Glioblastoma Multiforme (GBM) is the most aggressive and common form of primary brain tumor, characterized by rapid growth and a generally poor prognosis[1, 2]. Recent research has increasingly focused on the metabolic alterations associated with GBM[3], particularly those observed in cerebrospinal fluid (CSF)[4, 5], which provides a unique window into the metabolic state of the brain.

CSF metabolites reflect various biochemical processes occurring within the brain, and changes in these metabolites can provide critical insights into tumor metabolism[4, 5]. Studies have identified specific metabolic signatures in the CSF of GBM patients, including alterations in glucose metabolism[6], amino acid[7] levels, and lipid metabolism[8]. For instance, elevated levels of certain amino acids in CSF, such as glutamate, have been linked to the enhanced proliferative capabilities of tumor cells and the suppression of immune responses[9]. The altered metabolic profiles in CSF can be attributed to several pathophysiological mechanisms, such as the Warburg effect[10] (aerobic glycolysis), which is prominent in many cancers including GBM[11, 12]. This metabolic reprogramming supports rapid tumor growth by optimizing energy production and biosynthetic processes. Additionally, disruptions in lipid metabolism in GBM patients may influence membrane synthesis, signaling pathways, and energy storage, further contributing to tumor pathogenesis[13]. Analyzing CSF metabolites offers potential diagnostic and prognostic value. Specific patterns of metabolites can help in distinguishing GBM from other types of tumors or neurological conditions. Furthermore, changes in the metabolite profiles over time could potentially monitor tumor progression or response to therapy[14, 15]. Understanding the metabolic interactions between GBM cells and the CSF environment opens avenues for targeted therapies. For example, interventions aimed at normalizing specific metabolic pathways could potentially hinder tumor growth or enhance the efficacy of existing treatments[16].

Two-sample Mendelian Randomization (MR) is a statistical method that uses genetic variants as instrumental variables (IVs) to estimate the causal effect of an exposure on an outcome using summary data from different genetic association studies[17]. This approach leverages the fact that genetic variants are randomly assigned at conception, thereby providing a natural control for confounding variables and avoiding reverse causation[18]. Traditional observational studies in the context of GBM and CSF metabolites can be significantly confounded by various factors like age, lifestyle, or underlying health conditions[19]. Two-sample MR significantly minimizes these confounders because the genetic variants used as instruments are not typically influenced by these factors[20]. Meanwhile, it can be challenging to determine whether changes in CSF metabolites are a cause or a consequence of GBM development. Two-sample MR helps in establishing a clearer direction of causality, assuming that the genetic instruments affect the outcome only through their impact on the exposure[20]. In addition, two-sample MR can harness the power of large-scale GWAS datasets that may already exist, thus allowing researchers to investigate causal relationships without the need for extensive new data collection[21].

The primary aim of this study is to investigate the CSF metabolites potentially associated with GBM through the application of two-sample Mendelian Randomization analysis. By identifying metabolites in the CSF that may influence the development or progression of GBM, the study could provide insights into new diagnostic markers or therapeutic targets.

In our research, we utilized a GWAS dataset from the FinnGen database (https://www.finngen.fi) that focused specifically on Glioblastoma Multiforme (GBM). This dataset was derived from an extensive genome-wide association study of a European cohort, which included 243 patients diagnosed with GBM and compared them with a control group of 287,137 individuals. The analysis rigorously evaluated around 16.38 million genetic variants, each subjected to stringent quality control protocols and enhanced through sophisticated imputation techniques.

We sourced GWAS summary data on cerebrospinal fluid (CSF) metabolites from the MRC Integrative Epidemiology Unit GWAS database (https://gwas.mrcieu.ac.uk/). This data was gathered from participants involved in the Wisconsin Alzheimer's Disease Research Center (WADRC) and the Wisconsin Registry for Alzheimer's Prevention (WRAP)[5]. During these studies, CSF samples were obtained through lumbar punctures and adhered to standardized collection and preservation methods detailed in earlier research. A total of 689 subjects participated, with 532 from WADRC and 157 from WRAP, each providing distinct CSF samples for the study. Inclusion criteria for WADRC participants included being over 45 years old, having the capacity to make decisions, and the capability to abstain from eating for 12 hours. Participants with conditions such as kidney issues, congestive heart failure, or significant neurological disorders were excluded. This research was incorporated into the Generations of WRAP (GROW) initiative and was sanctioned by the University of Wisconsin Health Sciences Institutional Review Board.

