Morphological and Functional Characterization of IL-12Rβ2 Chain on Intestinal Epithelial Cells: Implications for Local and Systemic Immunoregulation

Interaction between intestinal epithelial cells (IECs) and the underlying immune systems is critical for maintaining intestinal immune homeostasis and mounting appropriate immune responses. We have previously showed that the T helper type 1 (TH1) cytokine IL-12 plays a key role in the delicate immunological balance in the gut and the lack of appropriate levels of IL-12 had important consequences for health and disease, particularly with regard to food allergy. Here, we sought to understand the role of IL-12 in the regulation of lymphoepithelial cross talk and how this interaction affects immune responses locally and systemically. Using a combination of microscopy and flow cytometry techniques we observed that freshly isolated IECs expressed an incomplete, yet functional IL-12 receptor (IL-12R) formed solely by the IL-12Rβ2 chain that albeit the lack of the complementary IL-12β1 chain responded to ex vivo challenge with IL-12. Furthermore, the expression of IL-12Rβ2 on IECs is strategically located at the interface between epithelial and immune cells of the lamina propria and using in vitro coculture models and primary intestinal organoids we showed that immune-derived signals were required for the expression of IL-12Rβ2 on IECs. The biological relevance of the IEC-associated IL-12Rβ2 was assessed in vivo in a mouse model of food allergy characterized by allergy-associated diminished intestinal levels of IL-12 and in chimeric mice that lack the IL-12Rβ2 chain on IECs. These experimental models enabled us to show that the antiallergic properties of orally delivered recombinant Lactococcus lactis secreting bioactive IL-12 (rLc-IL12) were reduced in mice lacking the IL-12β2 chain on IECs. Finally, we observed that the oral delivery of IL-12 was accompanied by the downregulation of the production of the IEC-derived proallergic cytokine thymic stromal lymphopoietin (TSLP). However, further analysis of intestinal levels of TSLP in IL-12Rβ2−/− mice suggested that this event was not directly linked to the IEC-associated IL-12Rβ2 chain. We interpreted these data as showing that IEC-associated IL12Rβ2 is a component of the cytokine network operating at the interface between the intestinal epithelium and immune system that plays a role in immune regulation.

Interaction between intestinal epithelial cells (IECs) and the underlying immune systems is critical for maintaining intestinal immune homeostasis and mounting appropriate immune responses. We have previously showed that the T helper type 1 (TH1) cytokine IL-12 plays a key role in the delicate immunological balance in the gut and the lack of appropriate levels of IL-12 had important consequences for health and disease, particularly with regard to food allergy. Here, we sought to understand the role of IL-12 in the regulation of lymphoepithelial cross talk and how this interaction affects immune responses locally and systemically. Using a combination of microscopy and flow cytometry techniques we observed that freshly isolated IECs expressed an incomplete, yet functional IL-12 receptor (IL-12R) formed solely by the IL-12Rβ2 chain that albeit the lack of the complementary IL-12β1 chain responded to ex vivo challenge with IL-12. Furthermore, the expression of IL-12Rβ2 on IECs is strategically located at the interface between epithelial and immune cells of the lamina propria and using in vitro coculture models and primary intestinal organoids we showed that immune-derived signals were required for the expression of IL-12Rβ2 on IECs. The biological relevance of the IEC-associated IL-12Rβ2 was assessed in vivo in a mouse model of food allergy characterized by allergy-associated diminished intestinal levels of IL-12 and in chimeric mice that lack the IL-12Rβ2 chain on IECs. These experimental models enabled us to show that the antiallergic properties of orally delivered recombinant Lactococcus lactis secreting bioactive IL-12 (rLc-IL12) were reduced in mice lacking the IL-12β2 chain on IECs. Finally, we observed that the oral delivery of IL-12 was accompanied by the downregulation of the production of the IEC-derived proallergic cytokine thymic stromal lymphopoietin (TSLP). However, further analysis of intestinal levels of TSLP in IL-12Rβ2 −/− mice suggested that this event was not directly linked to the IECassociated IL-12Rβ2 chain. We interpreted these data as showing that IEC-associated IL12Rβ2 is a component of the cytokine network operating at the interface between the intestinal epithelium and immune system that plays a role in immune regulation.

