Relationship between Helicobacter pylori Infection and Plasmacytoid and Myeloid Dendritic Cells in Peripheral Blood and Gastric Mucosa of Children

Purpose To investigate the frequency and activation status of peripheral plasmacytoid DCs (pDCs) and myeloid DCs (mDCs) as well as gastric mucosa DC subset distribution in Helicobacter pylori- (H. pylori-) infected and noninfected children. Materials and Methods Thirty-six children were studied; twenty-one had H. pylori. The frequencies of circulating pDCs (lineage−HLA-DR+CD123+) and mDCs (lineage−HLA-DR+CD11c+) and their activation status (CD83, CD86, and HLA-DR expression) were assessed by flow cytometry. Additionally, the densities of CD11c+, CD123+, CD83+, CD86+, and LAMP3+ cells in the gastric mucosa were determined by immunohistochemistry. Results The frequency of circulating CD83+ mDCs was higher in H. pylori-infected children than in the noninfected controls. The pDCs demonstrated upregulated HLA-DR surface expression, but no change in CD86 expression. Additionally, the densities of gastric lamina propria CD11c+ cells and epithelial pDCs were increased. There was a significant association between frequency of circulating CD83+ mDCs and gastric lamina propria mDC infiltration. Conclusion This study shows that although H. pylori-infected children had an increased population of mature mDCs bearing CD83 in the peripheral blood, they lack mature CD83+ mDCs in the gastric mucosa, which may promote tolerance to local antigens rather than immunity. In addition, this may reduce excessive inflammatory activity as reported for children compared to adults.


Introduction
Helicobacter pylori (H. pylori) colonizes the human stomach. Infection usually starts in early childhood and can persist for decades, sometimes even for the lifetime [1]. H. pyloriinfected children develop gastric mucosal inflammation [2] but with much lower infiltration of polymorphonuclear and mononuclear cells and decreased incidence of gastric and duodenal ulceration, gastric adenocarcinoma, and gastric mucosa-associated lymphoid tissue (MALT) lymphoma as compared to adults [3,4].
Dendritic cells (DCs) are professional antigen-presenting cells (APCs) that induce T cell response against a presented antigen. Their migration from the site of an antigen uptake to the regional lymph nodes and their subsequent maturation profile decide about the type of the subsequent adaptive immune response profiles against infection (protective or nonprotective responses). The DC maturation process is associated with the expression of several membrane molecules engaged in the DC migration (CCR7), antigen presentation abilities (MHC class II), and their costimulatory activities (CD80, CD83, and CD86) [5,6].
Human DCs are rare in peripheral blood [7] and consist of two major subsets, myeloid DCs (mDCs) and plasmacytoid DCs (pDCs) [8]. The mDCs express myeloid antigen such as CD11c and can be subdivided into CD1c-and CD141-bearing cells. pDCs lack these markers and instead express CD123, CD303, and CD304 molecules. Both mDC and pDC subsets display APC capacity, but this function is reduced in pDCs, which have relatively low expression of HLA-DR [9]. It is well documented that pDCs promote the generation of regulatory T cells (Tregs) [10].
The circulating DC number and their activation status may change during the course of infection [11,12] and may alter throughout a lifetime [13]. These cells play an important role in shaping the anti-infection response by constantly completing the population of tissue-residing DCs.
In contrast to many studies concerning DC participation in H. pylori-induced immune response in adults [23,24,[26][27][28][29][30], similar ones in children are still rare [31]. Additionally, estimation of the frequencies of mDCs and pDCs and the expression levels of CD83, CD86, and HLA-DR in the peripheral DC subsets have not yet been evaluated in the H. pylori-infected children. There is also a paucity of information regarding the expression profile of the human DC maturation markers (LAMP3 and CD83), the myeloid cell marker (CD11c), and the pDC marker (CD123) in the gastric mucosa of H. pylori-infected children. Therefore, the aim of this study was (a) to investigate the distribution and activation status of peripheral and gastric mucosa DC subsets (pDCs and mDCs) in the H. pylori-infected and noninfected children and (b) to evaluate correlations between the DC subset phenotype characteristics and the activity score on the basis of the updated Sydney gastritis classification system.

