Stana Tokić M.D., Department of Molecular Diagnostics and Tissue Typing, Osijek University Hospital, J. Huttlera 4, HR-31 000 Osijek, Croatia, phone: +38 531 514 312, fax: +38 531 514 206, e-mail: tokic.stana@kbo.hr

Introduction

Hashimoto’s thyroiditis (HT) is an organ-specific autoimmune disorder characterised by chronic thyroid inflammation, and subsequently, hypothyroidism [1]. Morphological and functional alterations in HT are predominantly mediated by T-helper (Th) 1 cytokines and cytotoxic T-lymphocytes through apoptotic cell death [2]. Recently, however, both CD4+CD25+ regulatory T-cells (Treg) [3–5] and interleukin IL-23-driven Th17 effectors [6–8] have been independently associated with thyroid inflammation, disease chronicity, and HT destructiveness.

CD28/T-cell receptor/CTLA4 complex controls T-cell homeostasis, tolerance, and autoimmunity in Hashimoto’s thyroiditis [9, 10]. In addition, many other molecules, such as CD45 protein tyrosine phosphatase (PTPase) [11, 12] and vitamin D receptor (VDR) [13, 14], cooperatively interact with the TCR complex to affect Th1/Th17/Treg processes central to the pathogenesis of HT. Nevertheless, their role has been comparatively less well established.

CD28 acts as potent regulator of T-cell mRNA splicing [15]. In addition, emerging data suggest that the identification of disease-specific splice variants may provide novel biomarkers and drug targets [16, 17]. CD45 PTPase activity is regulated by isoform-specific differential homodimerisation, and alternative splicing of exons 4–6 (A-C) within the CD45 extracellular domain results in the production of at least five different CD45 isoforms: CD45RABC (exons 4-6 included), CD45RAB, CD45RBC, CD45RB, and CD45R0 (exons 4–6 skipped). Naive T-cells express longer isoforms at higher levels, while activated and memory T-cells express shorter isoforms (CD45R0) [18]. The alternatively spliced CD45 variants differentially homodimerise in primary T-cells, with the smallest isoform (CD45R0) being able to dimerise more efficiently and rapidly than the larger isoforms [19]. The resulting CD45 dimers have significantly reduced phosphatase activity; thus, the splicing switch is potentially critical for preventing prolonged TCR signalling and undesirable tissue injury.

We therefore sought to investigate VDR, CTLA4, CD28, and CD45 mRNA expression profiles in peripheral T-cells of Croatian HT patients in relation to HT susceptibility and phenotypic characteristics of affected individuals.

Material and methods

Subjects

The collection consisted of 45 HT patients (3 males) and 13 healthy controls (2 male, Table I). Control subjects were characterised as euthyroid, thyroid peroxidase autoantibodies (TPOAb)-negative individuals, with normal ultrasound findings of the thyroid gland and no evidence or family history of autoimmune and endocrine disorders. HT was diagnosed as previously described [20], and patients were classified as hypothyroid (hypoHT, n = 15), spontaneously euthyroid (euHT, n = 18), and rendered euthyroid by hormone replacement therapy (substHT, n = 12). Overt hypothyroidism requiring L-thyroxine (T4) treatment was defined by thyroid-stimulating hormone (TSH) > 10 mIU/L and/or free-T4 < 10 pmol/L; otherwise, it was considered subclinical if TSH was > 4.7 mU/L and fT4 was ≥ 10 pmol/L on at least two consecutive visits ≥ 3 months apart. All participants were unrelated adults from eastern Croatia and had to be in self-proclaimed good health and free of any obvious medical illness for at least one month before the blood withdrawal, including acute infections and allergic reactions. None of the participants used vitamin D supplements or immunemodifying drugs. Informed consent in written form was obtained from all participants prior to the testing, and the study protocol was reviewed and approved by the institutional ethical committee (Document No. 29-1:9890-11/2007).

Thyroid function measurement

Thyroid stimulating hormone (TSH) (normal range: 0.46–4.7 mIU/L), free tri-iodothyronine (fT3) (1.9–5.7 pmol/L) and free thyroxine (fT4) (10–22 pmol/L) were measured in morning (8–11 a.m.) sera using immunoassay methods (Vitros fT3 Reagent Pack, Vitros fT4 Reagent Pack and Vitros TSH Reagent Pack, Ortho- Clinical Diagnostics, Amersham, UK) according to the manufacturer’s instructions. Maximum baseline serum TPOAb-IgG (negative < 125 kIU/L) levels were measured by ELISA (Anti-TPO Kit, Milenia Biotec, Germany).

