CD26/dipeptidyl peptidase 4-deficiency alters thymic emigration patterns and leukcocyte subsets in F344-rats age-dependently

Abstract

As CD26 (dipeptidyl peptidase 4/DPP4) rapidly truncates incretins N-terminally, including glucagon-like peptide-1, DPP4-inhibitors have been developed for treatment of diabetes type 2. To some extent this is surprising, as CD26/DPP4 is also deeply involved in immune regulation. Long-term pharmacological studies are hampered by off-target inhibition of DPP4-homologues. Therefore, we studied the effects of genetic CD26/DPP4-deficiency by investigating blood, spleen and thymus leucocyte subpopulations of wild-type and CD26-deficient F344-rats at different ages. In young animals at 1 and 3 months of age, there were no differences in leucocyte subsets, while in older animals the T cell composition was changed significantly. From the age of 6 months onwards, reduced numbers of recent thymic emigrants and memory T cells, and consequently an increased amount of naive T cells were observed in CD26-deficient rats. In addition, the architecture of the thymus was altered, as observed by a reduced density of lymphocytes in the medulla. Furthermore, the number of proliferating cells in the thymus was decreased in CD26-deficient rats at a higher age. Moreover, CD26-deficiency resulted in markedly reduced numbers of B cells in later life. Additionally, an age- but not CD26-dependent increase of regulatory T cells and a decrease of natural killer cell numbers were detected in the blood and spleen. Our findings indicate an important role of CD26 in maintaining lymphocyte composition, memory T cell generation and thymic emigration patterns during immunosenescence, with possible implications for using DPP4-inhibitors.

Introduction

CD26 is a unique and evolutionary highly conserved type II transmembrane glycoprotein. Its main functions are peptidase activity, known as dipeptidyl-peptidase 4 (DPP4), interaction with the extracellular matrix and T cell co-stimulation (for review see: [1–3]). It is expressed widely on endothelia and epithelia, such as the kidney brush border and intestine, and interestingly also on immune cells, e.g. T cells, activated B and activated natural killer (NK) cells [4]. Up to 70% of human peripheral blood lymphocytes express detectable CD26 protein levels [5]. As CD26 is described as a functional collagen receptor involved in thymic T cell ontogeny as well as T cell activation [6], it also plays a role in thymocyte maturation. CD26 expression in the thymus is regulated tightly in rodents [7], but also in humans [8]. As thymocytes mature, CD26 expression is usually up-regulated [9]. The highest level of CD26 expression is reached in mature CD4 or CD8 single-positive T cells within the thymus [9]. However, the functional role of CD26 expressed on maturing thymocytes remains unclear.

The recent introduction of DPP4-inhibitors into clinics aims to enhance the endogenous insulin secretion in diabetes mellitus type 2 via elevated levels of glucagon-like peptide-1 and gastric inhibitory protein [10,11]. This might represent a double-edged sword as, apart from the metabolic benefit, the associated immunological effects of long term DPP4-inhibition on regulatory processes such as T cell maturation are not understood fully at this stage [3,12].

Until now, long-term data focusing on the immunological naive conditions derived from in vivo experiments were available neither in CD26-deficient rat models nor in CD26^−/− mice. Therefore, the purpose of this study was to characterize and quantify lymphocyte subsets with a battery of antibodies not only in the blood, but also in the thymus as a primary lymphoid organ, and in the spleen as an example of a secondary lymphoid organ, in order to define potential long term effects of chronic DPP4-deficiency with respect to age under non-challenged conditions.

Materials and methods

Experimental groups and treatment of animals

Female wild-type F344/Ztm rats (DPP4pos) and CD26-deficient rats F344/Crl(Wiga)SvH-Dpp4^m expressing a mutated CD26 protein and lacking DPP4-activity (DPP4-def) [13] were studied at 1, 3, 6 and 12 months of age in groups of 10. Rats were maintained with food and water available ad libitum in a separated minimal-barrier sustained facility at the Central Animal Facility (Ztm) of the Hannover Medical School and monitored microbiologically according to FELASA recommendations [14]. The government of Lower Saxony, Laves, Oldenburg, Germany, approved all research and animal care procedures.