Materials and Methods
2.1. Ethics Statement. The study was undertaken according to Helsinki declaration with ethics committee approval of Collegium Medicum Nicolaus Copernicus University in Bydgoszcz, Poland, and informed consent was obtained from all the patients' parents and from the patients older than 16 years prior to blood and antral biopsy collection.

Patients.
In total, thirty-six subjects with dyspeptic symptoms, older than eleven years of age, were included in this study. The exclusion criteria involved the following: (1) a history of antibiotic use during 4 weeks for other reasons than H. pylori infection; (2) the presence of other inflammatory diseases, such as celiac disease, inflammatory bowel disease, or allergy; and (3) gastric perforation or hemorrhage, history of surgery, bleeding conditions, or evidence of other clinical conditions or intestinal parasites. Each subject or subject's parents provided the clinical history.
All the children were examined with gastroduodenal endoscopy due to permanent abdominal pain. Based on previous observation that H. pylori-associated gastritis of the corpus was present only when antral gastritis was present [32], three antral biopsies were taken from each patient. One biopsy specimen was subjected to a rapid urease test. The second specimen was used for histological analysis. The last specimen was subjected to immunohistochemical staining.
The urease test was performed for every patient. Furthermore, H. pylori infection was excluded in every subject by use of the 13 C urea breath test, performed within one week of undergoing endoscopy. 13 C concentration was measured with an infrared radiation analyzer (OLYMPUS Fanci2, Tokyo, Japan), assuming 4‰ as the cutoff point.
A patient was considered H. pylori-infected when the 13 C urea breath test and either the rapid urease test or microscopic evaluation were positive for H. pylori. A patient was considered noninfected when all three tests were negative. Twenty-one patients were found to be H. pylori positive, while fifteen patients satisfied the criteria for a negative H. pylori status. None of the patients had ulcer disease or macroscopic lesions of the duodenal mucosa upon endoscopic examination.
The study was conducted in the H. pylori-infected and noninfected children and adolescents. The group of noninfected children (controls) consisted of 7 boys and 8 girls (median age 15 years, range 11-18 years). The group of H. pylori-infected children comprised 9 boys and 12 girls with recognized gastritis and H. pylori infection (median age 15 years, range 12-18 years). None of the patients had a history of other inflammatory diseases or cancer.

Histological Assessment.
The specimen was formalinfixed and embedded in paraffin, sectioned and stained with hematoxylin and eosin for histological analysis, or Giemsa modified by Gray stain for H. pylori detection. Biopsy specimens were graded for gastritis by two independent pathologists based on the updated Sydney system.

Flow Cytometry
Analyses. During endoscopy, 5 mL of venous blood was taken for immunological testing. The frequencies of total DCs (tDCs), mDCs and pDCs, as well as the expression of CD83, CD86, and HLA-DR were determined in whole blood samples with the use of a five-color flow cytometry method. We followed the methods of Helmin-Basa et al. [33]. First, two tubes containing 200 μL whole blood were stained with FITC Lin-1 lineage cocktail (antibodies against human CD3, CD14, CD16, CD19, CD20, and CD56), PerCP-anti-human HLA-DR, APC-antihuman CD83, PE-Cy7-anti-human CD86, PE-anti-human CD11c, or PE-anti-human CD123 (BD Biosciences, Franklin Laker, New Jersey, USA) antibodies for 30 min at room temperature (RT). The erythrocytes were removed with the use of BD Bioscience FACS Lysing Solution (BD Biosciences). The stained leukocytes were suspended in phosphate-buffered saline (PBS) and analyzed immediately with flow cytometry (FACSCanto II, BD Biosciences). The instrument was set up using a BD Cytometer Setup and Tracking beads. Data acquisition was performed using BD FACSDiva software, version 6.1.3 (BD Biosciences) and the analysis with FlowJo 7.5.5 (Tree Star Inc., Ashland, Oregon, USA). Compensation settings were performed with a set of compensation tubes using the BD CompBead antibody-capturing particles and the compensation setup tool with BD FACSDiva software.
Fluorescence minus one (FMO) was used for gating ( Figure 1). The gating strategy applied for the DC subset enumeration in two separate tubes is shown in Figure 2. Doublets and aggregates were avoided by selecting singlet cells on a forward scatter: FSC (area)/FSC (height) plot. Logical gating was used to identify tDC, mDC, and pDC populations. tDCs were defined as Lin1 -HLA-DR + , mDCs were defined as Lin-1 − HLA-DR + CD11c + cells, while pDCs were defined as Lin-1 − HLA-DR + CD123 + cells. The geometric mean fluorescence intensity (gMFI) was determined to quantify cell surface expression of activation molecules. Statistical analyses were based on at least 5000 events gated on the mDCs and 2500 events gated on the pDCs. FFPE tissue sections were cut on a manual rotary microtome (AccuCut, Sakura, Torrance, USA). Paraffin sections (4 μm) were prepared and mounted onto extra adhesive slides (SuperFrost Plus, Menzel Glasser, Braunschweig, Germany).
The immunohistochemical staining was performed according to previously described procedures [34,35] and standardized using a series of positive and negative control reactions. The positive control reaction was performed on a model tissue selected according to reference sources (The Human Protein Atlas, http://www.proteinatlas.org) [36] and the antibody datasheets. The negative control reactions were performed on additional tissue sections, by substituting the primary antibody with a solution of 1% bovine serum albumin (BSA) diluted in PBS.

