Abstract Background Lipoxygenase family proteins (LOXs) are involved in various stages of tumor development, however, their specific roles in tumor-infiltrating lymphocytes (TILs) within colon adenocarcinoma (COAD) remain poorly defined. This study aims to comprehensively examine LOXs expression in COAD and evaluate their potential associations with immune cell infiltration and clinical outcomes. Methods We analyzed transcriptomic and clinical data from 477 tumor and 41 adjacent normal tissue samples in the TCGA-COAD dataset to evaluate the expression levels of LOX family genes and their associations with overall survival. To validate ALOXE3 expression, we performed RT-qPCR on fresh tumor and the corresponding matched adjacent tissues from six human colon carcinoma patients. Additionally, immune cell infiltration associated with LOX expression was explored using the TIMER and TISIDB databases. For functional validation, ALOXE3 was either overexpressed or silenced via shRNA in colon cancer cell lines, and its effects on tumor progression were assessed through in vitro proliferation assays and in vivo xenograft models. Results Among all LOX family members, only ALOXE3 expression was significantly associated with survival outcomes in COAD patients (overall survival: HR = 1.56, p < 0.05; disease-specific survival: HR = 2.12, p < 0.01; progression-free interval: HR = 1.55, p < 0.05). Functional assays showed that ALOXE3 overexpression significantly promoted tumor cell proliferation in vitro and enhanced tumor growth in vivo, whereas shRNA-mediated knockdown of ALOXE3 markedly suppressed cell proliferation. KEGG pathway analysis of genes co-expressed with ALOXE3 revealed a remarkable enrichment in mitogen-activated protein kinase (MAPK) signlaing pathway. Consistently, ALOXE3 overexpression resulted in activation of the ERK1/2 signaling pathway, as confirmed by Western blot analysis (p < 0.05). Furthermore, treatment with the ERK inhibitor SCH772984 effectively suppressed ALOXE3-induced tumor cell proliferation, suggesting that ALOXE3 may drive tumor growth via activation of the ERK1/2 signaling pathway. Conclusions ALOXE3 promotes tumor progression in COAD through activation of the ERK1/2 signaling pathway and exhibits a strong association with the immune cell infiltration of the tumor microenvironment. It may serve as a prognostic biomarker and a potential therapeutic target in COAD. Further studies are warranted to validate its clinical applicability and explore its role in immunotherapeutic approaches. Supplementary Information The online version contains supplementary material available at 10.1186/s12957-025-03939-3. Graphical Abstract [46]graphic file with name 12957_2025_3939_Figa_HTML.jpg Keywords: Colon adenocarcinoma, Lipoxygenase family proteins, Survival potential, Tumor cell growth, Signaling pathway Introduction Among all malignancies, colon adenocarcinoma (COAD) ranks third in incidence and fourth in cancer-related mortality globally [[47]1]. Over the next two decades, the global burden of cancer is expected to rise by 60%, reaching approximately 2.2 million new cases and 1.1 million deaths each year [[48]2]. COAD presents a diagnostic challenge due to its asymptomatic nature in the early stages, often leading to delayed diagnosis and contributing to high mortality. More than 600,000 deaths annually are attributed to COAD [[49]3]. Although targeted therapies have significantly advanced COAD management in recent years [[50]4], the 5-year survival rate for patients diagnosed at advanced stages remains poor. Consequently, identifying reliable prognostic biomarkers and therapeutic targets is critical for improving patient stratification and enabling personalized treatment strategies. Several molecular markers, including KRAS, BRAF, TP53, and microsatellite instability status, are currently used to guide COAD prognosis and treatment decisions [[51]5–[52]7]. However, these markers exhibit limited predictive accuracy across diverse patient cohorts and are frequently not clinically actionable. Moreover, their roles in modulating the tumor immune microenvironment—a critical determinant of immunotherapy response—remain poorly understood [[53]8]. These limitations highlight the urgent need for novel molecular biomarkers that not only provide prognostic value but are also mechanistically associated with immune regulation. The lipoxygenase (LOX) group comprises a set of enzymes that contain iron but lack heme. In humans, six LOX isoforms have been identified: Epidermis-type lipoxygenase 3 (ALOXE3), Arachidonate 5-lipoxygenase (ALOX5), Arachidonate 12-Lipoxygenase (ALOX12), Arachidonate 12-Lipoxygenase, 12R Type (ALOX12B), Arachidonate 15-Lipoxygenase (ALOX15), and Arachidonate 15-lipoxygenase type II (ALOX15B). In mammals, the eicosanoids’ metabolism (including leukotrienes, prostaglandins, and nonclassical eicosanoids) is mediated by various LOX isoforms [[54]9]. LOXs have been implicated in inflammation, dermatological disorders, and tumorigenesis [[55]10]. ALOX5 is upregulated in gastric cancer, and its inhibition suppresses tumor growth in N87 and AGS gastric cancer cell lines [[56]11]. Elevated expression levels of ALOX12 and ALOX15 have been reported in COAD [[57]12]. However, the roles of LOX family members in shaping the tumor immune microenvironment, particularly in COAD, remain largely unexplored. Among these isoforms, ALOXE3 has remained relatively understudied in the context of oncology. Initially identified in the epidermis and associated with skin barrier function, ALOXE3 has recently been linked to lipid metabolism, inflammation, and responses to metabolic stress [[58]13, [59]14]. Emerging evidence suggests it may also be involved in cancer-related pathways [[60]15–[61]17], yet its biological role in COAD, especially in relation to tumor immunity and prognosis, remains poorly defined. In this study, we determined the expression levels of LOX family members and investigated their prognostic relevance. Our findings highlight the potential prognostic significance of LOX expression in COAD. Additionally, we explored the association between LOX expression and tumor-infiltrating immune cells using the TISIDB and TIMER databases. We further characterized the potential relationship between the expression of LOX genes and infiltrating immune cells, which may influence the prognosis of COAD. Finally, we manipulated ALOXE3 expression in COAD cell lines via overexpression and shRNA-mediated silencing to elucidate its functional role in tumor cell proliferation. KEGG pathway