Pubertal stage specific changes in T-cell subpopulations in healthy individuals

Abstract

This study investigates how puberty affects T-cell subpopulations in healthy individuals, aiming to understand why kidney transplant outcomes are worse during adolescence. We hypothesize that pubertal maturation shifts the immune system towards a more pro-inflammatory phenotype. To explore this, we examined T-cell subtypes in individuals aged 8–30y. Pubertal maturation was assessed in 66 healthy individuals (median age: 17y, 42% male) using skeletal age and Tanner stage, and individuals were subsequently classified into one of four pubertal stages (pre-: n = 10, early-: n = 8, late-: n = 6, and post-puberty: n = 42). Multiple differentiation stages of CD4, CD8, and TCRγδ T-cells subpopulations were determined in peripheral blood samples using flow cytometry. Our results showed that absolute naïve CD4 T-cell and recent thymic emigrant CD4 T-cell counts generally decreased over the course of puberty (p = 0.004, p = 0.015), while absolute CD4 effector memory cell counts increased (p = 0.002). Notably, higher absolute naïve CD4 T-cell counts were observed during early-puberty compared to post-puberty which could not be explained by aging (p = 0.19). These findings indicate a developmental shift during puberty from a naïve to a more mature T-cell profile, supporting the idea that pubertal maturation affects the composition of the immune system.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-27212-5.

Introduction

In kidney transplant recipients (KTx-recipients), an increase in the hazard of graft loss, and in acute and chronic rejections^1–4 is observed over the course of adolescence. This increase is observed independently of therapy compliance and is more evident in girls⁵. Given the key role immune activation plays in graft rejections, we postulate that adolescence causes a shift in the immune system towards a more pro-inflammatory phenotype. This shift could potentially be explained by physical maturation and exposure to antigens during adolescence.

Physical maturation in adolescence is mediated through endocrinological changes, with consequent elevations of sex-hormones and upregulation of the growth hormone axis (GH-axis) having divergent effects on the immune system. Estrogens skew helper T-cell (CD4 +) populations from the pro-inflammatory Th1 type to the anti-inflammatory Th2 type⁶. Furthermore, androgens inhibit T-cell production, Th1 differentiation, and reduce thymus size⁷. Subsequently, the GH-axis, mostly mediated by IGF-1, promotes T-cell proliferation by stimulating thymus growth⁸. Beyond endocrinological effects, naïve T-cells (Tnaïve) differentiate into T-effector memory cells (TemRO)⁹ upon stimulation with novel pathogens. Prolonged stimulation of TemROs, as seen in latent CMV infections or in aging, leads to an increase of the more pro-inflammatory, terminally differentiated effector memory cells (TemRA)¹⁰. Exposure to new antigens might be elevated during adolescence due to expanding social interactions, exemplified by a rise in the incidence of “kissing disease”^11,12. These novel pathogens potentially cause a steeper rise in immune memory during adolescence compared to other periods in life. Ultimately, the overall effect of the endocrinological changes and exposure to novel antigens on T-cell immunity remains unclear. Previously, studies have established reference values of T-cell subpopulations in healthy children from birth to adulthood according to chronological age^13–15. However, these T-cell subpopulations might not reflect the effect of puberty, given that the age of onset and duration of puberty varies greatly among individuals^16,17.

Given the limited data on T-cell subpopulations in healthy individuals across different pubertal stages, our primary objective was to investigate these profiles in healthy immunocompetent individuals prior to studying kidney transplant patients. We hypothesize that T-cell subpopulations undergo changes during adolescence. Hence, in an initial first step towards this hypothesis, we aim to establish the peripheral proportions and absolute cell counts of T-cell subpopulations in healthy individuals relative to their stage of pubertal maturation.

Methods

Study population

Healthy individuals were included as part of the ADOLESCE-NT study between 2012 and 2021. The ADOLESCE-NT study is a prospective, cross-sectional, observational, multicenter study designed to explore the factors contributing to the accelerated decline in renal graft function observed during adolescence^1–4.

All participants gave their written informed consent to participate in this study. The ADOLESCE-NT study was carried out in compliance with the principles of the declaration of Helsinki. This study follows the guidelines of, and is approved by the medical ethical committee of the Erasmus MC, Rotterdam, the Netherlands (protocol code MEC 2011-499).