inTrODUcTiOn
The constant dialog between the intestinal epithelium and under lying immune system plays a key role in intestinal immune homeostasis by shaping the appropriate immunological microenvi ronment (1). Intestinal epithelial cells (IECs) signal to the immune system via the secretion of regulatory cytokines; in the presence of commensal bacteria, IECs secrete antiinflammatory molecules that include thymic stromal lymphopoietin (TSLP) (2) and trans forming growth factorβ (3). By contrast, pathogens cause epithelial cells to release proinflammatory factors, such as IL8 (CXCL8), MCP1 (CCL2) and MIP3α (4,5) and might trigger rapid response to clear the infection (6). In turn, cytokines produced by the intesti nal immune system may signal back to the epithelium and regulate epithelial functions such as, for example, antigen sampling by spe cialized M cell (7). The T helper type 1 (TH1) cytokine IL12 plays a variety of roles in intestinal immunity ranging from response to pathogens (8), interaction with group 1 innate lymphoid cells in maintaining barrier integrity (9) to protecting against food allergy (10). With regard to the latter we have previously shown that the intestinal levels of IL12p70 in the gut are critical for the resistance to food allergy (10). Indeed, sensitization to the food allergen peanuts was characterized by significantly reduced intestinal levels of IL12. One of the most convincing evidence for the regulatory role of IL12 in the intestinal mucosa was brought about by the observation that blocking IL12 in the gut at the time of antigen presentation by targeted oral delivery of microencapsulated neutralizing antibody significantly increased susceptibility to food allergy in mice (10). The notion that IL12 is critical in food allergy is further strengthened by the observation that oral delivery of liposome encapsulated IL12 suppressed allergic reaction (11). However, aspects of its mechanism(s) of action remained to be determined. We observed that IECs expressed only the low affinity IL12 binding site IL12Rβ2 chain, but not the complementary high affinity binding site IL12Rβ1 chain. IECassociated IL12Rβ2 operates as an incomplete, yet still operational IL12R the expression of which is under the control of immune cellderived signals. This observation prompted us to test the hypothesis that the presence of IECassociated IL12Rβ2 could directly affect the regulatory properties of IECs. This was done by adopting a strategy based on oral administration of a Lactococcus lactis strain genetically engineered to produce biologically active IL12 (rLcIL12) (12) in allergic wildtype (wt) and chimeric mice that lacked IL12Rβ2 on the intestinal epithelium. We observed that the lack of IL12Rβ2 on IECs significantly reduced the antiallergy activity of rLcIL12. Also, IL12 mediated suppression of food allergy was accompanied by the downregulation of the IECderived cytokine TSLP.
These data suggested that IECassociated IL12Rβ2 contributed to control the finely tuned regulatory cytokine network operating at the lymphoepithelial interface in the gut and impacted on the host's ability to mount appropriate immune responses.

Mice and Bacterial strain
Mice were maintained in an accessrestricted room in con ventional conditions. B6.129S1Il12rb2tm1Jm/J (IL12Rβ2 −/− ) and wt C57BL/6J mice that were previously used to generate chimeric mice (13) were purchased from (Jackson Laboratory, Bar Harbor, ME, USA). IL12Rβ2 −/− mice were maintained on BL/6 background. Chimeric mice were produced by adopting a slightly modified busulfanbased method that allowed a stable bone marrow (BM) chimerism (14). Briefly, recipient mice were treated with daily injections of 20 mg/kg busulfan (6 consecutive days for a total dose of 120 mg/kg busulfan). Twentyfour hours after the last BU injection, 5 × 10 6 BM cells collected from the femur and tibia of wt or IL12Rβ2 −/− donors and injected i.v. (tail vein) into recipient mice. Age and sexmatched busulfantreated IL12Rβ2deficient mice were reconstituted either with BM from wt mice (WT→ IL12Rβ2 −/− ) to generate mice lacking IL12Rβ2 in IECs (IL12Rβ2 ∆IEC ) or BM from syngeneic IL12Rβ2deficient mice (IL12Rβ2 −/− → IL12Rβ2 −/− ) to control for the effects of the busulfan treat ment and BM reconstitution. Mice were then allowed to rest for at least 8 weeks before the beginning of the experiments. Chimerism was functionally assessed by determining the abil ity to produce IFNγ in vivo following i.p. administration of mouse IL12 (1 μg/mouse for 5 consecutive days) as previously described (15). Furthermore, wt C57BL/6J and chimeric mice were sensitized according to a wellestablished model of food (peanuts) allergy (10,16). Ground whole peanuts were used as antigen and crude peanut extract (CPE) was prepared as described in detail previously (10,17). Briefly, mice received intragastric (i.g.) administration (oral gavage) of 1 mg/mouse of ground whole peanut together with 10 µg of cholera toxin (CT) (Quadratech Ltd., Surrey, UK) in PBS on days 0, 7, 14, 21 and 28. On day 35, all mice were challenged by oral gavage with CPE (1 mg/mouse divided into two doses at 30 min intervals). The administration regimen of the sensitizing mixture was based on previously established protocol (10,16). To assess the role of rLcIL12 in suppressing food allergy, wt and chimeric mice were gavaged with PBS (untreated) or L. lactis FI10611 secreting bioactive IL12p70 (rLcIL12) (8-10 mice/group). Description of the wt L. lactis strain, generation of the recombinant IL12 producing strain, levels and regulation of IL12 secretion and its biological activity were described by us in detail previously (12). Strains were grown to an OD600 of approx 1.0. Cell pellets were collected, washed with PBS and resuspended in 2% NaHCO3. A suspension of 100 µl containing 10 9 colonyforming units/ml was administrated by oral gavage in a single dose to the mice 24 and 2 h before administration of the sensitizing mixture. The percentage of mice that developed Type I hypersensitivity reaction was assessed as previously described (10). Serum levels of peanutspecific IgE antibody were determined by ELISA in all groups of sensitized and control mice as described previously in detail (18). Plasma levels of histamine were detected by using an enzyme immunoassay kit (Immunotech S.A.S., Marseille, France) whereas the vascular leakage was monitored by measur ing the hematocrit. Body temperature was monitored using a rectal probe (Kent Scient. Corp. Torrington, USA). Levels of the cytokines IL4 and IL5 produced by splenocytes in vitro follow ing antigen (CPE 20 µg/ml) recall challenge (72 h); plasma levels of IFNγ were determined by ELISA kits (ELISA MAX™ Deluxe Sets, BiolegendCambridge Bioscience, Cambridge, UK).

gene expression
Total RNA was extracted from gut tissues using RNeasy Mini Kit (Qiagen Ltd.). For cDNA synthesis, total RNA was reverse tran scribed with Quantitect Reverse Transcription Kit (Qiagen Ltd). QPCR was performed on 5 µl of diluted (1:5) template cDNA, using the ABI 7500 RealTime PCR System (Applied Biosystems, Warrington, UK) and SYBR Green Detection. Data were normal ized against an invariant endogenous control, 18S ribosomal RNA, and threshold cycle number (Ct) values obtained were converted into fold of relative induction using the ΔΔCt. Frozen sections (8-10 µm) were fixed with 10% buffered formalin for 30 min and endogenous peroxidases were quenched with 0.3% hydrogen peroxide for 1 h and stained with biotinylated goat antimouse TSLP (R&D Systems) and then treated with Vectastain Elite ABC kit (Vector Laboratories). Fluorescein deposit was achieved on the site of the immunoreaction with the TSA Plus fluorescence system (PerkinElmer). Sections were counterstained with mouse antiβactin Mab (SigmaAldrich, Milan) followed by TRITCconjugated donkey antimouse IgG antibody (Jackson ImmunoResearch LabStarfish, Milan, Italy). For TEM analysis, tissues were fixed in 4% paraformaldehyde and, following dehydration, embedded in Epon 812. Sections were incubated overnight at 4°C with rabbit antiIL12Rβ2 (Bioss Inc., Boston, MA, USA) followed by 15 nm goldconjugated anti rabbit IgG (BB International, Cardiff, UK) and examined with a 201 Philips TE microscope.
