Indirect demonstration of the lifetime function of human thymus

Abstract

The aim of this study was to test the hypothesis that human thymus maintains its function as the site of early T cell development throughout life, but to a progressively diminishing extent. Mononuclear cell suspensions prepared from the samples of 39 human thymuses were analysed for the total number of cells per gram of thymus tissue, percentage of single marker-positive CD2, CD4 and CD8 cells, percentages of double-positive CD4 CD8 and CD2 CD8 cells, double-negative CD4 CD8 cells, absolute numbers of these cells per gram of tissue, and extent of the in vitro proliferation upon stimulation with concanavalin A (Con A), phytohaemagglutinin (PHA) and pokeweed mitogen (PWM) mitogens. The main outcome measures were flow cytometric data on thymus lymphoid cell composition (according to CD classification), expressed as percentages and numbers of cells per gram of thymus tissue. The total number of mononuclear cells expressed per gram of thymus tissue exponentially decreased with age. The slope of none of the analysed cell subpopulations differed from the slope of the line constructed for age-related decline of the total number of mononuclear cells (−0.024 on a semilogarithmic scale). The thymuses of all ages contained all analysed cell subpopulations in approximately the same proportions: percentages of these cell subpopulations did not change with age, except for all CD4⁺ (P = 0.017) and double-positive CD4⁺ CD8⁺ (P = 0.016) cells, which tended to decrease with age. The extent of proliferation of thymus cells upon stimulation with T and B cell mitogens was unrelated to age. We conclude that the thymus retains its function as the site of differentiation of T lymphocytes throughout life. With respect to the number of involved lymphoid cells, the function exponentially decreases with age.

INTRODUCTION

Well documented age-related thymus atrophy [1–5], especially the replacement of its lymphoid tissue with fat tissue [5, 6], and weakening of thymus-dependent immune responses with age [7], leads to a conclusion that the thymus does not function after puberty. Small amounts of lymphoid tissue can be found in the gland of greater age [4, 8], but this was mostly ascribed to insignificant tissue remnants or lymphocyte infiltration unrelated to thymus function. On the other hand, Jerne [9] suggested that the lymphoid system ‘ossificates’ with age, because it is gradually ‘filled’ with long-living lymphocytes and needs progressively fewer newly differentiated T cells. With this idea in mind, we tested the hypothesis that the thymus retains its function throughout life, but to a quantitatively diminished extent.

We analysed human thymus tissue obtained during heart surgery, and expressed the number of thymus cells per gram of thymus tissue [8], because this analysis allows comparisons of cell contents of thymuses of any size, composition and weight. It was assumed that, if the thymus does function throughout life, the number of lymphoid cells per gram of thymus tissue will diminish with age [8] but the cell composition will remain the same, i.e. it will reflect the (permanent) thymus role in differentiation of T lymphocytes. In other words, the percentages of the relevant thymus lymphocytes (e.g. CD4⁺ CD8⁺ cells) within the total population of thymus lymphocytes should remain constant, whereas their concentration (absolute number per gram of thymus tissue) should follow the pattern of decline of the total number of thymus lymphoid cells per gram of tissue during ageing [8]. Also, we assumed that if the lymphocyte content in the thymus is associated with organ function only (and not with lymphocyte infiltration), proliferative capacity of thymus cells after stimulation with mitogens should not correlate with age. Our results indicate in an indirect manner that the thymus functions throughout life, albeit with considerably diminished capacity later in life.

MATERIALS AND METHODS

Two to four tissue samples from different locations were taken from thymuses of 39 individuals. Median weight of samples of 39 patients was 3.19 g (range 0.34−13.3 g). Cell suspensions were then analysed for the following parameters: total number of cells per gram of thymus tissue, percentages and absolute numbers per gram of thymus tissue of all CD2⁺, CD4⁺ and CD8⁺ cells, of double-positive CD4⁺ CD8⁺ and CD2⁺ CD8⁺ cells, of single-positive CD4 (CD4⁺ CD8⁻) and CD8⁺ (CD4⁻ CD8⁺) cells, and of double-negative CD4⁻ CD8⁻ cells. Also, for each thymus cell suspension, the in vitro proliferation of cells upon stimulation with concanavalin A (Con A), phytohaemagglutinin (PHA) and pokeweed mitogen (PWM) mitogens was determined.