Healthy participants, aged eight until 30 years, were recruited from the Erasmus MC Sophia Children’s Hospital, from the Erasmus MC, and the Erasmiaans Gymnasium (secondary school).Exclusion criteria were nephrological disease and precocious puberty, defined as the onset of puberty before the age of eight in girls and before the age of nine in boys. Individuals had one study visit after inclusion. The visit included an assessment of individual characteristics such as height, weight, medical history, tanner stage, and medication. An X-ray of the hand was performed to assess (bone) skeletal age. Pubertal maturation was assessed based on skeletal age and Tanner stage, and used to categorize participants into one of four pubertal stages (pre-, early-, late-, and post-puberty). Estrogens are known to influence T-cell subpopulations⁶. To account for hormonal variation, study visits for female participants using oral contraceptive were scheduled on the first day of starting a new blister packet. For females not using oral contraceptives, study visits were scheduled three to five days after the onset of their menstrual period.

Determination of pubertal stage

Pubertal maturation was assessed by subdividing healthy individuals into four different stages: pre-puberty, early-puberty, late-puberty, and post-puberty. In male participants, skeletal age and testicular volume, or Tanner genital (G) stage were assessed to determine their pubertal stage¹⁸. In female participants, skeletal age, Tanner breast (B) stage, and whether menarche had occurred were used as measures to determine pubertal stage¹⁹. The determination process is shown in Supplementary Fig. S1. In pediatric participants, Tanner stages were assessed by a physician, or through self-assessment with the help of Tanner stage images and an orchidometer (to estimate testicular volume). In participants older than 18 years Tanner stages were obtained through self-assessment. This approach allowed us to evaluate whether pubertal stage, as defined by Tanner score in combination with skeletal age, provides additional explanatory value beyond chronological age, given that established reference values for lymphocyte populations are based solely on age.

Lymphocyte immunophenotyping

Peripheral blood mononuclear cells (PBMC’s) were isolated from lysed, peripheral-drawn, venous blood samples within 24 h after sampling. Lymphocyte subpopulations were measured using an 8-colour flow cytometry staining procedure consisting of three tubes. The antibodies, fluorochromes, and antibody clones used are listed in Supplementary Table S1.

The following CD4+ and CD8+ T-cell subpopulations were defined in the first tube: naïve T-cells (Tnaïve) being CD27+ , CCR7+ and CD45RO-, central memory T-cells (Tcm) being both CCR7+ and CD45RO+ , Effector memory T-cells (TemRO) being CCR7- and CD45RO+ , and terminally differentiated effector memory T-cells (TemRA) CD27-, CCR7- and CD45RO-. In the second tube, T-cell receptor (TCR) γδ-cells were defined as TCRαβ- cells. TCRγδ-cells were further subdivided into Vδ1+ , Vδ2+ , and non-Vδ1 non-Vδ2 cells. Recent thymic emigrants (RTE) were defined in the third tube as being CD31+ and CD62Lhigh for CD4+ populations.

After addition of the reagents, PBMC’s were incubated for 10 min at room temperature, and subsequently washed and centrifuged for 5 min at 1455 rpm. After addition of FACSflow, PBMC’s were analyzed using a FACSCanto II machine using FACS Diva and Infinicyt software, version 10.0. The objective was to capture a total of 1 × 10E6 events per sample, after which lymphocytes and, subsequently, the CD4+ and CD8+ subpopulations within the lymphocytes were gated.

The proportions of lymphocyte subpopulations obtained from the flow cytometry analyses were converted into absolute counts by recalculating via a Trucount tube™. Whole blood was pipetted into a Trucount tube™, after which the reagents described in Supplementary Table S1 were added. After incubation at room temperature for 10 min, NH4Cl was added for lysing, the tube was vortexed and incubated for another 10 min at room temperature. Subsequently, the tube was analyzed using a FACSCanto II instrument. The proportions of CD3 populations (CD4+ , CD8+ , and TCRγδ) were then multiplied by the absolute CD3 count. In addition to the absolute counts of the CD3+CD19- T-cell population, absolute counts of NK-cells, defined as CD3-CD16+ CD56+ , and B-cells, defined as CD3- and CD19+ were determined. Gating and data analysis of the T-cell subpopulations was carried out by a single technician. The gating strategy is displayed in Supplementary Fig. S2.