isolation of iecs From Mouse small intestine and evaluation of il-12r expression and Functional integrity Mice were sacrificed; small intestines were removed, dissected longitudinally and washed with sterile PBS. Tissue was cut into 0.5 cm pieces and washed with sterile PBS. To remove IECs, the tissue was subjected to a first wash in PBS containing 10% FCS, 1 mM DTT (20 ml), and a second wash in PBS containing 10% FCS, 15 mM EDTA (20 ml) at 37°C, 250 rpm for 15 and 20 min, respectively. Preceding each wash, tissuebuffer mix was vortexed for 1 min. After washings the resultant pellets were resuspended in PBS containing 10% FCS and stored on ice. CD45 + cells were removed from crude IEC preparations using magnetic labeling with AntiPE Microbeads (Miltenyi Biotec) and magnetic separa tion with LD MidMACS depletion columns (Miltenyi Biotec). The expression and function of IL12Rβ1 and IL12β2 chains on viable IECs was assessed by flow cytometry. Briefly, isolated IECs were doublestained with antiIL12Rβ1FITC and antiIL12Rβ2PE or mouse IgG1PE (R&D Systems) and mouse IgG1FITC (eBio sciences, UK) isotype controls. After incubation for 1 h on ice, cells were washed once and pellets were resuspended and fixed with an equal volume of 4% formalin for 20 min at room temperature. Fixed cells were washed once with PBSA. The functional integrity of IL12Rβ2 through MAPKp38 phosphorylation was tested as follows. Freshly isolated IECs were placed in a 37°C water bath in the presence of 500 µM H2O2 or 2 µg/ml IL12 (Biolegend, UK) for various times. Control cells were left untreated. At the end of the incubation period, an equal volume of 8% buffered formalin was added. After 10 min, fixed cells were washed with PBSA and allowed to cool on ice; cells were then permeabilized on ice for at least 30 min with Perm Buffer III (BD Biosciences, UK). Cells were washed and stained for 30 min at room temperature with rabbit antiphosphoMAPKp38 (Cell Signaling Technology, Danvers, MA, US) or normal rabbit IgG (Senta Cruz Biotechnology, Dallas, TX, US) as control. Cells were then incubated for a further 30 min with donkey antirabbit IgGFITC (Santa Cruz). For viability assessment cells were washed and resuspended in 1 ml PBS. Dead cells were stained using Live/Dead Fixable Violet Dead Cell Stain Kit (Life Technologies, UK) according to the manufacturer's instructions and data were acquired by flow cytometry.

statistical analysis
The animal study data were analyzed by (ANOVA) Wilcoxon rank test. The results for the gene expression data show the average relative fold of induction of IL12Rβ2, TSLP, OX40L and IL33. Data were analyzed using a Student's ttest (twotailed, equal variance) followed by application of the Bonferroni correction the level of which was set at p ≤ 0.0125. In either case, pvalue of ≤0.05 was considered statistically significant. The results are expressed as mean ± SD. resUlTs iecs express the il-12rβ2 chain but not the complementary il-12rβ1 chain The TH1 cytokine IL12 binds to a heterodimeric IL12Rβ1/β2 complexes on a variety of immune cells (21). First, flow cytom etry analysis was carried out using epithelial cell suspension to identify intestinal cells expressing IL12R. We observed that approximately 58% of small intestine CD45 − cells and in a small proportion (>1%) of CD45 + cells ( Figure 1A) expressed IL12Rβ2 chain. IL12Rβ2 positive cells that did not express CD45 (IL12Rβ2 + CD45 − ) were identified and selected. The epithelial origin of the IL12Rβ2 + CD45 − cells was confirmed through expression of cytokeratin as seen in Figure 1B. Figure 1C demonstrated that these IL12Rβ2 + CD45 − cytokeratin + cells did not coexpress IL12Rβ1. Side scatter (SSC) analysis is also shown (Figures 1D,E). This showed that the IL12β2 chain is expressed in vivo by IECs in absence of the complementary IL12Rβ1 chain.
It has been shown that the presence of the IL12Rβ2 chain is indispensable for IL12 mediated signaling (15) although IL12Rβ2 alone was insufficient to induce the canonical signal transducer and activator of transcription (Stat) 4 activation in response to IL12 (22). However, an IL12Rβ1/STAT4independent pathway and based on MAP kinase (MAPK) activation in the presence of only the IL12Rβ2 chain was observed (22). Thus, to assess the function of IL12Rβ2 in the absence of the complementary IL12Rβ1 chain, freshly isolated IECs were cultured in the pres ence of IL12p70 and levels of phosphoMAPK38 were assessed. Rapidly, after the application of the stimulus, expression of phosphoMAPKp38 reached similar levels observed following challenge of IECs with hydrogen peroxide (H2O2) a nonspecific activation stimulus ( Figure 1F). By contrast, IECs isolated from IL12Rβ2 −/− mice did not respond to challenge with IL12p70 dem onstrating that activation of MAPK38 was an IL12Rβ2 specific event. The expression of the IL12Rβ2 chain on IECs was further confirmed by TEM by immunogold staining. Positive staining was observed within the Golgi area (Figure 2A), in the proximity of the basolateral membrane ( Figure 2B) and, most relevant to lymphoepithelial cross talk, at the interface between IECs and immune cells harboring the lamina propria (lp) (Figure 2C).