Our study drew upon a broad spectrum of GWAS data for 1,400 plasma metabolites, integrating various datasets[22]. We acquired the GWAS summary statistics from the EBI GWAS Catalog of European origin. Additionally, we accessed individual-level metabolite data from the Canadian Longitudinal Study on Aging (CLSA), available at the CLSA website, under strict guidelines that govern the access to anonymized CLSA data, complying with ethical and privacy standards.

The genetic IVs were selected based on their strong association with the plasma metabolite levels, as confirmed by GWAS. These IVs must also satisfy the MR assumptions of relevance (the IVs are associated with the exposure), independence (the IVs are not associated with any confounders that may influence both the exposure and the outcome), and exclusion restriction (the IVs influence the outcome solely through the exposure). To validate these IVs, we employed several statistical methods, including pleiotropy tests and heterogeneity assessments, to ensure no direct effects on exposures.

In our study, we applied Bayesian Weighted Mendelian Randomization (MR) analysis to verify the potential causal relationships between metabolites and GBM. This approach integrates a probabilistic framework, weighting each genetic instrument based on its reliability, and uses posterior probabilities to enhance the estimation of causal effects[23]. We used Bayesian weighted median methods, which are particularly effective when some instruments might be invalid or exhibit pleiotropy. Sensitivity analyses, including Bayesian MR-Egger regression, were conducted to ensure the robustness of our findings. This method allows us to provide more precise and credible estimates of the causal links. Statistical Analysis

The two-sample MR analysis was conducted using the inverse variance weighted (IVW) method as the primary analysis. This method combines the Wald ratio estimates (the ratio of the coefficient from the exposure GWAS to the coefficient from the outcome GWAS for each genetic variant) using a meta-analysis approach. We supplemented this with additional sensitivity analyses, including MR-Egger regression to test for directional pleiotropy and weighted median approaches, which provide valid estimates even if some of the instrumental variables are invalid.

Figure 4 illustrated the workflow of the present study, including conventional, reverse, and Bayesian MR analyses to assess causal relationships between CSF/plasma metabolites and GBM.

In this study, a two-sample MR analysis on a set of 292 CSF metabolites initially established a significant causal link between GBM and five specific metabolites, namely 3-methoxytyramine sulfate, caffeine, dimethyl sulfone, fructose, and phenol sulfate, which were rigorously validated using Bayesian methods. Interestingly, reverse causation analysis suggested a potential protective role for fructose levels in CSF against GBM, although these associations were not mirrored in related plasma metabolites, indicating a distinct metabolic interaction within the CSF milieu.

Metabolomic profiling has become an increasingly important aspect of cancer research, including GBM[15, 24]. In this study, we found GBM was positively correlated to 3-methoxytyramine sulfate, caffeine, dimethyl sulfone, and phenol sulfate levels. 3-Methoxytyramine is a metabolite of dopamine, and its sulfate form would be an example of phase II metabolism[25], potentially relevant in the context of GBM due to the brain's reliance on precise neurotransmitter signaling and metabolism. It seems that there’re no direct results regarding the specific relationship between caffeine and GBM or other tumors. However, caffeine, a central nervous system stimulant, has been extensively studied in oncology research due to its various biological effects that may influence cancer development and progression[26]. Epidemiological studies have investigated the consumption of caffeine and its correlation with the incidence of various types of cancer, including GBM[27, 28], although results have been mixed and sometimes contradictory. Similarly, it appears that specific research linking dimethyl sulfone directly with GBM or tumors was not readily available from current studies. Dimethyl sulfone, also known as methylsulfonylmethane (MSM)[29], is a sulfur-containing compound popularly used for its anti-inflammatory effects and is often explored in the context of joint health[30], oxidative stress[31], and immune function[32]. MSM’s anti-inflammatory effects might theoretically modulate the tumor microenvironment. By reducing inflammation, MSM could potentially affect tumor initiation, progression, or metastasis[32]. MSM is also recognized for its ability to reduce oxidative stress, which is a critical factor in the etiology and progression of various cancers[33]. By modulating oxidative stress, MSM might influence the stability of the tumor environment and the behavior of cancer cells. Phenol sulfate is a metabolite often associated with gut microbial activity[34], which has become a point of interest in cancer research due to the gut-brain axis and the potential systemic effects of gut microbiota on various diseases, including cancers[35]. Phenol sulfate is produced through the metabolism of phenolic compounds by gut bacteria. The dysregulation of microbial metabolism can lead to altered levels of such metabolites, which may influence systemic inflammation—a known risk factor for tumorigenesis[36]. As a phenolic compound, phenol sulfate could potentially contribute to oxidative stress. Elevated levels of oxidative stress have been linked to DNA damage and cancer progression, including in neural tissues[37]. When it comes to fructose, this common dietary sugar has been implicated in various studies concerning its effects on metabolic health and potential connections to cancer[38]. Fructose can be metabolized by cancer cells to fuel their rapid growth and proliferation[39]. It has been shown that certain cancer cells can utilize fructose to produce nucleotides and increase their proliferative capacity. This mechanism could potentially be relevant in the context of GBM, where high metabolic demands characterize the rapidly dividing tumor cells[40, 41]. On the other hand, chronic consumption of high levels of fructose can lead to inflammation[42] and oxidative stress[43]. Inflammatory microenvironments and oxidative damage to cellular components such as DNA may contribute to the mutagenesis and malignant transformation associated with GBM.

