Multivariate analysis of immunosenescence data in healthy humans and diverse diseases

Abstract

Introduction: Immunosenescence is a dynamic process, where both genetic and environmental factors account for the substantial inter-individual variability. This paper integrates all the data on immunosenescence markers generated in our laboratory and describes the differences and/or similarities between individuals based on their biological conditions (immunosenescence markers) and their associations with chronological age and health status.

Materials and methods: The dataset consisted of immunological data from healthy donors, centenarians, patients diagnosed with chronic kidney disease, COVID-19 and non-small cell lung cancer (NSCLC), treatment-naïve or treated with platinum-based chemotherapy. To determine whether there are groups of immunologically different individuals despite their age or clinical condition, cluster analysis was performed. Canonical discriminant analysis was performed to determine which variables characterize each cluster.

Results: There are differences in the expression of immunosenescence markers between healthy subjects and patients diagnosed with different pathological conditions, regardless of their age. Meanwhile, the distribution of the clusters indicates the presence of two separate groups of healthy participants, one of them characterized by a high frequency of naïve lymphocytes, and the other with high expression of terminally differentiated lymphocyte subsets. Advanced NSCLC treatment-naïve patients were in the same cluster as a group of healthy subjects. Additionally, centenarians belong to a different cluster than healthy subjects, suggesting they might have a unique immune signature.

Conclusion: The distribution of clusters appears to be more appropriate than univariate associations of single markers for health and disease research. The present work reveals which immune markers are relevant in different physiological and pathological contexts and indicates the need for deeper studies on the biological age of the immune system.

Keywords: centenarians; healthy subjects; immunosenescence markers; multivariate analysis; non-small cell lung cancer.

FIGURE 1

Flow diagram of subject selection, data processing and statistical analysis.

FIGURE 2

Frequency of naïve lymphocytes. (a) CD4 T cells (CD4+CD45RA+CCR7+), (b) CD8 T cells (CD4+CD45RA+CCR7+) (c) B cells (CD19+CD27−IgD+). The asterisks indicate statistically significant differences among the groups (*p < 0.05) using ANOVA test.

FIGURE 3

Frequency of late-stage differentiated lymphocytes. (a) EMRA CD4 T cells, (b) EMRA CD8 T cells (c) CD8+CD28− T cells. The asterisks indicate statistically significant differences among the groups (*p < 0.05) using ANOVA test.

FIGURE 4

K-means clustering performed for the 397 subjects represented as a 2-dimensional cluster plot based on a principal component analysis.

FIGURE 5

Canonical discriminant analysis biplot with Can1 and Can2 (the 2 firsts canonical dimensions) shows the correspondence of immunological profiles with each four clusters (showed in different colors). Points represent individuals. Vectors represent the correlations of immunological variables with the canonical dimensions.

FIGURE 6

(a) K-means clustering performed for the 239 subjects over 60 years old. (b) Canonical discriminant analysis biplot for the immunological profiles of the four clusters in patients over 60 years old.