Composition of peripheral blood immune cell compartment in stage 5 chronic kidney disease is affected by smoking and other risk factors associated with systemic inflammatory response

Abstract

Background: Stage 5 chronic kidney disease (CKD5) is linked to complex yet not fully understood disturbances in immune system. This study aimed to investigate these disturbances by exploring the detailed composition of peripheral blood immune cell compartments in CKD5 patients and to provide integrative, multivariable dissection of how common inflammatory risk factors shape the immune landscape.

Methods: This cross-sectional study included 107 patients with chronic kidney disease stage 5 (CKD5) and 29 healthy blood donors as controls. Peripheral blood B cells, T cells, and dendritic cells were measured using a standardized and validated flow cytometry panel. The impact of selected clinical factors on immune cell composition was initially evaluated using a robust multivariate method (PERMANOVA). Variables that significantly affected immune cell composition were then included in a subsequent series of Poisson regression models, assessing predictors influence on the counts of individual immune cell subpopulations.

Results: Compared to healthy controls, CKD5 patients presented with B cell lymphopenia across all measured subsets except for plasmablasts, T cell lymphopenia with an immunosenescent phenotype predominantly in the CD4⁺ compartment, and significantly higher counts of LIN-HLA-DR⁺ antigen-presenting cells, mainly due to an increase in myeloid dendritic cell subpopulations. PERMANOVA identified smoking, CMV seropositivity, age, dialysis treatment, and atherosclerotic cardiovascular disease as factors significantly influencing peripheral blood immune composition. Subsequent Poisson regression models revealed that smoking was associated mainly with an increase in switched memory B cells, CMV seropositivity with an increase in CD4⁺ and CD8⁺ TEMRA cells, age with a decrease in naive CD8⁺ T cells, and dialysis treatment with an increase in marginal-zone B cells.

Conclusions: Patients with CKD5 exhibit distinct composition of peripheral blood immune cells, further modified by other factors associated with systemic inflammatory response. These factors should be considered in immunomonitoring protocols and may enhance prediction of clinical outcomes such as vaccine responses.

Keywords: B cells; T cells; chronic kidney disease; dendritic cells; flow cytometry; inflammation; peripheral blood immune cell composition.

Figure 1

Matrix of Spearman correlation coefficients between clinical characteristics and counts of various peripheral blood immune cell subpopulations. The dendrogram illustrates similarities in cell counts, where closely linked tips indicate that the associated immune cell subpopulations are interrelated, with higher counts tending to co-occur. Correlations with a cluster of related co-occurring sub-populations suggest a specific influence (e.g., CMV infection), while correlations with individual subpopulations across different clusters indicate a more diffuse effect (e.g., smoking). The dendrogram is constructed using the unweighted pair group method with arithmetic mean (UPGMA) clustering. Clusters of subpopulations (as defined by UPGMA) are as follows: (1, grey) circulating plasmablast; (2, red) plasmacytoid dendritic cells (pDC); (3, purple) B cell compartment – including naive B cells, transitional B cells, class-switched memory cells, marginal-zone B cells, and class non-switched memory cells; (4, blue) myeloid dendritic cells (mDC1, mDC2, mDC3); (5, green) naive T cell compartment – CD4+ and CD8+ naive T cells; and (6, orange) memory T cell compartment – including CD4+ and CD8+ TEMRA, effector memory (EM), and central memory (CM) subsets.

Figure 2

The volcano plot illustrates the effects of clinical characteristics on the counts of various peripheral blood immune cell subpopulations, estimated using a generalized linear model (GLM) with a Poisson distribution. The x-axis represents the effect size of each predictor on the outcomes. For example, an effect size of 0.85 for smoking on switched memory cells suggests that smoking is associated with a 2.33-fold increase in the count of these cells, as determined by exp(0.85). The y-axis displays the −log10(P-value) for each association, before P-value adjustment. Only associations with unadjusted P-values below 0.05 are labelled.