Genetic susceptibility and lifestyle modify the association of long-term air pollution exposure on major depressive disorder: a prospective study in UK Biobank

Major depressive disorder (MDD) is among the most common mental illnesses, and it severely limits psychosocial functioning and negatively affects the quality of life [1]. MDD affects approximately 6% of the adult population worldwide each year [2], and patients with MDD are nearly 20-fold more likely to die by suicide than individuals without MDD [3]. According to the World Health Organization, MDD will be the leading disease burden worldwide by 2030 [4].

Evidence indicates that socioeconomic status, medical conditions, and family history play a major role in the development of mental health disorders and that environmental factors may also influence the development of such disorders through neuroinflammatory pathways and oxidative stress [5, 6]. Although air pollution is the most common environmental risk to human health, results on the correlation between air pollution and health are sparse and inconsistent. For example, a recent systematic review reported that long-term exposure to air pollution was associated with an increased risk of depression, but the association was not significant in more than half of the studies included in the review [7]. Furthermore, the size and quality of these studies varied considerably. Therefore, additional large population-based cohort studies are necessary to test the potential association between long-term exposure to air pollution and MDD.

Evidence suggests that genetic factors play a critical role in the development of MDD [8, 9]. A genome-wide association study (GWAS) identified some genetic variants associated with MDD risk [10]. Analyzing the cumulative genetic burden of these genetic variants by using polygenic risk scores (PRSs) could provide quantitative measures of genetic susceptibility and could help effectively identify individuals at high risk of MDD [11]. Recent studies have suggested that genetic susceptibility may influence the environment–disease relationship [12, 13]. However, the influence of genetic susceptibility on the association between exposure to air pollution and MDD risk is unclear. In addition, unhealthy lifestyle behaviors, such as smoking, excessive alcohol intake, and lack of physical activity, increase the risk of MDD [14‐16]. Several studies have revealed that healthy lifestyle habits played a pivotal role in attenuating the effect of air pollution on the risk of various diseases [17‐19]. However, whether a healthy lifestyle can mitigate the effect of air pollution on MDD risk is unclear.

To address the aforementioned questions, this study investigated the association between exposure to air pollution and the incidence of MDD in a large population-based cohort (UK Biobank). In addition, we examined the potential modifying effect of genetic susceptibility and lifestyle on this association.

During a total of 3,427,084 person-years (median follow-up 9.7 years), 14,710 incident MDD events were recorded. Table 1 displays the baseline characteristics of the study participants. Participants who developed MDD were more likely to have the least healthy lifestyle, high genetic risk, and higher air pollution exposure. The Pearson correlation coefficients and the summary statistics of air pollutants are shown in Additional file 1: Fig. S8 and Additional file 1: Table S4, respectively.

Table 2 presents the relations of each air pollutant with MDD risk. In the finally multivariate-adjusted model, PM_2.5 (HR: 1.16, 95% CI: 1.07–1.26; per 5 μg/m³) and NO_x (HR: 1.02, 95% CI: 1.01–1.05; per 20 μg/m³) were significantly associated with MDD. However, we did not observe an association between PM₁₀ (HR: 1.00, 95% CI: 0.92–1.09; per 10 μg/m³) and NO₂ (HR: 1.00, 95% CI: 0.98–1.02; per 10 μg/m³) with MDD. Moreover, compared with the lowest quartile, the HRs (95% CIs) of MDD were 1.12 (1.06, 1.17) and 1.07 (1.02, 1.13) for subjects with the highest quartile of exposure to PM_2.5 and NO_x, respectively. These results did not alter appreciably in the sensitivity analyses (Additional file 1: Table S5-S12). We then examined the association between principal components (PCs) and MDD risk (Additional file 1: Table S13). We found that the first PC (PC1) predominated by PM_2.5, NO₂, and NO_x was associated with an increased risk of MDD (HR = 1.02, 95% CI: 1.01, 1.04). However, no significant association was observed for PC2 which predominated by PM₁₀ (HR = 1.00, 95% CI: 0.98, 1.01). We also used the restricted cubic spline to assess the potential dose-response relationship of air pollutants with MDD in Fig. 1. We found nonlinear relationships between PM_2.5, NO₂, and NO_x with the risk of incident MDD (P-nonlinearity for PM_2.5: 0.006; NO₂: 0.003; NO_x: 0.003). Additional file 1: Table S14 shows results about the association between air pollution and MDD stratified by age and sex group. No significant interaction was found between air pollution and age on the risk of incident MDD, whereas significant interactions were observed between PM_2.5 with gender on the risk of incident MDD (P-interactions were 0.049 for PM_2.5). The association of air pollution exposure with MDD risk was stronger in men than in women (Additional file 1: Table S14).

