Reference values for T and B lymphocyte subpopulations in Turkish children and adults

Background/aim Established reference values are critical for the interpretation of immunologic assessments. In particular, the proportion and absolute counts of T- and B- cell subpopulations are subject to change with age and ethnicity. We aimed to establish age-specific reference values for lymphocyte subsets using updated immunophenotyping panels. Materials and methods We studied a total of 297 healthy Turkish subjects aged 0 to 50 years, stratified into major age brackets in a cluster factor of 10 per age-group. The predetermined age intervals contained randomly allocated participants enrolled over a period of 6 months, who were homogenously distributed by sex. We analyzed a complete blood count test and simultaneously with detailed immunophenotyping enumerated the percent and absolute cell counts of lymphocyte subsets. Results The percentage and absolute counts of lymphocyte subsets show a marked surge across the age-span. T helper, T cytotoxic, and the natural killer cell numbers were increasing from birth until 6 months, followed by a gradual decrease thereafter. B cell numbers were rising until 2 years, followed by a gradual decrease for the upcoming years, accompanied by a steady expansion of unclass-switched- and class-switched- B cells. Conclusion We provide updated extensive reference intervals for lymphocyte subpopulations in Turkish people.


Introduction
The determination of cell-surface markers that define functional subsets of lymphocytes has immensely contributed to the understanding of the human immune system. The utilization of flow cytometry in the clinical lab and applications of immune-based staining to phenotype the major subsets have improved the diagnostic precision of the immunologists with regards to immunological and hematological disorders [1][2][3][4][5][6][7][8][9][10][11]. Lymphocyte subpopulation analysis studies are broadly used for diagnosis and classification of over 430 different inborn errors of immunity as well as monitoring the treatment outcome [12]. The distribution of lymphocyte subpopulations in a variety of immune-mediated diseases is a recent sparkling area of research, but the variability of normal ranges requires well-established reference values.
Lymphocyte subset counts and percentages are significantly influenced by different environmental factors, antigen stimulation, ethnicity, vaccination schedule, and the characteristics of the flow cytometry laboratory [1][2][3][5][6][7][8][9][10][11]13]. Further, immune markers are in evolution and require updates to reflect the current understanding of lymphocyte biology. Therefore, reference determination is a dynamic process, in which population-specific references should be periodically updated to account for this evolution. Former studies have successfully determined the composition of T-and B-cell pools, yet few of them conducted a systematic approach to determine the distribution of each subset from birth to the entire childhood span [1][2][3][5][6][7][8][9][10][11]. The most recent reference determination study for the Turkish population was published in 2004, however, new immunologic subsets have been defined and implemented into medical practice since then, warranting updated investigation into the subject [3].
The current study establishes age-dependent reference values of lymphocyte subsets for both children and adults, with an updated panel that expands the repertoire of the examined subsets over the former work. We think that our results would be useful for both immunology and allergy specialists as well as physicians during the evaluation of primary and secondary forms of immune deficiencies.

Subjects and samples
This is a cross-sectional study involving healthy children and adults. Peripheral blood samples were provided from children aged 0 through 18 years who were seen in our general pediatric clinics, well-child outpatient clinics (<5 years of age). The cord blood was obtained from newborns who were delivered in the Marmara University Research and Training Hospital's Obstetrics Department. The adult volunteers who provided blood samples were the family members of pediatric participants. Ethical approval was obtained from Marmara University Medical Faculty Clinical Researches Ethics Committee with the number of 09.2015.239/70737436-0.50.06.04. Eligibility criteria included: any children or adults in good health; full-term newborn (39-41 gestational); children within the normal range of growth indices (in between 3-97 percentile). The eligibility was assessed through history taking and physical examination to exclude any current or chronic diseases or conditions that may interfere with the results of the study. The volunteers filled out a questionnaire concerning their health and they were considered as clinically healthy if they did not have fever, respiratory or gastrointestinal disease during the last 2 weeks, no medication use other than vitamins within the last 2 weeks. Individuals with chronic diseases, active infections, or individuals who are taking any kind of medications, or those with immune-deficient relatives were excluded from the study. Their immunizations were questioned, and those who did not receive a vaccine in the last month prior to blood withdrawal were included. For the newborns, exclusion criteria also included chorioamnionitis, premature rupture of membrane, chronic illness of the mother, in utero malnutrition, Rh incompatibility, signs of infection, suspicion of hematological and neurological diseases, neonatal malformation, and dysmorphic features. Blood draw was performed following informed consent, either from the subjects themselves (adult subjects) or from the parents of the pediatric participants.
The minimum number of 30 attendees in each age stratum was determined based on previous studies and the law of great numbers [1][2][3][5][6][7][8][9][10][11] We aimed to study equal numbers from each sex for every age bracket. Out of 307 individuals enrolled in the study, 49.5% (147) were females and 50.5% (150) males; with the total number of subjects permissive for data analyses being 297.