Evaluation of Protein Expression Based on
Immunohistochemical Staining. The evaluation of protein expression in the antrum area of obtained FFPE tissue sections was performed at 20x and 40x original objective magnifications, using the ECLIPSE E400 (Nikon Instruments Europe, Amsterdam, Netherlands) light microscope. For the evaluation of expression, immunohistochemical reactions were scored according to morphometric principles based on a Remmele-Stegner scale (IRS-Index Remmele-Stegner; immunoreactive score) [37], used in our previous publications [34,38].
The total immunoreactivity score was defined according to the scale obtained from the grade of intensity multiplied by the score of positively stained cells to give a total score from 0 to 12. The intensity of staining was scored as follows: 0, negative; 1, low staining; 2, moderate staining; and 3, strong staining. The number of positive cells was categorized as follows: (1)  2.6. Statistical Analyses. All data are expressed as medians with the first and third quartiles. The results were compared using Mann-Whitney U test calculated by the STATISTICA 12 software (StatSoft). For normal distribution, variables were analyzed using the Kolmogorov-Smirnov test with Lilliefors correction. The age and gastric inflammation scores were correlated with the frequency of DC and their subsets using the Spearman's rank correlation. Statistical significance was considered at p < 0:05.  (p < 0:001 and p = 0:003, respectively) when compared to the H. pylori noninfected ones (Table 1). Interestingly, three children with H. pylori infection had mild gastric mucosal atrophy in the antrum. However, no intestinal metaplasia was seen in the children's gastric mucosa.

Results
3.2. The Distribution of Circulating tDCs, mDCs, and pDCs and Their Activation Status in H. pylori-Infected and Noninfected Children. The percentages of tDCs, mDCs, and pDCs were similar between the two groups (Figures 3(a)-3(c)). While the percentages of CD83 expressing tDCs and Table 1: Gastric mucosa histology of H. pylori-infected and noninfected children (median score (range) according to the updated Sydney system: 0, none; 1, mild; 2, moderate; 3, marked).

Antral mucosa MN cells (intensity) PMN cells (activity) Atrophy
Intestinal metaplasia H. pyloriĤ  (Figures 4(a) and 4(c)). There were no statistically significant differences in CD83 and CD86 expression between the two groups (p > 0:05) (Figures 4(a)-4(c)). pylori infection on the local immune system, the gastric mucosa sections were labeled with the following antibodies: anti-CD11c (high expression on mDCs, but low on granulocytes, macrophages, and a subset of B cells), CD123 (a marker of pDCs), CD83 and LAMP3 (markers of mature DCs), and CD86 (costimulatory molecule upregulated on activated DCs and other APCs).
A statistically significant increase in the accumulation of CD11c-positive cells was noted in the gastric lamina propria mucosa of H. pylori-infected children (p = 0:023) (Figures 5(a) and 5(b)). There was no significant difference between CD83 + , CD86 + , and CD123 + cells in H. pyloriinfected and noninfected children (p > 0:05). However, the data display a tendency toward an increased density of CD123 + DCs in the gastric epithelium mucosa of H. pylori-infected children (Figures 5(a)-5(c)) (p > 0:05). Moreover, we used immunohistochemistry to detect LAMP3 + cells in an antrum. We observed these cells but

