The IgE production is initially induced in subcutaneous fat and depends on extrafollicular B cells

Background Growing body of evidence indicates that IgE production can be developed by mechanisms that differ from those responsible for IgG and IgA production. One potential possibility is generation of IgE producing cells from tissue-associated B-cells and/or through extrafollicular pathway. But the role of subcutaneous fat-associated B-cells in this process is poorly investigated. The aim of the present study was to investigate the role of different B- and T- cell subpopulations after long-term antigen administration in IgE response. Methods BALB/c mice were immunized 3 times a eeks for 4 weeks in withers region enriched with subcutaneous fat with high and low antigen doses as well as by intraperitoneal route in region enriched with visceral fat for comparison. Results After long-term antigen administration that promotes the type of immune response which is more similar to one observed in young allergic children, subcutaneous fat tissue B-cells generates more rapid and active IgE class switched and IgE-produced cells. Although IgE production at later time points was initiated also in regional lymph nodes, the early IgE production was exclusively linked with subcutaneous fat. We observed that low-dose induced strong IgE production accompanied by minimal IgG1 production was linked with extrafollicular B-2 derived plasmablasts as well as extrafollicular T- helpers accumulation. Delayed IgE class switching in regional lymph nodes and visceral fat tissue was characterized by the absence of both stable plasmablasts and T-extrafollicular helpers accumulation. Conclusion Extrafollicular B- and T-cell responses in subcutaneous fat are necessary for early IgE class switching and sensitization process in the case of allergen penetration through skin.

6 119 induces weak, if any, specific IgE production [28]. The mechanisms which mediate participation 120 of B and T-cell populations in such response, as well as more delayed IgE class switching in 121 regional lymph nodes and abdominal fat tissues are not fully understood. 122 In many allergic and asthma models, B-cell and T-cell subpopulations were thoroughly week for 4 weeks (28 days) ( Figure S1). OVA was administrated by subcutaneous route (s.c.) in 154 withers region (W) in low (100 ng) or high (10 µg) dose or by intraperitoneal route (i.p.) in low 155 dose. Antigen was administrated in sterilized saline solution in 100 µl volume. Intact mice or 156 saline-treated animals were used as control groups. There were 20 mice in each experimental 157 group. Every 7 days 5 mice from each three experimental group were challenged with 0.2 ml of 158 0.25% OVA solution to estimate anaphylaxis severity. Body temperature was measured by 159 infrared thermometer CEM DT-8806S (SEM Test Instruments, Moscow, Russia) as it was 160 performed in [33]. The temperature was measured every 15 minutes for 1.5 hour. We observed 161 that the most significant temperature decline was detected after 45 minutes, and the magnitude of 162 this decline was considered as a quantitative indicator of anaphylaxis severity. The magnitude of 163 this decline is always was not higher than 2.5 °C in the case of animal survival. In some cases, 164 however, we observed animals' death after 30-60 minutes upon challenge. In lethal cases, 165 anaphylaxis severity is believed to be higher than in survival cases, and the earlier death means 166 the higher severity. Therefore, we assigned the value of -dT «3» to death time point 1 hr, value After systematic anaphylaxis intensity measurement, mice were bled. The blood was taken 170 by retroorbital technique from living anesthesized animals and by cardiac puncture post mortem. 171 Serums were collected by centrifugation and store at -20 o C. Mice were sacrificed by isoflurane 172 («Aeran», Baxter) inhalation and perfusion through retroorbital sinus was performed. Withers 173 adipose tissue samples or abdominal adipose tissue and regional lymph nodes were collected. For 174 quantitative PCR samples from adipose tissue were homogenized in ExtractRNA (Evrogen) which 175 is Trizol analog. For flow cytometry homogenization was performed in PBS pH=7.2. 176 Homogenates were than centrifugeted (300 g) and washed 2 times with PBS. Regional lymph 177 nodes were initially homogenized in PBS following centrifugation. 5*10 5 cells were taken from 178 suspension, pelleted and resuspended in ExtractRNA for gene expression levels measurement. The   Amplificator (BioRad) according to the following protocol: +95°C initial denaturation for 3 207 minutes followed by 50 cycles: 5 s denaturation at +95°C; 20 s annealing and elongation at +64°C.

212
The following primers and probes were used:      animals which was comparable to specific IgE production in low dose group to 28 th day (Fig. 1G).

305
It should be mentioned that this production reached such level only at 28 th day and, that is more To investigate the relative contribution of IgE low and IgE high B-cells from tissue and lymph 344 nodes in IgE production, we verified the presence of correlations between IgE production and 345 percentage of these cells in tissue or lymph node respectively. Surprisingly, we have found that 346 the quantity of IgE low , but not IgE high , B-cells is linked with specific IgE production. Despite that 347 initial IgE class switching occurred in the site of antigen administration in tissue, the lymph node 348 B-cells also participated in IgE production (Fig. 2E). There were no significant correlations 349 between IgE production and quantity of IgE high B-cells either in tissue or lymph nodes (Fig. S2).

