27a-5p, and miR-1246. Fig. 5The drug sensitivity and regulatory network analysis of FAM110A. Results of the drug sensitivity analysis of FAM110A based on the GDSC (A), CCLE (B), and CellMiner (C) databases. The mRNA-TF (D) and mRNA-miRNA (E) regulatory networks of FAM110A The drug sensitivity and regulatory network analysis of FAM110A. Results of the drug sensitivity analysis of FAM110A based on the GDSC (A), CCLE (B), and CellMiner (C) databases. The mRNA-TF (D) and mRNA-miRNA (E) regulatory networks of FAM110A The functional enrichment analysis and CNV analysis of FAM110A and co-expressed genes The GO and KEGG enrichment analyses were further established based on FAM110A and co-expressed genes. In the cellular component (CC) aspect, FAM110A and the co-expressed genes were enriched in the cation channel complex, ion channel complex, calcium channel complex, transmembrane transporter complex, and transporter complex. In the molecular function (MF) aspect, FAM110A and the co-expressed genes were enriched in the SH2 domain binding, voltage-gated monoatomic cation channel activity, GTPase regulator activity, nucleoside-triphosphatase regulator activity, and superoxide-generating NADPH oxidase activator activity. Meanwhile, FAM110A and the co-expressed genes were enriched in the taste transduction, morphine addiction, leukocyte transendothelial migration, and purine metabolism pathways (Fig. 6A). In addition, based on the above results, the CC, MF, and pathway networks were established (Fig. 6B-D). It was shown that, in the CC aspect, three genes were respectively enriched in the cation channel complex, ion channel complex, transmembrane transporter complex, and transporter complex. In the MF aspect, three genes were respectively enriched in the GTPase regulator activity and nucleoside-triphosphatase regulator activity. Only one gene was enriched in each biological pathway. To determine the impact of genes in the TCGA-COAD dataset on COAD, the biological processes, CC, and MF of genes were analyzed using the GSEA analysis (Fig. 6E). It was shown that genes in the MAF110A-high expression group were enriched in multiple biological functions and signaling pathways, including the Rmts methylate histone arginines (Fig. 6F), apoptosis induced DNA fragmentation (Fig. 6G), pre-Notch expression and processing (Fig. 6H), and oxidative stress-induced senescence (Fig. 6I). The CNV of the FAM110A and co-expressed genes in the TCGA-COAD dataset were analyzed using GISTIC2.0. It was found that, apart from the ARHGEF10, FAM110A, MICU3, and SRXN1, other genes with unaltered CNV possessed a higher frequency (Fig. 6J). In addition, apart from the ARHGAP5, ARHGEF10, ENTPD1-AS1, FOXP1-IT1, IGLJ3, MICU3, and NPTN-IP1, the CNV of other genes was mainly for amplification rather than deep deletion (Fig. 6K). Fig. 6The functional enrichment analysis and CNV analysis of FAM110A and co-expressed genes. (A) The bubble plot of the enriched CC, MF, and biological pathways of FAM110A and co-expressed genes. The network diagrams of enriched CC (B), MF (C), and biological pathways (D) of FAM110A and co-expressed genes. (E) The GSEA enrichment results of genes in the TCGA-COAD. The genes in the FAM110A-high expression group were enriched in the Rmts Methylate Histone Arginines (F), Apoptosis Induced DNA Fragmentation (G), Pre Notch Expression and Processing (H), Oxidative Stress Induced Senescence (I). The frequency of alteration events (J) and CNV types (K) of FAM110A and co-expressed genes in the COAD sample of the TCGA-COAD dataset The functional enrichment analysis and CNV analysis of FAM110A and co-expressed genes. (A) The bubble plot of the enriched CC, MF, and biological pathways of FAM110A and co-expressed genes. The network diagrams of enriched CC (B), MF (C), and biological pathways (D) of FAM110A and co-expressed genes. (E) The GSEA enrichment results of genes in the TCGA-COAD. The genes in the FAM110A-high expression group were enriched in the Rmts Methylate Histone Arginines (F), Apoptosis Induced DNA Fragmentation (G), Pre Notch Expression and Processing (H), Oxidative Stress Induced Senescence (I). The frequency of alteration events (J) and CNV types (K) of