CMV-status

Qualitative CMV-IgG measurements (DiaSorin, Liaison XL®) were done in all participants, as CMV-positivity is a known confounder that can particularly increase CD8 TemRA counts through prolonged stimulation of effector memory cells²⁰. Results were stratified based on CMV-IgG positive or negative status.

Analysis

Statistical analysis was conducted using SPSS version 28. Figures were created using R, version 4.3.2, and RStudio, version 2023.12.0, with ggplot2 package, version 3.5.1. For all analyses, the level of significance was set at 0.05. Lymphocyte subpopulation proportions and absolute counts were tested using the Kruskal–Wallis, followed by Dunn’s post hoc test to identify differences among the four stages of puberty: p-values were adjusted for multiple comparisons using Bonferroni correction. Regression analyses, using generalized linear models, were performed to compare parameter coefficients of pubertal stage, age and, sex as predictors of T-cell subpopulations counts and proportions. This approach allowed us to assess whether pubertal stage provides additional explanatory value beyond age given that established reference values of lymphocyte populations are all in relation to age. The normality of the data was assessed using the Shapiro–Wilk test. Non-normally distributed proportions and absolute cell counts were transformed into a normal distribution, as assessed by the Shapiro–wilk test. The specific transformation used for the values of separate variables is displayed in Supplementary Table S2. Generalized linear models with a normal distribution and identity link function for both categorical and continuous predictors were fitted. To obtain interpretable effect size estimates for categorical predictors, the model was re-run using the univariate general linear models(GLM) analysis under identical assumptions. Partial eta squared (η²) values are reported as measures of effect size for categorical predictors and Cohen’s f² for continuous predictors. Other effect sizes were estimated using the Cramér’s V for Chi-square tests and eta squared(η²) for Kruskall-Wallis test.

Results

Participant characteristics

Informed consent was obtained from 69 participants. Flow cytometry data was missing in three participants, who were therefore excluded from further analysis. The baseline characteristics of the included individuals are described in Table 1. Percentages of male and female participants did not differ significantly across the different pubertal stages (Χ²(3) = 1.807, p = 0.613, V = 0.165). When comparing age and pubertal stage, age appeared limited in the approximation of pubertal stage (Fig. 1). For example, without having any additional knowledge, an individual aged 15 years can either have an early-, late-, or postpubertal stage. Hence, in the rest of our analyses we focused on pubertal stages. The total percentage of CMV-positive individuals was 18.2% which, according to a chi-square test, did not differ significantly among pubertal stage (Χ²(3) = 0.209, p = 0.976, V = 0.056).

Table 1.

Baseline characteristics of participants.

Participants, n.		66
Males, n. (%)		28		(42.4)1
Females, n. (%)		38		(57.6)1
Age, years (mean, SD, SEM, 95%CI)		17.7		(6.3, 0.78, 16.2–19.2)
Height, cm (mean, SD, SEM, 95%CI)		168		(15.4, 1.90, 164.3–171.7)
Weight, kg (mean, SD, SEM, 95%CI)		62.1		(18.7, 2.30, 57.6–66.6)
BMI, (mean, SD, SEM, 95%CI)		21.6		(4.4, 0.54, 20.5–22.6)
CMV-positive, n. (%)		12		(18.2)2
		n. , (%male)		Age, (min–max)
Pubertal stage	Pre-puberty	10	(60)	(8.3–11.2)
	Early-puberty	8	(37.5)	(9.4–15.3)
	Late-puberty	6	(50)	(11.4–16.9)
	Post-puberty	42	(38.1)	(15.1–29.4)

Abbreviations: SD, standard deviation; SEM, Standard Error of the Mean; CMV, cytomegalovirus.¹, Χ2(3) = 1.807, p = 0.613, V = 0.165; ², Χ2(3) = 0.209, p = 0.976, V = 0.056.

Fig. 1.

Bar chart and scatterplot of pubertal stage in relation to age. Bars represent minimum, median, maximum age within each pubertal stage.