Additional file 1: Table S15 presents the associations of PRS with the risk of MDD. We found a significant association of PRS with MDD in the multivariable-adjusted model (HR: 1.10, 95% CI: 1.07–1.12). Figure 2 shows the risk of incident MDD for combined air pollutants and genetic risk. Compared with participants with low genetic risk and low air pollution exposure, those with high genetic risk and high PM_2.5 exposures had the highest risk of incident MDD (PM_2.5: HR: 1.34, 95% CI:1.23–1.46). The interaction effects between PRS with PM_2.5, PM₁₀, and NO₂ on the risk of incident MDD were significant (P-interaction for PM_2.5: 0.036; PM₁₀: 0.025; NO₂: 0.030; NO_x: 0.080).

The risks of incident MDD according to lifestyle categories and lifestyle score are provided in Additional file 1: Table S16-S17. MDD risk increased monotonically across lifestyle categories and scores. The risk of incident MDD was 98% higher among those who have the least healthy lifestyle compared with those who have the most healthy lifestyle (HR: 1.98, 95% CI: 1.81–2.16). Figure 3 shows the risk of incident MDD for combined air pollutants and lifestyle. The highest MDD risk was observed in participants with the least healthy lifestyle and high PM_2.5, PM₁₀, NO₂, and NO_x exposures (PM_2.5: HR: 2.22, 95% CI:1.92–2.58; PM₁₀: HR: 2.09, 95% CI:1.78–2.45; NO₂: HR: 2.11, 95% CI: 1.82–2.46; NO_x: HR: 2.28, 95% CI: 1.97–2.64). The interaction effects between lifestyle with PM_2.5 in relation to incident MDD risk were significant (P-interaction for PM_2.5: 0.026; PM₁₀: 0.054; NO₂: 0.410; NO_x: 0.271). The potential effect modifications of individual lifestyle factors on MDD risk are presented in Additional file 1: Table S18. We observed a significant interaction between PM_2.5 and processed meat intake to MDD risk (P-interaction = 0.022).

In this large-scaled population-based longitudinal study, we found that long-term exposures to PM_2.5 and NO_x were significantly associated with the incident MDD risk. The risk of incident MDD was higher among those with high genetic risk or unhealthy lifestyle compared with those with low genetic risk or healthy lifestyle. We also found significant interactions of genetic susceptibility and lifestyles with exposure to air pollution. Increased genetic susceptibility and unhealthy lifestyles may intensify the impact of long-term exposure to air pollution on the risk of MDD.