Dissection of animals

The animals were dissected under isoflurane anesthesia. Briefly, the animals were killed by aortic exsanguination, thereby collecting ethylenediamine tetraacetic acid anti-coagulated blood samples for fluorescence activated cell sorter (FACS) analysis and cytospins. The spleen and thymus were removed, weighed, split and used for FACS analyses as well as histological examinations.

Analysis of the total cell numbers

Cell numbers of whole blood leucocytes were determined using a Coulter Counter (Beckman Coulter, Inc., Fullerton, CA, USA). Leucocyte numbers in the spleen and thymus were determined by staining with Tuerk's solution (Merck, Darmstadt, Germany) in a Neubauer counting chamber.

Analysis of leucocyte subsets by flow cytometry

Whole blood, splenic and thymic leucocytes were processed as described previously [15]. Analyses were run on a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA, USA) using commercially available monoclonal antibodies (mAbs): granulocytes were identified by forward/side-scatter and confirmed with RP1 antibody. CD26⁺ cells (mAb Ox61 and 5E8), B lymphocytes (mAb Ox12), CD4⁺ T cells (mAb W3/25/mAb R73), CD8⁺ T cells (mAb Ox8/mAb R73) and monocytes (mAb ED9/mAb W3/25) were identified as described previously [16]. NK cells were identified as CD161^high (mAb 10/78^high) CD4⁻ cells [17]. Recent thymic emigrants (RTE) (CD90⁺CD45RC⁻), memory T cells (CD4⁺CD90⁻CD45⁻) and CD4⁺ naive T cells (CD90⁻CD45RC⁺) were identified using a combination of the mAbs W3/25, Ox22 and Thy1 with the gating strategy published previously by Luettig et al.[18]. Regulatory T cells (T_regs) were identified as CD4⁺ (mAb W3/25), CD45RC⁻ (mAb Ox22) and CD25^high (mAb Ox39) cells and the results were confirmed later by intracellular staining for T_reg-specific forkhead box P3 (FoxP3) using a commercially available kit according to the manufacturer's instructions (Biolegend, München, Germany). All extracelluar antibodies were purchased from Serotec (Duesseldorf, Germany).

Analysis of leucocyte subsets by cytospins

Cytospins were prepared and analysed as described previously [15].

The DPP4 enzymatic activity

In order to determine plasma DPP4 enzyme activity, a microplate-based fluorescence assay was applied as described previously [19]. Briefly, the release of 4-nitroaniline from a substrate [glycyl-prolyl-4-nitroaniline (GPpNA) × HCl] cleaved by DPP4 was monitored at 405 nm and 37°C using the PowerWave XS spectral photometer (BioTek Instruments, Bad Friedrichshall, Germany).

Semiquantitative histology of the thymi

For histology, 10-µm thick sections were used. After dehydration and staining with Giemsa solution and differentiation in 100% methanol for 15 min and 96% ethanol (250 ml + 30 beads ice-acid) for 8 min, histological evaluation was carried out using a light microscope (Nikon Eclipse 80i). In order to quantify medullary thymocytes, a stereo investigator system (MicroBrightField, Inc., Williston, Vermont, USA) was used to define areas of interest in representative medullary slices. Thereafter, thymocytes were marked on screen followed by calculation of thymocyte counts/mm².

Detection of apoptotic and proliferating cells in the thymi via TdT-mediated biotin–dUTP nick-end labelling and Ki-67

A standard TdT-mediated biotin–dUTP nick-end labelling (TUNEL) method was employed on sections of the thymi to detect the fragmented nuclear DNA associated with apoptosis, as described previously [20]. For the detection of proliferating cells, a Ki-67 rabbit mAb (Epitomics, Burlingame, CA, USA) immunostaining (2 h incubation at room temperature in 2% bovine serum albumin, 0·3% TritonX in 0·1 M phosphate-buffered saline) was applied. For quantitative analyses of TUNEL or Ki-67-positive cells all tissue sections were scanned under low magnification (2×), and areas with positive stain were chosen for further analysis. All TUNEL⁺ or Ki-67⁺ cells within a grid on the ocular lens were counted (Nikon 10·0×; grid 0·75 × 0·75 mm = 0·5625 mm²/grid, using a Nikon PlanAPO VC objective, ×10, NA = 1·0). Each thymus was sectioned at three randomly chosen non-adjacent levels. From each level, three sections were evaluated. On average, 30 grids per section were examined (i.e. 0·5625 mm²/grid × 30 grids × 3 sections × 3 levels) resulting in a thymic area of 1·52 cm² per animal. The number of positive cells as well as total number of cells was determined by visual inspection and a resulting percentage value for TUNEL⁺ and Ki67⁺ cells calculated providing the apoptotic body index or proliferative body index. The sections were scored by two investigators blinded to the genotype of the rats.