FAM110A and co-expressed genes in the COAD sample of the TCGA-COAD dataset Results of the immune infiltration, TMB, MSI, ICG, and HLA analysis in the FAM110A-high and FAM110A-low expression groups After calculating the immune infiltration abundance of 28 immune cells in FAM110A-high and low expression groups of the COAD sample of the TCGA-COAD dataset using the ssGSEA algorithm, the immune infiltration abundance of immune cells between the FAM110A-high and FAM110A-low expression groups was compared. It was found that six types of immune cells possessed statistically significant differences between the two groups, including the CD56dim natural killer cells, activated B cells, CD56bright natural killer cells, monocytes, type 17 T helper cells, and activated CD8 T cells (Fig. 7A). Subsequently, the correlation among the six immune cells in the FAM110A-high and FAM110A-low expression groups was presented in the heat map, respectively. In the FAM110A-high group, the six immune cells were positively related to each other, among which monocytes were strongly related to CD56dim natural killer cells (r value = 0.87) and CD56bright natural killer cells (r value = 0.85) (Fig. 7B). The CD56dim natural killer cells achieve targeted killing by recognizing molecules that are abnormally expressed on the surface of tumor cells, and CD56bright natural killer cells exert immune regulatory effects by secreting cytokines and some immune regulatory molecules [26]. In the FAM110A-low group, the six immune cells were positively related to each other, among which activated CD8 T cells were strongly related to activated CD8 T cells (r value = 0.70) (Fig. 7C). Finally, the correlation between FAM110A and the six immune cells in the FAM110A-high and FAM110A-low expression groups was presented in the point plot. In the FAM110A-high group, FAM110A was positively related to the six immune cells (Fig. 7D). In the FAM110A-low group, FAM110A was weakly negatively or not related to the six immune cells (Fig. 7E). The TMB and MSI scores between the FAM110A-high and FAM110A-low expression groups were compared. As illustrated in Fig. 7F, there was no statistically significant difference in MSI score between the FAM110A-high and FAM110A-low expression groups. As illustrated in Fig. 7G, the TMB score in the FAM110A-high expression group was significantly lower than that in the FAM110A-low expression group. After intersecting the obtained ICG and HLA family genes with the TCGA-COAD dataset, 45 ICG, 27 HLA family genes, and their expression matrices were screened. The differences in 45 ICG and 27 HLA family genes between the FAM110A-high and FAM110A-low expression groups were respectively compared using the Mann-Whitney U test. As shown in Fig. 7H, there were statistically significant differences in the scores of CD44, LAG3, LGALS9, PDCD1, TNFRSF14, TNFRSF18, TNFRSF25, TNFRSF4, CD27, CD70, NRP1, TMIGD2, TNFRSF8, TNFRSF9, and TNFRSF15 between the FAM110A-high and FAM110A-low expression groups. As displayed in Fig. 7I, there were statistically significant differences in the scores of HLA-A, HLA-B, HLA-C, HLA-DQB1, HLA-DQB2, HLA-F, HLA-G, HLA-J, and HLA-S between the FAM110A-high and FAM110A-low expression groups. Fig. 7Results of the immune infiltration, TMB, MSI, ICG, and HLA analysis in the FAM110A-high and FAM110A-low expression groups. A The comparison of immune cell abundances between the FAM110A-high and FAM110A-low expression groups. The correlation among immune cells in the FAM110A-high (B) and FAM110A-low (C) expression groups. The correlation between FAM110A and immune cells in the FAM110A-high (D) and FAM110A-low (E) expression groups. The comparison of MSI (F), TMB (G), ICG (H), and HLA family genes (I) scores between the FAM110A-high and FAM110A-low expression groups. ***/**/*P < 0.001/0.01/0.05 Results of the immune infiltration, TMB, MSI, ICG, and HLA analysis in the FAM110A-high and FAM110A-low expression groups. A The comparison of immune cell abundances between the FAM110A-high and FAM110A-low expression groups. The correlation among immune cells in the FAM110A-high (B) and FAM110A-low (C) expression groups. The correlation between FAM110A and immune cells in the