Different Doses of Allergen Exposure on Dendritic Cells Determine Their Genetic/Epigenetic Regulation and T Cell Differentiation

Background: It has been reported that when DCs are exposed to high-dose antigens, they can induce Th0 cells into regulatory cells (Treg) and Th1 cells. When DCs are subjected to low-dose allergen, they can drive to Th2 cells. However, the mechanisms of such dose-effect relationship are poorly understood so far. Methods: Bone marrow immature DCs (imDCs) were generated from mice and stimulated with OVA of different concentrations (0, 10, 100, 1000, 10000 μg/ml, respectively). The mDCs were then seeded and cocultured with naïve T cells for 3 days, and then the markers of different T cell subgroups were flow cytometrically tested. RNA-seq detection and DNA methylation of DCs were performed. Results: When DCs were stimulated with low-dose (10ug/ml), 77 genes were up-regulated and 87 genes down-regulated. Most activated genes were related to ribosome synthesis and ion channel inhibition but not to the immune responses and Th2 activation. At the medium-dose (100μg/ml), 339 genes were up-regulated and 168 genes down-regulated. Most activated genes involved cytokine synthesis and regulation of immune responses. At high-dose (10000μg/ml), 2794 genes were up-regulated and 1156 genes down-regulated. Tumor necrosis factor signaling pathway, MAPK signaling pathway, antigen processing and presentation signaling pathway were mostly up-regulated. The related co-stimulators, co-inhibitory molecules, inhibitory cytokines, negative regulating enzymes were highly expressed. The monocarbate, coenzyme, fatty acid, glucolipid, starch, sucrose and other metabolism-related signaling pathways were down-regulated. Conclusion: The profiles of DNA methylation and RNA synthesis of DCs varied with different doses of OVA, which serves to induce T cells to differentiate in various directions. induced Treg cells differentiation in each concentration group C(1). The RPKM value of the gene induced Th2 cells differentiation in 10-0 group C(2). The RPKM value of the gene induced Th1 cells differentiation in 100-0 group C(3). The RPKM value of the gene induced Treg cells differentiation in 10000-0 group DC1/2/3 represents 0ugml DC4/5/6 represents 10ug/ml DC7/8/9 represents 100ug/ml DC10/11/12 represents 10000ug/ml Error bars represent standard deviations. ∗p<0.05 represents the diﬀerences between the treatment group and the control group. n = 3 independent experiments.

