Impact of cytosine and piperine on colorectal cancer progression based on Mendelian randomization and functional validation.

Introduction
Colorectal cancer (CRC) is one of the most common malignancies affecting the digestive system worldwide. Recent cancer statistics reveal that in 2020, there were over 1.9 million new diagnoses, resulting in about 935,000 deaths [1]. This accounts for approximately 10% of all cancer cases and mortality combined. The global rates of CRC rose alarmingly, from fifth to second place between 2018 and 2020 [1, 2]. Considering this alarming trend, improving prevention and screening methods for CRC has become a top priority. Multiple studies have identified several risk factors associated with CRC, such as smoking [3], alcohol use [4], Type 2 diabetes (T2D) [5], body mass index (BMI), waist-to-hip ratio (WHR) [6], and total cholesterol (TC) levels [7]. However, little research has focused on the metabolic changes linked to CRC, which indicates an important area for further investigation.
Recently, the rise of metabolomics within the systems biology framework has introduced a novel approach to exploring the mechanisms underlying various diseases. Metabolomics enables researchers to gain insights into biological processes by identifying altered metabolites and metabolic pathways associated with disease states [8, 9]. A growing body of evidence in recent years emphasizes the importance of metabolic alterations and energy processes in the proliferation and metastasis of cancerous cells [10, 11]. In normal cells, while altered metabolism can support proliferation and division, it may also disrupt differentiation, increasing susceptibility to cancer [10]. Furthermore, the targeted regulation of metabolites presents promising applications in cancer treatment. A notable example is dichloroacetate (DCA), which has been shown to inhibit the phosphorylation of pyruvate dehydrogenase (PDH) [12] and reverse the Warburg effect, thereby enhancing mitochondrial pyruvate oxidation and suppressing tumor growth [13, 14]. Modulating cellular metabolism has demonstrated the potential to increase the sensitivity of oncogenes to therapeutic interventions [15]. Combinations of metabolic inhibitors are emerging as a viable strategy to combat chemotherapy resistance, highlighting an area worthy of future research. Observational studies have been instrumental in identifying associations between circulating metabolites and colorectal cancer risk; however, these studies are inherently limited by potential confounding factors and reverse causation [16]. Confounders such as lifestyle, diet, and environmental exposures may bias the observed associations, making it challenging to establish causal relationships. Additionally, reverse causation where preclinical or undiagnosed cancer influences metabolite levels can further complicate interpretation of observational findings.
Mendelian randomization (MR) studies have recently gained traction as a method for examining the causes of diseases. In the absence of randomized controlled trials (RCTs), MR serves as a valuable method for exploring causal links between exposures and outcomes [17]. This technique employs genetically associated exposure through the selection of exposure-related single nucleotide polymorphisms (SNPs) as instrumental variables (IVs) [18]. This IV strategy effectively simulates RCTs, as SNPs are randomly distributed to offspring at conception, which helps reduce confounding factors; variables such as sex and age are less likely to bias the causal effects observed [19]. Additionally, the issue of reverse causality is diminished in MR studies, as genotype formation occurs before the disease manifests. Investigating metabolites linked to the onset of CRC is essential for enhancing early screening and prevention strategies and gaining insights into the biological processes involved in CRC treatment. The direct relationship between specific metabolites and CRC remains unknown, as there have been no prospective studies conducted thus far to explore this connection. Conventional observational studies often face design limitations that complicate the interpretation of causality. For instance, lifestyle changes following a diagnosis of cancer, long-term use of certain medications, and metabolic changes induced by tumors can all obscure the link between metabolites and CRC.
Due to the uncertainty surrounding the causal link between blood metabolites and CRC, further research is essential. In this investigation, we initially performed an extensive MR analysis to pinpoint possible causal metabolites linked to CRC. Following this, we will confirm these results through functional experiments, providing mechanistic insights into how these metabolites might affect CRC progression. Our study aims to enhance the understanding of the metabolic processes related to CRC and identify potential targets for preventive and therapeutic strategies.