Subtle changes observed in the major lymphocyte populations among pubertal stages

As a first step in evaluating lymphocyte dynamics during puberty, the association of absolute cell counts and proportions of the major lymphocyte populations (T-cells, B-cells, and NK-cells) with pubertal stage was determined (Fig. 2). Only the proportion of B-cells was significantly associated with pubertal stage (p = 0.033), with a significantly lower proportion during post-puberty compared to late-puberty (β = 0.167, SE = 0.0623, p = 0.007). A partial eta squared of 0.124 was obtained from the univariate GLM analysis, indicating a moderate to large effect size. However, no significant difference in absolute B-cell counts between the different pubertal stages was observed (p = 0.563), suggesting that the significant difference observed in B-cell proportions is either due to subtle changes in T-cells or NK-cells, or is caused by changes in cell populations that were not studied. Age had a similar effect with the proportion of B-cells showing a slightly decreasing trend with increasing age (β = − 0.007, SE = 0.0029, p = 0.01, f² = 0.1).

Fig. 2.

Lymphocyte proportions and absolute cell counts over the course of puberty. (A) Proportions of T-cells (CD3 +), B-cells (CD19 +), and NK-cells (CD3-/CD16 + /CD56 +) within the total lymphocyte population over the course of puberty. Exact values can be found in Supplementary Table S3. (B) Absolute cell counts of T-cells (CD3 +), B-cells (CD19 +), and NK-cells (CD3-/CD16 + /CD56 +) over the course of puberty. Exact values can be found in Supplementary Table S4. Bars represent, 25th percentile, median, 75th percentile.

Pubertal stage-specific changes observed in CD4 and CD8 Tnaïve and CD4 TemRO populations

Next, we analyzed proportions and counts of different maturation stages of CD4 and CD8 T-cells among the pubertal stages. As shown in Fig. 3A, proportions of CD4 Tnaïve cells were higher during pre-puberty and early-puberty compared to post-puberty (β = 16.62, SE = 3.89, p < 0.001; β = 19.06, SE = 4.09, p < 0.001). A large effect size was observed (partial η² = 0.345), based on the univariate GLM analysis. In contrast, CD4 TemRO proportions were significantly higher during post-puberty (β = − 1.48, SE = 0.39, p < 0.001; β = − 1.63, SE = 0.40, p < 0.001). Based on the univariate GLM analysis, a similar large effect size (partial η² = 0.290) was observed. In line with CD4 Tnaïve proportions, the absolute count of CD4 Tnaïve cells was also significantly associated with pubertal stage(p = 0.004), but not with age (p = 0.19, f² = 0.027). During early-puberty specifically, there was a higher CD4 Tnaïve cell count compared to post-puberty (β = 0.29, SE = 0.09, p < 0.001)(Fig. 3B). The absolute count of CD4 TemRO was lower during pre-puberty compared to post-puberty (β = − 0.29, SE = 0.08, p < 0.001). The effect size of pubertal stage on absolute CD4 Tnaïve cell count and absolute CD4 TemRO count was large (partial η² = 0.171; partial η² = 0.191), as obtained from the univariate GLM analysis. Age as a predictor had a similar negative effect on the proportions of CD4 Tnaïve and positive effect on CD4 TemRO populations, as can be seen from the regression lines (β = − 1.174, SE = 0.22, p < 0.001, f² = 0.451; β = 0.12, SE = 0.02, p < 0.001, f² = 0.522) (Fig. 3E). The absolute CD4 TemRO cell count also increased with age, illustrated by the regression line in Fig. 3F (β = 0.02, SE = 0.004, p < 0.001, f² = 0.471). Similar patterns were observed in CD8 Tnaïves and TemROs (Fig. 3C, D). However, no significant differences among the different pubertal stages were observed.

Fig. 3.

CD4 and CD8 Tnaïve, and TemRO proportions and absolute cell counts over the course of puberty. (A) Proportions of Tnaïve and TemRO within the CD4 population over the course of puberty. (B) Absolute cell counts of CD4 Tnaïve and TemRO over the course of puberty. (C) Proportions of Tnaïve and TemRO within the CD8 population over the course of puberty. (D) Absolute cell counts of CD8 Tnaïve and TemRO over the course of puberty. (E) Proportions of Tnaïve and TemRO within the CD4 population in relation to age. (F) Absolute cell counts of CD4 Tnaïve and TemRO in relation to age. (G) Proportions of Tnaïve and TemRO within the CD8 population in relation to age. (H) Absolute cell counts of CD8 Tnaïve and TemRO in relation to age. Bars represent, 25th percentile, median, 75th percentile. Dotted line represents regression line. Exact values displayed in Supplementary Tables S3 and S4.