It has been well reported and consensus has been reached that allergen dose can impact DC-mediated T cell differentiation. When DCs are exposed to high-dose antigen, they can induce Th0 cells into regulatory cells (Treg) and Th1 cells. Nonetheless, when DCs are subjected to low-dose antigen, they can induce Th0 to Th2 cells [3][4][5][6] . Clinically, exposure to allergens of different doses may lead to various outcomes of allergic conditions. For instance, a German multi-center study that had followed up 1,314 newborns for 20 year found that exposure to environmental dust mite allergen (HDM) was proportional to the incidence rate of allergic diseases. When the HDM concentration in the environment exceeded 10 μg/g, the incidence rate of dust mite allergy in families with allergy history could be as high as 5.5%. When the environmental concentration was lower than 0.1μg/g, the risk of allergic reactions was very low 7 .Tover ER et al. found that the highest and lowest exposure concentrations in the quintile of dust mite in asthma were relatively low in the in children aged 0-5 years and the incidence of asthma was highest when the exposure concentrations were at 3.5-23.4 μg/g 8 . The association between exposure doses and the incidence of allergic diseases might be ascribed to the aforementioned connection between antigen dose and the cell-differentiating effect of DCs in the upstream of immune responses. In the allergen-specific immunotherapy for allergic rhinitis and asthma, high concentrations of allergens were used to stimulate DCs in order to elicit the allergenspecific immune responses which will induce the conversion of T cells to Treg and Th1 cells instead of Th2 cells 9 . On the other hand, low-concentration allergens tend to be ineffective with such allergen-specific immunotherapy 10 .
Immature DCs (imDCs) are usually found in the immunostatic phase, expressing receptors which related to antigen recognition (such as Toll-like receptor, mannose receptor, etc.).
When exposed to the antigen, imDCs will engulf the external antigens and, after experiencing a series of changes, convert into mature DCs (mDCs). These changes involve: (1) loss of receptors which mediate endocytosis, (2) over-expression of MHC-Ⅱmolecules and costimulators, such as CD40, CD80, and CD86, etc and (3) over-expression of CCR7 and other related receptors. These changes help DCs migrate into secondary lymph nodes, and contact with naive T cells to present antigens 11 .
The studies are scanty concerning the differences in the signal molecules and cytokines produced during the DC activation and the association between such differences and the differentiating direction of the downstream T cells when imDCs are exposed to allergens of different doses. Huang et al. reported that DCs promoted the proliferation of Th1 cells, Th2 cells and Treg cells, respectively, by secreting different interleukines, such as IL-12, IL-6 and IL-23 (Th1 cells), IL-4 and IL-13 (Th2 cells), IL-10, IL-35 and TGF-β (Treg cells) 12,13 . Boonstra et al. demonstrated that when DCs were stimulated by low-dose of antigens, they expressed low-affinity MHC-Ⅱand present weak signals to Th cells, thereby inducing the differentiation of Th cells into Th2 cells; when DCs were exposed to high-dose antigens, strong signals were presented and, as a result, Th cells differentiated into Th1 cells 14 . Some studies reported that exposure to allergens of different doses could alter the expression of DC surface markers and related factors, such as CD40, CD86, IL-4, IL-12, TLR, and TCR, etc., which might change differentiating direction of T cells [15][16][17] .   DCs-induced differentiation of T cells involve multiple signaling pathways and a great   many factors, including CD40, CD80, CD86, ICAM-1, CCL17, RELM-a, ERK, c-Fos, NF-Kb,   Irf4, STAT5a/JAK2, KLF4, PDL2, Notch, DLL1 and DLL4, among others. [18][19][20]

DNA Methylation analysis of DCs
After the DCs were collected and centrifuged as aforementioned, the cells of each group were separated, and methylation analysis and sequencing were carried out (Wuhan

Statistical analysis
Data were processed using GraphPad Prism 7.0 statistical software, and the data were expressed as mean ± standard deviation. The differences between groups in findings of the flow cytometry were compared by one-way analysis of variance (ANOVA). P-

1.Morphological and flow cytometric analysis of DCs
At low magnifications, a small number of cell colonies were observed on day 1. Six days later, large number of colonies were seen. At high magnifications, no synapses were noticed on day 1. After day 6, the synapses were evident (Appendix Figure 1). On day 7, flow cytometric detection showed that the expression of CD40, CD80, CD86 and MHC-Ⅱon DC increased with the OVA dose. No difference was found in the expression of CD11c among the groups. (Figure 1).

Flow cytometrical analysis of Th cells
Flow cytometry showed that naive T cells could be induced to differentiate in different directions after co-cultured with DCs stimulated by OVA at different doses. In 10 DCs group, Th2 cells (CD4 + IL-4 + ) were dominant. The 100 and 1000 DCs group, show Th1 cells (CD4 + IFN-γ + ) dominance. In 10000 DCs group, Treg (CD4 + CD25 + Foxp3 + ) were dominant.
There was no significant difference in the number of Th17 cells (CD4 + IL-17 + ) among these groups. Naïve T cells could not be activated when co-cultured with imDCs. (Figure 2)