Materials and methods

Study design
The relationship between CRC and 1,400 metabolites was investigated utilizing two-sample MR analyses. This method used genetic variants as instrumental variables (IVs) to represent metabolic risk factors. To ensure that the causal conclusions were reliable, the IVs employed in the MR analysis needed to satisfy three essential criteria: (i) the genetic variants must show a direct link with the exposure (metabolites); (ii) the genetic variants should not be linked to confounding factors that could influence both the exposure and the outcome; and (iii) the genetic variants should affect the outcome exclusively through the exposure, without any alternative pathways involved (Fig. 1). This study used CRC data from the OpenGWAS, which included 387,318 Europeans, including 4,562 cases and 382,756 as the control group.

Metabolite GWAS data sources for exposure
The summary statistics for genome-wide association studies (GWAS) concerning each metabolite were retrieved from the European GWAS database (https://www.ebi.ac.uk/gwas/) under accession codes GCST90199621-90201020 [19]. To obtain cancer-specific data, relevant search terms were applied in the GWAS Catalog (https://gwas.mrcieu.ac.uk/). For CRC identification, the dataset labeled ukb-saige-153 was selected. This dataset originates from the UK Biobank (www.ukbiobank.ac.uk), a large-scale biomedical database and research resource containing detailed genetic, environmental, and lifestyle data on over 500,000 UK participants. The UK Biobank is one of the largest and most comprehensive biomedical datasets available globally.

Instrument selection
Considering the large number of SNPs showing genome-wide significance (P < 5 × 10^-8) for metabolite characteristics, a more stringent threshold (P < 5 × 10^-9) was applied to select genetic instrumental variables (IVs). These IVs were clumped based on linkage disequilibrium (LD) reference from the 1,000 Genomes Project. Due to the relatively small size of the GWAS dataset related to metabolites, a threshold (P = 5 × 10^-8) and a less strict clumping criterion (R² < 0.001, 10,000 kb) were utilized [20]. To ensure instrument strength, only IVs with an F statistic greater than 10 were retained for analysis. These IVs were then extracted from colorectal cancer outcome summary statistics, and any IVs showing potential pleiotropy (P < 1 × 10^-5) with CRC were excluded, following previously established protocols [21]. To ensure consistency throughout this analysis, any discrepancies in SNPs between the exposure and outcome datasets were aligned to guarantee uniformity in effect estimates for the same effect allele [22].

Functional experiments

Reagents and instruments
The reagents used included Dulbecco’s Modified Eagle Medium (DMEM; Cat# PM150210) and RPMI-1640 medium (Cat# PM150110), both purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China). Fetal bovine serum (FBS; Cat# 164210-50) was also from Procell. Penicillin–streptomycin solution (100×; Cat# C100C5) and 0.25% trypsin-EDTA (Cat# C100C1) were supplied by NCM Biotech (Suzhou, China). Phosphate-buffered saline (PBS; Cat# G4202) was acquired from Servicebio (Wuhan, China). Piperine (Cat# HY-N0144) and cytosine (Cat# HY-I0626) were purchased from MedChemExpress (Shanghai, China). Cell Counting Kit-8 (CCK-8; Cat# C0039) was obtained from Beyotime Biotechnology (Shanghai, China). Paraformaldehyde 4% (Cat# BL539A) was purchased from Biosharp (Hefei, China). Crystal violet solution (0.2%; Cat# C805211) was purchased from Macklin Biochemical Co., Ltd. (Shanghai, China).
Optical density measurements were performed using microplate readers (BioTek Instruments, USA; Agilent Technologies, USA). Cell morphology and wound healing assays were visualized using an inverted microscope (Nikon, Japan). Image analysis was conducted using ImageJ software (Fiji, version 2023).

Cell lines
Human colorectal cancer cell lines HCT-8 and RKO were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). HCT-8 cells were cultured in DMEM, and RKO cells in RPMI-1640 medium. Both media were supplemented with 10% FBS and 100 U/mL penicillin-streptomycin. Cells were incubated at 37 ℃ in a humidified atmosphere containing 5% CO2. At 80–90% confluence, cells were detached using trypsin-EDTA and subcultured. All cell lines were authenticated by short tandem repeat (STR) profiling and confirmed free of mycoplasma contamination before use.

