Data correlations between gender, cytomegalovirus infection and T cells, NK cells, and soluble immune mediators in elderly humans

We describe a cohort of 50 elderly subjects, age at least 70 years. We present gender-specific findings in T lymphocyte markers and soluble immune mediators. We show the correlation between cytomegalovirus infection status with CD56dim NK cell responses to a variety of stimuli and with CD56bright/CD56dim NK cell ratio. We also present the correlation of retinol binding protein (RBP)−4 plasma levels with NK cell responses and we explore the relationship between gender and adiponectin, 25(OH)D (vitamin D), and RBP4 in affecting CD56dim NK cell responses. These data are discussed in Al-Attar et al. (2016) [1].


Subject area
Immunology More specific subject area

Sex differences in immune cells and soluble mediators
Type of data Tables How data was acquired Flow cytometry and enzyme-linked immunosorbent assay Data format Analyzed Experimental factors Immune cells were analyzed immediately ex vivo or were stimulated in vitro Experimental features Correlations and mean differences were calculated Data source location Lexington, KY -USA Data accessibility Data are within this article

Value of the data
Researchers need to be aware of the gender differences in NK cell responses to various stimuli. Exposure to cytomegalovirus (CMV) affects immune responses from T and NK cells and could therefore be an important factor to consider when performing research in elderly human subjects.

Levels of soluble plasma immune mediators adiponectin and vitamin D (25(OH)D) which affect NK
cell development and activity are higher in elderly women compared to men, and are important factors to consider when studying human NK cells.

Data
Enclosed are data concerning T cell markers and subsets found in elderly women and men (Table 1). Also shown is the effect of cytomegalovirus (CMV) infection on CD56 dim NK cell responses to a variety of stimuli and on the CD56 bright /CD56 dim NK cell ratio in blood lymphocytes ( Table 2). We present the levels of various plasma immune mediators and their levels in elderly women and men (Table 3). We show how plasma RBP4 level correlates with NK cell responses in vitro (Table 4) and we explore if the interaction between gender and plasma adiponectin,RBP4, and 25(OH)D (vitamin D) affects CD56 dim NK cell responses in vitro (Table 5). Full details of the data can be found in Al-Attar et al. [1].

Experimental design, materials and methods
Male and female subjects 470 years were recruited from volunteer donor pools at the University of Kentucky Sanders-Brown Center on Aging and by advertisements. Venous blood from 26 males (age mean7standard error of the mean (SEM), 77.870.31, range 70-90 years) and 24 females (age mean7-SEM, 77.070.91, range 70-85 years) were analyzed between October 2012 and April 2014. Prospective donors were screened by telephone interview to exclude those with conditions previously demonstrated to affect NK cells. Exclusion criteria included hospitalization or serious illness in the prior year, history of immunologic illness (rheumatoid arthritis, systemic lupus, scleroderma, polymyositis, Sjögren's syndrome, transplantation, etc), current use of immunomodulatory medications (e.g., corticosteroids), inability to walk one city block, regular consumption of two or more alcoholic beverages per day (28 g ethanol), diabetes, and a history of cancer within the last 10 years, except non-melanoma skin cancers. Two female subjects (but no male subjects) received hormone replacement therapy, one with topical estrogen and one with topical estrogen, progesterone, and testosterone. The gender differences affecting CD56 dim NK cell CD38 density and the CD56 bright to CD56 dim NK cell ratio were no longer significant when the hormone replacement subjects were excluded from analysis. All other gender differences reported below remained significant after exclusion of these two subjects. Blood samples were obtained from people without acute For each tested analyte on gated T cell subset, the mean and standard error of the mean (SEM) is shown for each gender. All values are given as percentage of the gated population, except when geometric mean fluorescence intensity (gMFI) antigen level is given. Statistical significance of mean differences was determined by student's T test.
illness in the preceding week between 9:30 a.m. and 12:45 p.m. Lymphocytes and serum 25hydroxyvitamin D (25(OH)D) were analyzed immediately; plasma was aliquoted and stored at À 80°C. All human subjects research was approved by the University of Kentucky Institutional Review Board.