350
One could suppose that there are at least two pathways of IgE + cells generation in our model. The  In contrast to IgE, IgG1 class switching was initiated simultaneously both in subcutaneous 374 adipose tissue and regional lymph nodes (Fig. 3A)  pool and in regional lymph nodes ( Fig. 3B-C). However, only lymph node IgG1 + B-cell quantity 381 correlated with specific IgG1 production (Fig. 3D). So, IgG1 + production induced by low dose 382 antigen administration occurred mainly in regional lymph nodes.  subcutaneous adipose tissue compared to regional lymph nodes. for GC persistence in subcutaneous fat were unfavorable, and in regional lymph nodes this 423 induction was detected only at 28 th day. Instead, significant accumulation of CD19 + B220 -424 plasmablasts was observed. Most of these plasmablasts were CD19 + B220 -CD38 -CD95 + (Fig. 5C-425 D) and the amount of these cells in subcutaneous fat or lymph nodes directly correlated with 426 specific IgE production (Fig. 5C, E). The absence of CD38 and presence of CD95 may indicate 427 that these cells are closely relative to classical GCs differing from that only by the absence of B220 428 expression. The other possibility is that this phenotype could simply reflect full activated B-cell 429 state.

430
The other plasmablasts subpopulations also accumulated in adipose tissue and regional 431 lymph nodes and the earlier subpopulation in adipose tissue apparently represents B220 -

432
CD38 + CD95resembles naïve B-cells (Fig. S6A-B). It is likely that these cells later differentiate 433 into other plasmablast subpopulations. The percentage of all of these subpopulations in CD19 + B-434 cells, however, did not significantly correlate with IgE production with exception of CD19 + B220 -

442
The percentage of GC B-cells in withers adipose tissue inversely correlated with specific 443 IgE production. There was no functional association between the number GC B-cells in local 444 lymph nodes and IgE levels (Fig. S7). As seen in gating strategy plots (Fig. S5), these plasmablasts plasmablasts no increase in B-1a or MZ-B B-cells was detected before day 21 upon immunization.

447
Only at later time points there was transient increase in amount of B-1a B-cells (Fig. S8). However, 448 this increase did not correlate with IgE production (data not shown).

449
In addition, we observed that the percentage of different plasmablasts subpopulations in 450 regional lymph nodes and withers tissue started to increase at the same time points. It means that 451 antigen even at low doses is rapidly delivered in regional lymph nodes. So, one can suppose that 452 extrafollicular plasmablasts accumulation per se is necessary but not sufficient for IgE production, 453 and the impact of different types of T-helpers on these cells in withers adipose tissue and regional 454 lymph nodes could result in delayed IgE switching in regional lymph nodes. isotype switching in regional lymph nodes vs. subcutaneous adipose tissue B-cells.

462
Gating strategy for T-helper cell subsets, NK-cells and ILC2 cells is shown of Fig. S9.

463
Presuming that extrafolliculary proliferating plasmablasts account for specific IgE production, it 464 is logical that this production is also directly linked with extrafollicular T-helpers which are T-follicular helpers was also seen in regional lymph nodes at 21th day in high dose group (Fig. 6).

480
T-cells, which stabilize GCs, and B-cell follicles, respectively, results in accelerated specific IgE 481 production in subcutaneous fat. CXCR4 + CXCR5extrafollicular T-helpers accumulation could 482 support specific IgE but not specific IgG1 production after long-term antigen administration at the 483 levels compared to high dose immunized mice.

484
It is also interesting that we did not observe significant accumulation of classical T-helper 485 2 cells (CXCR4 -CXCR5 -ST2 + ) which resided mainly in T-cell zone [39] after low dose antigen 486 administration. Even high antigen doses induce their accumulation only at 28 th day and mainly in 487 subcutaneous adipose tissue but not in regional lymph nodes (Fig. S10). These cells could support 488 later stages of IgE production.

489
ILC2 cells are usually mentioned in context of immune reaction to tissue damage and sterile 490 inflammation [48]. Although ILC2 showed some tendency to accumulate in subcutaneous fat, they 491 were too rare and this tendency was insignificant (Fig. S11). Still we observed accumulation of  (Fig. S12 A-B). IgG1 + cells accumulated more 522 rapidly in subcutaneous fat than in abdominal (Fig. S12 D).

523
Delayed IgE and IgG1 production in abdominal fat tissue could be caused by reduced of 524 extrafollicular B-cell activation in this site compared to subcutaneous fat tissue. Indeed, in 525 abdominal fat tissue, the accumulation of CD19 + B220 -CD38 -CD95 + plasmablasts which were 526 responsible for IgE production in subcutaneous fat, was transient and unstable (Fig. 7 F-G). Such In our study, we did not use antibody-based or small molecular inhibitors of GCs. It is not 624 evident that IgE isotype switching per se occurs in early GC B-cells which emigrate from these 625 structures soon afterwards. First, in our work, the majority of IgE + cells remained CD95 -CD38 -

626
/low after beginning of isotype switch which happened 2 weeks after the immunization start. This 627 means that despite becoming activated these cells did not acquire full GC phenotype. Second, we 628 did not observe any increase in GC B-cells after low dose antigen administration either in tissue 629 or in regional lymph nodes and did not detect any positive correlations between these cells' week 4, they differentiated into fully activated plasmablasts (B220 -CD95 + ).

640
The low dose administration induced the extrafollicular but not GC-associated antibody 641 switching, and, therefore, turned to be not favorable for somatic hypermutation, which is essential . So, our extrafollicular IgE production must be T-cell dependent.

668
In our work, we have clearly shown that sustained plasmablasts activation is linked with 669 extrafollicular T-helpers accumulation and rapid IgE class switching in subcutaneous adipose 670 tissue in comparison to abdominal fat and regional lymph nodes. We could not completely identify 671 cell populations which produced type 2 cytokines, such as IL-4, for the switching per se at the 672 early stages. The current data did not provide evidence for the contribution of Th2 (Fig. S10) address these questions.

678
One of the most importance output from our work is that in allergically predisposed 679 subjects, humoral immune response to low antigen doses entering due to defects in barrier tissues