The expression of il-12rβ2 chain on iecs is controlled by immune system-Derived signals
The in vivo observation that IL12Rβ2 was expressed by IECs at the interface with local lpimmune cells prompted us to assess whether its expression was under the control of immunederived signals. Quantitative RTPCR analysis showed that the IL12Rβ2 was expressed on freshly isolated primary IECs; however, its expres sion was negligible on the murine IEC line MODEK ( Figure 3A). By contrast, when using coculture systems ( Figure 3B) we observed that PPlymphocytes seeded in the bottom compart ment of Transwells culture system induced the expression of the IL12Rβ2 in MODEK (1.5fold increase). Yet, a more pronounced increase (3.2fold) was seen when PPlymphocytes were allowed celltocell interaction with the MODEK cells. The importance of immunederived signals was further investigated in primary small intestine organoids that were developed in vitro. In the complete absence of immune cellderived signals organoids IECs lacked the expression of the IL12Rβ2 chain, however, when the organoids were grown in the presence of PPderived conditioned medium (PPCM) significantly higher number of IL12Rβ2 + cells were observed (Figure 3C; Figure S1 in Supplementary Material). This experimental approach also showed that in the presence of PPCM intestinal organoids also tend to increase the expression of cytokeratin ( Figure S1 in Supplementary Material). Taken together these data demonstrated that crosstalk between IECs and immune cells is instrumental to trigger the expression of the IL12Rβ2 on IECs.
iec-associated il-12rβ2 chain Plays a role in lympho-epithelial cross Talk To test the biological role of IECassociated IL12Rβ2 chain, we analyzed the ability of a recombinant bacterial vector secreting bioactive IL12p70 (rLcIL12) (12) to suppressing food allergic response in wt and chimeric IL12Rβ2 −/− and IL12Rβ2 ∆IEC mice that lack IL12Rβ2 on IECs. First, we assessed the produc tion of IFNγ in response to systemic administration of IL12 in IL12Rβ2 −/− mice following BM transplant from wt mice (WT→IL12Rβ2 −/− ) at week 8 posttransplant. (Figure 4) In agree ment with a previous report (15), the presence of an intact IL12R enabled wt mice to produce in vivo significant amount of IFNγ following i.p. administration of IL12. By contrast, IL12Rβ2 −/− and chimeric IL12Rβ2 −/− (IL12Rβ2 −/− →IL12Rβ2 −/− ) mice did not produce detectable levels of IFNγ. However, IFNγ production was completely restored in IL12Rβ2 −/− mice by BM transplant from wt mice (WT→IL12Rβ2 −/− ). The result of the functional assay was paralleled by the observation that postBM transplant mice displayed similar levels of MHCII + cells in the lp compared to wt mice ( Figure S2 in Supplementary Material). Similar results were obtained at week 28 posttrans plant (not shown). Taken together these data show a success ful BM engraftment. Subsequently we investigated the role of IECassociated IL12Rβ2 in the regulation of immune response using a wellestablished mouse model of food allergy. The IL12 secreting bacterial vector rLcIL12 was delivered prior to each administration of the sensitizing mixture and its effects on food . Flow cytometry analysis was carried out using epithelial cell suspension to identify intestinal cells expressing IL-12R. IL-12Rβ2 chain was detected on approximately 50% of small intestine CD45 − (IECs) cells and in a small proportion (>1%) of CD45 + cells (a). IL-12Rβ2 positive cells that did not express CD45 (IL-12Rβ2 + CD45 − ) were identified and selected. Their epithelial origin was confirmed through expression of cytokeratin (B). Furthermore, we showed that the expression of IL-12Rβ2 on these cells was not paralleled by the presence of the