Cell proliferation assay
HCT-8 cells in logarithmic growth phase were trypsinized, resuspended, and counted using a hemocytometer. 5,000 cells per well were seeded into 96-well plates. Cells were treated with piperine or cytosine at 0 and 50 µM. Cell viability was measured at 4, 24, 48, 72, and 96 h post-treatment by adding 10 µL of CCK-8 solution per well, followed by 1.5 h incubation at 37 ℃ in the dark. Optical density at 450 nm was read using a microplate reader. Each condition was performed in triplicate.

Wound healing assay
HCT-8 and RKO cells in logarithmic phase were trypsinized, counted, and seeded at 1.0 × 10^5 cells per well in 6-well plates. After reaching about 80% confluence, a scratch was made using a sterile 10 µL pipette tip. Detached cells were removed by washing twice with PBS. Medium containing 2% FBS and indicated metabolite concentrations was added. Images were captured at 0 and 24 h using an inverted microscope. Wound closure percentage was calculated by ImageJ as: [(initial wound area - final wound area) / initial wound area] × 100%. Experiments were done in triplicate.

Colony formation assay
HCT-8 and RKO cells were seeded at 500 cells per well in 6-well plates and cultured for 10 days. Colonies were fixed with 4% paraformaldehyde for 15 min, stained with 0.2% crystal violet for 10 min, then washed with PBS. Colonies were counted manually or by ImageJ. Experiments were performed in triplicate.

Statistical analysis
All statistical analyses were conducted with R 4.3.2 (https://www.Rproject.org). To assess the causal relationship between 1,400 metabolites, and colorectal cancer, we completed inverse variance weighting (IVW), MR Egger, Weighted median, and Weight mode, using the “TwoSampleMR” package [23]. Cochran’s Q statistical measurements and the appropriate p-values were used to test for heterogeneity among the selected IVs. If the null hypothesis was rejected, random effects IVW replaced fixed effects IVW [24]. To rule out the effect of horizontal multidimensionality, we used a frequently utilized approach (MR-Egger), which implores the presence of horizontal multidimensionality if the interception term is statistically significant [24]. We also used the powerful MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) method to eliminate potential horizontal pleiotropic outliers, which could have affected the estimation results of the MR-PRESSO package [25]. We also employed scatterplots, funnel plots, and leave-one- out. The scatter plots were used to demonstrate that the results were unaffected by outliers, and the funnel plots showed that the correlations were robust and not heterogeneous. Data from functional experiments were examined utilizing ImageJ software (version 2023; Fiji) and GraphPad Prism 8 (GraphPad; Dotmatics). Variations between the two groups were assessed using unpaired Student’s t-tests, whereas comparisons across multiple groups were performed using one-way ANOVA, followed by Bonferroni post hoc tests. A p-value of less than 0.05 was set as the threshold for statistical significance.

Results

Causal relationship between metabolites and CRC
We began by assessing the causal impact of different metabolites on CRC using the inverse variance weighted (IVW) approach in a MR analysis. This assessment revealed 58 metabolites significantly related to CRC (Figs. 2 and 3). Notably, piperine exhibited a strong link to CRC, with an odds ratio (OR) of 0.862, a confidence interval (CI) of 95% at 0.768–0.967, and a p-value of 0.011. In addition, cytosine was significantly associated with an increased risk of CRC, showing an OR of 1.166, 95% CI of 1.020–1.333, and a p-value of 0.023 (Fig. 4). A sensitivity analysis was also performed. Although some heterogeneity and significant findings were observed from Cochran’s Q test (P < 0.05) (see Table SI), the causal estimates remained consistent using the random-effects IVW model. The p-values for the MR-Egger intercept were > 0.05, highlighting the absence of significant pleiotropy effects (see Table SII). Furthermore, we conducted additional evaluations through scatter plots (Fig. 5A-B), funnel plots (Fig. 5C-D), and leave-one-out plots (Fig. 5E-F), which helped exclude the impact of outliers and horizontal pleiotropy on the identified key metabolites.