analysis revealed that genes co-expressed with ALOXE3 were significantly enriched in mitogen-activated protein kinase (MAPK) activity, suggesting a possible role of the MAPK signaling pathway in ALOXE3-mediated COAD progression. Materials and methods Data collection and analysis We retrieved RNA-Seq level 3 transcriptomic profiles and corresponding clinical data of COAD patients from the TCGA database ([62]https://portal.gdc.cancer.gov/). A total of 518 COAD samples were initially obtained, including 477 primary tumor tissues and 41 adjacent normal (paratumor) tissues. Samples were included based on the following criteria: (1) histologically confirmed colon adenocarcinoma; (2) availability of complete RNA-Seq expression data; and (3) accompanying clinical information including survival status and follow-up duration. Cases with incomplete clinical metadata, duplicate entries, or non-primary tumor specimens (e.g., metastases or recurrences) were excluded from further analysis. Gene expression data were normalized and analyzed using the DESeq2 package (version 1.26.0) in R (version 3.6.3) to identify differential expression patterns of lipoxygenase (LOX) family genes between tumor and normal tissues [[63]18]. Expression differences were statistically assessed using the Wilcoxon rank-sum test for non-parametric comparison between groups. Data visualization was performed using the ggplot2 package (version 3.3.3) [[64]19]. KM survival and receiver operating characteristic (ROC) analysis The term “survival analysis” refers to a group of methods for studying data with a time-to-event structure that is often used in various research domains. The TCGA COAD data were used for standard survival analysis, and the Kaplan-Meier (KM) plots were generated using the survminer and survival packages [[65]20]. The log-rank p-value, hazard ratios (HRs), and corresponding 95% confidence intervals (CIs) were calculated.The diagnostic receiver operating characteristic (ROC) curve was generated using the ‘pROC’ R package (v1.17.0.1). The enrichment analysis of biological functions To comprehend the biological functions and signaling routes linked with genes co-expressing with ALOXE3, we employed the R project for annotation and visualization, gene ontology (GO) for analyzing differential genes, and the Kyoto Encyclopedia of Genes and Genomes (KEGG) for pathway analysis. Evaluation of invasion by immune cells Gene expression patterns of TCGA COAD samples were analyzed using the Gene Set Variation Analysis (GSVA) method (R version 3.6.3) to estimate the levels of immune cell infiltration [[66]21]. This research analyzed 24 types of immune cells, such as Th1, Th17, Th2 cells, Tregs, activated dendritic cells (aDC), B lymphocytes, cytotoxic cells, dendritic cells (DC), eosinophils, immature dendritic cells (iDC), macrophages, mast cells, neutrophils, and NK CD56^bright cells [[67]22]. We employed Spearman’s correlation analysis to determine the relationship between immune infiltration cells and gene expression levels. The TISIDB database ([68]http://cis.hku.hk/TISIDB) was used to examine the associations between ALOXE3 and lymphocytes, immunomodulators, and chemokines were examined. TIMER web server ([69]https://cistrome.shinyapps.io/timer/) was used to explore the relationship between immune infiltrating cells/immune cell markers and gene expressions relevant to COAD prognosis. Gene expression levels were presented as log2-transformed RSEM values. Human colon tumor sample collection From January 2015 to January 2022, tissue samples were collected from COAD patients (3 men and 3 women, aged 39 to 77) diagnosed at the Affiliated Hospital of Nantong University, China. This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Nantong University’s Affiliated Hospital (No. 2022-K183-01). Informed consent was obtained from all patients. The inclusion criteria required a confirmed COAD diagnosis based on clinical examination, imaging studies, and pathological assessment. Patients without a confirmed diagnosis based on the established protocols or with a history of other malignancies were excluded from the study. Tumor tissues were collected from the lesion sites, and paired normal colon tissues were obtained from areas located 15 cm from the tumor margin to serve as controls. RT-qPCR and WB Fresh tumor tissues and adjacent normal tissues were collected from six patients diagnosed with colon adenocarcinoma at the Affiliated Hospital of Nantong University, China, following informed consent and institutional ethical approval. All samples were immediately snap-frozen in liquid nitrogen and stored at -80 °C until further analysis. Total RNA was extracted using the RNAiso Plus kit (Takara, Japan) according to the manufacturer’s instructions. The concentration and purity of RNA were determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific, USA), ensuring A260/A280 ratios between 1.8 and 2.0. RNA integrity was assessed by agarose gel electrophoresis. cDNA was synthesized from 1 µg of total RNA using a reverse transcription kit (AG Biotechnology, China). Quantitative real-time PCR (RT-qPCR) was carried out using the AG RT-qPCR Master Mix with SYBR Green dye (AG Biotechnology) on a QuantStudio™ 5 Real-Time PCR System (Applied Biosystems, USA). The thermal cycling conditions were as follows: initial denaturation at 95 °C for 3 min, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s. The relative gene expression levels were calculated using the 2^−ΔΔCT method, with GAPDH as the internal control for normalization. All reactions were performed in triplicate, and each sample was analyzed in at least three independent experiments to ensure reproducibility. The following primers were used in this experiment. Human ALOXE3 forward primer TGGAGCGGAAAGAATGCCAC and reverse primer GGAGCCTGGGTTCCACTCA were used as the RT-qPCR primer sequences. Human GAPDH forward primer GGAAGCTTGTCATCAATGGAAATC and reverse primer TGATGACCCTTTTGGCTCCC. For Western Blot experiments, proteins were extracted from tissues or cells using chilled RIPA lysis buffer (Beyotime, Shanghai, China) supplemented with phosphatase and protease inhibitors. The protein concentration was then measured using a BCA assay kit (Pierce), and the samples were heated to 100 °C for 5 min. Equal amounts of proteins were separated on a 10% SDS-PAGE gel and subsequently transferred to a 0.45 μm PVDF membrane for 90 min at a current of 165 mA. The membranes were blocked using 5% skim milk powder and incubated with the primary antibody overnight at 4 °C. After