FAM110A-high (D) and FAM110A-low (E) expression groups. The comparison of MSI (F), TMB (G), ICG (H), and HLA family genes (I) scores between the FAM110A-high and FAM110A-low expression groups. ***/**/*P < 0.001/0.01/0.05 Knocking down FAM110A inhibited HCT-116 and SW620 cell growth Before investigating the role of knocking down FAM110A on the growth of HCT-116 and SW620 cells, we performed IHC staining on 15 pairs of COAD tumor tissues and their corresponding paracancerous normal colonic mucosa samples to validate the abnormal expression of FAM110A in COAD at the clinical level. As presented in Fig. 8A, the positive rate of FAM110A staining intensity in COAD tumor tissues was significantly higher than that in paracancerous normal colonic mucosa tissues. As the above results predicted that FAM110A was highly expressed in COAD, we subsequently investigated the FAM110A expression levels in the human intestinal epithelial cell line and human colon cancer cell lines through qRT-PCR and Western blot. The mRNA and protein expression levels of FAM110A in DLD-1, HCT-116, SW480, and SW620 cells were dramatically higher than those in the HIEC-6 cells, among which the mRNA and protein expression levels of FAM110A in the HCT-116 and SW620 were higher (Fig. 8B and C). Therefore, the HCT-116 and SW620 cells were used for subsequent studies. Meanwhile, the qRT-PCR and Western blot technologies were adopted to select the optimal transfectant and verify transfection efficiency, respectively. As illustrated in Fig. 8D, the FAM110A mRNA expression level of HCT-116 and SW620 cells in the sh-FAM110A-1 and sh-FAM110A-2 groups was prominently lower than that in the sh-NC or control group, in which the FAM110A mRNA expression level of HCT-116 and SW620 cells in the sh-FAM110A-1 group was the lowest. Therefore, the sh-FAM110A-1 was used for the subsequent study. In addition, the FAM110A protein expression level of HCT-116 and SW620 cells in the sh-FAM110A group was markedly lower than that in the sh-NC group, indicating that the sh-FAM110A was successfully transfected (Fig. 8E). After that, the effects of knocking down FAM110A on the cell viability, migration, invasion, cell cycle, and apoptosis of HCT-116 and SW620 cells were investigated. As shown in Fig. 9A-E, knocking down FAM110A significantly inhibited the cell viability and migration rate and reduced the number of invaded cells of HCT-116 and SW620 cells. In addition, knocking down FAM110A increased the proportion of the G1 and G2 phases and decreased the proportion of the S phase, which might contribute to inhibiting the proliferation of HCT-116 and SW620 cells (Fig. 9F). Moreover, knocking down FAM110A substantially elevated the apoptosis rate of HCT-116 and SW620 cells (Fig. 9G). Fig. 8FAM110A is highly expressed in colorectal cancer tissues and cell lines. A The IHC staining results for FAM110A expression between the paracancerous tissue and tumor tissue. The mRNA (B) and protein (C) expression levels of FAM110A in the HIEC-6, DLD-1, HCT-116, SW480, and SW620 cells. The mRNA (D) and protein (E) expression levels of FAM110A in the HCT-116 and SW620 cells. ****/***/**/*P < 0.0001/0.001/0.01/0.05 vs. the first group FAM110A is highly expressed in colorectal cancer tissues and cell lines. A The IHC staining results for FAM110A expression between the paracancerous tissue and tumor tissue. The mRNA (B) and protein (C) expression levels of FAM110A in the HIEC-6, DLD-1, HCT-116, SW480, and SW620 cells. The mRNA (D) and protein (E) expression levels of FAM110A in the HCT-116 and SW620 cells. ****/***/**/*P < 0.0001/0.001/0.01/0.05 vs. the first group Fig. 9Knocking down FAM110A inhibited HCT-116 and SW620 cell growth. The cell viability of HCT-116 (A) and SW620 (B) cells. The migration rate (C), invaded cell number (D), cell cycle proportion (E), and apoptosis rate (F) of HCT-116 and SW620 cells. ****/***/**/*P < 0.0001/0.001/0.01/0.05 vs. the sh-NC group Knocking down FAM110A inhibited HCT-116 and SW620 cell growth. The cell viability of HCT-116 (A) and SW620 (B) cells. The migration rate (C), invaded cell number (D), cell cycle proportion (E), and apoptosis rate (F) of HCT-116 and SW620 cells. ****/***/**/*P < 