Peripheral blood mononuclear cell (PBMC) isolation

PBMC were isolated by density gradient centrifugation on a LymphoPrep (Axis Shield, Oslo, Norway) as described by Böyum [21]. After collection, heparinised blood samples were diluted 1:1 with 0.9% NaCl and following application to the LymphoPrep centrifuged for 20 minutes at 1000 g. Fraction of total lymphocytes and monocytes was harvested and rinsed with saline and then pelleted by centrifugation for 10 minutes at 550 g. The procedure was repeated twice. Cells were resuspended in 1 mL of Isolation buffer (PBS without Ca2+ and Mg2+ with 0.1% BSA and 2 mM EDTA) and counted after trypan-blue staining using the Bürker- Türk counting chambers and light microscope.

Lymphocyte separation

T-lymphocytes were separated from PBMC by negative selection using a Dynabeads untouched human T-cells isolation kit (Invitrogen, Paisley, UK) according to the guidelines in the manufacturer’s leafl et. Briefly, PBMC (1 × 107) were incubated with FBS and mouse monoclonal antibodies specific for CD14, CD16, CD19, CD36, CD56, CDw123, and CD235 (20 minutes at 2–8°C). Following incubation, cells were washed with Isolation buffer and pelleted by centrifugation (300 g, 8 minutes). The pellet was resuspended and incubated for 15 minutes with pre-washed magnetic beads coated with monoclonal human anti-mouse IgG antibody. The bead-bound cells were subsequently separated on a DynaMagTM magnet, leaving the cell suspension free from B-lymphocytes, NK cells, monocytes, platelets, dendritic cells, granulocytes, and erythrocytes. The remaining untouched T-lymphocytes were transferred to a new tube. Steps involving washing and binding of bead-bound cells to a magnet were repeated twice. Residual Dynabeads were removed by placing the tube in a magnet for two minutes and then transferring the supernatant to a new tube.

Total RNA extraction and cDNA synthesis

Total RNA was extracted from T-lymphocytes using TRI REAGENT (Sigma, USA) solution as described by Chomczinsky and Sacchi [22]. After isolation, RNA integrity was electrophoretically verified by ethidium bromide staining and OD260/OD280 nm absorption ratio > 1.9 determined by UV spectrometry using a Shimadzu-UV-1601 instrument (Shimadzu Scientific Instruments, SAD). Total RNA (500 ng) was reverse transcribed to cDNA using the iScript Select cDNA Synthesis Kit (Bio-Rad Laboratories, California, USA) and the provided oligo(dT) primer mix according to the manufacturer’s instructions.

Gene expression profiling

Following the RT step, mRNA expression analysis of target (VDR, CTLA4, CD28, CD45) and reference (GADPH) genes was carried out in a thin-wall PCR tubes using a 96-well Veriti thermal cycler (Applied Biosystem). Sequence-specific PCR primers for all assays were designed using PrimerQuest software (http://eu.idtdna.com/Scitool/Applications/PrimerQuest) and NCBI reference sequences (Table II). All PCR experiments were performed in 25 μl final volume reactions comprising 2 μl of cDNA template, 1.5 mM of MgCl2, 200 μM of each dNTP (Roche Diagnostics, Mannheim, Germany), 0.25 μM of each primer (TIB MOLBIOL, Berlin, Germany), and 1U of EUROTAQ polymerase (EuroClon, Italija, EU). Following initial denaturation (two minutes at 95°C), 35 repeating cycles with the same denaturation (95°C for 30 seconds) and elongation (72°C for 1 minute) steps were performed for all genes. The annealing step for GAPDH was set at 68°C for 90 seconds, for VDR at 56°C for 30 seconds, and for CTLA4, CD28, and CD45 at 55°C for 30 seconds. All reactions were terminally extended at 72°C for 15 minutes. All RT-PCR experiments were performed in duplicate, and no template negative control was included in each experiment. To check for DNA contamination, the reverse transcription negative controls with no added RT enzyme were run in parallel and did not yield any PCR product.

