Introduction

Depression is a prevalent mental disorder characterized by persistent low mood, reduced interest, and diminished pleasure, which can disrupt eating and sleeping patterns and, in severe cases, lead to suicide. The global incidence of depression has increased by almost 50% over the past 30 years, affecting more than 264 million people worldwide1. Of particular concern is the recurrent nature of depression2. Tryptophan metabolism, particularly involving serotonin (5-HT) and kynurenine (KYN), plays a crucial role in the development of depression. It is responsible for synthesizing essential neuroactive compounds, 5-HT, and KYN3,4,5,6. Antidepressants that target the reuptake of 5-HT in the synaptic cleft have proven effective and are among the most widely used medications worldwide for treating depression, highlighting the significance of 5-HT in depressive disorders4. In contrast, research on KYN is relatively limited. As investigations progress, the metabolic connections between tryptophan, 5-HT, and KYN are becoming increasingly intricate in the context of depression. Despite the extensive research on the relationship between KYN and depression, conclusions are inconsistent. Some studies indicate lower levels of KYN in individuals with depression compared to control groups, while others report no association between depression and KYN levels7,8. Additionally, certain studies suggest elevated KYN levels in individuals with depression9.

Mendelian randomization (MR) is a method that utilizes data from genome-wide association studies (GWAS) and leverages genetic variants, specifically single nucleotide polymorphisms (SNPs), strongly associated with an exposure (such as KYN) as instrumental variables (IVs) to determine the causal relationship between the exposure and the outcome of interest (such as depression). By relying on the random distribution of genetic variants, MR is less prone to confounding factors and reverse causality, reducing bias and resembling randomized controlled trials. Hence, the objective of this study is to elucidate the causal relationship between KYN and depression through a stringent two-sample MR analysis.

Method

Data source

The GWAS data for depression and KYN: Both are derived from the European population’s GWAS from IEU Open GWAS (https://gwas.mrcieu.ac.uk/), with GWAS IDs for depression and KYN being ieu-b-102 and met-a-375, respectively. The GWAS data for depression originate from a study involving 170,756 depression patients and 329,443 controls10. The GWAS data for KYN originate from a human blood metabolite study with a sample size of 781611. Refer to Table 1 for specific details.

Table 1 Sample sizes, age groups and the proportion of males and females of the study cohorts.
Full size table

This study reanalyzes public data and does not require ethical approval.

Method

Selection of instrumental variables

To perform two-sample MR analysis and investigate the causal relationship between KYN and depression, IVs need to meet certain assumptions: (1) The SNPs were significantly correlated with the exposure; (2) SNPs can only influence the outcome through exposure. The following steps were taken to select suitable IVs:

Single nucleotide SNPs associated with exposure: SNPs were selected from the GWAS data for KYN based on their association with KYN at a significance threshold of P < 5 × 10–6 (If SNPs are filtered based on a threshold of P < 5 × 10–8, after the exclusion of weak instruments and the removal of confounding steps, we only retained 2 SNPs, making it unfeasible to conduct MR analysis). SNPs exhibiting genomic linkage imbalance (distance window < 10,000 kb, linkage disequilibrium coefficient r2 < 0.001) were removed. The remaining SNPs were assessed for their F-statistic, and SNPs with F < 10, indicating weak instrumental variables, were excluded. Only SNPs with F > 10, demonstrating a significant association with the exposure, were retained12. The F-statistic was calculated using the formula F = (N−K−1) * R2/(1−R2), R2 = β2 (1—EAF) * 2EAF / SD2, where R2 represents the proportion of variation explained by each SNP, K is the number of SNPs, N is the total sample size, EAF is the allele frequency of the mutation, β is the beta coefficient related to the exposure factor, and SD is the variance13.

SNPs independent of outcome: SNPs associated with the outcome (depression) at a significance threshold of P < 5 × 10–8 were removed from the IVs. Palindromic variants with identical base pairs were also excluded to avoid potential double-counting of their variation. The remaining SNPs were verified against Phenoscanner (http://www.phenoscanner.medschl.cam.ac.uk/) to identify and exclude those influenced by confounding factors. The remaining SNPs meeting the MR assumptions were used as instrumental variables.