washing, the secondary antibody was incubated for two hours at room temperature. Primary antibodies included ERK (137F5) (4695, 1:1000, Cell Signaling Technology, USA), p-ERK (Thr202/Tyr204) (9101, 1:1000, Cell Signaling Technology, USA), p38 (9212, 1:1000, Cell Signaling Technology, USA), p-p38 (Thr180/Tyr182) (4511, 1:1000, Cell Signaling Technology, USA), JNK (9252, 1:1000, Cell Signaling Technology, USA), p-JNK (Thr183/Tyr185) (9251, 1:1000, Cell Signaling Technology, USA), and GAPDH (ab181602, 1:5000, Abcam, UK).The secondary antibody was anti-rabbit IgG (7074,1:2000, Cell Signaling Technology, USA). To assess total protein levels, the membranes were stripped for 15 min at room temperature using RestoreTM Western Blot Stripping buffer (21059, Thermo Fisher Scientific, USA). Cell culture COAD cell lines (CCD841 CoN, SW480, SW620, HCT116, and HCT15) were obtained from Xuezhen Ma’s Laboratory in the Department of Oncology at Qingdao Central Hospital, affiliated with the University of Health and Rehabilitation Sciences and Qingdao University. The RKO cell line was purchased from Servicebio Co. (Wuhan, China). All cells were cultured in DMEM with 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin inside a humidified incubator set to 5% CO2 and 37˚C. Overexpression and knockdown of ALOXE3 To overexpress ALOXE3, we constructed lentiviral vectors encoding the human ALOXE3 gene in the CMV-MCS-EF1a-ZsGreen1-PGK-Puro vector (ZebraBox Bio-technology Company, Nanjing, China). SW480 and SW620 cells were transduced with 0.5 mL of lentiviral supernatant containing 8 µg/mL polybrene, supplemented with 1.5 mL of DMEM. 48 h post-transduction, cells were selected with 4 µg/mL puromycin. ALOXE3 overexpression efficiency was confirmed by using RT-qPCR. To knock down ALOXE3, we constructed ALOXE3 shRNA lentiviral expression vector (hU6-MCS-CMV-mCherry-PGK-Blast (ZBFV-014)) and a negative control vector (ZebraBox Bio-technology Company, Nanjing, China). The shRNA target sequences were as follows: ALOXE3 shRNA-1: CCGGCACACTGGACAACAT. ALOXE3 shRNA-2: GTGGGCACGTGTGGTGAAA. shRNA-NC: TCTCCGAACGTGTCACGT. Twelve hours post-transduction, the medium was replaced with fresh complete medium. HCT15 and HCT116 cells were used for lentiviral transduction. The cells were finally selected using blasticidin (10 µg/mL). ALOXE3 knockdown efficiency was validated by RT-qPCR. CCK8 assay To evaluate the effect of ALOXE3 on cell proliferation, COAD cells were seeded in 96-well plates at a density of 1000 cells per well in 100 µL of culture medium. In line with the manufacturer’s guidelines, the viability of cells was assessed by using the Cell Counting Kit-8 (CCK-8, MedChemExpress, USA). From days 1 to 4, each well was added 10 µL of CCK8 reagent within 90 µL of serum-free medium. After a 2-hour incubation, the absorbance was recorded at 450 nm with a microplate reader, using 630 nm as the reference wavelength. Colony formation assay To assess the effect of ALOXE3 on colony formation efficiency, COAD cells were seeded into 6-well plates at a density 500 cells per well. After one week, cells were washed with PBS, fixed in ethanol for 30 s, and stained with crystal violet for 20 min. After washing with PBS, colonies containing at least 50 cells were counted, and images were captured using a digital camera. Wound healing assay COAD cells (8 × 10⁵) were seeded into a 6-well plate. Wounds were scratched onto the monolayer of cells with a 10 µL pipette tip. The separated cells were rinsed with PBS, and then a medium without serum was added. Subsequently, photographs were taken at the appropriate time under an inverted phase-contrast microscope, and the area occupied by cell migration was estimated using Image J. Transwell assay For migration and invasion assays by using transwell chambers, serum-starved COAD cells (3 × 10⁴ per well) were resuspended in 200 µL of serum-free medium and placed into the upper chambers (Corning, USA), with Matrigel (ABW Biosciences, USA) for the invasion assay or without it for the migration assay. For the invasion assay, the upper chambers were coated with 30 µL of Matrigel for one hour, and the lower chambers were filled with 500 µL of complete medium with 10% fetal bovine serum. After 24 h of incubation at 37 °C, cells were fixed with 4% paraformaldehyde for 20 min and stained with 0.1% crystal violet for 15 min. Stained cells were counted under an inverted light microscope with 3 randomly selected fields per chamber. Tumor xenografts in a nude mouse model Six-week-old female BALB/c-nude mice, weighing 18 and 20 g, were purchased from Jinan Pengyue Experimental Animal Breeding Co. in Jinan, China. All animal experiments were approved by the Animal Ethics Committee of Qingdao University (No.AHQU-MAL20230728). Male mice, five weeks old, were maintained under standard housing conditions (20–26 °C, 12-h light/dark cycle) with ad libitum access to food and water. To assess the effect of ALOXE3 on tumor growth, SW480, SW620, HCT15, and HCT116 cells (4 × 10^6 cells per mouse) either overexpressing ALOXE3 or stably transduced with ALOXE3-targeting shRNA, were resuspended in 200 µL of PBS and subcutaneously injected into the right flank of mice. Body weights were recorded twice a week using an electronic weighing scale. During the study, tumor sizes were measured every three days with a caliper. Tumor volume was calculated using the formula: volume = (length) × (width) ^2/2. Mice were euthanized via CO₂ inhalation either 28 days post-tumor implantation or upon meeting humane endpoint criteria, including tumor size exceeding 2 cm³, severe lethargy, > 20% body weight loss, respiratory distress, impaired mobility, or hypothermia. The evaluation of mouse fatalities was categorized into three groups: tumor-related, technical, or of unknown cause. Classification was based on parameters such as gross appearance, weight loss, and the quantity and timing of deaths in each group. At the end of the experiment, tumors were excised and weighed. Analytical statistics Differential gene expression was assessed using the Wilcoxon rank-sum test. Survival analyses were conducted using Kaplan-Meier curves with the log-rank test to compare survival distributions, and univariate Cox proportional hazards regression models were applied to estimate HRs and 95% CIs. For multivariate analysis, Cox regression was adjusted for relevant clinical covariates, including age, gender, and tumor stage. All p values under 0.05 were regarded as significant. Results Expression levels of LOXs in COAD Humans express six types of LOXs, which