0.0001/0.001/0.01/0.05 vs. the sh-NC group Knocking down FAM110A promoted ferroptosis of HCT-116 and SW620 cells by regulating the Nrf2/HO-1 pathway As the above results predicted that FAM110A was enriched in superoxide-generating NADPH oxidase activator activity and oxidative stress-induced senescence pathway, we subsequently investigated the effects of knocking down FAM110A on the oxidative stress (OS) of HCT-116 and SW620 cells. It was demonstrated that knocking down FAM110A significantly up-regulated the ROS, MDA, 4-HNE, and Fe2+ levels of HCT-116 and SW620 cells, and significantly down-regulated the GSH level of HCT-116 and SW620 cells (Fig. 10A-E). As the accumulation of ROS is one of the inducers of ferroptosis, we next investigated whether knocking down FAM110A regulated the ferroptosis of HCT-116 and SW620 cells [27]. As illustrated in Fig. 10F-H and K-M, knocking down FAM110A prominently down-regulated the protein expression levels of GPX4 and SLC7A11 of HCT-116 and SW620 cells, indicating that knocking down FAM110A aggravated the ferroptosis of HCT-116 and SW620 cells. In addition, the above results suggested that FAM110A was associated with HDAC1, EP300, and NRF1, and previous studies confirmed that these proteins were involved in regulating Nrf2 [21, 28, 29]. Therefore, we subsequently investigated whether knocking down FAM110A exacerbated the ferroptosis of HCT-116 and SW620 cells by regulating the Nrf2/HO-1 pathway. As presented in Fig. 10I, J and N, and 10O, knocking down FAM110A significantly down-regulated the protein expression levels of Nrf2 and HO-1 of HCT-116 and SW620 cells, indicating that knocking down FAM110A exacerbated the ferroptosis of HCT-116 and SW620 cells by suppressing the Nrf2/HO-1 pathway. Fig. 10Knocking down FAM110A promoted ferroptosis of HCT-116 and SW620 cells by regulating the Nrf2/HO-1 pathway. The ROS (A), MDA (B), Fe2+ (D), and GSH (E) levels in HCT-116 and SW620 cells. (C) The 4-HNE level in the supernatant of HCT-116 and SW620 cells. The blots (F) and the statistical analysis of the protein expression levels of GPX4 (G), SLC7A11 (H), Nrf2 (I), and HO-1 (J) of HCT-116 cells. The blots (K) and the statistical analysis of the protein expression levels of GPX4 (L), SLC7A11 (M), Nrf2 (N), and HO-1 (O) of SW620 cells. ****/***/**/*P < 0.0001/0.001/0.01/0.05 vs. the sh-NC group Knocking down FAM110A promoted ferroptosis of HCT-116 and SW620 cells by regulating the Nrf2/HO-1 pathway. The ROS (A), MDA (B), Fe2+ (D), and GSH (E) levels in HCT-116 and SW620 cells. (C) The 4-HNE level in the supernatant of HCT-116 and SW620 cells. The blots (F) and the statistical analysis of the protein expression levels of GPX4 (G), SLC7A11 (H), Nrf2 (I), and HO-1 (J) of HCT-116 cells. The blots (K) and the statistical analysis of the protein expression levels of GPX4 (L), SLC7A11 (M), Nrf2 (N), and HO-1 (O) of SW620 cells. ****/***/**/*P < 0.0001/0.001/0.01/0.05 vs. the sh-NC group Overexpressing Nrf2 reversed the effects of FAM110A knockdown on HCT-116 cell growth As the above results demonstrated that knocking down FAM110A promoted ferroptosis of HCT-116 cells by regulating the Nrf2/HO-1 pathway, we further investigated whether Nrf2 overexpression could rescue the effects of FAM110A knockdown on HCT-116 cell growth. First, the qRT-PCR and Western blot were performed to determine whether oe-Nrf2 was successfully transfected. As presented in Fig. 11A and B, compared with the sh-NC group, the mRNA and protein expression levels of Nrf2 in HCT-116 cells in the sh-FAM110A group were significantly downregulated. Meanwhile, compared with the sh-FAM110A + oe-NC group, the mRNA and protein expression levels of Nrf2 in HCT-116 cells in the sh-FAM110A + oe-Nrf2 group were significantly upregulated, suggesting that Nrf2 was successfully overexpressed in HCT-116 cells. Subsequently, the cell viability, migration, invasion, apoptosis, and cell cycle of HCT-116 cells were detected to determine whether Nrf2 overexpression could rescue the effects of FAM110A knockdown on HCT-116 cell growth. Compared with the sh-NC group, FAM110A knockdown distinctly suppressed cell viability, migration, and invasion, triggered cell cycle arrest, and