Plasma carotenoids are inversely correlated with granulocyte counts and soluble inflammatory markers in a middle-aged population: a cross-sectional study with mediation analysis

Abstract

Background: High intake of fruits and vegetables is generally associated with reduced levels of inflammation. In line with this, plasma levels of carotenoids have shown inverse associations with inflammatory markers, in particular C-reactive protein (CRP) and leukocyte counts. However, it remains unclear to what extent carotenoids are associated with specific leukocyte subsets or other inflammatory markers. This study systematically assessed the inter-relationships among total and individual carotenoids, circulating leukocyte subsets, and soluble inflammatory markers in a middle-aged population.

Methods: A subcohort of 1078 subjects, aged 50-64, was recruited from the Swedish CArdioPulmonary bioImage Study (SCAPIS) cohort. Leukocyte subsets were determined by whole blood flow cytometry. Five major carotenoids, namely lutein, β-cryptoxanthin, lycopene, α-carotene and β-carotene, and inflammatory markers including CRP, interleukin (IL)-6 and interleukin-18, matrix metalloproteinase (MMP)-9, and myeloperoxidase (MPO), were measured in plasma. Nutrient intake was estimated by validated food frequency questionnaires.

Results: Among leukocyte subsets, only granulocytes showed independent and inverse associations with all carotenoids after adjustment. CRP, IL-18, and MMP-9 exhibited similar inverse relationships with most carotenoids. Mediation analysis revealed that the associations of carotenoids with CRP and MMP-9 were mediated by granulocyte counts. Lutein and β-cryptoxanthin remained independently associated with MMP-9 after accounting for the mediation effects of granulocyte counts. No estimated nutrient intake showed comparable associations with leukocyte subsets or inflammatory markers.

Conclusions: To our knowledge, this is the largest population-based study investigating relationships between plasma carotenoids, leukocyte subsets, and soluble inflammatory markers. It provides evidence that low levels of carotenoids in plasma are linked to low-grade chronic inflammation and, furthermore, that this relationship is mediated by granulocyte numbers.

Keywords: C-reactive protein; Carotenoid; Granulocyte; Interleukin-18; Leukocyte; Matrix metalloproteinase-9; Population-based cohort.

Fig. 1

A heatmap indicating the Spearman correlation coefficients between nutritional parameters and counts of leukocyte subtypes and inflammatory markers. Nutritional parameters include plasma carotenoids and estimated dietary intake of macro- and micronutrients derived from food frequency questionnaires. Nutrients that showed no significant correlation with any plasma carotenoid were excluded from this figure. The color gradient reflects the strength and direction of correlations, with blue indicating negative correlations and purple indicating positive correlations. CRP, C-reactive protein; IL, interleukin; MPO, myeloperoxidase; MMP-9, matrix metalloproteinase-9; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids. *** p < 0.001; ** p < 0.01. p < 0.05 not shown

Fig. 2

Linear regression models between tertiles of carotenoids and A immune cell counts or B inflammatory markers. Regression models with inflammatory markers were adjusted for age, gender, total cholesterol, antihypertensive medication, smoking status, abdominal obesity, moderate and vigorous physical activity, self-reported type-2 diabetes, and the proportion of the total daily energy intake from polyunsaturated fat (model 2). Data represents the effect size, which represents the percentage change of the immune cell counts or inflammatory markers with each increasing tertile of carotenoids. All values were log-transformed before regression analysis. Error bars represent the 95% confidence interval. CRP, C-reactive protein; IL, interleukin; MPO, myeloperoxidase; MMP-9, matrix metalloproteinase-9; CI, confidence interval. *** p < 0.001; ** p < 0.01; * p < 0.05

Fig. 3

Mediation models. A The association between total plasma carotenoids as exposure variable (X), CRP as outcome variable (Y), and cell types that qualified for mediation as mediator variables (M). B The association between total plasma carotenoids as exposure variable (X), MMP-9 as outcome variable (Y), and granulocytes as mediator variable (M). All regression analyses were conducted on log-transformed variables. The β-coefficient represents the effect size of the regression model. Significant mediation is denoted by # (Bonferroni corrected threshold p < 0.017). CRP, C-reactive protein; MMP-9, matrix metalloproteinase-9