Relative quantification of gene expression

RT-PCR products (5 μl) were applied to 2% agarose gel, separated in an electric field, and visualised by ethidium bromide staining (Fig. 1). The amount of cDNA present in agarose bands was quantified by measuring fluorescence band intensities using a UV Transilluminator (HVD life sciences, Vienna, Austria), Olympus digital camera, and computer image processing software ImageQuant™. The fluorescent signal ratio of target and GAPDH transcripts in each sample were used for normalisation of the mRNA expression levels according to the formula:

Each sample was subjected to two cycles of RT-PCR experiments, followed by quantification and normalisation of the results. Collected data were described by arithmetic mean and standard deviation. Finally, the fold difference in mRNA levels in unknown samples was calculated relative to healthy controls.

Statistical analysis

Normality of distributions was tested by Anderson-Darling test. Data are presented as medians with interquartile ranges. Mann-Whitney and Kruskall-Wallis test with Dunn’s post-hoc analysis were used for group comparisons. Pair-wise correlations were determined by Spearman rank-test. Two-tailed P < 0.05 was considered significant. Uncorrected P-values are presented throughout the text. For correlation data analysis across eight transcripts, Bonferroni’s correction was used. Italicised P-values indicate those that remained significant after sequential Bonferroni’s corrections (P < 0.0063). All statistical analyses were performed with NCSS2007 (v07.1.20, NCSS LLC, Kaysville, Utah, USA).

Results

Patients

Patients’ characteristics are presented in Table I. Median age, sex, thyroid volume, and fT3 levels were similar in all groups. As expected, hypothyroid patients exhibited significantly higher serum TSH levels and lower fT4 levels compared to the untreated, spontaneously euthyroid HT patients, HT patients undergoing LT4 treatment, and healthy controls.

Expression profile of peripheral T cells in HT patients vs. controls

No significant differences in mRNA expression profiles were found among patient subgroups (data not shown), which prompted us to further analyse HT data as one group. Subsequently, a nominally higher VDR [median expression levels (IQR), HT vs. controls 46 (30–69) vs. 80 (62–105), P = 0.021, Fig. 2a], CTLA4 [96 (59–114) vs. 113 (98–132), P = 0.018, Fig. 2b], CD28 [35 (23–49) vs. 50 (37–62), P = 0.0074, Fig. 2c], and CD45RAB [10 (8–14) vs. 13 (10–19), P = 0.042, Fig. 2e] mRNA expression was observed in unsorted peripheral T-lymphocytes of healthy controls when compared to HT patients. However, no significance persisted upon correction for multiple testing (P < 0.0063). Concerning other CD45 mRNA isoforms, no changes across the studied groups [HT vs. controls] were observed when CD45RABC [16 (12–19) vs. 17 (12–23), P = 0.272, Fig. 2d], CD45RBC [11 (8–14) vs. 13 (10-19), P = 0.062, Fig. 2f], CD45RB [45 (36–63) vs. 63 (40–74), P = 0.126, Fig. 2g], or CD45R0 transcripts were measured [35 (21–52) vs. 23 (13–42), P = 0.168, Fig. 2h].

Correlation analysis

Pair-wise correlation analysis of pooled samples (HT patients + controls, n = 58) was preformed to examine the possible association between mRNA expression profiles, age, and thyroid hormone levels measured at the time of T-cell sampling. Firstly, an inverse relationship between serum fT3 and age was found (Spearman’s ρ = –0.307, P = 0.018). Furthermore, VDR mRNA expression was related to fT3 levels (ρ = 0.352, P = 0.0062). A weak, but nominally significant positive correlation was observed between VDR and CTLA4 mRNA expression levels (ρ = 0.326, P = 0.012). In contrast, CTLA4 mRNA expression was strongly related to CD28 mRNA abundance (ρ = 0.425, P = 0.00087). CD45RABC mRNA levels weakly paralleled serum fT4 levels (ρ = 0.31, P = 0.021), whilst CD45RB mRNA expression was independently and positively related to CD28 (ρ = 0.52, P = 0.000033) and CTLA4 mRNA levels (ρ = 0.583, P = 0.000002). In contrast, CD45R0 mRNA levels were inversely related to both VDR (ρ = –0.318, P = 0.017) and fT3 levels (ρ = -0.433, P = 0.0012). Upon correction for fT3 levels, the association between CD45R0 and VDR mRNA was lost (data not shown). There was no effect of gender, and age did not affect expression of VDR. Subgroup analysis revealed an additional, positive association between CD45R0 mRNA and age in HT patients (n = 45, ρ = 0.33, P = 0.025), predominantly in spontaneously euthyroid HT cases (n = 18, ρ = 0.5, P = 0.036), but not in healthy controls. Furthermore, a positive association was observed between CD45R0 transcript levels and serum logTSH in untreated, hypothyroid cases (n = 15, ρ = 0.593, P = 0.02). Finally, a negative association between transcript levels of CTLA4 and age (ρ = -0.63, P = 0.028), between CD45RB mRNA levels and age (ρ = –0.74, P = 0.0056), and a positive relation between CD45RBC mRNA and serum fT4 levels (ρ = 0.73, P = 0.0074) was found in the treated patients group. However, a minority of these correlations would have remained statistically significant after Bonferroni’s correction for multiple testing.