Research on the role of air pollution exposure in the development of MDD is limited. We comprehensively examined the association between long-term exposure to various air pollutants and the risk of incident MDD. We used air pollution data from 2010 to capture long-term air pollution exposure, consistent with previous research [12, 36‐38]. Air pollution levels in England have been relatively stable over these years [39, 40]. As the fluctuation of the temporal trend of most air pollution was generally stable during the study period, the average values of air pollution could be used as a surrogate measure of long-term exposure [41]. The use of the average air pollution levels as the mean estimates in our sensitivity analysis further demonstrated the robustness of this approach. Our findings are consistent with those of previous studies. For example, in a prospective study of 27,270 participants in Korea, the researchers observed a 44% increase in the risk of MDD for each 10-μg/m³ increase in PM_2.5 levels [42]. Similarly, a nationwide prospective cohort study in the USA found an association between exposure to PM_2.5 and the onset of depression [32]. Meanwhile, there are some other studies that are not consistent with our findings. For example, Zhang et al. discovered that long-term exposure to outdoor PM₁₀ was associated with the development of depression; nevertheless, they did not find this association for PM_2.5 [43]. Notably, previous studies lacked large-scale cohort settings and primarily examined particulate matter exposure (e.g., PM_2.5), excluding other gaseous pollutants (e.g., NO_x), which led to inconsistent results regarding the role of long-term exposure to air pollution in the development of MDD. Our study provides new evidence for epidemiological studies on the association between air pollutants and MDD. Our findings may have implications for policy regulation and clinical trials because changes in policies and individual behavior may help reduce air pollution and thus help mitigate symptoms of MDD.

There are some underlying mechanisms that could shed light on the associations of air pollution with MDD. One of the underlying mechanisms is that oxidative stress and neuroinflammation pathways induced by air pollutants could stimulate the onset and progression of MDD [44]. Air pollutants can penetrate the lung tissue compartments, enter the circulatory system, and reach the brain, causing oxidative stress and inflammation of the central nervous system [45]. In addition, experimental and animal studies have observed that inhalation of particulate matter could stimulate increased expression of redox/glucocorticoid-sensitive genes in rats, suggesting the involvement of the hormonal pathway in mental health disorders associated with particulate matter [46]. However, determining which pathway offers the most critical link is difficult because of the existing scarcity of particle-specific translocation kinetics and exposure levels [47]. In addition, vascular disease is an essential intermediate factor in the association between air pollution exposure and an increased risk of subsequent MDD. Increasing bodies of evidence demonstrate that exposure to air pollution leads to cerebrovascular disease, which may affect the central nervous system and the brain, contributing to an increased risk of depression and other related conditions [48]. Previous studies have also indicated that vascular disease is associated with inflammatory pathway activation, leading to MDD or dysthymia [49]. Additional studies are necessary to determine the precise mechanisms underlying the air pollution–induced pathogenesis of MDD.

Previous research demonstrated that the etiology of MDD is multifactorial and that its heritability is approximately 35% [9]. Research also demonstrated that PRS may serve as an early indicator of clinically significant levels of depression and be associated with the risk of depression [50]. Our results are consistent with these findings. Additionally, we investigated the contribution of genetic susceptibility to the association between air pollution and MDD and found that air pollution may increase the risk of MDD, particularly among individuals with high genetic susceptibility. Li et al. explored how PM_2.5 exposure interacts with polygenic risk in the development of MDD across multiple levels of brain network function [51]. They observed that a combination of exposure to high levels of air pollution and a relatively high polygenic risk for MDD disproportionately augmented stress-related effects on the brain circuitry. Working memory and stress-related information transfer across cortical and subcortical brain networks were influenced by PM_2.5 exposures to differing extents depending on the polygenic risk for MDD in gene-by-environment interactions [51]. However, other explanations for these mechanisms can be applied when they are separated into particular variants. Previous studies have revealed that patients with psychiatric disorders had a higher mRNA expression level of vaccinia-related kinase 2 (VRK2) than did healthy individuals [52, 53]. In addition, a randomized crossover study suggested that higher PM_2.5 exposure was positively associated with the mRNA expression of cytokine [54]. Therefore, air pollution may interact with rs1518395 located in VRK2 to jointly affect the onset of MDD. In addition, some SNPs from an MDD GWAS, such as rs10514299, could be enriched in genes expressed in the central nervous system and function in transcriptional regulation related to neurodevelopment [10]. Because of their toxicity to the central nervous system, air pollutants may also contribute to the development of mental diseases [55]. Therefore, by affecting the central nervous system, SNPs and air pollution may contribute to the onset of depression. Accordingly, elucidating the pathophysiology of MDD is imperative.