Statistical analysis

Statistical analysis was performed using two-factorial analysis of variance (anova) for repeated measurements (genotype as between-subject factor, increasing age as within-subject factor) in order to test for age-dependent changes. This was followed by one-way anova split by age, if appropriate, and followed by Fisher's tests for protected last significant differences as a post-hoc test. All data are given as arithmetic mean ± standard error of the mean. P-values of ≤ 0·05 were considered significant and indicated by asterisks in the figures.

Results

CD26-deficiency results in reduced weight gain

The body weight development of wild-type as well as CD26-deficient F344-rats was monitored at 1, 3, 6 and 12 months of age, revealing a reduced body weight gain in CD26-deficient animals (Fig. 1a). The body weight gain in wild-types was not accompanied by increases of DPP4 expression levels, as neither mean fluorescence on lymphocyte subsets (data not shown) nor plasmatic DPP4-activity showed age-dependent changes (Fig. 1b). CD26 expression on T cells was highest in the thymus compared with blood and spleen (Fig. 1c). Nearly all T cells in the thymus expressed CD26 (Fig. 1d), with no age-dependent changes (not shown). Circulating T cells in the blood still showed a very high frequency of CD26 expression (95%), while expression on splenic T cells dropped to about 80% (Fig. 1d). The anti-CD26 mAbs 5E8 (not shown) and Ox61 failed to recognize the mutant CD26 protein in the deficient rat substrain (Fig. 1d).

Fig. 1.

Body weights of wild-type (WT) (open squares) and dipeptidyl-peptidase 4 (DPP4)-deficient rats (black squares) develop parallel in the early life span but diverge significantly in older individuals (a) (n = 10). The DPP4-activity in the serum assessed via microplate-based fluorescence assay of WT (open squares) and mutant rats (black squares) showed no significant alterations over life (b). (c) Mean fluorescence of the anti-CD26 monoclonal antibody (mAb) Ox61 in the T cell receptor (TCR)⁺CD4⁺ population in different tissues of WT animals exemplified in animals at the age of 3 months. (d) Representative flow cytometry staining of thymus, blood and spleens of WT (left column) and DPP4-deficient (right column) rats aged 3 months. Plots are gated on TCR⁺ lymphocytes. For (c) and (d) no age-dependent differences could be observed (n = 10).

Age-dependent decrease of absolute leucocyte number, but no genotype-dependent differences

Determination of absolute leucocyte numbers and percentage composition in the blood, spleen and thymus revealed no CD26-dependent differences at any investigated time-point (data not shown). In the investigated time-span, mononuclear cells decreased in the blood from 66 ± 5% (1·7 ± 0·1 × 10⁶/ml) to 49 ± 4% (0·8 ± 0·1 × 10⁶/ml) (P < 0·001) and in the spleen from 75 ± 5% (3·4 ± 0·2 × 10⁶/mg) to 68 ± 3% (2·5 ± 0·4 × 10⁶/mg) (P = 0·08). In the thymus, however, the percentage of mononuclear cells remained constant at 93 ± 3%, but a severe decrease in absolute numbers from 15·1 ± 1·9 × 10⁶/mg to 1·7 ± 0·2 × 10⁶/mg became apparent (P < 0·001). No changes in the percentile composition of CD4 and CD8 single-positive or CD4/CD8 double-positive T cells were observed between groups (data not shown).