Functional experiment validate cytosine and Piperine in CRC
To assess the impact of two metabolites, cytosine and piperine, on the growth and movement capabilities of CRC cells, we developed in vitro CRC models using HCT-8 and RKO cell lines. The CCK-8 assay indicated that cytosine stimulated the proliferation of HCT-8 CRC cells, while piperine demonstrated a reduction in proliferation within the same cell line (Fig. 6A-B). The scratch wound assay revealed that CRC cells (both HCT-8 and RKO) treated with cytosine exhibited a significantly quicker closure of the scratch compared to the control group, signifying improved migration abilities (Fig. 7A-B). In contrast, cells treated with piperine showed considerably slower wound healing compared to the control, indicating an inhibitory effect on cell migration (Fig. 8A-B). The colony formation assay also showed that cytosine enhanced both the proliferation and migration of HCT-8 and RKO CRC cells. Conversely, piperine suppresses both proliferation and migration of these cancer cell types (Fig. 9A-B).

The selection of cytosine and piperine for our study is based on multiple key factors.

Biologically, cytosine plays an essential role in nucleotide metabolism, DNA synthesis, and repair, while piperine is involved in antioxidant, anti-inflammatory, and immune-regulatory pathways, with growing evidence supporting its potential anti-tumor and immune-modulatory effects. These biological properties make both metabolites highly relevant to our research, particularly in exploring the tumor immune microenvironment and potential therapeutic interventions.

From a practical perspective, both cytosine and piperine are cost-effective, which is crucial given the limited budget of our project. Their affordability enables us to conduct a broad range of experiments while maintaining financial feasibility. Furthermore, these metabolites are readily available from reliable suppliers, ensuring smooth and uninterrupted experimental operations.

Both cytosine and piperine are easy to handle and have well-established methodologies for their use in experiments. This simplicity enhances the reliability, stability, and reproducibility of our results. Moreover, considering our budget constraints, these metabolites provide an efficient solution to maximize the impact of our research within our resource limitations.

Finally, we believe that cytosine and piperine not only hold promise for the current study but also offer significant potential for future research, particularly in cancer immunotherapy and metabolic intervention. Their mechanisms could contribute to breakthroughs in clinical applications, making them invaluable for long-term scientific progress.

Discussion
Cancer cells have a unique ability to thrive in environments that often lack essential nutrients. They can extract the nutrients they need to survive and grow, allowing them to continue multiplying and forming new cells [26]. This adaptability is crucial for tumor growth, as it involves a significant shift in how these cells manage their energy and resources to support their rapid division and the creation of new tissue [27], resulting in the production of different metabolites. The alterations in intracellular or extracellular metabolites associated with tumor metabolic reprogramming significantly influence gene expression, cellular differentiation, and the tumor microenvironment. The advancement of metabolomics technology has progressively heightened people’s curiosity in investigating the clinical usefulness of circulating metabolites as non-intrusive biomarkers. It has the advantages of convenience, easy access, and minimal harm and is a promising biomarker. Increasing evidence indicates that plasma metabolites in cancer patients are significantly linked to both the development and prognosis of cancer [28–30]. In this investigation, we examined the association between various circulating metabolites and colorectal cancer through MR. We discovered that higher levels of Cytosine were positively associated with an elevated risk of colorectal cancer. Meanwhile, Piperine levels showed a negative correlation with colorectal cancer risk, functioning as a protective factor. These findings enhance our knowledge of the intricate relationships between metabolites and genetics in the progression of colorectal cancer, paving the way for innovative ideas to cancer prevention and treatment.
Cytosine levels are key components in nucleic acids (DNA and RNA). In the double helix of DNA, cytosine on one strand pairs with guanine on the other, establishing three hydrogen bonds between molecules. This Hydrogen bonding between complementary base pairs is crucial for stabilizing the DNA double helix structure. Cytosine methylation naturally exists in the genomes of different organisms and is a well-known epigenetic mechanism that regulates gene expression. Meanwhile, the primary DNA modification in mammalian genomes is cytosine methylation, which involves adding a methyl group [5-methylcytosine (5mC)] [31] to the C5 of the cytosine ring and is catalyzed by DNA methyltransferases (Dnmts) [32, 33]. The conversion of 5mC to 5-hydroxymethylcytosine (5hmC) is carried out by the 10 to 11 translocation proteins, which also transform 5mC into 5-formylcytosine (5fC) and 5-carboxycytosine (5caC) through enzyme-dependent activities [34]. Cytosine methylation can inhibit gene expression by silencing the promoters [33]. Prior research has identified key biomarkers for colon cancer prognosis by measuring the levels of 5mC, 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxycytosine (5caC) in colon cancer tissues in comparison to adjacent nontumor tissues. The results implied that in patients with varying levels of 5mC (65.2 vs. 95.2 months, P = 0.014) with 5hmC (71.2 vs. 97.5 months, P = 0.045), the disease-specific survival of the moderate positive group was considerably shorter than that of the strong positive group [35] indicating that cytosine methylation deficiency is associated with poor prognosis. We observed that an overall rise in cytosine levels increases the risk of colorectal cancer, but we cannot determine the degree of methylation occurring in overall cytosine. Based on previous research, we speculate that although cytosine increases, the degree of cytosine methylation may not be high, so it is impossible to silence oncogenes, leading to an elevated risk of colorectal cancer. Thus, further experimental verification is needed.
Piperine is a bioactive phenolic compound with antioxidant, anti-inflammatory, and biologically enhanced activities [36–38]. Piperine also exerts chemopreventive effects and induces various effector proteins involved in apoptosis in cancer cells through multiple mechanisms, thereby exerting anticancer effects [39]. “Studies have shown that Piperine can serve as an enhancer of sorafenib (SOR) for its anti-proliferative and cytotoxic effects. By promoting the arrest of G2/M phase, triggering apoptosis, downregulating DNMT3B and HDAC3 expression, and upregulating tumor suppressor miRNA-29c, Piperine significantly boosts SOR’s efficacy against hepatocellular carcinoma [40]. Moreover, Piperine enhances doxorubicin’s effectiveness in triple-negative breast cancer by targeting the PI3K/Akt/mTOR pathway and cancer stem cells [41]. Additionally, Piperine inhibits cell growth and promotes apoptosis in human gastric cancer cells by downregulating the PI3K/Akt pathway [42]. In a study conducted by Jianyu Xia and his colleagues, it was discovered that piperine triggers autophagy-dependent cell death in CRC cells by increasing reactive oxygen species (ROS) production and inhibiting Akt/mTOR signaling [43]. This is consistent with our results, which show that Piperine can hinder the growth of colorectal cancer cells. As the levels of Piperine rise, the likelihood of developing colorectal cancer diminishes, potentially due to Piperine’s role in inducing autophagy-dependent cell death in CRC cells.