include ALOXE3, ALOX5, ALOX12, ALOX12B, ALOX15, and ALOX15B. Using the R programming language, we conducted a comparative analysis of LOXs expression between COAD and adjacent normal tissues, based on 477 tumor and 41 normal samples from the TCGA database (Fig. [70]1A to F). Patient characteristics from the TCGA dataset are summarized in Table [71]S1. The results showed that only ALOXE3 was highly expressed in COAD (Fig. [72]1A, p < 0.001). Conversely, ALOX12, ALOX12B, and ALOX15B were significantly downregulated in COAD samples compared to adjacent normal tissues (Fig. [73]1C, D and F, p < 0.05, p < 0.01). No significant differences were observed in ALOX5 or ALOX15 expression between COAD and adjacent normal tissues (Fig. [74]1B and E, p >0.05). Fig. 1. [75]Fig. 1 [76]Open in a new tab The expressions of human LOXs in Colon cancer. (A-F) Comparative analysis of the expression levels of ALOXE3 (A), ALOX5 (B), ALOX12 (C), ALOX12B (D), ALOX15 (E), and ALOX15B (F) in colon cancer tissues to paratumor tissues (RNA-seq data from TCGA CORD). The number in the paratumor group is 41, and the number in the tumor group is 477. (G) The expression of ALOXE3 in tumor tissues and paratumor samples from COAD patients was determined by using qPCR. Compared with the indicated group, * p < 0.05, ** p < 0.01, *** p < 0.001, n.s.: no significant difference Additionally, RT-qPCR was performed on 6 pairs of human COAD and adjacent normal tissues. ALOXE3 mRNA levels were significantly increased in tumor tissues (Fig. [77]1G, p < 0.001). Our experimental results corroborated our computational predictions, confirming markedly elevated ALOXE3 expression in COAD. Prognostic significance of LOXs in COAD patients KM survival analyses based on the TCGA dataset were conducted for overall survival (OS), disease-specific survival (DSS), and progression-free interval (PFI). Patients were stratified into high and low expression groups based on median gene expression. High ALOXE3 expression was significantly associated with poor OS (HR = 1.56, 95% CI: 1.05–2.31, p = 0.028), DSS (HR = 2.12, 95% CI: 1.26–3.58, p = 0.005), and PFI (HR = 1.55, 95% CI: 1.09–2.21, p = 0.014) (Fig. [78]2A to C). Fig. 2. [79]Fig. 2 [80]Open in a new tab Prognostic significance, ROC analysis, functional enrichment, and immune relevance of ALOXE3 expression in COAD. (A-C) The overall survival (A), disease-specific survival (B), and progress-free interval (C) of ALOXE3 mRNA level in Colon Cancer Patients (Kaplan-Meier Plotter, tumor samples: n = 477). (D) ROC analysis of ALOXE3 in COAD was performed. (E) GO/KEGG pathways of ALOXE3 in colon adenocarcinoma cohort. (F) The correlation between ALOXE3 expression and infiltrating immune cells in human colon cancer. (G) High ALOXE3 level enriched in naïve CD4^+ T cells had worse OS in COAD. Compared with the indicated group, * p < 0.05, ** p < 0.01, *** p < 0.001, n.s.: no significant difference Additionally, ROC analysis revealed an AUC of 0.832 for ALOXE3 predicting COAD (Fig. [81]2D), indicating good prognostic performance. None of the other LOX genes (ALOX5, ALOX12, ALOX12B, ALOX15, ALOX15B) showed statistically significant associations with survival outcomes (Figures [82]S1-S3). The relationship between clinical parameters and ALOXE3 expression in COAD To further investigate the clinical significance of ALOXE3 expression in colon adenocarcinoma, we analyzed its association with various clinicopathological features from the TCGA-COAD cohort (Table [83]S1). The clinicopathological variables assessed included T stage (tumor invasion depth), N stage (lymph node involvement), M stage (distant metastasis), pathological stage, age, and gender. Our analysis revealed that ALOXE3 expression was significantly associated with more advanced pathological stages. Specifically, ALOXE3 expression showed positive correlations with higher N stage (p < 0.05), M stage (p < 0.01), pathologic stage III and IV (p < 0.01), suggesting a potential role in promoting lymphatic and distant metastasis. Moreover, patients aged over 65 years exhibited significantly higher ALOXE3 expression compared to younger individuals (p < 0.05), suggesting a possible age-related increase in ALOXE3-associated tumor aggressiveness. No statistically significant difference was found with respect to T stage (p = 0.091), there was a trend toward increased ALOXE3 expression in patients with T4 tumors, which may reflect a contribution to local tumor progression. In contrast, no significant differences in ALOXE3 expression were found between male and female patients (p = 0.627), suggesting that ALOXE3-driven tumor behavior is likely independent of gender. Collectively, these findings indicate that high ALOXE3 expression is closely associated with more aggressive clinical features in COAD, reinforcing its potential as both a prognostic biomarker and a therapeutic target. The co-expression genes with ALOXE3 in COAD To elucidate the biological functions of ALOXE3 in COAD, we used clusterProfiler and ggplot2 packages of the R project. GO term annotation showed that the top 50 genes most positively correlated with ALOXE3 predominantly function in MAP kinase activity (Fig. [84]2E and Table [85]S2, Pearson r > 0.5). Additionally, KEGG pathway analysis showed significant enrichment in proteoglycans in cancer and the relaxin signaling pathway, suggesting ALOXE3 may regulate signaling cascades associated with cell proliferation and tumor progression. These results underscore the potential role of ALOXE3-related gene networks in regulating COAD progression and prognosis. ALOXE3 expression is linked with immune infiltration, immune stimulators, and chemokines in COAD Tumor-infiltrating lymphocytes (TILs) are independent predictors of cancer prognosis [[86]23]. In this work, we utilized the GSVA package from the R project to explore the association between ALOXE3 expression and immune infiltration in COAD. ALOXE3 expression was significantly positively correlated with the infiltration of various immune cells, including aDC cells (r = 0.239, p < 0.001), CD8 T cells (r = 0.148, p = 0.001), cytotoxic cells (r = 0.286, p < 0.001), iDC (r = 0.148, p = 0.001), macrophages (r = 0.237, p < 0.001), neutrophils (r = 0.213, p < 0.001), NK CD56^bright cells (r = 0.224, p < 0.001), NK CD56^dim cells (r = 0.213, p < 0.001), Tem (r = 0.107, p = 0.019), Tgd (r = 0.14, p = 0.002), Th1 cells (r = 0.26, p < 0.001), and Th2 cells (r = 0.148, p = 0.001) in COAD (Figure [87]S4B, Table [88]S1). Moreover, a significant negative correlation was observed between ALOXE3 