Isolation of thymocytes

Thymocytes were obtained from specimens of normal thymuses removed at the time of cardiac surgery from 39 individuals (23 males and 16 females) aged 1−73 years (median age 46 years). None of the subjects had been previously treated with any immunosuppressive therapy. All patients or their parents gave informed consent. The study was approved by the Ethics Committee of the University Hospital Centre Zagreb.

Single-cell suspensions of thymus lymphoid cells were obtained by mincing the thymus tissue and then pressing the fragments through a stainless steel mesh. Viable thymocytes were isolated by Ficoll−Hypaque (Pharmacia, Uppsala, Sweden) density gradient centrifugation. The number of nucleated cells was determined and related to a gram of the original sample. Cell viability, determined as the ability of intact cells to exclude trypan blue dye, was 93% (range 80−100%). There was no correlation between cell viability and patients' ages (data not shown).

Flow cytometry

A two-colour flow cytometric analysis was performed using MoAbs for CD4 (T4) and CD2 (T11) markers conjugated with PE (Coulter, Hialeah, FL) and CD8 (T8) antibody conjugated with FITC, as previously described [10]. Thymocyte suspensions (100 ml) were incubated in PBS containing 5% bovine serum (BS) with previously defined optimal concentrations of PE-conjugated anti-CD4 and anti-CD2 and FITC-conjugated anti-CD8 antibodies. The samples were incubated for 45 min at 4°C and washed twice in PBS. After additional washing in PBS, cells were analysed on an EPICS-C (Coulter) flow cytometer. Thymocytes were gated on a two-parametric histogram L90°LS X FALS, and then the log integrated fluorescence (LGFL and LRFL) was analysed on a Quad-Stat program to obtain the percentage of cells positive for green fluorescence, red fluorescence and both.

Proliferative assays

Triplicate cultures of thymocytes in 200 ml of M166 medium with 20% decomplemented and pooled human AB sera were prepared in flat-bottomed 96-well microculture trays (Becton Dickinson, Lincoln Park, NJ). Cultures were stimulated with PHA 10 ng/ml (Welcome, Beckenham, UK), Con A 20 ng/ml (Sigma, St Louis, MO) or PWM 10 ng/ml (Sigma). Cultures were incubated at 37°C in a humidified atmosphere of 5% CO₂ in air for 72 h. During the last 18 h of incubation, they were labelled with 3 mCi ³H-thymidine (³H-TdR) (Radiochemical Centre, Amersham, UK). Radioactive material was collected on Millipore filters (Titertek Cell Harvester Filter; Flow Labs, Irvine, UK) and counted in a scintillation counter (214, RackBeta; Wallac, Turku, Finland). Results were expressed as Δct/min (difference of ct/min in stimulated and unstimulated cultures) and stimulation index (SI; ratio of ct/min in stimulated and unstimulated cultures).

Statistical analysis

Numerical data are presented as mean values and s.d. when the data distribution was normal, and as median value and 10th to 90th percentile range in all other cases. Correlation analysis and linear regression were used to test the association between lymphoid cell expression of thymus antigens and patients' ages. When testing the absolute numbers of thymus cells, logarithmic data transformation was performed before analysis. Results of correlation analysis are presented with Pearson's correlation coefficient. In linear regression analysis, age was considered as the dependent variable and antigen expression or absolute number of cells as independent. P < 0.05 was considered significant. Statistics were calculated using MedCalc software (MedCalc, Mariakerke, Belgium).