CD26-deficiency results in an altered composition of CD4⁺ T cell subpopulations

Although no differences in the absolute and relative number of T cells or their main two subsets CD4⁺ and CD8⁺ became apparent, CD4⁺ subpopulations (memory T cells, RTE, naive T cells and T_regs) of both substrains were investigated further as to a possible involvement of CD26 in T cell turnover. In the blood, the percentage of CD4⁺ T cells remained constant from the age of 3 months onwards, whereas the absolute number of this population decreased over time, but no differences with respect to genotype were observed (Fig. 2a and b).

Fig. 2.

CD4⁺ T cells in the peripheral blood (left column) and spleen (right column) across age were analysed by flow cytometry (n = 10). First, percentage contingent [left ordinate: solid lines; black solid squares WT, triangles dipeptidyl-peptidase 4 (DPP4)-deficient] of all CD4⁺ T cells of all mononuclear cells as well as the absolute cell numbers (right ordinate: broken lines; grey open squares WT, open triangles DPP4-deficient) are given (a, b). Consecutively, the percentage composition in blood and spleen of memory T cells, recent thymic emigrants (RTE), naive T cells (c, d) and T regulatory cells (T_regs) (e, f) are displayed. Calculated ratios of naive versus memory T cells numbers are given in the upper row for peripheral blood (g) and spleen (h) over all time-points investigated (n = 10). The dashed line in (g) indicates a 1 : 1 ratio. Therefore, this line depicts the turning point, when naive T cell numbers fall below the numbers of memory T cells, as can be seen at the age of 12 months in the blood. (i, j) Ratio of naive T cells versus RTE, analogously.

Amount of CD4⁺CD25^high T_regs increases during ageing

The CD4⁺CD45RC^low population includes memory T cells and T_regs, which were differentiated by additional staining for CD25 and FoxP3. The percentage of CD4⁺CD25^high T_regs in blood and spleen doubled from nearly 4% at the age of 1 month to nearly 8% at the age of 12 months (Fig. 2c and d). In the blood, their absolute number increased from 3 × 10⁴/ml to 4 × 10⁴/ml, whereas the overall lymphocyte number decreased as mentioned above. No differences in respect to genotype were observed (Fig. 2e and f).

CD26-deficiency results in a reduced memory T cell pool

Starting at the age of 6 months and becoming even more evident at the age of 12 months, rats lacking CD26 showed a significantly reduced percentage of CD4⁺CD45RC^low T cells in the blood and spleen (Fig. 2c and d). As there were no genotype-specific differences in the T_reg population, the reduction of memory T cells must account for the changes observed. In absolute numbers, this difference became even more apparent as memory T cells in the blood of 12-month-old wild-type animals were 2·6 ± 0·2 × 10⁵/ml compared with only about 60% (e.g. 1·6 ± 0·2 × 10⁵/ml) in DPP4-deficient animals.

CD26-deficiency results in lower numbers of RTE and elevated levels of naive T cells

To differentiate further additional CD4⁺ T cell subpopulations, RTE and naive T cells were investigated, as indicated by higher expression of CD45RC. At higher age, CD26-deficient animals showed a decrease in relative and absolute numbers (not shown) of RTE in the blood as well as in the spleen (Fig. 2c and d). By contrast, the relative and absolute numbers (not shown) of naive T cells in CD26-deficient animals were elevated significantly with a twofold increased number in the blood and threefold increased number in the spleen at the age of 12 months (Fig. 2c and d). In CD26-deficient animals, the ratios of naive T/memory T cells shifted towards the naive T cells in the blood as well as in the spleen (Fig. 2g and h). A similar pattern became apparent in the ratios of naive T cells/RTE. Again, in CD26-deficient animals the ratios were increased towards the naive T cells (Fig. 2i and j).

CD26-deficiency results in altered thymus architecture

Because the findings of lowered amounts of RTE may result from a differential maturation of T cells because of changes in CD26 expression in the thymus, a stereological approach was chosen to perform a semi-quantitative histological analysis of thymi. It was thus possible to show a clearly reduced density of lymphocytes in the medulla of CD26-deficient animals (Fig. 3a–h). Because of the vast amount of cortical thymocytes, these significant changes in medullar density did not become apparent in FACS analyses of the thymi. Already, at the age of 1 month, CD26-deficient thymi showed a significant reduction with 11 081 ± 1627 lymphocytes/mm² compared with 15 187 ± 2500 lymphocytes/mm² in wild-type animals (P < 0·02).