expression and Th17 cells (r = -0.201, p < 0.001) in COAD (Fig. [89]2F, Table [90]S3). We further assessed the relationship between ALOXE3 expression and immune stimulators and chemokines in COAD by the TISIDB database. ALOXE3 expression was found to be significantly correlated with multiple immune-related molecules, including immune inhibitors, immunological stimulators, and MHC molecules, and ALOXE3 expression levels in COAD tissues. In addition, ALOXE3 expression was also associated with various chemokines (Fig. [91]S4). Furthermore, analysis via the TIMER web server revealed a significant correlation between ALOXE3 expression and multiple immune cell markers in COAD. Specifically, ALOXE3 expression positively correlated with CD8^+ T cell markers (CD8A: r = 0.203, p < 0.001; CD8B: r = 0.192, p < 0.001), general T cell markers (CD3E: r = 0.107, p < 0.05; CD2: r = 0.112, p < 0.05), neutrophil markers (CXCR4: r = 0.278, p < 0.001; CD44: r = 0.152, p < 0.01; CD33: r = 0.11, p < 0.05), dendritic cell markers (BST2: r = 0.276, p < 0.001; CD24: r = − 0.356, p < 0.001), and macrophage markers (CD14: r = 0.181, p < 0.001; CD68: r = 0.194, p < 0.001; CCR5: r = 0.137, p < 0.01) (Table [92]S4). These findings suggest that the elevated expression of ALOXE3 was associated with tumor immune infiltration in COAD. Immune cell-based prognostic value of ALOXE3 expression in COAD As demonstrated in this study, ALOXE3 expression was significantly associated with immune cell infiltration in COAD. Additionally, COAD patients with increased ALOXE3 expression had a poorer prognosis. This led to the hypothesis that ALOXE3 may influence patient prognosis partly through its effect on immune infiltration. ALOXE3 expression showed a significant correlation with seven types of infiltrating immune cells in COAD tissues, such as CD4^+ T cells, CD8^+ T cells, macrophages, monocytes, neutrophils, dendritic cells, NK cells, and Tgd. We employed a Cox proportional hazards model using data from the TIMER database to assess the prognostic relevance of ALOXE3 expression and the abundance of tumor-infiltrating naïve CD4⁺ T cells. The analysis revealed that both naïve CD4⁺ T cell infiltration and ALOXE3 expression were significantly associated with clinical prognosis in COAD (Fig. [93]2G, HR = 3.07, p < 0.01). Nevertheless, no significant prognostic differences were observed for enriched levels of CD8^+ T cells, macrophages, neutrophils, monocytes, myeloid and plasmacytoid dendritic cells, NK cells, and Tgd between groups with high and low ALOXE3 expression levels (Fig. [94]S5 and [95]S6, p >0.05). The studies presented above imply that immune infiltration may, in part, influence the prognosis of COAD patients with high ALOXE3 expression. These findings imply that immune cell infiltration may partially mediate the prognostic impact of elevated ALOXE3 expression in COAD, highlighting a potential immunomodulatory role of ALOXE3 in the tumor microenvironment. ALOXE3 promotes COAD progression both in vitro and in vivo Next, we assessed the expression levels of ALOXE3 in COAD cell lines. Results indicated that ALOXE3 expression was higher in HCT15 and HCT116 cell lines compared to RKO, SW480, and SW620 cell lines (Figure [96]S7). In contrast, the expression of ALOXE3 was lower in SW480 and SW620 cell lines compared to other COAD cells (Figure [97]S7). To evaluate the role of ALOXE3 in regulating COAD phenotype, we generated two ALOXE3 shRNAs and overexpression lentiviruses. Lentiviral vectors carrying ALOXE3 were stably transduced into SW480 and SW620 cells to induce overexpression, as confirmed by RT-qPCR (Figure [98]S8 illustrates the overexpression efficiency in these RCC cells, p < 0.01). Conversely, ALOXE3 knockdown was achieved by stable transduction of shRNA-expressing lentiviruses into HCT15 and HCT116 cells (Figure [99]S9 displays the knockdown efficiency in these COAD cells, p < 0.01). To evaluate cell proliferation, CCK-8 assays were conducted over a 4-day period. ALOXE3-overexpressing SW480 and SW620 cells exhibited significantly higher proliferation rates compared to vector controls (n = 3 per group, p < 0.01; Fig. [100]3A and B). Colony formation assays showed that ALOXE3 overexpression increased the number of colonies (Fig. [101]3C to E), with quantification revealing a ~ 2-fold increase (p < 0.001). Fig. 3. [102]Fig. 3 [103]Open in a new tab Overexpression of ALOXE3 promoted COAD progression. (A-B) Cell viability of SW480 and SW620 cells after ALOXE3 overexpression, assessed by the CCK-8 assay. (C-E) Colony formation assay analyzing cell proliferation in different cells. (F-I) Wound healing assay measuring the migratory ability of ALOXE3-overexpressing and control cells. Scale bar, 400 𝜇m. (J-M) Transwell assay assessing cell migration and invasion abilities in SW480 cells and SW620 cells. Scale bar, 400 𝜇m. Data (means ± SEM, n = 3) were representative of three separate experiments with similar results. Compared with the indicated group, **p < 0.01, ***p < 0.001 Wound healing assays demonstrated enhanced migration in ALOXE3-overexpressing cells, with ~ 43% wound closure at 24 h compared to ~ 21% in controls (p < 0.01, Fig. [104]3F to I). In Transwell migration and invasion assays, ALOXE3 overexpression significantly increased the number of migrated and invaded cells per field (average increase of 1.6-fold; p < 0.01, Fig. [105]3J to M). Conversely, ALOXE3 knockdown in HCT15 and HCT116 cells led to suppressed proliferation (CCK-8, colony formation, p < 0.001, Fig. [106]4A to E), decreased migration (wound healing, p < 0.001, Fig. [107]4F to I), and significantly reduced migrasive and invasive capacity (Transwell assay, p < 0.05, Fig. [108]4J to M). Fig. 4. [109]Fig. 4 [110]Open in a new tab Knockdown of ALOXE3 inhibited the development of COAD. (A-B) CCK-8 assays determine the growth curves of the ALOXE3 stably knockdown HCT15 and HCT116 cells. (C-E) Colony formation assay analyzing cell proliferation in different cells. (F-I) Wound healing assay assessing the migratory ability of ALOXE3-knockdown and control cells. Scale bar, 400 𝜇m. (J-M) Transwell assay measuring migration and invasion in ALOXE3-knockdown HCT15 and HCT116 cells. Scale bar, 400 𝜇m. Data (means ± SEM, n = 3) were representative of three separate experiments with similar results. Compared with the indicated group, **p < 0.01, ***p < 0.001 To further evaluate the tumor-promoting role of ALOXE3 in vivo, we established subcutaneous xenograft models in nude mice (n = 5 per group). Mice were injected with 4 × 10⁶ SW480 or SW620 cells stably