Abstract
Natto, known for its high vitamin K content, has been demonstrated to suppress atherosclerosis in large-scale clinical trials through a yet-unknown mechanism. In this study, we used a previously reported mouse model, transplanting the bone marrow of mice expressing infra-red fluorescent protein (iRFP) into LDLR-deficient mice, allowing unique and non-invasive observation of foam cells expressing iRFP in atherosclerotic lesions. Using 3 natto strains, we meticulously examined the effects of varying vitamin K levels on atherosclerosis in these mice. Notably, high vitamin K natto significantly reduced aortic staining and iRFP fluorescence, indicative of decreased atherosclerosis. Furthermore, mice administered natto showed changes in gut microbiota, including an increase in natto bacteria within the cecum, and a significant reduction in serum CCL2 expression. In experiments with LPS-stimulated macrophages, adding natto decreased CCL2 expression and increased anti-inflammatory cytokine IL-10 expression. This suggests that natto inhibits atherosclerosis through suppression of intestinal inflammation and reduced CCL2 expression in macrophages.
Introduction
Atherosclerosis is a chronic, multifactorial disease characterized by the accumulation of lipids, inflammatory cells, and extracellular matrix components within the arterial wall1. The resulting atherosclerotic plaques can progress over time, eventually leading to cardiovascular events such as myocardial infarction and stroke2. Despite advances in atherosclerosis prevention and treatment, it remains the leading cause of death worldwide3 that also induces dysfunction within the cardiovascular system, jeopardizes overall health, significantly diminishes the quality of life, and increases the cost of healthcare, making it a crucial, global public health issue.
A key aspect of atherosclerosis research involves the localization of macrophages in atherogenic lesions, making them valuable markers for in vivo imaging4,5. Exploiting this phenomenon was detailed in our recent publication, which highlighted the utility of near-infrared fluorescent protein (iRFP) to identify potential drugs or foods capable of reducing atherosclerotic lesions. This non-invasive imaging approach does not require any injections in mice, making it an attractive tool for evaluating therapeutic interventions5,6.
Natto, a traditional Japanese food made from fermented soybeans, is a rich source of vitamin K2 and has been repeatedly shown to benefit the cardiovascular system7. Epidemiological studies have suggested a possible inverse association between vitamin K2 intake and cardiovascular disease risk8 by specifically inhibiting arterial calcification, enhancing arterial elasticity, and modulating inflammation9. However, the underlying mechanisms remain unclear and additional preclinical and clinical studies are needed to evaluate the potential benefits of vitamin K2 against atherosclerosis. However, patients on anticoagulation therapy are given warfarin, which inhibits vitamin K, thereby necessitating a limited intake of this vitamin. This creates a dilemma, as natto, a food beneficial against atherosclerosis and rich in vitamin K2, should otherwise be an ideal dietary choice.
Natto additionally benefits the gut microbiota, an emergent factor in the development and progression of atherosclerosis10. Recent studies have shown that changes in gut microbiota composition and diversity can impact immune response, inflammation, and lipid metabolism, all of which are implicated in atherosclerosis11. Furthermore, gut microbiota-derived metabolites, such as trimethylamine N-oxide (TMAO), have been shown as drivers of atherosclerotic pathogenesis12. Additionally, chemokines, such as CCL2/MCP1 (a key regulator of macrophage recruitment in atherosclerotic plaques), are modulated by gut microbiota composition and metabolites10. Thus, understanding the complex interplay between gut microbiota and atherosclerosis may lead to new therapeutic strategies.
The current study evaluated the impact of natto consumption on atherosclerotic progression using an in vivo murine imaging model and strains of natto with varying vitamin K2 levels developed originally to accommodate patients on anticoagulation therapy. These strains facilitated the creation of three types of natto: high vitamin K natto (HVK), normal natto (NN), and low vitamin K natto (LVK). The influence of each type on atherosclerosis was then assessed using iRFP in a non-invasive imaging method6.
Findings suggest that natto intake therapeutically affects atherosclerosis by modulating gut microbiota composition and regulating the expression of pro-atherosclerotic cytokines and chemokines, such as CCL2. Macrophage gene expression analysis indicated that vitamin K2, surfactin, and the bacteria themselves play key roles in these effects. Notably, employing iRFP-expressing hematopoietic cells enabled direct visualization of natto’s heightened efficacy in treating atherosclerosis.
Results
Quantitative analysis of natto variants
Initially, three distinct natto variants were developed, distinguished by their vitamin K2 concentrations. These variants were created using different strains of bacteria involved natto fermentation and were comprised of high vitamin K2 natto (HVK), normal natto (NN), and low vitamin K2 natto (LVK). Each was tested against high-cholesterol diet (HCD) controls for impact on atherosclerotic pathogenesis with regard to modulation of macrophage activation and lesion size.
Our results showed no significant differences among the three natto types in terms of water, protein, fat, fiber, ash, and soluble non-nitrogenous substances. However, striking disparities were apparent in the vitamin K2 content: HVK contained the highest quantity with 199 μg/100 g, followed by NN with 93 μg/100 g, and LVK with 30 μg/100 g. No vitamin K2 was detected in the HCD (Supplementary Fig. 1A). Furthermore, nattokinase activity was highest in HVK at 110 FU/g, compared to NN at 82 FU/g, and LVK at 38 FU/g.
