Immune Phenotype of Children with Newly Diagnosed and Gluten-Free Diet-Treated Celiac Disease
- First Online:
- Cite this article as:
- Cseh, Á., Vásárhelyi, B., Szalay, B. et al. Dig Dis Sci (2011) 56: 792. doi:10.1007/s10620-010-1363-6
- 126 Views
Recent data suggest the involvement of both the adaptive and the innate immune system in celiac disease (CD). However, little is known about the immune phenotype of children with CD and its alteration upon dietary intervention.
We characterized the prevalence of major interacting members of the adaptive and innate immune system in peripheral blood of newly diagnosed children with CD and tested its alteration with the improvement of clinical signs after the introduction of gluten-free diet (GFD).
Peripheral blood was taken from ten children with biopsy-proven CD at the time of diagnosis and after the resolution of clinical symptoms following GFD. As controls, 15 children with functional abdominal pain were enrolled. The prevalence of the cells of adaptive and innate immunity was measured with labeled antibodies against surface markers and intracellular FoxP3 using a flow cytometer.
Patients with CD were found to have lower T helper, Th1 and natural killer (NK), NKT and invariant NKT cell prevalence and with higher prevalence of activated CD4+ cells, myeloid dendritic cells (DC) and Toll-like receptor (TLR) 2 and TLR-4 positive DCs and monocytes compared to controls. After resolution of symptoms on GFD, the majority of these changes normalized, although the prevalence of NK and NKT cell, DC and TLR-2 expressing DCs and monocytes remained abnormal.
The immune phenotype in childhood CD indicates the implication of both adaptive and innate immune system. The normalization of immune abnormalities occurs on GFD, but the kinetics of this process probably differs among different cell types.
KeywordsCeliac diseaseLymphocyteRegulatory T cellDendritic cellMonocyteToll-like receptor
Celiac disease (CD) is an inflammatory disease of the small bowel that affects up to 1% of the worldwide population . Its main clinical symptoms in childhood include intestinal (diarrhoea, constipation, chronic abdominal pain, malabsorption) and extraintestinal symptoms (iron deficiency, hepatitis, osteopenia, migraine, occipital epilepsy) . In genetically susceptible individuals, signs and symptom of CD are triggered by an immune response against dietary proteins normally present in wheat, rye, and barley involving adaptive immune system . Recent observations, however, raise the notion that the innate immune system is also implicated in pathomechanism of CD [4–6].
Most studies performed in CD highlighted the involvement of adaptive immunity. Indeed, total T lymphocyte prevalence was lower at the periphery  while increased in biopsy samples [3, 8, 9] in CD compared to age-matched controls. This finding was accompanied by an increased prevalence of activated lymphocytes in CD irrespective of whether it was measured in peripheral blood [7, 10–13] or biopsy specimens [14, 15]. Simultaneously with these alterations marked abnormalities of function and prevalence of regulatory T cells (Tregs), the major suppressor cells of adaptive immune system were observed in blood and biopsy specimens of patients with CD, respectively [16–18].
Less is known about the prevalence of the major members of innate immune system in CD, although some reports suggested their implication in CD [4, 5, 19]. Indeed, the prevalence of NK, NKT, and iNKT cells decreased in CD subjects’ blood [7, 20–22] and biopsy specimens [14, 21]. More limited data are available about the prevalence of antigen-presenting cells (APCs). There is just one study demonstrating lower prevalence of peripheral dendritic cells (DCs)  while another found no alteration in peripheral monocyte prevalence in CD . In addition, the expression of pattern-recognition molecules of innate immunity, namely Toll-like receptor (TLR) 2 and 4, also increased in biopsy samples of patients with CD .
Our aim was to investigate in a comprehensive manner the prevalence of major interacting members of the adaptive and innate immune system in newly diagnosed CD and to test whether immune phenotype is changing with the improvement of clinical signs and symptoms after the introduction of gluten-free diet (GFD).
We enrolled ten children in our study (four boys and six girls, age 3 [2–5] years), duration of major clinical symptoms 2 [2–4] months leading to the establishment of the diagnosis of CD. The diagnosis of CD was made according to the criteria of the European Society of Paediatric Gastroenterology, Hepatology and Nutrition  and according to Marsh criteria . Transglutaminase (TG) IgA antibody levels were above 200 U/l and histological investigation of biopsy samples demonstrated subtotal or total villous atrophy of intestinal mucosa with increased number of intraepithelial lymphocytes (>40%) in all patients. Strict GFD was introduced in all patients diagnosed with CD. Clinical symptoms of the enrolled children fully resolved after 3 [2.5–4] months of GFD and TG IgA levels returned to normal (10–40 U/l). As controls, 15 children (six boys and nine girls, age 3 [2–6] years) with functional non-organic abdominal pain were recruited. All patients and controls were diagnosed, treated and followed up in the Outpatient Clinic of the First Department of Pediatrics, Semmelweis University between December 2007 and December 2009. The Institutional Ethical Committee approved our study; written parental informed consent was obtained.