Discussion

Vitamin D deficiency is widespread among patients with HT [23, 24], but how vitamin D metabolism affects HT pathophysiology is not well understood. Most effects of vitamin D are mediated by the binding of 1,25(OH)2D3 to the intracellular vitamin D receptor (VDR), which promotes or inhibits transcription of vitamin D responsive genes [25]. In thyroid, VDR ligands directly affect thyroid cell function [26, 27] and suppress tissue inflammatory processes by inhibiting Th1 [28], Th17, Tc17 [14], CD8+Tc, and invariant natural killer T-cells [29]. Despite this, no data currently exist on VDR mRNA expression in target cell populations from HT patients. In the present study, a nominally higher expression level of total VDR mRNA was observed in peripheral T-lymphocytes of healthy controls when compared to the HT patients; however, the mechanisms that mediate the apparent down-regulation of VDR transcript levels in T-cells of HT patients remain to be clarified.

A putative positive relation between VDR mRNA expression in T-cells and thyroid functional status has been observed. Nevertheless, no causal inference was possible because several mechanisms, such as VDR-mediated protection against autoimmune thyroid dysfunction [26, 30], T3-dependent shift in T-cell composition/vitamin D3 metabolism [31], and immune effects on both T3 conversion and thyroid activity [32], may all provide parts of conjecture. Furthermore, a question remains about the exact source of these mRNAs. Peripheral T-cell populations exhibit high levels of cellular complexity; thus, lineage-specific up- or down-regulation, altered composition of T-cell compartments, a developmental shift – from naïve to memory cells – or a combination – may be responsible for observed mRNA data. Finally, the number of VDR binding sites in primary CD4+ cells is strongly correlated with 25 hydroxy vitamin D3 levels [33]. Thus, it would also be useful to obtain 25(OH)D3 and 1,25(OH)2D3 measurements, environmental ultraviolet radiation exposure, and seasonal data for future study.

Tregs, mostly defined as CD4+CD25+FoxP3+ expressing T-lymphocytes, constitutively express CTLA4 [34] and play an important role in HT aetiology. In the present study, transcript levels of CTLA4, a major gatekeeper of thyroid autoimmunity, have been nominally reduced in T-cells of HT patients. Decreased CTLA4 levels were also detected by Kucharska [9], although another report failed to find a clear decrease in CD4+CTLA4+ lymphocytes [35]. In addition, a weak relation between VDR and CTLA4 mRNA was also seen in our sample set, but no simple explanation could be provided. Both VDR and CTLA4 expression are tightly controlled and rapidly co-induced by T-cell activation [13]; in addition, CTLA4 is constitutively expressed on CD4+CD25+ Treg cells [34], shows higher levels in Th2 [36] and Th17 clones [37], and is expressed signifi- cantly higher in the CD4+ than in CD8+ T-cells [38]. Lastly, VDR binding has recently been described within the CTLA4 region in lymphoblastoid cell lines upon ligand stimulation [39], and a vitamin D-dependent increase in CTLA4 expression and the frequency of FoxP3+CTLA4+ T-cells has been reported through direct VDR binding to the FOXP3 gene [14].

In addition to vitamin D3, thyroid hormones and TSH are also known to influence the function, composition, and development of lymphoid cells and organs [31, 40]. During these processes, alternative splicing of pre-mRNA encoding the transmembrane phosphatase CD45 marks the transition from naïve to activated T-cells [18]. In the present study, CD45 composition was apparently affected by age, thyroid hormones, and case-control status, suggesting a shift from long/CD45RA+ isoforms to short/CD45R0 transcripts in an inverse, age- and T3-dependent manner. Higher proportions of both CD4+ and CD8+ T cells expressing CD45R0 have previously been detected in HT patients [41]; moreover, these cells have been functionally characterised as memory/effector T-cells and high producers of proinflammatory cytokines TNF and IFN [42]. Here, older age and lower T3 levels have been associated with the accumulation of CD45R0 isoform in HT patients, but not controls, a pattern suggestive of accelerated T-cell ageing in elderly, hypothyroid HT subjects. The characterisation of other functional markers, an aim that seems worth pursuing in future studies, is necessary to delineate the subsequent stage of maturation for various CD45R0+ T-cell compartments in those patients.