promoted apoptosis of HCT-116 cells. Notably, overexpression of Nrf2 observably elevated cell viability (Fig. 11C), increased the number of invaded and migrated cells (Fig. 11D-F), relieved cell cycle arrest (Fig. 11G), and reduced the apoptosis rate (Fig. 11H) in FAM110A-knockdown HCT-116 cells. These data indicated that overexpressing Nrf2 could reverse the effects of FAM110A knockdown on the cell viability, migration, and invasion, cell cycle distribution, and apoptosis of HCT-116 cells. Fig. 11Overexpressing Nrf2 reversed the effects of FAM110A knockdown on HCT-116 cell growth. The mRNA (A) and protein (B) expression levels of Nrf2 in HCT-116 cells. (C) The cell viability of HCT-116 cells. (D) The representative images of the invasion and migration of HCT-116 cells. The number of invaded (E) and migrated (F) HCT-116 cells. The cell cycle proportion (G) and apoptosis rate (H) of HCT-116 cells. ****/***/**P < 0.0001/0.001/0.01 vs. the sh-NC group; ####/###/##P < 0.0001/0.001/0.01 vs. the sh-FAM110A + oe-NC group Overexpressing Nrf2 reversed the effects of FAM110A knockdown on HCT-116 cell growth. The mRNA (A) and protein (B) expression levels of Nrf2 in HCT-116 cells. (C) The cell viability of HCT-116 cells. (D) The representative images of the invasion and migration of HCT-116 cells. The number of invaded (E) and migrated (F) HCT-116 cells. The cell cycle proportion (G) and apoptosis rate (H) of HCT-116 cells. ****/***/**P < 0.0001/0.001/0.01 vs. the sh-NC group; ####/###/##P < 0.0001/0.001/0.01 vs. the sh-FAM110A + oe-NC group Overexpressing Nrf2 reversed the effects of FAM110A knockdown on the ferroptosis of HCT-116 cells To further verify whether Nrf2 overexpression could abolish the ferroptosis-promoting effects induced by FAM110A knockdown in HCT-116 cells, rescue experiments were conducted. Compared with the sh-NC group, FAM110A knockdown significantly increased the levels of ROS, MDA, 4-HNE, and Fe2+, while decreasing GSH levels in HCT-116 cells (Fig. 12A-E). Importantly, overexpression of Nrf2 markedly reversed these changes, reducing ROS, MDA, 4-HNE, and Fe2+ accumulation and restoring GSH levels in FAM110A-knockdown HCT-116 cells. Western blot analysis demonstrated that knocking down FAM110A notably downregulated the protein expression of HO-1, GPX4, and SLC7A11, whereas Nrf2 overexpression evidently upregulated the expression of these proteins (Fig. 12F-I). Collectively, these findings confirmed that overexpressing Nrf2 reversed the ferroptosis-promoting effects of FAM110A knockdown in HCT-116 cells. Fig. 12Overexpressing Nrf2 reversed the effects of FAM110A knockdown on the ferroptosis of HCT-116 cells. The ROS (A), MDA (B), Fe2+ (D), and GSH (E) levels in HCT-116 cells. (C) The 4-HNE level in the supernatant of HCT-116 cells. The blots (F) and the statistical analysis of the protein expression levels of HO-1 (G), GPX4 (H), and SLC7A11 (I) in HCT-116 cells. ***/**P < 0.001/0.01 vs. the sh-NC group; ###P < 0.001 vs. the sh-FAM110A + oe-NC group Overexpressing Nrf2 reversed the effects of FAM110A knockdown on the ferroptosis of HCT-116 cells. The ROS (A), MDA (B), Fe2+ (D), and GSH (E) levels in HCT-116 cells. (C) The 4-HNE level in the supernatant of HCT-116 cells. The blots (F) and the statistical analysis of the protein expression levels of HO-1 (G), GPX4 (H), and SLC7A11 (I) in HCT-116 cells. ***/**P < 0.001/0.01 vs. the sh-NC group; ###P < 0.001 vs. the sh-FAM110A + oe-NC group Discussion COAD, one kind of CRC, is a major health concern worldwide, being the third most common cancer and the second leading cause of cancer-related deaths [1]. Due to the early symptoms of COAD mainly include abdominal pain, weight loss, and changes in intestinal habits, COAD is often screened out in the later stages [30]. Despite advances in screening and treatment, the prognosis for advanced-stage COAD patients remains poor. Therefore, there is an urgent need for improved diagnostic markers and therapeutic targets to enhance early detection rates and develop personalized treatment strategies. Recent studies have highlighted the potential of molecular profiling in revolutionizing the management of COAD by enabling more precise diagnoses and