Higher CD4 RTE subpopulation counts in female individuals during puberty

The pattern of a rise in absolute Tnaïve cell counts in early-puberty was also seen in the absolute CD4 RTE cell counts. Absolute CD4 RTE counts differed significantly among pubertal stages, with Dunn’s test revealing a significant higher count in early-puberty compared to post-puberty (H(3) = 9.683, p = 0.021, p = 0.025, η² = 0.107). Furthermore, both pubertal stage and sex were significantly associated with CD4 RTE proportions and absolute cell counts (pubertal stage: p = 0.002; p = 0.015; sex: p = 0.024; p = 0.005) (see Table 2). Moderate to large effect sizes of pubertal stage (partial η² = 0.187, partial η² = 0.136), and moderate effect sizes of sex (partial η² = 0.072; partial η² = 0.105) on CD4 RTE proportions and absolute cell counts were observed in the univariate GLM analysis. Illustrated in Fig. 4A, B, females had higher proportions and absolute counts compared to males (β = − 8.615, SE = 3.82, p = 0.024; β = − 0.2, SE = 0.071, p = 0.005). The pattern in proportions and absolute cell counts between males and females during puberty differs visibly. However, when both pubertal stage and sex were tested in a multivariate analysis there was no significant interaction. In early-puberty, the bar representing absolute CD4 RTE cell counts for male individuals (n = 3) visually collapses due to low variability.

Table 2.

Univariate analyses of pubertal stages, age, sex, and gender in relation to proportions and absolute cell counts.

Cell proportions					Absolute cell counts
	Puberty	Age	Sex	CMV		Puberty	Age	Sex	CMV
CD3T	NS	NS	S	NS	CD3	NS	NS	S	NS
CD19T	S	S	NS	NS	CD19	NS	NS	NS	NS
CD16CD56T	NS	NS	NS	NS	CD16CD56	NS	NS	NS	NS
CD4 T-cells1	NS	NS	NS	NS	CD4 T-cells	NS	NS	NS	NS
CD4 Tnaïve*	S	S	NS	NS	CD4 Tnaïve	S	NS	NS	NS
CD4 central memory*	NS	NS	S	NS	CD4 central memory	NS	NS	S	NS
CD4 TemRO*	S	S	NS	S	CD4 TemRO	S	S	NS	NS
CD4 TemRA*	S	S	NS	S	CD4 TemRA	S	S	NS	S
CD8 T-cells1	NS	NS	NS	NS	CD8 T-cells	NS	NS	NS	NS
CD8 Tnaïve*	S	S	NS	S	CD8 Tnaïve	NS	S	NS	NS
CD8 central memory*	NS	NS	NS	NS	CD8 central memory	NS	NS	S	NS
CD8 TemRO	S	S	NS	S	CD8 TemRO	NS	S	S	S
CD8 TemRA*	NS	NS	NS	S	CD8 TemRA	NS	NS	NS	S
TCRγδ T-cells2	NS	S	NS	NS	TCRγδ T-cells	NS	NS	NS	NS
Vδ1+ TCRγδ T-cell*	S	S	NS	NS	Vδ1+ TCRγδ T-cell	S	S	NS	NS
Vδ2+ TCRγδ T-cell*	S	S	NS	S	Vδ2+ TCRγδ T-cell	NS	NS	NS	NS
TCRγδ T-cell Vδ1-, Vδ2-*	S	S	NS	S	TCRγδ T-cell Vδ1-, Non Vδ2-	S	S	NS	NS
CD4 CD31+CD62Lhi (RTE)*	S	S	S	S	CD4 CD31+CD62Lhi (RTE)	S	S	S	NS

^T, Proportion of cell-type within Trucount Tube™; ¹, Proportion of cell-type within first tube; ², Proportion of cell-type within second tube; ³, Proportion of cell-type within third tube; ^*, Proportion of T-cell-subtype within their respective population (i.e. CD3, CD8, and TCRγδ); -, omitted because of duplicate results; Not Significant (NS), p > 0.05; Significant (S), p ≤ 0.05. Bolded and underlined results indicate the superior predictor between puberty and age based on goodness-of-fit statistics.