Fig. 3.

The micrographs in panel (a–h) show representative Giemsa-stained areas of thymi of both rat substrains at the age of 1 and 12 months, indicating a reduced density of thymocytes in dipeptidyl-peptidase 4 (DPP4)-deficient animals. In the upper row, an overview (bar = 500 µm) as well as a higher magnification (bar = equals 25 µm) of juvenile thymi of wild-type (WT) (a, b) and matched DPP4-deficient rats (c, d) is given. In the lower row, thymi of 12-month-old WT (e, f) and DPP4-deficient rats (g, h) are displayed. (i) Level of proliferation assessed by a Ki-67 immunostaining and calculated proliferative body index (PBI) over all time-points investigated in WT (white bars) and DPP4-deficient rats (black bars). Micrographs are representative stainings from rats aged 1 year. (j) TdT-mediated biotin–dUTP nick-end labelling (TUNEL) methodology was applied to measure the amount of apoptotic cells (e.g. apoptotic body index, ABI) and depicted in analogy to (i).

CD26-deficiency reduces thymocyte proliferation at higher age

The TUNEL⁺ cells undergoing apoptosis and proliferating Ki-67⁺ cells were found in all animals. TUNEL⁺ cells were present both in the medulla and cortex of both strains, while Ki67⁺ cells were located primarily in cortical areas and around the vessels (Fig. 3i). A trend towards a higher number of TUNEL⁺ cells were observed in CD26-deficient rats at 6 and 12 months of age (Fig. 3j). The percentage of Ki-67 positive cells in the thymus was significantly lower in CD26-deficient animals at the age of 12 months (Fig. 3i).

CD26-deficiency influences B cell levels in later life, but not NK levels and granulocytes

To study further the impact of CD26 on other lymphocyte populations expressing CD26, besides T cell subpopulations, B cells and NK cells were also investigated. B cells displayed a significantly reduced percentage and absolute number in the blood of CD26-deficient rats at the age of 1 year (Fig. 4a–d). The absolute number of B cells decreased over time, but no differences between CD26-deficient and wild type animals became apparent at earlier ages (data not shown). The number of NK cells showed no CD26-dependent differences, but decreased significantly in a linear fashion over age (Fig. 4e and f). For monocytes and NK T cells, CD26-dependent trends of elevated levels in the blood and spleen were found (not shown). A trend to reduced relative and absolute numbers of granulocytes was found at all time-points investigated (not shown).

Fig. 4.

B cells show reduced numbers in dipeptidyl-peptidase 4 (DPP4)-deficient animals at the age of 12 months in the relative number of peripheral blood mononuclear cells (% PBMC) as well as in absolute numbers (×10⁶/ml) in the blood (a, c), as identified by flow cytometry. In the spleen, the relative number of B cells in all spleen mononuclear cells (% SMC) is decreased (b), whereas differences in absolute numbers (×10⁶/mg) fail to be significant (d). At earlier time-points, no differences become apparent (n = 10). The percentage composition of mononuclear cells shows a significant CD26-independent decrease of natural killer (NK) cells over age. The white (WT) and black bars (DPP4-deficient) depict the percentage of peripheral blood mononuclear cells (% PBMC) in (e) and splenic mononuclear cells (% SMC) in (f).

Discussion

This is the first report demonstrating that CD26-deficiency affects the composition of lymphocytes, and in particular that of T cell subpopulations during immunosenescence in rats. At younger age, the immune phenotyping of naive rats revealed no differences related to CD26. However, in contrast, memory T cells and RTE were profoundly reduced and consequently levels of naive T cells were elevated at higher ages.