tailored therapies [31, 32]. In particular, identifying DEGs and understanding their role in tumorigenesis provides novel biomarkers for diagnosis and prognosis. Furthermore, investigating gene expression patterns concerning drug sensitivity offers promising avenues for optimizing chemotherapeutic regimens. Our research leverages these insights by focusing on the FAM110A gene’s expression differences in COAD, its diagnostic utility, prognostic significance, association with drug response, and involvement in immune processes. These comprehensive approaches have the potential to significantly improve patient outcomes by informing clinical decision-making with robust molecular data. In addition, FAM110A’s role in centrosome dynamics suggests that its knockdown may disrupt mitotic spindle formation, synergizing with ferroptosis induction—a hypothesis warranting further investigation. Recent studies suggested that FAM110A might play a role in cell division and spindle organization [9, 11]. Our findings demonstrated that its upregulation in tumor samples compared to normal tissues suggested a potential oncogenic role, which was supported by its association with higher AUC values for diagnostic utility. The differential expression analysis reveals thousands of DEGs, providing insights into the molecular complexity underlying COAD. Among these genes, those significantly up-regulated or down-regulated DEGs might be involved in key pathways driving tumorigenesis or tumor suppression. The identification of DEGs and their pathways opens avenues for further functional studies to elucidate their roles in COAD progression and response to treatment. Our multifactorial Cox regression model incorporating clinical stages and FAM110A expression indicated the significance of FAM110A’s expression as an independent prognostic factor, displayed its potential as a biomarker for patient stratification and targeted therapy, and highlighted the importance of integrating molecular data with traditional clinicopathological factors. This approach enhances the predictive accuracy for patient outcomes, suggesting that precision medicine strategies should consider both molecular markers and clinical characteristics when devising prognostic models. The drug sensitivity analysis linking FAM110A expression with responsiveness to various anticancer agents suggests its potential influence on chemotherapeutic efficacy. The negative correlations with some drugs suggested that patients with high FAM110A expression might benefit from alternative treatments, paving the way for personalized medicine approaches in COAD. The IHC analysis visually confirmed the increased presence of FAM110A protein in COAD tissue, further confirming the discovery of differential expression. In addition, the clinical IHC staining also affirmed that the expression of FAM110A in the tumor tissue was higher than that in the paracancerous tissue. This approach provided tangible evidence supporting the FAM110A as a biomarker and might guide histopathological evaluations during diagnosis or treatment monitoring of COAD. Immune cells play a critical role in the tumor microenvironment, influencing tumor progression and curative effects [33]. For instance, high cytotoxic T lymphocyte infiltration was generally associated with better prognosis and enhanced response to immunotherapy, while increased presence of regulatory T cells (Tregs) suppressed anti-tumor immunity, often correlating with poor outcomes [34, 35]. In our study, the immune infiltration analysis revealed significant differences in the abundance of various immune cell types between high- and low-FAM110A expression groups in COAD. The observed differential immune cell infiltration concerning FAM110A expression levels could have implications for understanding the immune landscape of COAD. The association between FAM110A expression and specific immune cell types might indicate a potential role of this gene in modulating the immune microenvironment. This is particularly important because emerging treatments for COAD are increasingly focused on utilizing the patient’s immune system to target cancer cells [36]. Furthermore, our findings suggested that