Immunophenotyping
Peripheral blood from the antecubital region or dorsum of the hand or umbilical cord was taken in an amount of 0.5 mL (for complete blood count) and 2 mL (for flow cytometry) into dipotassium ethylenediaminetetraacetic acid tubes. For further contribution to the homogeneity of the sampling data, blinded specimens were studied with routine procedures for immunophenotyping and complete blood count analyses. Leucocyte differential including, absolute lymphocyte count was quantified by a complete blood count autoanalyzer (Unicell DxH 800, Beckman Coulter ABD). Relative percentages of T-, B-, and NKcells, and the T-and B-cell subsets were measured using flow cytometry, which was then used to calculate the absolute counts of respective immune subsets.  Table 1 [14]. Flow analysis was performed by BD Calibure and quadrant statistics were made by BD FACS CellQuest Pro software. Following the gating process performed by CD45 positivity and CD14 negativity for all leucocytes and the explained markers above for particular subsets, histograms indicated the percentage of different cell types after the correction. Absolute counts of different lymphocyte subtypes were calculated by (leucocyte count × % lymphocyte × % antibody positivity)/10,000 = Absolute number/μL formula [15].

Statistical analysis SPSS (Statistics Program for Social Sciences) program
(v: 24.0) and Microsoft Excel were used for storage and analysis of data. Our assumption for defining reference values is that almost all of our data should be embraced concerning the sample size, mean, standard deviation, and confidence interval; in this sense, tolerance interval gives limits with specified confidence that a specified proportion of the population lies within these limits, was chosen as a method of analysis [7,16]. The upper and lower limits of the two-sided tolerance interval were set so that it enclosed at least 90% of the population, with a 95% of confidence interval. Shapiro-Wilks normality test and descriptive statistical histograms were used to define the homogeneity of values. On the other hand, most of our data were positively skewed like the previous studies so the logarithmic transformation was used to reduce skewness and normalize distribution for better statistical purposes [7]. Yet, even after logarithmic transformation some values were not normally distributed so nonparametric test results were calculated to define limit values. Both original and retransformed logarithmic data along with parametric and nonparametric limits indicated by asterisks are given in Tables 2-6.
Statistical analysis to measure the significance of the differences between each age stratum for all 35 different peripheral blood lymphocyte subgroups was also performed. Homogeneity of variances was compared by Levene's test, which indicated variance heterogeneity (p < 0.05). A Welch one-way analysis of variance (ANOVA) was applied to all 35 different peripheral blood lymphocyte   Abbreviations: d: days, mo: months, yr: years, *Two-sided parametric tolerance interval of log-transformed data, **Two-sided nonparametric tolerance interval of log-transformed data.  Abbreviations: d: days, mo: months, yr: years *Two-sided parametric tolerance interval of log-transformed data **Two-sided nonparametric tolerance interval of log-transformed data. subgroups to verify the statistical difference among the 10 age groups. Further, to compute the multiple pairwise comparison between the individual means of each age group, for which the homogeneity of variance has not been met, Games-Howell post hoc test was applied.

Results
We determined the percentage of lymphocyte subsets and calculated absolute cell counts as presented in Tables 2-6. We calculated the median and the lower and upper limits for each immune subset for every age-stratum. One-way ANOVA analysis showed the age difference was statistically significant (p-values < 0.05) for all 35 different lymphocyte subgroups, except for the terminal effector memory CD4+ T cells. Further, the Games-Howell post hoc test results showed that most individual age groups were also statistically different in their mean values (data not presented).
The absolute cell counts of CD3+ T lymphocytes gradually increased from birth up until 6 to 9 months when a maximum is reached (age-specific median = 4469/mm 3 which then declines to values that are below the newborn levels ( Table 2). We found that the absolute cell counts of CD4+ and CD8+ T lymphocytes, as well as CD45RA+ naive T cells follow a similar pattern, with a gradual rise until 9 months followed by a linear reduction over the subsequent ages (Tables 2 and 3). When investigating the trends in relative percentages, the CD3+, CD4+, and CD8+ T cells all showed stable characteristics at ranges between 65.6%-74.9%, 35.8%-50.7%, and 17%-27.9%, respectively ( Table 2). CD19+ B cells doubled at 1 to 2 years of age interval which decreased by half after 2 to 5 years of age and remained stable throughout the life span. The relative percentage rates of the medians of B cells were distributed between 9.4%-22.7% (Table 2). Increasing NKcell counts until the first 6 months showed a continuous decline afterward and their relative percentages increased to adulthood, however, a significant change pattern was not noted ( Table 2).
The percentages of central memory and effector memory CD4+ T cells tended to increase during childhood in contrast to naive CD4+ and CD8+ T cells which decreased (Table 5).
Absolute cell counts of naive B cells tended to increase in the first 6 months followed by a decrease thereafter; this pattern was just in contrast to that of unclass-switched B cells. The absolute number of class-switched B cells increased in the first 2 years, then started decreasing; however, the percentage values for these cells were on a continuous rise. Autoreactive B cells (CD21lowCD38low) remained relatively stable throughout the lifespan at a fixed ratio of 2%-5% order. Autoreactive B cells that were relatively high at birth decreased in number till 40 days, then increased up to 2 years of age and showed a decrease again after 2 years of age (Table 4). CD3+ TCR alpha-beta cells increased through adulthood; by contrast, the CD3+ gamma-delta cells decreased in percentage. The percentage of double-negative T cells (DNT) remained in the range of 0.2%-5.5% with no age-specific pattern. The percentages of recent thymic emigrants (RTE) cells increased in the first 9 months but then decreased gradually.

