A "crossomics" study analysing variability of different components in peripheral blood of healthy caucasoid individuals

Abstract

Background: Different immunotherapy approaches for the treatment of cancer and autoimmune diseases are being developed and tested in clinical studies worldwide. Their resulting complex experimental data should be properly evaluated, therefore reliable normal healthy control baseline values are indispensable.

Methodology/principal findings: To assess intra- and inter-individual variability of various biomarkers, peripheral blood of 16 age and gender equilibrated healthy volunteers was sampled on 3 different days within a period of one month. Complex "crossomics" analyses of plasma metabolite profiles, antibody concentrations and lymphocyte subset counts as well as whole genome expression profiling in CD4+T and NK cells were performed. Some of the observed age, gender and BMI dependences are in agreement with the existing knowledge, like negative correlation between sex hormone levels and age or BMI related increase in lipids and soluble sugars. Thus we can assume that the distribution of all 39.743 analysed markers is well representing the normal Caucasoid population. All lymphocyte subsets, 20% of metabolites and less than 10% of genes, were identified as highly variable in our dataset.

Conclusions/significance: Our study shows that the intra-individual variability was at least two-fold lower compared to the inter-individual one at all investigated levels, showing the importance of personalised medicine approach from yet another perspective.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. Overview of the experimental design.

Sixteen gender and age matched healthy individuals were enrolled in a study. Fasting morning blood samples were taken on three days within one month and analysed using different omics approaches. Fasting morning samples were collected from 16 volunteer three times within a one month period.

Figure 2. Overview of metabolite profiles variability found in plasma samples of healthy volunteers by principle component analysis score plot.

Triangle – male volunteer, circle – female volunteer, coloured according to volunteers, labels shows sampling number within one volunteer. PC - principal component.

Figure 3. Average concentrations of two soluble sugars, glucose and mannose, are significantly increased in plasma samples of healthy volunteers with higher BMI values.

pGlucose = 0.0014, pmannose = 0.0001. Error bars represent standard error within monthly measurements.

Figure 4. Variability of different lymphocyte subsets cell counts and antibody concentrations assessed in blood samples of healthy individuals.

The HCA was performed both, for the analysis of samples and the variation of each parameter. The colour scale denotes increasing cell counts/concentrations of the measured parameters, with both extremes being low (blue) and high (yellow). The total numbers of leukocytes (WBC), lymphocytes (LYM), natural killer cells (NK), CD8+ and CD4+ T lymphocytes and their activated forms (aNK, aCD8+ T, aCD4+ T), were determined by flow cytometry. The concentrations of IgG and IgM antibodies are also included. Samples are marked by consecutive number of volunteer (P1–P17), volunteer gender (F, M) and sampling day (D1–5).

Figure 5. Intra-individual variability in gene expression compared to inter-individual variability.

Variabilities are represented as boxplots of gene expression CV for each individual and in complete dataset. A) CD4+ T and B) NK cells. For subject P9 this evaluation was not possible for CD4+ T cell dataset as only one sample out of the three passed the technical quality control in all steps of the procedure. The same was true for subjects P9, P13 and P12 in NK cells dataset.

Figure 6. The pattern of transcripts with variable day-to-day expression seems to be in majority individuum-specific.

Distribution of transcripts with high day to day variability in expression among healthy individuals for A) CD4+ T, and B) NK cells is shown. Transcripts identified as highly variable (CV>0.5) in at least one individual are ordered in rows of the heat map (1952 transcripts for CD4+ T and 2635 transcripts for NK cells). Individuals included in our study are listed in columns. Grey colour – the transcript was not identified as highly variable in the particular volunteer. Black colour – the transcript was identified as highly variable in particular volunteer.

Figure 7. The most prominent BMI-related changes in gene expression were found for interleukin 7 receptor (IL7R), insulin-like growth factor binding protein 3 (IGFBP3), defensin 3 (DEFA3) and DO beta major histocompatibility complex, class II (HLA-DOB).

CD4+ T cells - open squares, NK cells - closed circles. Error bars represent intra-individual standard error.