overexpressing ALOXE3 or with vector control cells. At the endpoint, tumors were excised and weighed. Mice injected with ALOXE3-overexpressing cells developed significantly larger tumors compared to controls (p < 0.01). For SW480 cells, the average tumor volume at day 28 was 1146 ± 266 mm³ vs. 274 ± 51 mm³ in controls (Fig. [111]5A and B, p < 0.01), with tumor weights of 0.263 ± 0.032 g vs. 0.114 ± 0.001 g (Fig. [112]5C, p < 0.01). For SW620 cells, the average tumor volume at day 28 was 799 ± 170 mm³ vs. 242 ± 30 mm³ in controls (Fig. [113]5D and E, p < 0.01), and tumor weight averaged 0.271 ± 0.033 g vs. 0.099 ± 0.013 g (Fig. [114]5F, p < 0.01). Fig. 5. [115]Fig. 5 [116]Open in a new tab ALOXE3 promoted colon cancer progression in a mouse tumor model. (A-F) The in vivo effects of ALOXE3 overexpression on the progression of COAD. (A, D) Representative gross images of subcutaneous tumors derived from COAD cells with or without ALOXE3 overexpression. (B, E) Tumor growth curves of indicated groups. (C, F) The tumor weights of different groups. (G-L) In vivo effects of ALOXE3 knockdown on COAD development. (G, J) Gross image of subcutaneous tumors derived from a xenograft model using different COAD cells. (H, K) Tumor growth curve of different groups. (I, L) The tumor weights of different groups. Data (n = 4 mice) were presented as the mean ± SEM from three independent experiments. Compared with the indicated group, ** p < 0.01, *** p < 0.001 In contrast, mice injected with ALOXE3-knockdown HCT15 or HCT116 cells developed smaller tumors. HCT15-sh1 tumors showed a final volume of 77 ± 9 mm³ vs. 309 ± 66 mm³ in control mice (p < 0.01, Fig. [117]5G and H), and tumor weight was reduced by ~ 33% (Fig. [118]5I, p < 0.01). HCT15-sh2 tumors showed a final volume of 103 ± 9 mm³ vs. 309 ± 66 mm³ in control mice (p < 0.01, Fig. [119]5G and H), and tumor weight was reduced by ~ 34% (Fig. [120]5I, p < 0.01). HCT116-sh1 tumors showed a final volume of 81 ± 10 mm³ vs. 266 ± 40 mm³ in control mice (p < 0.01, Fig. [121]5J and K), and tumor weight was reduced by ~ 34% (Fig. [122]5L, p < 0.01). HCT116-sh2 tumors showed a final volume of 74 ± 11 mm³ vs. 266 ± 40 mm³ in control mice (p < 0.01, Fig. [123]5J and K), and tumor weight was reduced by ~ 32% (Fig. [124]5L, p < 0.01). Collectively, these findings suggest that ALOXE3 plays a tumor-promoting role in COAD by enhancing proliferation, migration, and invasion in vitro and by accelerating tumor growth in vivo. Meanwhile, ALOXE3 may serve as a potential oncogenic driver and therapeutic target in COAD. ALOXE3 promoted colon cancer progression via the ERK1/2 signaling pathway Based on KEGG pathway enrichment analysis, we observed a strong association between ALOXE3 expression and the MAPK signaling cascade, which lays a critical role in regulating proliferation, migration, and survival. To experimentally validate whether ALOXE3 regulates COAD progression through the MAPK/ERK pathway, we performed western blotting to evaluate the phosphorylation status of ERK1/2 in colon cancer cell lines following ALOXE3 modulation. Our results revealed that overexpression of ALOXE3 in SW480 cells led to a marked increase in phosphorylated ERK1/2 (p-ERK1/2) levels, while total ERK1/2 levels remained unchanged (Fig. [125]6A to C and Figures S10, p < 0.05). This suggests that ALOXE3 selectively activates ERK1/2, rather than other MAPK signlaing pathway. Conversely, ALOXE3 knockdown in HCT116 cells resulted in a significant reduction of p-ERK1/2, further supporting its role in regulating ERK1/2 activity (Fig. [126]6D to F and Figures [127]S11, p < 0.05). Fig. 6. [128]Fig. 6 [129]Open in a new tab ALOXE3 promoted colon cancer progression via activation of ERK1/2. (A-F) The phosphorylation level of ERK, p-38, and JNK in the SW480 cells with ALOXE3 overexpression and in the HCT116 cells with ALOXE3 knockdown. The data are presented as the means ± SEM (n = 3). Compared with the indicated groups, *p < 0.05, **p < 0.01, ***p < 0.001 To further determine whether the oncogenic effects of ALOXE3 were mediated through ERK1/2 activation, we treated ALOXE3-overexpressing cells with SCH772984, a highly selective ERK1/2 inhibitor. Functional assays demonstrated that ERK1/2 inhibition substantially attenuated ALOXE3-induced cell proliferation (CCK-8 assay: Figure [130]7A and B, p < 0.05; colony formation: Figure [131]7C to E, p < 0.001), wound healing ability (Fig. [132]7F to I, p < 0.05), and matrigel migration and invasion in both SW480 and SW620 cells (Fig. [133]7J to M, p < 0.05). Specifically, the growth-promoting effects were reversed, and the motility and invasiveness of tumor cells were significantly suppressed upon ERK blockade. Fig. 7. [134]Fig. 7 [135]Open in a new tab ALOXE3 promoted colon cancer progression via ERK1/2 signaling pathway. (A-B) Cell viability of SW480 and SW620 cells following ALOXE3 overexpression and treatment with ERK inhibitor SCH772984 was detected by the cell counting kit-8 assay. (C-E) Colony formation assay showing the proliferative ability of SW480 and SW620 cells following ALOXE3 overexpression and SCH772984 treatment. (F-I) Wound healing assays were performed in different groups. Scale bar, 400 𝜇m. (J-M) Transwell assay assessing cell migration and invasion abilities in different groups. Scale bar, 400 𝜇m. Data (n = 3) were presented as the mean ± SEM from three independent experiments. Compared with the indicated group, * p < 0.05, ** p < 0.01, *** p < 0.001 Together, these findings indicate that ALOXE3 exerts its tumor-promoting functions, at least in part, through activation of the ERK1/2 signaling pathway. The downstream effects of ALOXE3-mediated ERK activation likely contribute to enhanced cell proliferation, migratory capacity, and invasiveness in COAD, highlighting ALOXE3-ERK1/2 axis as a potential therapeutic target in patients with high ALOXE3 expression. Discussion ALOXE3 is a member of the LOXs family, whose other members—such as ALOX5, ALOX12, and ALOX15—have previously been implicated in cancer development. In our study, we systematically analyzed the expression profiles and prognostic relevance of LOX family genes in COAD using data from TCGA. Remarkably, the expression of all LOX genes, except ALOX5 and ALOX15, were significantly changed in COAD compared to adjacent normal samples. Among these, ALOXE3 exhibited the strongest association with patient survival. Survival analysis revealed that elevated ALOXE3 expression was significantly correlated with poor prognosis in COAD. To validate these findings, we conducted RT-qPCR