Colorectal cancer (CRC), encompassing both colon and rectal malignancies, is prevalent cancer affecting the digestive system, ranking as a leading cause of cancer-associated mortality and morbidity globally (Siegel et al., 2020). Despite a stabilized or decreased incidence in recent decades, attributed to enhanced screening practices, such as colonoscopic polypectomy, and alterations in risk factors, including a reduction in smoking and increased aspirin consumption, the 5-year survival rate remains unsatisfactory (Edwards et al., 2014; Favoriti et al., 2016; Brenner and Chen, 2018; Keum and Giovannucci, 2019). The widely utilized tumor-node-metastasis (TNM) system plays a crucial role in clinical decision-making for risk assessment and treatment planning. However, owing to the high molecular heterogeneity of CRC, patients with seemingly identical clinicopathological features may exhibit considerable variation in the risk of recurrence and death (Lin et al., 2021). Thus, the identification of effective methods for early diagnosis has become a focal point of research endeavors.

Emerging evidence has highlighted the significance of the nervous system in the pathogenesis of malignancies, with nerves being identified as a critical component of the tumor microenvironment (Zahalka and Frenette, 2020). And as important neural signaling messengers, studies have suggested that neurotransmitters and their receptors play a crucial role in tumor proliferation, angiogenesis, and metastasis, contributing to cancer development. For example, activation of the β2-adrenoceptor has been shown to promote tumor growth and angiogenesis through increased expression of vascular endothelial growth factor, metalloproteases 2, and metalloproteases 9, which further enhance angiogenic and metastatic processes in ovarian, lung, and breast cancers (Thaker et al., 2006). Furthermore, neurotransmitter receptors are widely expressed on the surface of immune cells and are regulated by their corresponding neurotransmitters, thus affecting tumor immune responses (Jiang et al., 2020b; Cervantes-Villagrana et al., 2020).

The role of neurotransmitter receptors in CRC is also complex, as noted in the literature. Li et al. (Li et al., 2021) demonstrated that the overproduction of 5-hydroxytryptamine overproduction promotes colitis-associated CRC progression by enhancing NLRP3 inflammasome activation. In addition, another study found that atropine and muscarinic receptor 3 blockers reduced tumor weight, volume, and enhanced antitumor immune responses by increasing infiltration of CD4⁺ and CD8⁺ T cells and significantly reducing PD-L1 expression in a CRC mouse model (Kuol et al., 2022). Gamma-Aminobutyric Acid Type B Receptor (GABABR) also plays a pivotal role in CRC progression. GABABR1, a central component of GABABR, shows significantly lower expression in tumor tissues than in non-tumor normal tissues. It impairs the migration and invasion of CRC cells by inhibiting EMT and the Hippo/YAP1 pathway (Wang et al., 2021). However, studies focusing on subtype characterization and risk signatures based on neurotransmitter receptor-related genes in CRC remain limited.

Herein, we conducted Weighted Gene Co-Expression Network Analysis (WGCNA) and machine learning to establish a reliable signature rooted in neurotransmitter receptor-related genes (NRGs). And its prognostic utility was systematically evaluated in CRC patients. Additionally, we revealed the importance of these gene signatures in the immune microenvironment of CRC. The associations of antineoplastic drugs with MRS were also explored. Overall, this study provides a research basis for exploring the potential pathogenesis of CRC and offers new ideas for treating this disease.

Neurotransmitters are traditionally known as nerve-secreted substances that modulate excitatory or inhibitory neuronal functions by binding to specific receptors. And our understanding of the regulatory role of the neurotransmitter system in tumor initiation and progression continues to advance. Neurotransmitters exhibit varying effects on the numerous functions of cancer cells, endothelial cells, and immune cells across various human cancer types. The aberrant expression of neurotransmitter signaling genes in colorectal cancer underscores the potential of neurotransmitters to enhance tumor growth and metastasis by stimulating processes such as cell proliferation, migration, invasion, and angiogenesis. In addition, neurotransmitters can influence immune cells and endothelial cells in the tumor microenvironment, fostering inflammation and contributing to the advancement of tumor growth (Battaglin et al., 2022). Nevertheless, the precise impact of various neurotransmitter receptors on colorectal cancer progression remains poorly understood. Therefore, we have identified a signature consisting of four genes associated with neurotransmitter receptors: CHRNA3, GABRD, GRIK3, and GRIK5 to predict prognosis and treatment response in CRC patients.