Six milliliters of lithium-heparin anticoagulated venous blood were taken from controls during their routine assessment and from patients at the time of diagnosis and at the second visit, when clinical symptoms were resolved. After isolation of peripheral blood mononuclear cells (PBMCs), the cellular markers (6B11, CCR4, CD3, CD4, CD8, CD11c, CD14, CD25, CD45RA, CD45RO, CD62L, CD69, CD123, CD161, CXCR3, HLA-DR and Lin-1, BD Biosciences Pharmingen, San Diego, CA, USA; TLR-2, TLR-4 and FoxP3 assay, eBioscience, San Diego, CA, USA) were measured with a BD FACS Aria according to the manufacturers’ protocols. We identified the members of adaptive immunity, such as helper T (i.e., CD4+), cytotoxic T (i.e., CD8+), Th1 (i.e., CXCR3+), Th2 (i.e., CCR4+), naïve (i.e., CD45RA+), effector/memory (i.e., CD45RO+), early activation marker expressing (i.e., CD25+ and CD69+), late activation marker positive (i.e., CD62L+ and HLA-DR+) and regulatory T lymphocytes (i.e., CD4+CD25hiFoxP3+). Simultaneously, we identified the members of innate immunity including NK (CD3−CD161+), NKT (i.e., CD3+CD161+), iNKT (CD3+6b11+), DC (i.e., Lin1−HLA-DR+), mDC (i.e., CD11c+), pDC (i.e., CD123+) and peripheral monocyte cells (i.e., CD14+), along with the TLR-2 and TLR-4 expressing APCs (DCs and monocytes).
All data were expressed as median (interquartile range). Mann–Whitney’s, Wilcoxon’s, and Spearman’s correlation tests were used for the comparison of data between groups, the comparison of data within CD group before and after GFD and for the regression analysis between cell prevalence and TG IgA levels, respectively. The study was not powered for the multitude of statistical tests we performed; therefore, our results should be regarded as exploratory ones.
The prevalence and ratio of the members of adaptive immunity
Cell prevalence in parent population
Newly diagnosed celiac patients
Celiac patients on GFD
(n = 15)
(n = 10)
CD4+ in PBMC
CD8+ in PBMC
Th1 in CD4+
Th2 in CD4+
Naive in CD4+
Effector in CD4+
CD25+ in CD4+
CD69+ in CD4+
CD62L+ in CD4+
HLA-DR+ in CD4+
Treg in CD4+
The prevalence and ratio of the members of innate immunity
Cell prevalence in parent population
Newly diagnosed celiac patients
Celiac patients on GFD
(n = 15)
(n = 10)
NK in PBMC
NKT in PBMC
iNKT in PBMC
DC in PBMC
mDC in DC
pDC in DC
TLR-2+ in DC
TLR-4+ in DC
Monocyte in PBMC
TLR-2+ in monocyte
TLR-4+ in monocyte
Major differences in the prevalence of cells of adaptive immune system were observed between newly diagnosed children with CD and controls. These include lower-than-normal prevalence of CD4+, while CD4+ cells expressing early (i.e., CD69) and late (i.e., HLA-DR) activation markers were more prevalent in CD. In addition, Th1 cell populations were smaller in CD in comparison to controls. The prevalence of Tregs was comparable.
The phenotype of innate immunity was also abnormal in CD as NK, NKT and iNKT cells were less prevalent in newly diagnosed patients with CD than in controls. On the other hand, DCs, particularly mDCs, were more prevalent in CD than in healthy control subjects. APCs expressing TLR-2 and TLR-4 receptors were also more prevalent in CD. After introduction of GFD and cessation of CD symptoms the immune phenotype started to normalize. CD4 lymphocytes’ prevalence increased along with that of Th1 committed CD4 cells. The prevalence of CD4+ cells expressing early and late activation markers (i.e., CD69+ and CD62L+, respectively) tended to normalize.
Interestingly, the abnormalities in innate immune system characteristic for newly diagnosed CD were maintained after GFD; only the prevalence of TLR-4 expressing APCs normalized.
Our exploratory analysis revealed important correlations between TG IgA levels and immune phenotype including the prevalence of CD8+, CD4+CD69+, CD4+HLA-DR+ cells, pDCs and TLR-2 expressing monocytes in patients with CD on GFD (p = 0.048, r = −0.79; p = 0.011, r = −0.86; p = 0.022, r = −0.81; p = 0.025, r = −0.80 and p = 0.020, r = −0.85, respectively).