In addition, we also confirmed that unhealthy lifestyles were associated with higher risks of MDD. Considering the complexity of health behaviors and that most health behaviors are interconnected, a comprehensive analysis of healthy lifestyles may better capture the impact of lifestyle than an analysis based on a single factor. Our findings are in concert with the previous studies. Adjibade and colleagues formulated a healthy lifestyle index that incorporates multiple lifestyle factors and discovered that combined healthy lifestyles were associated with a lower risk of depressive symptoms [56]. We also observed that the deleterious associations between PM_2.5 and MDD were stronger among individuals who led unhealthy lifestyles. Indeed, besides long-term air pollution exposure may reach the brain through the lung–brain axis and induce systemic inflammation [57], unhealthy lifestyle factors have also been associated with elevated inflammation levels [58, 59]. Conversely, higher levels of systemic inflammation marker may contribute to the development of different neuropsychiatric disorders including depression [60]. Therefore, when air pollution and unhealthy lifestyle are employed together for MDD, it is reasonable to appear enhanced effect. These findings emphasize the importance of lifestyle changes. The benefit of air pollution exposure reduction in lowering the risk of MDD is expected to be greatest among individuals with healthy lifestyles; this finding can inform the establishment of personalized preventive strategies for reducing the risk of MDD.

To the best of our knowledge, our study is the first to evaluate the modifying effect of genetic susceptibility and lifestyles on the association between air pollution exposure and the risk of MDD. The main strengths of our study are its inclusion of a large sample size, prospective design, and consistent results in several sensitivity analyses. Nevertheless, we also acknowledge several limitations of our study. First, an exposure assessment based on a single address does not eliminate the possibility of exposure misclassification caused by outside activities. Further studies with more accurate estimates are needed to confirm the present findings. In addition, we had to admit that the effect of the collinearity cannot be ruled out, single-pollutant associations may be not independent, and the results should be interpreted with caution. Second, common to most previous environmental epidemiology studies [12, 36‐38], we used the annual average air pollution concentration in 2010 as a proxy for the long-term air pollution exposure, which might induce the exposure misclassification. However, previous studies have suggested that the fluctuation of the temporal trend of most air pollution was generally stable during the study period in UK Biobank [12, 61, 62]. Furthermore, of note, similar results were found when a time-varying air pollution exposure was used in the sensitivity analysis, supporting the validity of using the baseline concentration. In addition, the UK Biobank lacks data on air pollution composition; therefore, there is still uncertainty as to which components are the most harmful. Third, incident MDD cases are not always well captured through hospital inpatient records and death registries. Although diagnosis by a doctor is a more common and precise way, some mildly depressed people may do not go to the hospital, resulting in MDD cases that were likely to be underreported. Fourth, a sample of 500,000 was recruited in UK Biobank with remarkable speed and efficiency, but this efficiency was achieved at the expense of a response rate (5.5%) and was subject to selection bias. Nevertheless, the absolute difference in these estimates was low and lead to the practical importance of such risk underestimation is likely to be small [63]. Fifth, additional MDD-related variants may be identified in future GWAS, the inclusion of additional SNPs in further study may help to further refine the estimation of genetic risk. Sixth, although we have adjusted for a series of potential confounders in our analysis, potential residual confounding from unmeasured or unknown variables might still be present. Finally, because the majority of our study’s participants were of European descent, the generalization of our findings regarding the associations of air pollution exposure and genetic susceptibility with MDD to other populations should be interpreted with caution.

In summary, based on this large prospective cohort study, we found that long-term exposure to ambient air pollution was associated with a higher risk of MDD. High genetic risk and unhealthy lifestyle may intensify the impact of air pollution on MDD risk, highlighting the importance of identifying individuals with high genetic risk and developing healthy lifestyles for reducing the harm of air pollution to public mental health.