Fig. 4.

CD4 RTE proportions and absolute cell counts over the course of puberty by sex. (A) Bars of the proportions of CD31/CD62L^high cells within the CD4 population over the course of puberty. Bars represent, 25th percentile, median, 75th percentile. (B) Bars of absolute cell counts of CD4/CD31/CD62L^high cells over the course of puberty. Bars represent, 25th percentile, median, 75th percentile.

Converging patterns in TCRγδ T-cell populations over the course of puberty

Subsequently, we focused our analysis on TCRγδ T-cells. Absolute cell counts of Vδ1+ and Vδ2+ cells were compared at every pubertal stage and were found to differ significantly from each other in pre-puberty, late-puberty, and post-puberty (W = 45.00, p = 0.008; W = 20.00, p = 0.046; W = 901.00, p < 0.001). However, no significant differences were observed during early-puberty (W = 22.00, p = 0.176). When the TCRγδ Vδ1 and TCRγδ Vδ2 subpopulations were analyzed separately, a significant difference in both the proportions and the absolute cell counts of TCRγδ Vδ1 was found (H(3) = 17.428, p < 0.001, η² = 0.24; H(3) = 14.131, p = 0.003, η² = 0.186). Dunn’s test revealed a significant higher proportion in pre-puberty and especially early-puberty (p = 0.021; p = 0.018). The proportion of Vδ1+ followed the pattern of an increase during early-puberty followed by a decrease in post-puberty (Fig. 5A). In contrast, a more linear pattern was observed in absolute Vδ1+ counts (Fig. 5B), with a significant difference between late- and post-puberty counts (p = 0.037). In the Vδ2+ population proportions differed signficantly among pubertal stages (H(3) = 14.390, p = 0.002, η² = 0.190), with lower proportions seen in pre-puberty and early-puberty compared to post-puberty (p = 0.041; p = 0.035). No significant difference among the pubertal stages was found in absolute Vδ2+ counts. Age had a similar effect as pubertal stage on the Vδ1+ and Vδ2+ cell proportions (Fig. 5C, D). Compared to pubertal stage, age was less strongly associated with Vδ1+ proportions, a more strongly associated with Vδ2+ proportions, and a comparably associated with absolute Vδ1+ counts (Table 2).

Fig. 5.

TCRγδ T-cell population proportions and absolute cell counts over the course of puberty. (A) Proportions of Vδ1+ , and Vδ2+ cells within the TCRγδ T-cell population over the course of puberty. Exact values can be found in Supplementary Table S3. (B) Absolute cell counts of Vδ1+ TCRγδ, and Vδ2+ TCRγδ cells over the course of puberty. Exact values can be found in Supplementary Table S4. (C) Scatter plot with dotted regression line of the proportions of Vδ1+ and Vδ2+ cells within the TCRγδ population in relation to age. (D) Scatter plot with dotted regression line of absolute cell counts of Vδ1+ TCRγδ and Vδ2+ TCRγδ cells in relation to age. Bars represent, 25th percentile, median, 75th percentile.

Pubertal stage, age, sex, and CMV-status in relation to lymphocyte subpopulations proportions and absolute counts

Table 2 displays the results of univariate analyses on whether pubertal stage, age, sex, and CMV-status were significant predictors of a variable. Exact coefficients and p-values can be found in Supplementary Table S5. If sex was a significant predictor, all proportions and absolute cell counts, regardless of cell type, were significantly higher in female participants (e.g. Figure 4). When CMV-status was a significant predictor, CMV-positive individuals had higher proportions of CD4 TemRO and TemRA, CD8 TemRA and TemRO, and TCRγδ T-cell non-Vδ1, non-Vδ2 cells. In absolute cell counts CMV-positive individuals had higher CD4 TemRA, and CD8 TemRO and TemRA counts. In CMV-negative individuals higher proportions of CD8 Tnaïve, Vδ2+TCRγδ, CD4 RTE, CD8 RTE, and TCRγδ RTE cell populations were found. When parameters were used in multivariate testing, no significant interactions were observed.