RESULTS

The specimens of thymus tissue from each subject were pooled and further treated as a single sample. Table 1 shows percentages of the respective cell populations in the thymus samples (mean + s.d.) and their relations to the age of the patients. After the regression lines were constructed, further analysis was performed in order to test the hypothesis that the slopes of constructed lines did not differ from zero, i.e. that the proportions of cells did not change with age. Only when both correlation proved to be significant (P < 0.05, third column) and the corresponding linear regression showed that the slope significantly differed from zero (P < 0.05, far right column) could it be concluded that the proportion of cells changed with age. In this respect, only the percentages of all CD4⁺ (single- and double-positive) and of double-positive CD4⁺ CD8⁺ cells correlated with patients' ages, showing a significant tendency to decrease. The finding is illustrated in Fig. 1: none of the slopes of regression lines, except for all CD4⁺ and CD4⁺ CD8⁺ cells, differed significantly from zero, i.e. the percentages of all other cells did not change with age.

Table 1.

Statistical analysis of data on percentages of T cell subpopulations in 39 human thymuses. Correlation and linear regression analyses were calculated according to the patients’ ages (years)*

Fig. 1.

Linear regression with 95% confidence intervals of percentages of thymus lymphoid cell subpopulations according to patients' ages. For statistical details, see Table 1.

Table 2 shows statistically identical data analysis on the absolute numbers of cells per gram of thymus tissue for the respective cell populations (median, 10−90% range). Number of cells per gram of thymus tissue significantly correlated with age, showing negative correlation coefficients, i.e. the cell quantity significantly decreased with age. Regression lines for all cell populations were then compared with the regression line obtained for all mononuclear thymus cells to test whether the dynamics of cell loss with age affected all cell populations equally. Apart from CD4⁻ CD8⁻ cells, the analysis revealed that neither slopes nor intercepts of the regression lines for all cell populations differed from the slope/intercept of the regression line found for the total number of mononuclear cells per gram of thymus tissue (columns p(b_o = 7.42) and p(b₁ = − 0.024)). Accordingly, the age-related regression lines with 95% confidence limits, shown in Fig. 2, illustrate that the line slopes for all cell populations analysed did not differ from that calculated for the total number of all mononuclear cells per gram of thymus tissue (top left). Only the percentages of CD4⁻ CD8⁻ cells did not correlate significantly with age, which made further analysis futile.

Table 2.

Statistical analysis of data on absolute numbers of T cell subpopulations per thymus tissue mass (g) in 39 human thymuses. Correlation and linear regression analyses were calculated according to the patients' ages (years)*

Fig. 2.

Linear regression with 95% confidence intervals of absolute numbers (×10⁶) of thymus lymphoid cell subpopulations per gram of thymus tissue according to patients' ages. For statistical details, see Table 2.

Our final analysis concentrated on testing the hypothesis that mitogen-induced proliferation of thymus cells should not reveal an age-related pattern if the thymus cell composition did not reflect the presence of infiltrating mature lymphocytes but its function of T cell differentiation only. In as much as the mature lymphocytes do and thymus lymphocytes do not proliferate [11], the results of mitogen-driven in vitro proliferation of thymus cells should provide the respective answers. Table 3 shows the data of thymus cell proliferation after in vitro cell stimulation with PHA, Con A and PWM. The data on Δct/min and SIs are arranged according to patients' ages in ascending order. It can be seen that thymus cells from some patients did proliferate upon stimulation with some mitogens, but from most did not. No age-related pattern of reactivity could be revealed, either by patient grouping or by regression analysis (data not shown). Also, there was no correlation between positive reactivity to a mitogen and CD4/CD8 cell composition of the thymuses (data not shown). The only significant correlation was that of simultaneous reaction to stimulation with PHA and Con A: P < 0.05 both for Δct/min (r = 0.671) and SI (r = 0.834).

Table 3.

Mitogenic responses of thymocytes in 39 subjects aged 1–73 years*

DISCUSSION

CD4 and CD8 single T cell markers, and especially their double-marker analysis (e.g. that of CD4 CD8 positive and negative cells) appeared to be the most relevant possible approach to test our hypothesis. The reason for including a pan-T (CD2) marker is related to the possible contamination of the samples with non-lymphoid cells, mostly with respect to the analysis of proportions of double-negative T cells. The mean percentage of CD2 cells was > 94% (Table 1), which corresponds to literature data [11] and shows that most of the analysed cells were lymphocytes.