Correlation Analysis.
To evaluate whether H. pylori-associated gastritis influences DC frequency or whether these values depend more on age, we correlated age, gastric inflammation, and H. pylori density scores with the circulating and gastric DC frequencies.

Discussion
We described here peripheral DC subset distribution (mDCs and pDCs) and their maturation status (expression of CD83, CD86, and HLA-DR) in the H. pylori-infected and noninfected children. Additionally, the densities of CD11c + (mDCs), CD123 + (pDCs), CD83 + or LAMP3 + (mature DCs), and CD86 + (activated) cells in the gastric mucosa were determined by immunohistochemistry. Similar to the other studies on H. pylori infection, our control group consisted of dyspeptic, noninfected patients [39,40], because gastroduoendoscopy is not considered an ethical procedure in the noncomplaining subjects.
H. pylori-infected children exhibited (1) an increased percentage of circulating CD83 + mDCs, but with normal expression of CD86 and HLA-DR, (2) upregulation of HLA-DR expression but unchanged expression of CD83 and CD86 on pDCs subset, (3) increased densities of CD11c + cells in the lamina propria and CD123 + cells in the epithelium, and (4) a significant association between frequency of circulating CD83 + mDCs and lamina propria infiltrating CD11c + DCs.
The gastric mucosa, which does not initially have organized and diffuse lymphoid tissue, acquires MALT-like structures with follicles and germinal centers following contact with H. pylori [41]. H. pylori infection in children is associated with a higher accumulation of lymphoid follicles containing CD4 + T cells, B cells, and DCs [42] in the lamina propria of the gastric mucosa, identified by macroscopically evident nodularity.
H. pylori bacteria attach to the gastric epithelium and contact the gastric epithelium and subepithelial lamina propria DCs (direct or indirect), which rapidly increase in the H. pylori-infected mucosa compared with an uninfected mucosa [18,23,24]. DCs loaded with H. pylori likely migrate to gastric lymphoid follicles or draining paragastric lymph nodes to present H. pylori antigens to naive CD4 + T cells [43,44], thus initiating and regulating the host H. pylori-specific immune response.
Three subsets of circulating DCs exist: the CD123 + pDCs and the two types of CD11c + mDCs:CD1c + mDCs with CD11c low and CD141 + mDCs with CD11c high [8]. Since DCs are rare in human peripheral blood [7], we focused on (a) the enumeration of the two major DC subsets (CD11c + mDCs and CD123 + pDCs) and (b) the estimation of the correlation between these DC phenotypes, H. pylori-induced inflammation and H. pylori density scores, according to the updated Sydney system of gastritis classification. Additionally, we used immunohistochemistry to assess DC subset distribution in antral biopsy specimens. The distribution of circulating DC subsets in H. pyloriinfected patients has not yet been previously evaluated, and to our knowledge, this is the first study to demonstrate an increased percentage of CD83 + mDCs, but with no significant differences in surface expression of activation markers (CD86 and HLA-DR). Terminally differentiated DCs should induce CD83-specific maturation markers and increase the expression of CD86 costimulatory molecule and HLA-DR. Thus, an increase in the frequency of CD83-expressing mDCs along with the absence of upregulation of CD86 and HLA-DR molecules on mDCs might suggest that the maturation status of circulating mDCs in H. pylori-infected children was changed but only within the small part of mDCs. However, the rest of circulating mDCs still expressed immature/tolerogenic phenotype (HLA-DR low CD86 low ).
The peripheral blood is not the best source for the studying of DC maturation status in the course of H. pylori infection; hence, we studied also the gastric epithelial and lamina propria DC subset distribution and its maturation markers.
We observed high levels of CD11c-positive cells and a tendency toward increased density of CD83-positive cells in the gastric lamina propria of H. pylori-infected children. CD11c is a marker for DCs of myeloid origin (mDCs) which are potent APCs and can activate T cells. CD83 is a more selective marker, present on mature DCs. Hence, we used CD11c and CD83 antigens for detecting mDCs and mature DCs in a gastric mucosa.