In our study, we employed WGCNA analysis and identified 17 modules to help explore the characteristic relationship between neurotransmitter scores and gene clusters. Additionally, we utilized machine learning algorithms to enhance the accuracy of biomarker screening. Three machine learning algorithms (LASSO logistic, SVM-RFE and RF) were primarily utilized to screen feature variables and establish the best classification model. As a result, we identified CHRNA3, GABRD, GRIK3, and GRIK5 as biomarkers by combining the machine learning algorithm and WGCNA. And these biomarkers were well-validated in the external validation cohorts.

CHRNA3 is a member of the nicotinic acetylcholine receptor (nAChRs) gene cluster, which serves as the “gateways” through which nicotine exerts its effects on the brain (Ware et al., 2012). PPI analysis reveals that CHRNA3 has a solid physical interaction with CHRNA5 and CHRNB4. Previous studies have reported that variations in CHRNA3-A5-B4 are independently and additively associated with increased cigarette consumption, nicotine dependence, and lung cancer risk (Chmielowiec et al., 2022). Despite the absence of evidence on the role of CHRNA3 in CRC progression, we have identified, for the first time, that high CHRNA3 expression is associated with poor prognosis in CRC patients. Gamma-aminobutyric acid (GABA) is the principal inhibitory neurotransmitter in the adult mammalian central nervous system. Its receptors, expressed in various tumor tissues, play a crucial role in regulating tumor cell proliferation and migration (Joseph et al., 2002; Kanbara et al., 2018; Jiang et al., 2020a; Zhang et al., 2023). The γ-aminobutyric acid type A receptor d subunit (GABRD), encoded in the human chromosome 1p36 region, has yet to be fully elucidated regarding its involvement in cancers (Zhang et al., 2019). Recently, Wu et al. conducted a study that found enhanced expression of GABRD to be predictive of poor prognosis in CRC patients (Yan et al., 2020), consistent with our results. And Huang et al. confirmed that GABRD receptors indicated by T cells directly inhibit CD8⁺ T cells by participating in signal regulation (Huang et al., 2022). Glutamate ionotropic receptor kainate type subunit 3 (GRIK3) and glutamate receptor ionotropic kainate-5 (GRIK5) are members of the glutamate kainate receptor family and play crucial roles in the neuroactive ligand-receptor interaction pathway (Fang et al., 2021; Minoza et al., 2022). There is compelling evidence suggesting that GRIK3 participates in cancer progression. For example, GRIK3 was reported to mediate the function of CircRNA and promote the proliferation and metastasis of colon cancer cells (Fang et al., 2021). Xiao et al. found that GRIK3 promotes epithelial-mesenchymal transition in breast cancer cells by regulating SPDEF/CDH1 signaling (Xiao et al., 2019). Furthermore, GRIK5 has been identified as a potential biomarker for melanoma metastatic progression (Minoza et al., 2022). Although limited research exists on the relationship between GRIK5 expression and colorectal cancer (CRC), our study fills this gap by identifying a significant association between high GRIK5 expression and CRC progression.

Functional enrichments were performed to gain insights into the biological processes in which these hub genes may be involved. Our results suggest that these hub genes may play critical roles in cellular metabolic processes, particularly in organic acid, inorganic acid, and lipid metabolism. It is well-established that the hostile tumor microenvironment surrounding cancer cells drives metabolic changes that impact tumorigenesis and metastatic potential. Previous studies have validated the prominent status of lipid metabolism in cancer progression. Moreover, targeting dysfunctional lipid metabolism has shown promising results as an approach to impede tumor growth (Bian et al., 2021). Amino acids also represent a crucial aspect of the tumor microenvironment, which can significantly affect cancer cell metabolism and overall tumor development (Stepka et al., 2021). Our findings suggest that these biomarkers may contribute to the construction of a TME that favors tumor development. Thus, the current study investigated the correlation between tumor infiltrating cells and neurotransmitter receptor-related gene prognostic signatures. Using the CIBERSORT algorithm, we comprehensively evaluated the abundance and infiltration of twenty-two immune cells in COAD patients. The results showed that the high expression of gene signature groups exhibits elevated levels of immunosuppressive cells, such as Tregs and macrophages M0, and low infiltration of anti-tumor cells, including CD4⁺ and dendritic cells. Previous studies have demonstrated that the significant infiltration of M0 macrophages in the tumor microenvironment may predict poor prognosis (Zheng et al., 2021), while Tregs can induce immune tolerance and facilitate immune escape and tumor metastasis (Bauer et al., 2014). These findings are consistent with our study, supporting that high expression of hub genes with increased immunosuppressive cell infiltration is associated with poor prognosis in COAD patients. The B cell populations in the TME exhibit significant heterogeneity in surface immunophenotype and function (Downs-Canner et al., 2022). Memory B cells are found in higher numbers in tumors than peripheral blood, accounting for 34% of B cells in tumors compared to 14% in peripheral blood, regardless of tumor grade. Interestingly, patients who respond to immune checkpoint inhibition therapy exhibit increased memory B cells, CXCR3+ cells, and germinal center-like B cells in the TME (Helmink et al., 2020). Here, we found a positive association between hub genes and naive B cells, indicating that the neurotransmitter receptor-related gene signature may serve as a predictive marker for the effectiveness of immunotherapy in CRC patients.