Strengths and limitations
The novelty of our study lies in the implementation of a MR approach, which allows for the simultaneous investigation of causal relationships between metabolites and colorectal cancer. Unlike traditional observational studies, which are susceptible to confounding and reverse causation, our MR design leverages genetic variants as instrumental variables to infer causality with greater robustness. Furthermore, compared to previous genetic studies that identified associations without establishing causality, our framework provides a more comprehensive understanding of the causal interplay between metabolites and colorectal cancer. This approach can elucidate potential feedback loops and clarify the directionality of effects, thereby offering novel insights into the underlying biological mechanisms and informing more targeted intervention strategies.
However, several limitations should be taken into account when interpreting these findings. The MR results need to be validated through animal experimental models, as population-level analyses have inherent constraints. we did not analyze the reverse MR in this study. Additionally, since our study population was predominantly of European ancestry, caution is advised when generalizing these results to other ethnic groups. Additionally, the focus narrowed on gene-metabolite pairs derived from contemporary expression data and biological insights, specifically on effector genes. Yet, one cannot discount the potential importance of other highly heritable metabolites or their ratios in the context of disease, inviting future inquiries to broaden this scope by integrating further expression data and metabolic revelations to uncover the role of additional metabolites and the ratios. Another challenge faced by this MR analysis is its reliance on the fact that most metabolites and metabolite ratios were tethered to a single instrumental variable (IV). While this study examined a comprehensive array of metabolites, the pathways and mechanisms of specific metabolites in the context of the disease remain shrouded in ambiguity, ultimately constraining the interpretative power of the MR finding.

Conclusions
This comprehensive MR analysis of metabolites has uncovered new gene-metabolite associations, offering an enhanced understanding of genetic regulation in human metabolism. These discoveries are anticipated to facilitate the identification of potential markers and targets that could aid in precision prevention and lead to behavioral modifications, innovative pharmaceutical interventions, and personalized treatments. Such strategies would be specifically designed to address the unique metabolic vulnerabilities associated with CRC.

Electronic Supplementary Material
Below is the link to the electronic supplementary material.