analysis, which confirmed significantly higher ALOXE3 mRNA expression in human COAD tissues relative to normal counterparts. To further elucidate the functional role of ALOXE3, we performed a series of in vitro and in vivo assays, including CCK-8, colony formation, wound healing, transwell, and tumor xenografts in a nude mice. KEGG pathway analysis showed that genes co-expressed with ALOXE3 predominantly participate in MAP kinase activity. In line with this, overexpression of ALOXE3 in colon cancer cells led to enhanced activation of the ERK1/2 signaling pathway, whereas treatment with the ERK inhibitor SCH772984 effectively abrogated the tumor-promoting effects of ALOXE3. Taken together, the findings demonstrate that ALOXE3 promotes COAD progression through the ERK1/2 pathway activation. These results suggest that ALOXE3 as a potential oncogenic driver and prognostic biomarker in colon adenocarcinoma. Previous studies have reported conflicting findings regarding the roles of other LOX family members in COAD [[136]24]. For example, researchers observed that ALOX15 levels were elevated in colorectal cancer tumor samples compared to adjacent non-tumorous tissues using Western blotting and immunohistochemistry [[137]25]. However, Shureiqi et al. found that a marked downregulation of ALOX15 expression in colorectal cancer tissues and altered colon cell lines, along with decreased levels of 13-HODE, a key ALOX15 byproduct, as revealed through immunohistochemical analysis [[138]26]. But Western blotting showed no difference in ALOX15 expression [[139]27]. Notably, 13-HODE has been shown to induce apoptosis in COAD cells and is upregulated following NSAID treatment, while pharmacological inhibition of ALOX15 abrogated the effects of NSAIDs [[140]27]. The results indicated that ALOX15 could serve as a molecular target in cancer treatment. Furthermore, ALOX5 has been implicated in COAD progression. The genetic variation of ALOX5 was associated with the risk of COAD [[141]28]. Suppressing ALOX5 expression inhibited the proliferation, migration, and invasion of COAD cells in vitro [[142]29, [143]30]. Our results also showed that the expression levels of ALOX5 and ALOX15 were markedly altered in COAD, compared with the paratumor samples. Although we observed differential expression of ALOX5 and ALOX15 in COAD, neither gene was significantly associated with patient survival, underscoring the distinct prognostic relevance of ALOXE3 in COAD. Previous studies indicated that ALOXE3 expression is significantly reduced in human glioblastoma (GBM) [[144]15], and silencing this gene in GBM cells led to increase tumor growth and decrease survival in mice [[145]15]. Ruan et al. used multivariate survival analysis and comprehensive prognostic assessment to establish a correlation between ALOXE3 and ALOX12 and overall survival in colon cancer. Notably, low expression levels of both ALOXE3 and ALOX12 were associated with improved prognosis in COAD. Furthermore, the combined expression of ALOXE3 and ALOX12 might serve as a promising predictive biomarker for COAD [[146]31]. Similarly, our study found that ALOXE3 expression was uniquely related to COAD survival prognosis among all LOXs. Additionally, we showed that elevated ALOXE3 expression promoted tumor growth, while its knockdown suppressed COAD progression, both in vitro and in vivo. We next explored the association between ALOXE3 expression and immunotype indicators in COAD for the first time. The research indicates that ALOXE3 expression is associated with immune cell infiltration in COAD, involving dendritic cells (DCs), macrophages, neutrophils, and naïve CD4^+ T cells was related to ALOXE3 expression. DCs, a heterogeneous population of antigen-presenting cells, can infiltrate tumors and modulate immune responses. The recruitment of immunosuppressive DC subsets into the tumor microenvironment has been shown to restrict T cell infiltration and promote tumor progression [[147]32]. Additionally, immunosuppressive DCs promoted CD4^+ T regulatory cell (Tregs) expansion in the tumor microenvironment. The local immune milieu can also promote the differentiation of naïve CD4⁺ T cells into induced Tregs [[148]33, [149]34], which were associated with poorer COAD patient prognoses. These predicted correlations in our study suggested the underlying mechanisms by which elevated ALOXE3 expression may be associated with infiltrated immune cell function in COAD. As a result, the poor prognosis of COAD may be associated with the presence of infiltrating immune cells within the tumor microenvironment. In COAD’s naïve CD4^+ T cell groups, elevated ALOXE3 expression was prevalent, and these cells exhibited a worse prognosis based on Kaplan Meier-Plotter analysis with the TIMER data. Given that ALOXE3 expression was mainly linked to naïve CD4^+ T cells, it is hypothesized that these genes were more common in naïve CD4^+ T cells from COAD patients, this hypothesis requires further validation through functional bioassays. Additionally, ALOXE3 co-expressed genes were primarily involved in the relaxin signaling pathway, proteoglycans-related pathways, and MAP kinase activity. Related genes of MAP kinase activity [[150]35–[151]38] and proteoglycans [[152]39] have identified as possible biomarkers in COAD. The relaxin signaling pathway has been implicated in promoting tumor proliferation, angiogenesis, extracellular matrix remodeling, and tumoral blood flow [[153]40–[154]42]. Future research is thus required to determine whether ALOXE3-mediated immune infiltration in COAD is mechanistically linked to the MAPK and relaxin signaling pathways. These insights may further elucidate the contribution of immune infiltration to poor prognosis in COAD patients with elevated ALOXE3 expression. ERK1/2 activation often stimulates cell proliferation, and its uncontrolled activity is a characteristic feature of several types of cancer [[155]43–[156]46]. In human colorectal cancer specimens, elevated levels of phosphorylated ERK have been observed prior to radiation therapy [[157]47]. Moreover, sustained ERK1/2 activation has been implicated in promoting the progression of colon cancer [[158]48–[159]51]. Our study found that the ALOXE3 co-expressed gene network predominantly participates in MAP kinase activity. Consistently, both in vitro and in vivo experiments demonstrated that ALOXE3 overexpression significantly promoted the proliferation, motility, and invasive potential of COAD cells through the ERK1/2 signaling pathway. Despite these findings, our study has several limitations