MR analysis

MR analysis was performed using five methods: Weighted median (WM), MR-Egger, inverse variance weighted (IVW), simple mode, and weighted mode. The IVW method served as the primary analysis. A significant result from the IVW analysis (P < 0.05) along with consistent effect directions (positive or negative) across the other four methods supports the inference of a causal relationship. This means that the results of these five algorithms are consistent, supporting the remarkable results of IVW. The IVW algorithm is a commonly used two-sample MR analysis method, which can make full use of the sample information of each study, thereby improving statistical power. When using other algorithms, some preprocessing of the data may be required, such as removing outliers, performing data transformation, etc., and these preprocessing may affect the results. The IVW algorithm requires less data and does not require special preprocessing, so the results may be more accurate. In addition, if the sample size is small or the sample distribution is uneven, the results of other algorithms may be more affected, and the IVW algorithm can maintain high statistical power in this case. When all SNPs adhere to the assumptions of MR, only under such conditions can the inverse variance weighted (IVW) method yield accurate estimates of causal relationships. The WM method leverages a majority of effective instrumental variables (IVs) for causal inference. MR-Egger, on the other hand, allows the inclusion of IVs that exhibit pleiotropy. The intercept is used to gauge horizontal pleiotropy, while the slope serves to ascertain causal relationships. A P-value less than 0.05 signifies that the results of the MR analysis carry significant statistical importance. The “TwoSampleMR” package in R 4.1.2 was utilized for all data processing.

Sensitivity analysis

Sensitivity tests were carried out to gauge the robustness of the findings. Cochran’s Q test was implemented to assess the heterogeneity among SNPs, with a P-value of less than 0.05 indicating significant heterogeneity. Techniques such as the MR-Egger intercept test and MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) were deployed to investigate horizontal pleiotropy. If the intercept in the MR-Egger intercept test is statistically significant (P < 0.05), it implies a significant presence of pleiotropy. If the P-value of the MR-PRESSO result falls below 0.05, it suggests that the MR analysis presents considerable horizontal pleiotropy, which means that the IVs can directly influence the outcome without acting through the exposure factor. A leave-one-out analysis was conducted, which involved the sequential exclusion of each SNP to examine the impact on the combined effect of the remaining SNPs. Any notable alterations in the results due to the exclusion of a particular SNP suggests that the SNP has a substantial influence on the result, which infringes on the assumption of MR. The final results were presented using scatter plots, forest plots, and funnel plots.

Institutional review board statement

This study was based on a published database and did not require ethical approval.

Results

The MR analysis examined the causal relationship between KYN as the exposure and depression as the outcome. After applying the selection criteria, a total of 17 SNPs were identified as IVs that met the MR assumptions. Eight SNPs with F < 10 were excluded from the analysis, namely, rs3809198 (F = 0.83), rs16924894 (F = 1.46), rs16974854 (F = 1.74), rs11593042 (F = 3.50), rs511797 (F = 4.70), rs3789978 (F = 6.28), rs6575634 (F = 9.63), and rs2651516 (F = 9.97). No SNPs significantly associated with depression (P < 5 × 10–8) were found among the KYN SNPs extracted from the depression GWAS data. Palindromic variants were not detected. However, the SNP rs3184504 was identified as confounded by body mass index, which has been causally linked to depression in previous MR studies. Hence, this SNP was removed from the analysis14. Detailed results are presented in Table 2.

Table 2 Information on the SNPs for the final screened KYN (n = 17).
Full size table

The IVW method, serving as the primary analysis, indicated that each standard deviation increase in KYN was associated with an approximately 1.4-fold increase in the risk of depression (OR = 1.351, 95% CI 1.110–1.645, P = 0.003). The effect directions from the other four methods were consistent with IVW. Further details can be found in Table 3.