In accordance with our hypothesis, general thymic cell composition (percentages of analysed cell subpopulations) did not change with age and the age-related decline of cells' absolute numbers per gram of thymus tissue was identical to that of all nucleated cells in the thymus. Only the proportions of all CD4⁺ and CD4⁺ CD8⁺ cells appeared to decrease with age (Table 1, Fig. 1), and the decrease in total number of CD4⁻ CD8⁻ cells did not correlate significantly with age (Table 2, Fig. 2). The latter may be ascribed to the small numbers of double-negative cells found in the thymus and the subsequent lack of statistical significance. On the other hand, the age-related decrease of the proportions of CD4⁺ and CD4⁺ CD8⁺ cells is a deviation from the data predicted by our hypothesis. However, with respect to the fact that the proportions of all other cells did not change with age, and especially that the absolute numbers of these two cell populations did not deviate from the pattern of all thymus lymphoid cells and other cell subsets (Fig. 2), we believe that our data still confirm both our hypothesis and Jerne's proposal [9]. The percentages of both CD4⁺ and CD4⁺ CD8⁺ cells found in this study (Table 1) were somewhat lower than those reported in the literature [11], which may be the consequence of the quality of anti-CD4 serum used. Indeed, the proportions of all cells analysed did not differ between younger and older patients, regardless of the age at which the year breakpoint was set (e.g. 10, 12 and 20 years, analysis not shown).

The analysis of the proliferative capacity of thymus cells upon in vitro stimulation with mitogens PHA, Con A and PWM served to test an additional deductive consequence of our hypothesis. Namely, if the cell composition of the ageing thymus reflected a maintained, albeit diminished, lifetime function, then the proliferation pattern of thymus cells should not change with age. Our data (Table 3), as well as that of others [12], showed that the cells of some thymuses did proliferate upon stimulation with some of the tested mitogens, but no age-related pattern of proliferation could be deciphered. It is rational to assume that the occasional proliferations were related to factors other than age. These may be associated with the technique of thymus tissue harvesting (e.g. unintentional harvesting of larger amounts of peripheral blood) or other factors, like hormonal or stress status of the organism.

Our investigation yielded the results and conclusions which support Jerne's proposal that the immune system as a whole functions throughout life in the same qualitative fashion but, owing to the accumulation of the long-living T cells, the demand for cell production diminishes, and fewer and fewer new thymus cells are released into the system [9]. It seems that the functioning of the thymus at greater ages is not the consequence of its physiological reserve [13], but a reflection of life history of the immune system, with the thymus as its essential component [9]. In the context of the process of ageing and increased human longevity, thymus involution should not be viewed as an uncontrolled organ failure, but rather an evolutionary outcome of an organism's optimization of investment in maintenance, i.e. decreased need at greater age [14]. Indirectly proved by our present study, this physiological picture may have far-reaching theoretical and practical implications.

The exponential decrease in cell number per gram of thymus tissue has been quantified earlier [8], and our present data confirm the finding. In addition, this study shows that ‘concentrations’ of all function-relevant cell subpopulations also decrease in parallel with the total cell number in the ageing thymus. The relatively high intra-observer variability of our model (biological variation, sample harvesting, cell isolation and dying, etc.) provided the data which in several cases (Tables 1 and 2) yielded marginal statistical significance (or insignificance), prompting us to draw rather cautious conclusions. Yet the study clearly showed that immature and relatively short lived double-positive and double-negative cells were present in the thymus throughout adult life, implying that it continues to produce immature lymphocytes. Additional, differentiation-related lymphocyte markers like CD1, CD25 and/or CD44, together with immunomagnetic purification of the lymphocytes before staining, or cell selection based on CD45 antigen expression, would be more accurate tools for a definite quantification of age-related cellular turnover in the thymus and a consequent inference of its physiological significance.