Additionally, the bacterial count was greatest in HVK natto at 19,500 (×106 cfu/g), followed by NN natto at 13,700 (×106 cfu/g) and LVK natto at 638 (×106 cfu/g). In terms of polyglutamic acid (PGA) content, HVK also showed the highest quantity at 12.2 mg/g, with NN at 9.3 mg/g and LVK at 4.9 mg/g (Supplementary Fig. 1B).
Our findings showed that HVK natto contained the highest amount of vitamin K2, exhibited the greatest nattokinase activity, had the largest bacteria count, and had the highest PGA content. These parameters, except for PGA, have previously been shown to influence atherosclerosis and inflammation suppression13. Conversely, LVK natto was found to have the lowest values for each surveyed attribute. To ascertain the influence these parameters could have on atherosclerotic development, each natto type was tested against a control diet in a murine atherosclerosis model. Between these analyses, we compared the dietary intake and body weight measurements of the different feeding groups (HCD, HCD + HVK, HCD + LVK, HCD + NN), taken weekly (Supplementary Fig. 2A, B). The results did not reveal any statistically significant differences among the groups, which suggests that neither natto in the diet nor the differences in vitamin K2 content had a discernible impact on body weight or food consumption.
Effect of natto types on atherosclerotic formation in a murine model
Next, we utilized bone marrow-transplanted mice expressing iRFP (iRFP → LDLR−/−) generated using a previously described method for imaging atherosclerosis6. In these mice, iRFP is expressed in blood cells, and the fluorescence of atherosclerotic lesions has been previously reported to precisely indicate the size of lesions6. We induced atherosclerosis in these mice by a high-cholesterol diet (HCD) for nine weeks, during which we performed weekly in vivo imaging to observe iRFP signal. At the end of the ninth week, we harvested aortas after sacrifice and performed Oil red O staining to visualize arterial lipid accumulation (Supplementary Fig. 2C).
Evaluation of thoracic aortas by Oil red O staining revealed a significant reduction in atherosclerotic lesions in all natto groups (HVK, LVK, NN) compared to the HCD group (p < 0.05) (Fig. 1A, B), with a particularly pronounced decrease in the HCD + HVK group.
Furthermore, weekly IVIS imaging of the thoracic region (Fig. 1C, Supplementary Fig. 2C, D) revealed an increase in iRFP fluorescence over time in all groups. Figure 1D illustrates the weekly changes in thoracic iRFP signals, comparing HCD vs HCD + HVK, HCD vs HCD + NN, and HCD vs HCD + LVK. To ascertain whether there are differences in these results, it was determined to analyze them using Bayesian inference. Figure 1E displays the weekly expression intensity of iRFP estimated using Bayesian inference. In the HCD, HCD + NN, and HCD + LVK groups, the expression intensity reached around 100 pixel/area at week 5 and exceeded 300 pixel/area at week 10. However, in the HCD + HVK group, the intensity was approximately 75 pixel/area at week 5 and around 200 pixel/area at week 10. This suggests that the inhibitory effect of HVK on atherosclerosis was observed at an earlier time point than 5 weeks, as indicated by non-invasive observations (Fig. 1E left panel). Additionally, comparison of iRFP signals between groups at 5, 8, and 10 weeks after induction (Fig. 1F) demonstrated significant differences between HCD and HCD + HVK. The graphs represent the difference from the HCD control values, with the estimated mode (dashed line) and 89% confidence interval (black line) displayed.
In parallel, we conducted serum analyses to further elucidate the potential benefits of natto by examining impacts on liver function and lipid profiles. Each group demonstrated an initial increase in liver function markers (AST and ALT) in response to the high-fat diet, but there was no significant subsequent deterioration or improvement in these parameters over the time course. Furthermore, variable K2 content did not appear to significantly influence these liver function indicators (Supplementary Fig. 3A).
As for the lipid profiles, although certain statistical significances were found in serum lipid concentrations, these findings did not establish a consistent pattern (Supplementary Fig. 3B). Our investigation into the activity of lipoprotein lipase (LPL)—an enzyme often associated with suppression of atherosclerosis onset—showed that neither natto nor variable vitamin K2 content had a substantial impact on LPL activity (Supplementary Fig. 3C).
Effects of natto intake on gut microbiota diversity
Natto contains Bacillus subtilis var. natto, and we studied its effects on the gut microbiota in atherosclerosis-prone mice fed HCD. HCD + HVK, HCD + NN, and HCD + LVK natto groups were compared to the HCD group by cecal collection at 5 and 9 weeks. These contents were subjected to an alpha diversity (chao1) analysis and a rarefaction analysis to evaluate microbiota diversity (Fig. 2A, B). As a result, compared to the HCD group, the number of microbial species significantly increased by roughly 20% in the HCD + NN and HCD + LVK groups, while no significant increase was observed in the HCD + HVK group (Fig. 2A, B). We next conducted a UniFrac analysis to evaluate similarities in gut microbiotas between the natto intake and HCD groups (Fig. 2C). A two-dimensional scatter plot created by the principal coordinate analysis revealed significant differences between the natto intake (red, green, and yellow) and control groups (blue) (Fig. 2C). In addition, when we observed the specific distribution of microbial species in the gut microbiotas, we found differences in microbial composition, particularly in a trend toward decreased Firmicutes in the HCD + HVK and HCD + LVK groups, although it was not significant. A trend towards increased Epsilonbacteraeota was also observed, but it was also not significant (Fig. 2D, E)….