Values of proportions and absolute cell counts over the course of puberty are found in Supplementary Table S3 and Supplementary Table S4. These proportions and absolute cell counts are intended as descriptive reference values.

Discussion

In this study, we discovered changes in naïve T-cell subpopulations related to pubertal stage. While age-related pediatric reference values of different lymphocyte populations have been established by a few other studies, to the best of our knowledge, our study is the first to define descriptive pubertal-stage-specific reference values^13–15. This is an important step towards our hypothesis that T-cell subpopulation proportions and absolute counts change during adolescence influenced by the pubertal endocrinological shift.

Age is often used in other studies to approximate puberty. The onset and progression of puberty is different in every child^16,17, with children of the same age showing great variety in puberty progression as was observed in our own results. Our approach categorized puberty by using the well-established Tanner scale^18,19, as well as skeletal age to define four pubertal stages. Although our approach still simplifies the inherently continuous and gradual process of pubertal maturation, we argue that it provides a more biologically grounded framework for interpreting pubertal maturation than reliance on chronological age alone.

Increases in absolute counts and proportions of the more differentiated T-cell subpopulations can be seen as an enrichment of immune memory. This memory develops through exposure of an individual to novel antigens. Our results displayed a linear pattern in both CD4 and CD8 TemROs over the course of puberty, in contrast to the peak observed in Tnaïves during early-puberty. Our results corroborate previous studies which showed similar patterns in relation to age^13–15. These results suggest that although puberty may transiently boost Tnaïve counts, this effect may be insufficient to counterbalance the ongoing differentiation into effector memory T-cells. Which results in the observed decline of Tnaïve to TemRO cell count ratio over the course of puberty. Importantly, the concept of a “more active” immune system during puberty does not necessarily require an accelerated increase of TemRO subpopulations. Instead, functional activation, such as increased cytokine production, proliferation, or responsiveness to antigens, can reflect heightened immune activity even in the context of a steady, linear expansion of TemRO subpopulations. This functional activation may be further enhanced by increased social interactions and antigen exposure during adolescence, which provide additional stimuli for immune responsiveness without necessarily accelerating the absolute increase of TemRO subpopulations.

RTEs and Tnaïves peak during early-puberty and decrease towards post-puberty. Previous studies reported a similar decrease in naïve T-cell subpopulations towards adulthood in relation to age^13–15, but did not observe a distinct peak at a specific age. Our observation of a temporary increase in CD4 RTE and Tnaïve during early-puberty underscores the importance of incorporating pubertal stage, in addition to age, when establishing reference values for T-cell subpopulations. Secondly, TCRγδ proportions change in favor of Vδ1-cells during early-puberty. This is similar to the Vδ1-Vδ2 ratio in early childhood, where the Vδ1 phenotype is more dominant. When looking at absolute counts, this change seems mostly driven by expansion of the Vδ1-population. This expansion could lead to an enhanced engagement of the immune system when encountering stress signals, such as tissue damage. Thirdly, we expected higher proportions and counts of CD8 TemRA in CMV-seropositive individuals because of the prolonged stimulation of TemROs. The percentage of CMV-seropositivity in our population was below the average seropositivity in the Netherlands adjusted for age²¹. An explanation for this could be that the population we recruited from had an above average social economic status. Although the number of CMV-seropositive individuals was small, we did observe higher proportions and counts of CD8 TemRA in these individuals.

In addition to statistically significant findings, effect sizes were estimated to assess the relevance of the results. A large effect size was observed when pubertal stage significantly predicted T-cell proportions or counts, indicating that pubertal development has a substantial effect on the T-cell compartment. Similarly, age had a large effect size in relation to the observed changes in T-cell subpopulations. When effect sizes were converted and compared for CD4 Tnaïve and TemRO proportions for example, pubertal stage had a larger effect on CD4 Tnaïve proportions (pubertal stage: f² = 0.527 vs. age: f² = 0.451). However, the effect size of age was bigger on CD4 TemRO proportions (pubertal stage: f² = 0.408 vs. age: f² = 0. 0.522). This suggests that changes in specific segments of the T-cell compartment may be more strongly influenced by chronological age, while others may be influenced more by the processes of going through puberty.