Stromal cells and endothelial cells are another important component of TME. Interestingly, in the current study, increased expression of CHRNA3, GRIK3, and GRIK5 was highly associated with stromal cells, especially CAFs. CAFs play a critical role in CRC progression and are instrumental in shaping the tumor-promoting immune microenvironment (Kobayashi et al., 2022). Additionally, endothelial cells have been identified as one of the primary sources of CAFs and play a vital role in promoting tumor metastasis (Yan et al., 2020). Notably, our results demonstrated that increased GABRD expression was highly associated with endothelial cells.

Based on the neurotransmitter receptor-related gene prognostic signatures, a prognostic nomogram was developed. And We have validated the accuracy of the nomogram through calibration plots and decision curve analysis, which supports its potential as a valuable instrument for personalized risk management. We also utilized the CellMiner database to examine the relationship between FDA-approved drugs and these four targets. Our analysis of drug sensitivity revealed that the expression of CHRNA3, GABRD, GRIK3, and GRIK5 in cancer cells significantly impacted their response to chemotherapy. Our discoveries could provide novel insights into the appropriate selection of drugs, offering guidance for forthcoming studies in oncology. Finally, we performed a simple validation of the expression of hub genes in CRC tissues and found that, compared to normal tissues, the mRNA levels of CHRNA3, GRIK3, and GRIK5 were decreased. CHRNA3, GRIK3, and GRIK5 are primarily expressed in connective tissue. Due to the predominance of tumor cells in tumor tissues, they are downregulated in tumor tissues compared to normal tissues. Both immunohistochemistry and RT-PCR confirmed the high expression of GABRD in tumor tissues. Interestingly, among these four hub genes, three of them exhibit relatively lower expression in colorectal cancer tissue compared to normal tissue. This observation indeed presents a complex and multifaceted biological phenomenon. We suggest that gene function can be independent of expression levels. The phenomenon of oncogene downregulation in tumor tissues contradicts common expectations, which generally predict higher expression levels of cancer genes. We delved into the literature and found similar instances where gene expression is influenced by various mechanisms, such as: The B-cell lymphoma 2 (BCL-2) family of proteins regulates apoptosis in normal cells. In various cancers, increased expression of BCL-2 protein is associated with enhanced drug resistance of tumor cells, although in some cases, its expression may be lower in tumor tissues compared to the surrounding normal tissues (Adams and Cory, 2007). And Hexokinase 2 (HK2) is involved in glucose metabolism in normal tissues, while in cancer cells, increased activity of HK2 is associated with cancer survival and growth, even if its expression levels are lower than in normal tissues, representing the role of metabolic reprogramming in tumors (Patra and Hay, 2014). Despite these genes being less commonly discussed, their existence and paradoxical behavior in cancer biology are well-documented. To address the question of why these three genes are expressed at lower levels in colorectal cancer tissue but are associated with a poorer prognosis in patients with high expression, further investigations in proteomics and epigenetics are required.

Though we have identified a significant gene signature in CRC based on the NRGs, there were some limitations. Firstly, our findings were based on public databases, and thus, it is crucial to validate these results in a prospective cohort from our hospital. Additionally, we need to investigate the functions of the hub genes implicated in CRC progression using cell lines and/or mouse models.

In our study, we employed several bioinformatics approaches to identify a 4-gene signature related to neurotransmitter receptors to evaluate the prognosis of CRC patients. Our results indicated a significant association between the signature and the clinical features and immune system of colorectal cancer. Thus, the gene signature in our study could function as an independent prognostic indicator for CRC patients.