ex vivo staining
Whole blood was diluted 1:1 with PBS and overlaid on Lymphoprep s lymphocyte separation medium (Axis-Shield, Oslo, Norway) according to the manufacturer's protocol. Peripheral blood mononuclear cells (PBMC) were collected and washed twice. For ex vivo staining, $ 0.5 Â 10 6 PBMC were washed and incubated with human IgG for 15 min at room temperature to block Fc-receptor binding and then stained on ice for 30 min with combinations of fluorescently labeled mAb (See Supplementary Table S1 in Ref. [1]). After washing, the cells were analyzed on a LSR-II flow cytometer (BD, Franklin Lakes, NJ). CD4 and CD8 T cells were divided into four subpopulations, T N : CD62L þ CD45RA þ , T CM : CD62L þ CD45RA -, T EM : CD62L -CD45RA -, and T EMRA : CD62L -CD45RA þ . CD56 dim NK cells that did not stain positive for CD158b, CD158f, CD158e/k or NKG2A were considered unlicensed. CD56 dim NK cells were gated and tested for the indicated analyte in response to the designated stimulus. Shown is the mean and standard error of the mean (SEM) for subjects who were CMV infected ( þ ) or not CMV infected ( À ). Bold print denotes significant mean differences as determined by student's T test. * CD56 bright /CD56 dim ratio, tested immediately ex vivo.

Cytokines, soluble immune mediators, and antibodies
Following manufacturer instructions, enzyme-linked immunosorbent assay was used to measure plasma C-reactive protein (CRP) and adiponectin levels (eBioscience) and retinol-binding protein 4 (RBP4) levels (Abcam, Cambridge, England). Plasma IL-15 was quantified with the QuantiGlo Chemiluminescent Immunoassay kit (R&D Systems). Reported values are the average of a single measurement of each serum sample tested in two (IL-15), three (Adiponectin, CRP), or four (RBP4) independent experiments. For sphingosine-1-phosphate (S1P) and dihydrosphingosine-1-phosphate (dhS1P) measurements, plasma samples with added deuterated standards [2] were extracted using acidified organic solvents [3]. S1P and dhS1P 1evels were measured using high pressure liquid chromatography-electrospray ionization tandem mass spectrometry, and quantified by comparing levels to the internal standard [4]. The University of Kentucky Clinical Chemistry Laboratory measured serum 25(OH)D by liquid chromatography-tandem mass spectrometry using atmospheric pressure chemical ionization in positive ion mode and quantitated by comparison to a deuterated internal standard. Plasma samples were diluted 100-fold and anti-cytomegalovirus (CMV) IgG was measured in duplicate via binding to immobilized CMV antigen. Bound CMV IgG was detected by horseradish peroxidase conjugated anti-human IgG antibody (DRG CMV IgG ELISA, Springfield, NJ). None of the positive or negative anti-CMV IgG levels were close to the intermediate range. Shown are nonparametric Spearman's correlation coefficients (ρ) and significance (Sig.) between RBP4 level and the indicated parameter. Bold print denotes significant correlations.

Statistical methods
For means comparisons, data were first analyzed with Levene's test for equality of variance using SPSS, version 22 (IBM, Armonk, NY). When gender based variances were significantly unequal, means were compared by the nonparametric Mann-Whitney U test. None of the means analyzed by the Mann-Whitney test significantly differed by gender. Otherwise, means were compared using the 2tailed student's t test. Levels are displayed as means 7 SEM. Multivariate analysis of sex-specific NK cell responses. Each dependent variable (NK cell response) was compared with the independent variable, sex (Model 1, Pearson correlation), or with sex along with other independent variables, either adiponectin, RBP4, or 25(OH)D (vitamin D) plasma levels (Model 2, multiple linear regression). For scoring purposes, males¼ 1; females¼0. Therefore, a negative correlation with sex indicates stronger responses in women than in men. 25(OH)D levels were not available for 1 female and 2 male subjects. Therefore, the significance (Sig.) of the correlation between sex and the NK cell response is different than when adiponectin and RBP4 effects are analyzed. The value of the unstandardized coefficient (B) reflects the amount of change in the predicted preference ranking. Using the standardized coefficient (Beta), interpretations are based on the standard deviation (SD) of the variable. Each standardized coefficient (Beta) indicates the number of SD that the dependent variable changes for a 1 SD change in the independent variable, the other independent variable remaining constant. For both B and Beta, the higher the absolute value (positive or negative), the greater the effect of the given independent variable on the dependent variable. For example, in the first table (Model 2), the absolute value of B and Beta for Sex was significantly greater than B and Beta for Adiponectin, indicating that Sex had a stronger (inverse) correlation with the NK cell response than did Adiponectin. Significant values (as determined by Pearson correlation, Model 1, and multiple linear regression, Model 2) are shown in bold print.