FAM110A could serve as a biomarker for predicting response to immunotherapies. Given that some immunotherapeutic agents have shown efficacy in patients with specific immune profiles, characterizing the relationship between FAM110A expression and immune cell infiltration could help identify patients who are more likely to benefit from such treatments. In conclusion, our study underscored the importance of integrating molecular characteristics, such as FAM110A expression, with comprehensive immune profiling. This integrated approach not only enhanced the understanding of COAD biology but also paved the way for personalized medicine strategies that could improve clinical outcomes for COAD patients. In our studies, the GO and KEGG enrichment analyses revealed significant pathways associated with FAM110A expression in COAD. Notably, the genes co-expressed with FAM110A were enriched in cell cycle regulation, DNA repair, and apoptosis pathways. The cell cycle pathway is critical for tumor growth and progression, whose dysregulation usually leads to uncontrolled cellular proliferation [37]. Our findings suggest that FAM110A might play a role in modulating the cell cycle in COAD, which aligns with one previous study indicating its involvement in chromosomal stability and segregation [11]. Furthermore, DNA repair pathways are essential for maintaining genomic integrity [38]. The enrichment of DNA repair-related genes in the high FAM110A expression group might imply a compensatory mechanism for the increased genomic instability observed in COAD. This relationship between FAM110A expression and DNA repair capacity could be pivotal for understanding therapeutic responses since chemotherapeutic agents often target these pathways [39]. Apoptosis was another key process that was highlighted by our enrichment analysis. Apoptosis is a programmed form of cell death, whose evasion is a hallmark of cancer [40]. The association between FAM110A expression and apoptotic signaling suggested that it might contribute to the survival of COAD cells by influencing apoptotic resistance. In conclusion, our bioinformatics analysis has provided valuable clues about the potential roles of FAM110A in COAD pathogenesis. By linking FAM110A to crucial cancer-related pathways, such as cell cycle control, DNA repair mechanisms, and apoptotic processes, we have laid a foundation for future experimental validation and therapeutic exploitation of COAD. In addition, our studies indicated that FAM110A was also enriched in superoxide-generating NADPH oxidase activator activity and oxidative stress-induced senescence pathway, which belongs to the OS-related pathway. Multiple risk factors, such as age, environmental factors, lack of exercise, smoking, alcohol abuse, dietary habits, and obesity, interact with each other, leading to excessive ROS production and OS generation [41]. OS exists in multiple stages of cancer development and plays a crucial role in the transformation of normal cells into cancer cells and the proliferation, angiogenesis, and metastasis of tumors [42]. As mentioned, HDAC1, EP300, and NRF1 were the downstream transcription factors of FAM110A, among which HDAC1 regulates the stability and activity of Nrf2 through deacetylation, EP300 regulates Nrf2 acetylation through regulating histone acetyltransferase activity, and NRF1 regulates Nrf2 function by modulating mitochondrial respiratory chain function [43–45]. Therefore, after determining that knocking down FAM110A inhibited cell viability, migration, and invasion, arrested the cell cycle, and elevated apoptosis of HCT-116 and SW620 cells, the effects of FAM110A on the Nrf2/HO-1 pathway were investigated. More and more studies have affirmed that the activation of the Nrf2/HO-1 pathway, which was negatively related to the ROS level, exerted the antioxidative effects during the CRC progression [46, 47]. Our studies demonstrated that knocking down FAM110A increased the ROS, MDA, 4-HNE, and Fe2+ levels, reduced GSH levels, and suppressed the Nrf2/HO-1 pathway of HCT-116 and SW620 cells, which were consistent with the previous studies mentioned above. Meanwhile, the accumulation of ROS is one of the key factors in the