Regarding the situation in young animals, previous data in mice suggest that over-expression of CD26 interferes with transduction pathways needed for the maturation of T cells, as their homeostasis in peripheral blood is impaired and their subset distribution is altered [21]. In addition, knock-out experiments investigating only one time-point in mice showed genotype-dependent shifts in lymphocyte populations such as CD4⁺, NK and NK T cells, while in vitro studies revealed altered cytokine secretion, T cell-dependent antibody production and impaired immunoglobulin isotype switching of B cells in CD26-deficient mice [22]. Therefore, it is surprising that young rats expressing a mutant CD26 show the same immunological baseline as wild-type animals. In vitro investigation of functional aspects in the same F344-rat strain used as in the study at hand demonstrated that there are no differences between mitogen- or antigen-driven responses [23]. Lymphocyte proliferation as well as the ability to elicit secondary responses by means of immunoglobulin and cytokine production proved to be equally potent in CD26-deficient and wild-type rats in vitro[23]. We therefore conclude that up to an age of young adulthood this rat model of CD-26 deficiency is characterized by immunological equal baselines.

With regard to older animals, we found a CD26-independent increase of CD4⁺ T_reg numbers over age, replicating previous findings in humans [24]. Most intriguingly, however, impairment of CD26 has indeed a remarkable effect on T cell subpopulations, as the CD4⁺ pool is depleted of memory T cells as well as RTE and filled with naive T cells instead. An explanation is offered by one of the CD26 key functions, its co-stimulatory potency: CD26 on T cells interacts directly with antigen-presenting cells via Caveolin-1, resulting in an up-regulation of the co-stimulatory molecule CD86, which enhances the bond of the immunological synapse [25]. In the past, there has been a controversial debate as to what extent CD26 and its catalytic region are important for T cell co-stimulation [26,27]. Recent findings, however, demonstrate that CD26 is able to trigger direct T cell activation and proliferation via CARMA1 mediated NF-κB activation in the T cell in vitro[28]. Therefore, the clinical use of DPP4-inhibitors could be critical, as the catalytic centre of CD26/DPP4 is part of the linking site required for co-stimulation [25]. In consequence, the lack of memory T cells can be explained at least partly by the loss of an important pathway of memory T cell generation.

Furthermore, CD26-deficiency also results in a reduced number of RTE in older individuals, pointing to an involvement of the thymus where bone marrow-derived progenitor cells undergo maturation (for review see: [29,30]). The vast majority of cells in the thymus express CD26 (Fig. 1d and [5]), which is also thought to be a thymic maturation marker [8]. CD26 has been described as a mediator of lymphocyte migration through the thymus, where it is up-regulated as thymocytes mature and down-regulated on cells that undergo apoptosis [9,31]. Its enzymatic activity degrading immunomodulatory peptides is controlled ontogenically during T cell maturation in the thymus [9,32]. The thymus undergoes an age-dependent involution, but remains active up to a high age playing a central role in replenishing the peripheral T cell pool [33–35]. This function is impaired in DPP4-deficient rats, as the number of RTE is lowered in later life. Impressively, we found the architecture of the thymus altered much earlier, before peripheral differences become apparent.

Apart from effects on T cells, we found decreased B cell numbers in CD26-deficient animals only at higher age. On one hand, B cells express CD26 per se[36], which might effect B cell maturation. On the other hand, CD26 might have an impact on lymphoproliferation via its enzymatic DPP4-activity on its substrate neuropeptide Y (NPY), which we have shown previously to affect B cell numbers [17]. Because ageing plays an important role in NPY-mediated lymphoproliferation [37], it is tempting to link changes in NPY levels to the observed B cell reduction. However, linking age-dependent NPY-mediated lymphoproliferation to the observed effect is merely speculative at this point, and requires further investigation. The third lymphocyte population expressing CD26 are NK cells, which are also decreased in older animals. Interestingly, NK cells of older individuals are less capable of destroying tumour cells [38], and as we have shown previously DPP4-deficient F344-rats have blunted NK cell cytotoxicity against tumour cells [13,39], which possibly result in a higher likelihood of NK cell-dependent diseases in older individuals with impaired DPP4-function.

In conclusion, our findings point strongly to an altered immunosenescence in rats lacking DPP4-activity. Despite translational differences and a different set-up compared with pharmacological studies, our finding may also have implications for humans in the case of long-term treatment with anti-diabetics using the recently introduced DPP4-inhibitors. We therefore propose to monitor immunological parameters closely in patients receiving DPP4-inhibitor treatment.