Discussion
In this study, we established the normative reference ranges for peripheral blood lymphocyte subsets in a healthy Turkish population. The investigated immune subsets included an updated profile over the former studies that examined subjects in the same region. We think that these references will be useful for interpreting the immunologic subsets, that are often ordered for the evaluation of immune-mediated diseases.
The methodology we employed in this study was similar to previous work, in which lymphocyte subsets were investigated with respect to varying ages [1][2][3][5][6][7][8][9][10][11]. The trend in the absolute numbers of CD3+ T lymphocytes, cytotoxic T cells, helper T cells was similar to İkincioğulları et al. [3] and Schatorjé et al. [7] with a gradual rise during the first months of life, followed by a progressive decline in the subsequent years. This observation was in contrast to Comans-Bitter et al. [1] and Tosato et al. [10] studies. Major differences were observed among the subpopulations of these cells. We observed a general trend in the cells with regards to their differentiation into memory phenotype: while the percentage of the naive compartment of the helper and cytotoxic T cells showed a negative trend with increasing age, the memory cells were increasing in relative size as the subjects grew to adults. This is unsurprising as one could expect that immune maturation reflects exposure to microorganisms over time [10]. Likewise, naive CD4+ and CD8+ T cells decreased by half in the lymphocyte maturation cycle. In contrast to Schatorjé et al. 's findings [7], which showed stable values for the size of the central memory pool of CD4+ T cells, in our study, the percentages were threefold higher in adulthood compared to infancy period. The effector and terminal effector memory pools of CD4+ and CD8+ T cells showed a significant rise with advanced age. Possible lymphocyte development defects have not emerged only with deficiencies but also with unexpected growth of specific clones. Therefore, an exaggerated increase in memory T-cells would be an alarming finding for T-cell dysfunction [11]. Furthermore, all these subgroups are especially important for newborns too, as for which hypomorphic defects or the presence of maternal engraftments could easily mislead the diagnosis [11]. Therefore, the standard immunophenotyping studies must be carefully examined in the light of expected ranges.
DNT cells, which are related to autoimmunity and one of the diagnostic criteria for the autoimmune lymphoproliferative disorder, varied in the range of 0.2% to 5.5% among all age groups [17]. This finding contrasts with the former reports which describe greater variations [7,10]. RTE cell production correlates with thymic output and is an important marker for diseases such as severe combined immunodeficiencies, rheumatoid factor-negative polyarticular juvenile idiopathic arthritis, systemic lupus erythematosus, psoriasis, type 1 diabetes have defects in T-cell development in the thymus [7]. Similar to the findings of Schatorjé et al. [7] the relative size of this group was increasing with age followed by a gradual regression.
Within the B-cell pool, all of the examined subsets including class-switched B cells, unclass-switched B cells, and CD21lowCD38low B demonstrate a rise in number during the infancy period, followed by a decline after the age of 1-2 years. This differs slightly from previous observations, in which the major decline was observed in different age intervals. As for the relative percentages, naive cells had the highest numbers during the newborn period, then showed a gradual decline. By contrast, the memory B cell compartment, the cells with known functions in humoral immunity, autoimmunity, and chronic infectious diseases, demonstrated a gradual increase through life span [8].
In summary, we established the reference values for lymphocyte subsets stratified by age groups. Although this is not a national multicentre study, our study was conducted in one of the reference hospitals of a cosmopolitan city with different ethnic and cultural components. In light of previous literature and statistical analyses, we have reached a relatively large number of healthy participants and achieved our statistical goals with meaningful statistical reliability. We think the data presented can be readily adapted to medical practice.