In this study we comprehensively tested the prevalence of interacting cells of adaptive and innate immunity at the diagnosis of CD and its alteration upon the introduction of a GFD.
First, we found marked abnormalities in peripheral prevalence of adaptive immune cells. These included lower-than-normal CD4+ lymphocyte prevalence in newly diagnosed childhood CD, in line with a recent observation in newly diagnosed adult CDs . In addition, our children with CD also exhibited a decreased Th1/Th2 ratio in concordance with others . Of note, these values are in contrast to those regarding to small bowel mucosa samples demonstrating higher than normal lymphocyte and Th1 cell counts [9, 29]. These data raise the notion that the lower-than-normal CD4+ and Th1 cell prevalence in blood may be, at least partly, due to their accumulation in the affected organs [30, 31].
While the ratio of naive/memory lymphocytes was comparable in patients with CD and controls, in contrast with the data of others , marked variations were observed in the prevalence of activated lymphocytes depending on the activation markers measured. In line with previous data , the expression of CD62L, a late activation marker was lower than normal in CD. On the other hand, lymphocytes with early activation marker CD69 were more prevalent as in previous reports [7, 10, 13], while the prevalence of CD4+ expressing CD25 markers was comparable in each group studied, which is in contrast with adult CD [7, 10, 12]. The differences in activated lymphocyte prevalence values indicate a systemic inflammation in CD and supports a notion of a general alteration of adaptive immunity in children with CD. However, while our observations correspond to previous reports about adult CD, they apparently contradict to those obtained in CD biopsy samples indicating a marked difference between local immune milieu and systemic immune phenotype [30, 31].
Of note, our study does not indicate that abnormal systemic Tregs numbers would be present and contribute to CD. Previous data obtained in children and adolescents are contradictory as a recent study reported no difference in FoxP3+ Tregs number , while others found decreased prevalence of CD4+CD25hi positive Tregs . In addition, the prevalence of FoxP3+ Tregs increased in adults .
Interestingly, the members of innate immunity are also affected in CD. We found lower-than-normal NK, NKT, and iNKT prevalence at the periphery in line with previous reports [7, 20–22]. In contrast with others, however, we demonstrated an increase in peripheral DC prevalence and a shift to mDCs [23, 24]. While we found no association between mDC and Th1 cell prevalence values, this finding may form a basis for the hypothesis that altered mDC numbers may contribute to the alterations in Th1/Th2 ratio. Recent reports demonstrated a specific role of mDC in Th2 skewness of lymphocytes .
Another important finding is that TLR-2 and TLR-4 expression is increased on APCs in CD, reflecting locally increased expression of TLR-2 and TLR-4 receptors already observed in CD [14, 20, 25]. As TLRs are transiently expressed upon activation this alteration may indicate that systemic APCs are permanently activated in CD and, therefore, may further contribute to the activation of adaptive immunity presented as higher prevalence of activated CD4+ cells.
The second phase of our study was to evaluate the impact of GFD on immune phenotype. In blood samples taken at the complete resolution of clinical signs and symptoms, while at the presence of some remnant TgA levels we observed the normalization of CD4, Th1 and the majority of activated CD4+ subsets in children with CD. Interestingly, we still found a correlation between remnant TgA levels and the prevalence of CD8+ and CD4+CD69+ and CD4+HLADR+, indicating an ongoing susceptibility to immune dysregulation even at the presence of GFD [33, 34]. This normalization of adaptive immunity is also suggested by previous studies [7, 16, 34], but these studies were limited by their cross-sectional nature and compared separate patient populations with newly diagnosed CD and patients on GFD instead of a follow-up design that was applied in our study. A possible bias of our study is that participants were children and ageing on immune cell prevalence may have a dramatic impact on lymphocyte prevalence and function [36–38]. However, as we resampled our patients after 3 months, it is unlikely that the individual maturing of immune system would largely interfere our results.
Interestingly, the normalization of adaptive immunity was not accompanied by that of innate immunity as the prevalence of NK and NKT cells was still decreased, and that of DCs and TLR-2 and 4 expressing APCs remained permanently high. In addition, the abnormally high prevalence of TLR-2 and 4 expressing cells and pDCs correlated with remnant TG IgA levels that may suggest the general disturbance of innate immune system even in the absence of marked clinical signs and symptoms in CD. These findings indicate a delayed change of innate immune system, if any, upon therapeutic interventions in CD.
Summarizing our results, we concluded that immune phenotype in childhood CD exhibits several abnormalities: these include the increased activation of both adaptive and innate immunity. The alterations of adaptive immunity are mainly normalized with GFD, while those of innate immunity still persist in spite of the clinical improvement of CD-associated signs and symptoms.
This work was supported by grants TÁMOP-4.2.2-08/1/KMR-2008-0004, OTKA-76316, OTKA-K81117 and ETT-028-02.