A critical phenotypic and genetic change as T-cells becomes late-differentiated and progress to senescence is the loss of gene and surface expression of CD28 [43]. CD28 surface expression in HT is not steady and is influenced by several mechanisms [41]. In this study, nominally reduced levels of CD28 mRNA expression and the peripheral loss of CD45RAB mRNA were associated with HT development. Other analyses support these results: a parallel loss of CD28 expression and the peripheral reduction of CD45RA+ T-cells have recently been described in HT [41]. Several groups have reported an expansion of CD4+ T-cells lacking CD28 in chronic inflammatory disorders [44–46], as well as in some normal-aged subjects [47]. Such CD4+CD28- T-cells display an effector/memory CD45R0+ phenotype with increased replicative history and oligoclonality but reduced apoptosis [48]. An increased proinflammatory and cytotoxic capacity of these cells has been demonstrated [46], suggesting a pathogenic role of CD4+CD28 null T-cells in the development of chronic immune-mediated disorders.

The transitions of CD45R isoforms are rapidly induced upon TCR activation [11, 12]. In the present study, the CD45RB mRNA expression coincided most consistently with CTLA4 mRNA levels, but with wide inter-individual variations. As opposed to CD45R0+ T-lymphocytes, T-cells expressing CD45RB isoform have been shown to be less susceptible to T-cell activation [16]; moreover, different cytokine profiles have been observed between CD4+CD45RBlow (IL-4/IL-10 producing) and CD4+CD45RBhigh (TNFα producing) T-cells, independent of the CD45RA or CD45R0 expression [49]. Finally, the CD45RB-inducible regulatory region had been recently discovered within the CTLA4 promoter [17], lending mechanistic support to the tolerance mediated by anti-CD45RB through CTLA4 up-regulation [50]. Nevertheless, the functional significance of these observations for HT and L-T4 treatment remains unknown.

There are several caveats that deserve further attention.

1. Our findings emphasise that examination of transcript abundance alone in unsorted T-cells provides only a partial picture.
2. The correlation between RNA and protein expression levels of the selected target mRNAs is uncertain. Lack of difference in mRNA profiles does not preclude differences in protein expression, protein kinetics, and intracellular trafficking.
3. T-cell activation is followed by complex, contextand lineage-specific dynamics of CD28, CTLA-4, VDR, and CD45R expression. In the present study, expression data from vitamin D-stimulated cells were not available.
4. Studies of peripheral T-cell mRNAs may have been confounded by the composition of T-cell subpopulations and the different roles they play in HT. T-cells exhibit vast cellular complexity and plasticity; thus, only most robust changes are readily detected.
5. Statistical power remains limited in small data sets; thus, the cohort was exploratory in nature. Nevertheless, strict selection criteria were applied, resulting in a well-characterised cohort regarding demographics, treatment exposures, and outcomes. Detailed knowledge of disease status and residual thyroid function is fundamental in order to elucidate any pathogenetic component of HT, and may be an explanation for the contradictory results in several similarly sized but phenotypically mixed studies. Finally, the majority of patients in this study were female Caucasians of European ancestry, thus limiting generalisations to males and individuals of other ethnicities.

In conclusion, an altered peripheral blood T-cell mRNA expression profile, encompassing nominally reduced VDR, CTLA4, CD28, and CD45RAB mRNA levels, occurs in HT. The current study also reports associations between free T3 and VDR/CD45R transcripts, particularly in older individuals, evidence that may contribute to our understanding of senescence and immunoregulation in HT. Our data provide preliminary evidence that CD28/CTLA4-CD45R axis might contain relevant investigational targets whose identification could improve our understanding of thyroid-immune interactions and treatment effects in HT. Nevertheless, methodological refinements, encompassing cell sorting, qPCR, and protein expression data are necessary in the confirmatory step.

Acknowledgments

This study was supported by the Croatian Ministry of Science, Education, and Sport (grant No. 219-0982914- 2176, 219-2190372-2068, and 0127-0000000-3420). The authors wish to thank Ms T. Stipanović and Ms. K. Krizmanić for valuable technical assistance.