complementary IL-12Rβ1 chain (c). Also, plots of SSC vs. IL-12Rβ2 are shown (D,e); Cells were stained with PE-coupled isotype control antibody to facilitate gating and enable detection of non-specific fluorescence (D). Following staining with PE-coupled anti-IL-12Rβ2 antibody, positive cells were identified (e) within the gate R-1 constructed in D above. Finally, ex vivo stimulation of freshly isolated IECs from the mouse small intestine with IL-12p70 led to a rapid increase in intracellular phosphorylation of MAPKp38 only in IEC from wt mice but not IL-12Rβ2 −/− mice showing that the upregulation of phospho-MAPK38 is an IL-12Rβ2-specific event; stimulation with H2O2 was used as a positive control (F). Experiments on MAPKp38 were performed in triplicate using freshly isolated IECs (five mice/group). (*) indicates statistical difference. p < 0.05.  Figure 5A). Also, in these mice we observed that rLcIL12 prevented the drop in temperature associated with allergic reaction following the final challenge with the causal allergen ( Figure 5B). By contrast, IgE response was not affected by the administration of rLcIL12 in chimeric IL12Rβ2 −/− mice ( Figure 5C); also, these mice showed a markedly higher production of IgE compared to wt throughout the allergen sensitization period (summary in Table 1). In this group of mice, rLcIL12 also failed to prevent the allergy associated drop in body temperature ( Figure 5D). The pattern in IL12Rβ2 ∆IEC appeared to be different. In this group, the administration of rLcIL12 significantly reduced plasma levels of IgE but the magnitude of the reduction was smaller compared to wt mice (41 and 93%, respectively) ( Figure 5E). A similar profile was observed in regard to the effects of rLcIL12 on body temperature ( Figure 5F). Here too the treatment with rLcIL12 significantly prevented the drop of body temperature observed in wt mice but not as much as observed in the wt cohort. A similar trend was observed for all the other parameters of Type I allergic reactions monitored (Figure 6). Plasma levels of hista mine in sensitized wt mice reached 2,100 ± 185 ng/ml and were reduced to 320 ± 21 ng/ml (p < 0.01) in rLcIL12treated mice leading to a 75% reduction ( Figure 6A). Furthermore, while no effects were seen on histamine levels in IL12Rβ2 −/− mice, the administration of rLcIL12 led to a reduction of plasma levels in IL12Rβ2 ∆IEC from 2,600 ± 212 to 1,450 ± 104 ng ml totaling to a 45% reduction. Treatment with rLcIL12 also affected IL4 production following in vitro recall antigen chal lenge of splenocytes in both wt and chimeric IL12Rβ2 ∆IEC but not in chimeric IL12Rβ2 −/− mice. Levels of IL4 in culture supernatants following splenocytes challenge with the causal allergen were significantly reduced (p < 0.01) by treatment with rLcIL12 ( Figure 6B). By contrast, production of IL5 was downregulated by rLcIL12 only in wt mice but not in the two chimeric groups (Figure 6C). Hematocrit was utilized to moni tor vascular permeability leakage, an additional feature of type I hypersensitivity reaction (Figure 6D). Hematocrit levels were increased in sensitized untreated wt mice (60.5%) compared to baseline levels (41.5%) and, in these mice was significantly reduced by rLcIL12 (43.5%). Administration of rLcIL12 did not suppress vascular leakage in chimeric IL12Rβ2 −/− mice, however, rLcIL12 treatment reduced hematocrit levels in chimeric IL12Rβ2 ∆IEC mice compared to untreated syngeneic mice (42.2% and 61.2%, respectively). Finally, we observed that the expression of the IL12Rβ2 on IECs did not differ between sensitized and control naïve wt mice (Figures 6E,F) suggesting that allergyassociated skewed TH1/TH2 ratio did not affect the expression of the IL12Rβ2 chain on IECs.