These metabolites may serve as potential biomarkers for early detection and monitoring, enhancing our understanding of GBM pathophysiology and providing new targets for therapeutic intervention. Furthermore, the use of CSF as a medium surpasses other bodily fluids in neuro-oncology due to its proximity to the tumor environment and its capacity to provide a clearer picture of the metabolic state of brain tumors without the confounding effects of the blood-brain barrier[44, 45]. While some studies have noted an increase in metabolite production as an indicator of malignant biological behaviors in tumors, incorporating metabolomics into tumor diagnostics has remained underexplored[46]. Our findings underscore the potential of CSF metabolites to serve as indispensable components of diagnostic and prognostic frameworks, offering insights into the metabolic underpinnings of tumor malignancy and progression. The significant causal relationship between fructose levels in CSF and GBM presents intriguing insights into both the etiology and progression of this malignant brain tumor. Fructose, a simple carbohydrate known primarily for its role in energy production and metabolism, has recently been implicated in various pathophysiological conditions including cancer[47]. In the brain, fructose can be metabolized by neurons and astrocytes and may influence cellular processes such as inflammation, oxidative stress, and energy homeostasis[48]. These processes are critical in the context of GBM, where altered metabolic environments can drive tumor growth and resistance to therapies[3]. The presence of elevated fructose levels in the CSF of patients with GBM might indicate an adaptive metabolic response of the tumor environment or a dysregulated metabolic pathway associated with the tumor itself. Fructose might be utilized by tumor cells to support their high energy demands under hypoxic conditions or to mitigate oxidative stress[39]. However, in such a scenario, fructose would likely promote, rather than inhibit, tumor progression. The observed protective effect suggests a more complex interaction. High levels of fructose in the CSF could disrupt signaling pathways that are normally exploited by GBM cells for proliferation and invasion. For example, fructose could interfere with growth factor signaling or cellular adhesion processes, indirectly reducing tumor aggressiveness[49]. Understanding these interactions requires comprehensive metabolic profiling of GBM tissues and correlational studies involving CSF fructose levels and clinical outcomes. Further experimental studies using in vitro and in vivo models of GBM could elucidate the specific metabolic pathways through which fructose impacts tumor biology. Additionally, advanced imaging and metabolomic technologies could help visualize and quantify the dynamic changes in fructose metabolism within the tumor environment.

The study's primary limitations include potential concerns over the sample size and generalizability, the specificity of fructose among other common metabolites, inherent assumptions of Mendelian randomization, uncertainties in the temporal relationship between fructose levels and GBM development, and a lack of detailed insights into the biological mechanisms underlying this association. Despite these constraints, the research holds considerable significance. It introduces an innovative genetic approach to investigating CSF metabolic influences on GBM, offers potential diagnostic and prognostic enhancements, and opens new avenues for therapeutic interventions focusing on dietary and metabolic modifications.

In conclusion, this research demonstrates the role of specific CSF metabolites in influencing GBM, thereby providing a novel insight into the metabolic interactions associated with this malignancy. As such, these findings not only enhance our understanding of GBM pathophysiology but also open avenues for the development of innovative diagnostic and treatment modalities that could potentially impact patient management and outcomes.