CD11c + mDCs could be recruited from the blood in response to H. pylori infection, but we could not observe any association between frequency of circulating mDCs and gastric mucosa mDC density. However, we noted association between frequency of circulating mature CD83-bearing mDCs and gastric mDC density. The accumulation of mDCs in the gastric lamina propria of H. pylori-infected patients is necessary for mounting adaptive immune response needed for efficient bacterial clearance. The increased number of mDCs but not CD83-positive DCs in the H. pylori-infected gastric lamina propria mucosa could indicate that H. pylori infection in children may be associated with local accumulation of immature mDCs (lack of CD83-expressing DCs). The presence of these cells might initiate low protective immune response to H. pylori.
We had too small biopsy sections to evaluate other activation marker expressions (HLA-DR and CD80). Instead, we studied the expression of more specific DC maturation markers like LAMP3. We could not detect LAMP3 + cells in the gastric lamina propria by immunohistochemistry. However, in six H. pylori-infected children, we observed these cells in mucosal epithelium. This marker is absent on immature DCs but rapidly increase upon DCs activation [45], so our results might suggest that some H. pylori-infected children have small numbers of activated DCs in gastric epithelium. In H. pylori-infected adults, DC-LAMP + cells were also observed but they were located only in the gastric lamina propria [26].
Immature human DCs constitutively express intermediate amounts of CD86 and lack of CD80, CD83, and DC-LAMP [45]. CD80 is exclusively induced on mature DCs while CD86 is already present on immature cells and further upregulated upon stimulation [46]. Terminally differentiated DCs induce specific maturation markers including CD83 and DC-LAMP. Recent studies in adults also showed an increased population of HLA-DR-positive [23], DC-SIGN-positive [26], or DC-LAMP-positive DCs with a semimature phenotype (CD83 + , CD80 -CD86 low ) [23,26] in lymphoid follicles of the gastric lamina propria in the H. pylori-infected human gastric mucosa, and these cells were in the same location as follicular FoxP3 + Tregs [26]. The presence of CD11c-bearing mDCs but not mature DCs expressing CD83 or LAMP3 molecules in gastric lamina propria in pediatric may promote tolerance to local antigens rather than immunity. The presence of immature DCs and other immature APCs could increase H. pylori density in the gastric mucosa. On the other hand, it can also be beneficial, as it may reduce excessive inflammatory activities responsible for gastric ulcers and premalignant gastric lesions in some individuals. The absence of mature DCs in a pediatric gastric mucosa might also explain the lower local proinflammatory responses to H. pylori infection in children than in adults [47].
Our study also demonstrated a tendency to increase the density of CD123-(marker of pDCs) positive cells in gastric epithelium as compared to normal controls. This pDC subset may mount both protective and tolerogenic immune responses [48] and can induce Treg differentiation [49]. A previous study in adults showed that gastric CD303expressing DC (other marker of pDCs) cells were present in a very low number in both H. pylori-infected and uninfected individuals however with no significant difference between the groups. Thus, the presence of pDCs in the site of H. pylori infection (gastric epithelium) may initiate low protective immune response to H. pylori often observed.

Conclusion
This study shows that although H. pylori-infected children had an increased population of mature mDCs bearing CD83 in the peripheral blood, they lack mature CD83 + mDCs in the gastric mucosa, which may promote tolerance to local antigens rather than immunity. In addition, this may reduce excessive inflammatory activity as reported for children compared to adults.

Data Availability
The all data used to support the findings of this study are included within the article.

Disclosure
No company had any input or influence into the design, analyses, interpretation, or content of this manuscript.