Table 3 Results of the five MR algorithms.
Full size table

Sensitivity analysis demonstrated no heterogeneity among the SNPs, as indicated by Cochran’s Q test (IVW: Q = 12.982, P = 0.674; MR-Egger: Q = 12.905, P = 0.610). The MR-Egger intercept test showed no evidence of horizontal pleiotropy (Intercept = − 0.001, P = 0.785). MR-PRESSO analysis did not identify any outliers (P = 0.719). The scatter plot illustrated the stability of the SNPs strongly associated with both KYN and depression (Fig. 1). The funnel plot displayed a symmetrical distribution of the SNPs, indicating no pleiotropy in the MR analysis (Fig. 2). Leave-one-out analysis demonstrated that the exclusion of individual SNPs did not significantly impact the overall results (Fig. 3).

Figure 1
figure 1

Scatter plot (n = 17).

Full size image
Figure 2
figure 2

Funnel plot (n = 17).

Full size image
Figure 3
figure 3

Forest plot of the leave-one-out analysis (n = 17).

Full size image

Discussion

Emerging evidence points to the involvement of KYN, a metabolite of tryptophan, in the development of depression, with disruptions in the kynurenine pathway playing a significant role in the etiology of depression15,16. However, the relationship between KYN and depression remains unclear. This study endeavored to explore the causal relationship between KYN and depression using a two-sample MR approach. The research results provide robust evidence, indicating that an elevation in KYN levels significantly increases the risk of developing depression.

The chronic inflammation hypothesis of depression has garnered substantial literature support, positing depression as a neuroimmune disorder where the activation of the immune system plays a pivotal role in the onset and progression of depression17,18,19. Neuroinflammation is implicated in the regulation of the tryptophan-kynurenine (TRP-KYN) pathway, particularly modulated by various pro-inflammatory cytokines such as TNF-α, IL-6, INF-γ20,21. Previous observational studies have reported associations between abnormalities in tryptophan-kynurenine metabolism and depressive disorders, particularly in scenarios involving inflammation. These studies align with our findings, which demonstrate an increase in KYN levels in individuals with depression9,11,22,23,24,25. Some scholars have proposed a novel subtype of depression termed “immune-metabolic depression”, wherein alterations in inflammation, metabolism, and bioenergetic pathways are observed in the majority of individuals with depression26. Regrettably, given the current unavailability of GWAS data for immune-metabolic depression, we are unable to elucidate the causal relationship between KYN and immune-metabolic depression. If GWAS data for immune-metabolic depression and non-immune-metabolic depression become accessible in the future, a renewed MR analysis could be conducted to investigate the causal relationship between KYN and these two subtypes of depression. KYN may lean towards having a causal relationship with immune-metabolic depression. Future research might explore common genetic features between KYN and immune-metabolic depression in this direction.

Potential mechanisms underlying the link between KYN and depression may involve the induction of depressive-like behavior through increased expression of Nod-like receptor protein 2 (NLPR2) in astrocytes by KYN, mediated by proinflammatory cytokines4. Cytokines such as IL-1β, IL-6, and TNF-α can also inhibit the conversion of KYN to 5-HT by activating the kynurenine pathway, resulting in a decrease in 5-HT synthesis and the manifestation of depressive symptoms27.

Several limitations should be considered in interpreting the findings of this study. First, the data used in this analysis were derived from individuals of European ancestry, potentially limiting the generalizability of the results to other populations. Second, while various sensitivity analyses were conducted to assess the assumptions of the MR study, it is challenging to completely rule out the possibility of pleiotropy influencing the instrumental variables. Last, the sample size of the available GWAS datasets was limited, emphasizing the need for further research using larger and more diverse datasets.

Conclusions

In this study, we utilized a two-sample MR approach to investigate the causal relationship between KYN and depression. The findings provide compelling genetic evidence supporting a causal link between elevated levels of KYN and an increased risk of depression. This provides genetic evidence for understanding the etiological mechanisms of depression and potential interventions. Specifically, it prompts further exploration into whether the dysregulation of the KYN metabolic pathway is a pathogenic mechanism for depression, whether KYN can serve as a potential therapeutic target, or whether it can be used as a diagnostic or screening marker for depression. However, these findings warrant further investigation and validation.