that warrant consideration. First, the experimental validation of ALOXE3 expression was performed on a limited number of human COAD tissue samples (n = 6), which may reduce the statistical power and generalizability of our conclusions. Although the RT-qPCR results were consistent with our bioinformatics predictions, further validation using larger and independent clinical cohorts is necessary to confirm the diagnostic and prognostic significance of ALOXE3 in colon cancer. Second, much of our analysis is based on large-scale transcriptomic datasets and computational predictions, which, while informative, are inherently correlative. For example, the association between ALOXE3 expression and immune infiltration does not confirm direct mechanistic relationships. Thus, while these results generate important hypotheses, they require experimental validation in well-controlled functional assays. Third, although we found that ALOXE3 expression is significantly associated with several types of immune cells, particularly naïve CD4^+ T cells and dendritic cells, the precise mechanisms by which ALOXE3 modulates immune cell behavior remain unclear. It is unknown whether ALOXE3 acts directly on immune cells or indirectly through tumor-derived cytokines or chemokines. Future studies using single-cell RNA sequencing, spatial transcriptomics, and immune co-culture systems are necessary to identify the cellular sources of ALOXE3 and determine whether it directly influences Treg differentiation, dendritic cell activation, or other aspects of the tumor immune microenvironment. Furthermore, while we propose that ALOXE3 plays an oncogenic role specifically in COAD, the molecular basis underlying this tissue-specificity remains unclear. It is possible that ALOXE3 cooperates with COAD-specific genetic or epigenetic alterations—such as KRAS, APC, or BRAF mutations—which may create a permissive environment for its tumor-promoting activity. However, these potential interactions were not investigated in the present study. Future integrative analyses combining ALOXE3 expression with mutational and epigenomic data may uncover genetic contexts that modulate its activity. Additionally, our study does not address whether ALOXE3 expression is causally linked to COAD prognosis in patients with distinct genomic backgrounds, such as those with KRAS or BRAF mutations. Integrating genomic and transcriptomic profiles in future studies will help determine whether ALOXE3 is an independent prognostic factor or part of a broader molecular network. Likewise, the roles of other LOX family members, such as ALOX5 and ALOX15, remain ambiguous due to conflicting findings in the literature. While we observed differential expression, these genes were not prognostically significant in our dataset, contrasting with earlier studies suggesting tumor-promoting or suppressive roles. These discrepancies emphasize the need for functional validation in multiple datasets and model systems. From a translational standpoint, the oncogenic potential and immunomodulatory functions of ALOXE3 make it an attractive candidate for therapeutic targeting. Its role in activating the ERK signaling pathway suggests that patients with high ALOXE3 expression may benefit from MEK/ERK-targeted therapies. Moreover, its association with immune infiltration raises the possibility of combining such treatments with immunotherapy to enhance anti-tumor responses. However, these therapeutic implications remain hypothetical and require validation in clinical studies. Taken together, our bioinformatics and experimental analyses provide compelling evidence that ALOXE3 acts as a novel oncogene in COAD, promoting tumor progression through both ERK1/2 activation and modulation of immune cell infiltration. Moreover, this study underscores the potential of cross-disciplinary tools and technologies—such as AI-assisted imaging and multi-omics integration—for advancing cancer research. For example, artificial intelligence-based anatomical imaging, which has significantly enhanced diagnostic accuracy in veterinary medicine, could be leveraged in digital pathology for tumor segmentation, subtype classification, and survival prediction in COAD [[160]52–[161]55]. Similarly, transcriptomic profiling and network analysis techniques used in stem cell research offer valuable frameworks for uncovering critical regulatory pathways and therapeutic targets in colorectal cancer [[162]56–[163]59]. Overall, integrating advanced computational tools with experimental validation accelerates our understanding of tumor biology and supports the development of precision oncology strategies targeting ALOXE3 and its associated pathways. Conclusions In conclusion, our study reveals that ALOXE3 is significantly upregulated in COAD and is associated with poor patient prognosis, supporting its potential as a novel prognostic biomarker. By integrating bioinformatics analyses with experimental validation, we demonstrated that ALOXE3 promotes tumor progression via activation of the ERK1/2 signaling pathway and may be associated with immune infiltration within the tumor microenvironment, particularly through its association with naïve CD4^+ T cell infiltration. These findings suggest that ALOXE3 may serve not only as a prognostic indicator but also as a potential therapeutic target in COAD. From a clinical perspective, the identification of ALOXE3 as a biomarker opens new avenues for personalized treatment strategies. Its expression could be used to stratify patients by risk and guide clinical decisions regarding surveillance and therapy. Notably, the activation of the ERK1/2 signaling pathway in ALOXE3-overexpressing tumors suggests that these patients may benefit from ERK-targeted therapies, such as SCH772984. Thus, ALOXE3 may have dual utility as both a biomarker and a therapeutic target in combination treatment regimens. However, this study represents an preliminary investigation, and further research is essential to validate these findings. Future investigations should aim to elucidate the underlying biological mechanisms of ALOXE3, confirm its prognostic and predictive value in larger, multi-center cohorts, and evaluate its role in therapeutic response through prospective clinical trials. Ultimately, incorporating ALOXE3 into molecular diagnostic panels could enhance precision oncology and improve individualized management of COAD patients. Electronic supplementary material Below is the link to the electronic supplementary material. [164]Supplementary Material 1^ (10.5MB, docx) Acknowledgements