occurrence of ferroptosis, an important pathological mechanism of COAD [48]. The previous studies confirmed that the occurrence of ferroptosis was usually characterized by the down-regulation of GPX4, an antioxidant enzyme, and SLC7A11, a cysteine transporter protein [49, 50]. In this study, we demonstrated that knocking down FAM110A down-regulated GPX4 and SLC7A11 protein expression levels, suggesting that knocking down FAM110A induced the ferroptosis of HCT-116 and SW620 cells. Furthermore, rescue experiments confirmed that overexpressing Nrf2 reversed the effects of FAM110A knockdown on HCT-116 cell growth, including elevating cell viability, increasing invaded and migrated cell numbers, relieving cell cycle arrest, and reducing apoptosis rate. Meanwhile, Nrf2 overexpression also abolished the ferroptosis-promoting effects induced by FAM110A knockdown, including reducing ROS, MDA, 4-HNE, and Fe2+ accumulation, restoring GSH levels, and upregulating HO-1, GPX4, and SLC7A11 protein expression in FAM110A-knockdown HCT-116 cells. In conclusion, our studies offer the potential therapeutic pathways related to FAM110A, namely, FAM110A could serve as a novel target for drugs aimed at aggravating OS and ferroptosis of COAD. Reflecting on the limitations of this study, it is important to acknowledge that our findings were primarily based on bioinformatics analyses and in vitro experiments and have not been complemented with in vivo experiments. Although our studies suggested significant roles for FAM110A in COAD, the in vivo experimental validation is necessary to confirm these functions in a biological context. Additionally, the sample size used in this study might limit the generalizability of our results. A larger sample size would enhance the robustness of our conclusions and provide a more comprehensive understanding of COAD. Furthermore, using datasets from multiple sources might affect the consistency and reproducibility of our results. Looking forward, these results lay the groundwork for future investigations into the molecular underpinnings of COAD and highlight FAM110A as a promising target for therapeutic intervention and biomarker development. The in vitro experimental validation studies confirm these bioinformatics predictions and further explore the clinical utility of FAM110A in the treatment stratification for patients with COAD. Conclusions In summary, the comprehensive bioinformatics studies have identified key differentially expressed genes in COAD, with a particular focus on FAM110A, which was found to be significantly upregulated and potentially served as a diagnostic marker and prognostic indicator. We constructed a prognostic model incorporating FAM110A expression and demonstrated its predictive accuracy through time-dependent ROC curves. The drug sensitivity analysis linked FAM110A expression to responses against some chemotherapeutic agents. The immune infiltration analysis revealed differences between high- and low-FAM110A expression groups, suggesting an immunological component to its role in COAD. Additionally, we observed significant CNV events and associations with MSI and TMB levels, which could guide personalized treatment strategies. The regulatory networks involving FAM110A provided insights into its potential mechanisms of action in COAD pathogenesis. In addition, the in vitro experiments demonstrated that knocking down FAM110A reduced cell viability, migration, and invasion, arrested the cell cycle, and elevated apoptosis by suppressing the Nrf2/HO-1 pathway to elevate ferroptosis of HCT-116 and SW620 cells. Furthermore, rescue experiments verified that overexpressing Nrf2 could reverse the effects of FAM110A knockdown on HCT-116 cell viability, migration, invasion, cell cycle progression, and apoptosis. Meanwhile, overexpressing Nrf2 could abolish the ferroptosis-promoting effects induced by FAM110A knockdown by restoring the expression of HO-1, GPX4, and SLC7A11 and regulating ROS, MDA, 4-HNE, Fe2+, and GSH levels. These findings support FAM110A as a promising therapeutic target for COAD treatment and solidify the critical role of FAM110A in COAD progression by regulating the Nrf2/HO-1 axis-related ferroptosis.