Our findings thus suggests that going through puberty possibly causes a more active immune system. In turn, this more active immune system could be related to the observed early graft loss in adolescents. While the findings of our current study only concern healthy individuals, our future research will focus on the maturation of the immune system in adolescent end-stage renal disease patients and KTx-recipients. Additionally, our findings could also be extended to other immune related disease, such as autoimmune diseases, that have an increase in onset and severity during puberty.

We hypothesized that pubertal maturation causes a shift in the immune system. One key aspect of pubertal maturation are the endocrinological changes occurring during this period. Although providing a causal relationship between hormones and the immune system is beyond the scope of this article, puberty does coincide with a steep rise in estrogens, androgens, and within the GH-axis^22,23. Higher levels of sex-hormones persist after puberty, while the GH-axis is downregulated. The peak in absolute CD4 Tnaïve counts during early-puberty is likely caused by an increase in thymic output, indicated by the observed increase in CD4 RTE counts during that same pubertal stage. Similarly, there is a distinct peak during puberty in the GH-axis, which has proliferative effects on thymus size²⁴. Furthermore, CD4 RTE counts differed significantly between male and female participants, potentially resulting from differing levels of androgens and estrogens between the groups. Both androgens and estrogens decrease thymic size and output^25,26 through different mechanisms of action²⁷. However, an important difference is the oscillation of estrogen levels resulting from the menstrual cycle. The combination of this fluctuation and observed effects on T-cell maturation and function, could explain the sex-based difference in T-cell counts, and subsequent differences in immune activity and pathology. Additionally, T-cell differentiation and function can be influenced through epigenetic mechanisms caused by sex hormones. Hormonal signals can induce differential methylation of genes involved in immune regulation. Subsequently, this would lead to changes in gene expression that modulate the balance between RTE, Tnaïve, and TemRO cell populations, hereby contributing to the shifts in T-cell subpopulations observed during adolescence^28,29. Further research should therefore focus on the specific effect of changes in distinct hormones during pubertal maturation on T-cells and immune activity.

A limitation of our study is that participants were not equally distributed among the different stages of puberty. A limited number of children in early- and late-puberty was included, resulting in small group sizes for these stages. To increase power, the number of pubertal stages could be brought down to three stages. However, as we showed in our results, crucial puberty-specific changes would then have been missed. Another solution would be to have multiple measurements per participant. This approach would also allow the observation of changes within the T-cell compartment of a single individual. However, as puberty progression differs greatly among individuals^16,17, including each individual in every pubertal stage would be challenging and resource intensive.

In conclusion, this study demonstrates higher levels of Tnaïves during early-puberty and a shift to a more experienced and active immune profile in post-puberty in healthy individuals. While further studies are needed to establish causal relations between puberty and the immune system, our study is the first to give an impression of the changes that happen in the immune system of adolescents.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A.H.M.B., A.W.L., E.A.M.C., F.H.M.P., and H.d.J. conceived and designed the study. A.A.v.d.E. was involved with virological considerations and D.v.d.K. was involved with endocrinological considerations. Data was collected by A.M.T., F.H.M.P., H.d.J., and K.M.E.M. Experiments were designed and conducted by A.W.L.,and K.C.H. Statistical analysis was performed by H.d.B. and interpreted by A.W.L., F.H.M.P., H.d.B., H.d,J., and M.H.G.B. The manuscript text was drafted by H.d.B., H.d.J., and F.H.M.P., and an initial draft was critically reviewed by A.W.L., A.M.T., E.H.H.M.R., and M.H.G.B. All authors reviewed and approved the final manuscript.

Funding

The ADOLESCE-NT study was initiated in April 2012, with funding provided by “Exercising for Sophia” (Sporten voor Sophia) (Step 2011/project 982), an initiative supported by the “Friends of Sophia Foundation” (Stichting Vrienden van het Sophia), a charitable foundation that raises funds to support the Erasmus MC-Sophia Children’s Hospital.

Data availability

The data supporting our findings is available from the corresponding author upon reasonable request. Due to institutional restrictions, the entire data set cannot be made publicly available.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

In order to participate in the ADOLESCE-NT study, all subjects gave their fully informed, written consent. The study was carried out in compliance with the principles of the Declaration of Helsinki, and follows the guidelines of and is approved by the medical ethical committee of the Erasmus MC, Rotterdam, the Netherlands (protocol code MEC 2011-499).