RNA-seq and DNA methylation of DCs
Since 100 DCs group and 1000 DCs group both induce Th1 dominant, we chose the dose of 100 μg/ml in the ensuing experiment. We used 0, 10, 100, 10000 μg/ml OVA to stimulate DC then conduct RNA-seq and methylation analysis. The results showed that there existed a strong correlation between the gene expression level and methylation level in each group, and the experimental results were reliable and the sample selection was reasonable as shown by the Pearson correlation coefficients (> 0.8). We compared groups of different doses with the control, and conducted clustering analysis on the upregulated and down-regulated genes by using RPKM analysis for RNA-seq and DMRs analysis for methylation, respectively. The results showed that the samples had a high homogeneity among samples in each group (Appendix Figure 2).
Our study showed that DNA methylation, rather than demethylation, was the main expression profile in different regions of the genome in all dose groups ( Figure 3C,4C,5C).
In this study, the gene expression had four patterns and the methylation presented six patterns with the increasing dose of OVA. Clustering analysis showed that as the OVA dose increased, the differential gene expression level was obviously lowered or raised, and the differential methylation level showed obvious methylation or demethylation (Appendix Figure 8).Our analysis showed that some activation signaling pathways shared by various dose groups mainly involved ribosome synthesis (e.g. Gm13423, Rps4x-ps genes), Jak-stat signaling (e.g. Hamp, Socs1 genes), pantozoate, CoA biosynthetic and other cellmetabolism-related processes (e.g. Vnn3 gene). In the study, the differentially expressed genes were subjected to GO-KEGG enrichment analysis, and the gene body area and overlapping genes with DMRs to functional enrichment analysis. As a result, the top 20 genes with the highest significance were selected to make a plot according to the order of P values. If the numbers of enriched pathways in the sample was less than 20, they were all shown in a scatter plot ( Figure 3EF 4EF 5EF). The results of inter-group analyses of 10-0 DCs, 100-0 DCs, 10000-0 DCs are shown below.

Intergroup analysis of 10-0 DCs
The RPKM results of 10 DCs group showed that 77 genes were up-regulated and 87 genes down-regulated ( Figure 3A)

Intergroup analysis of 10000-0 DCs
The RPKM results of 10000 DCs group showed that 2794 genes were up-regulated and 1156 genes down-regulated ( Figure 5A). The key up-regulated genes included Acod1, Ccl5,

Functional analysis of DCs
We analyzed the activation of the functional genes in different groups according to RPKM.

Discussion
In this study, we cultured imDCs from murine bone marrow mononuclear cells, stimulated them with OVA of different doses, and then co-cultured the mature DCs with naïve T cells.
Although the DCs from different sources were functionally different, the response patterns are similar when exposed to antigens of gradually-increasing dose 14 Table 1).  DCs-induced Th1 polarization, HMGN1 R848, MAPKs, IRF3, IRF7 and IFN-a were significantly up-regulated, which is also consistent with our results 26 . Nonetheless, we also found that some genes involving Th1 polarization were down-regulated (such as Cxcr3

Meanwhile, in order to observe the effect of the duration of DC-T cells co
Itgax Ptprc, Ccr6, etc.). We speculated that at certain concentrations, although the differentiation of DCs may overall goes to a certain direction (polarization), there is still a certain mechanism of positive regulation and negative feedback, which precisely controls the direction immune reaction. There should be a balance between positive and negative feedbacks, which dictates the expression of the Th1-related genes. Briefly, when exposed reported that the major up-regulated genes of tolerance DCs derived from mouse mononuclear cells included Socs2, CD83, Cd150, Cd200, Cd274, Aldh1a2, etc., which was in line with our results 27 . It was reported that imDCs could induce tolerance, that is, imDCs processed and presented antigens without costimulatory molecules, leading to the T cell dysfunction or deficiency. However, in the past few years, fully mature toleranceregulating dendritic DCs were found in various environments and they highly expressed co-inhibitory molecules (PD-L1, PD-L2) and secreted inhibitory factors (Il10, PGE2, TNF-a, TGF-b, Il27 and Il13), and induce T cells towards Treg 28 . In current study, we found that DCs were highly activated, with most Treg related genes being up-regulated (Il18, Ccl17 Ccl1, Tgfb, Cd274, Il6, Ticam1, Cd1d, Il23, Ctla4, Stat3, Pdcd1lg2, Il27, Itgal, Il10 and Il13 et al). , That is, DCs not only showed high activation status under high concentration OVA stimulation, but also related co-stimulatory molecules, co-suppressor molecules, inhibitory cytokines and negative regulatory enzymes were highly expressed. Therefore, it can be inferred that DC tolerance is a highly active behavior. By altering the potential activation