Downregulation of iec-Derived TslP Following Oral Delivery of rlc-il12 Did not Occur Directly via iec-associated il-12rβ2
The overexpression of the IECderived cytokines TSLP and IL33 is critical in food allergy (23)(24)(25). We then tested the hypothesis that the engagement of IL12 with the functional IECassociated IL12Rβ2 chain could play a role in the regulation of these proal lergic cytokines. First, quantitative RTPCR analysis ( Figure 7A) showed that TSLP gene expression was significantly upregulated in the gut epithelium of sensitized mice as well as in mice treated with CT alone, compared to control nonsensitized (PBStreated) mice. Furthermore, administration of rLcIL12 led to a significant downregulation of TSLP gene expression in the gut epithelium of sensitized mice. The regulatory effect of rLcIL12 on IEC derived TSLP was further confirmed by immunohistochemistry. The latter approach did not detect TSLP neither in PBS nor in rLcIL12treated mice (Figures 7B,C) but only in sensitized mice (Figures 7D,E). Taken together, these results concurred with reports that suggested a critical role of the epitheliumderived TSLP in allergic reactions (23,24) and pointed to an important role of IL12 in the regulation of the epitheliumderived TSLP. Furthermore, immunofluorescence staining was carried out in both IL12Rβ2 −/− and IL12Rβ2 ∆IEC chimeric mice undergoing allergic sensitization and treated with rLcIL12. Both sensitized chimeric mice retained the ability to produce TSLP by IECs ( Figure  S3 in Supplementary material). The latter observation further sug gested that IECassociated IL12Rβ2 plays a role in the control of TSLP production by IECs. However, when the expression of TSLP was assessed in the epithelium of naïve IL12Rβ2 −/− mice we did not observe an increase in the level of TSLP compared to naïve wt counterparts ( Figure 7F). Taken together these data point to a sce nario suggesting that IECassociated IL12Rβ2 is an important, but not the only factor controlling TSLP production by IECs. Overall, taken together these data demonstrated that the downregulation of IECderived TSLP following oral administration of IL12 is not directly linked to IL12 engagement with the IL12Rβ2 chain on IECs, thus suggesting that IL12 might trigger the production of additional factor(s) that in turn controlled the production of TSLP by the intestinal epithelium. Finally, as a corollary parallel analysis showed that administration of rLcIL12 suppressed allergic reac tion without downregulating levels of IECderived IL33 while at the same time partially suppressing allergyassociated expression of OX40L ( Figure S4 in Supplementary Material). The latter observation pointed to a synergistic action of TSLP and IL33 on the expression of the allergycritical OX40L rather than an exclusive role of IL33 as previously suggested (25).

DiscUssiOn
Here, we report that mouse IECs expressed the IL12Rβ2 but not the complementary IL12Rβ1 chain that functions as an incom plete, yet functional IL12R that plays a role in lymphoepithelial cross talk. In the mouse system, the two β1 and β2 components of the IL12R have specific role in IL12mediated binding and signaling. IL12Rβ1 chain is the primary binding site confer ring both high and low affinity, instead the IL12Rβ2 only binds IL12 with low affinity (26). However, these studies carried out in IL12Rβ1 −/− mice initially suggested that the β1 component was required for both high affinity binding and for immune cells response to IL12 it was later showed that only the IL12Rβ2 chain was strictly required for IL12 signaling (15). Indeed, no IL12 mediated biological functions could be detected in IL12Rβ1 + /β2 − mice. Combined, these studies suggested that in the mouse system the IL12Rβ1 is primarily responsible for binding and IL12Rβ2 is further required for IL12 signaling. Our study extends this concept showing that while this is the case for immune cells the expression of the sole β2 chain of the IL12R on IECs might be an important component of the lymphoepithelial cross talk. The expression of IL12Rβ2 on IECs depended on immunederived stimuli. This concept stemmed from several observations. A mouse smallintestine derived epithelial cell line and primary intestinal organoids cultured in the absence of gutderived immune cells did not express the IL12Rβ2 chain. By contrast, the presence of gutderived immune cells triggered the expression of the IL12Rβ2 subunit by both an IEC line and primary intestinal organoids. This is in line with previous reports showing that the expression of the IL12Rβ2 but not of the IL12Rβ1 is regulated by signals (i.e., cytokines) originating from the immune system (22). Although the nature of the signals remains to be determined it appeared that both soluble factor(s) and celltocell interaction are important to trigger the expression of IL12Rβ2. This was also suggested by the observation that in vivo, the expression of the IL12Rβ2 chain was located at the basolateral domain of IECs at the interface with local lpimmune cells.
IL12 ligation with the conventional IL12Rβ1/β2 complex signaled through the Janus kinase (JAK)signal transduce pathway in both human and mouse (27,28); in particular, the critical role of Stat4 in IL12R signaling has been demonstrated in vivo via the use of Stat 4deficient mice (29). On the other hand, although in IL12Rβ1deficient humans the presence of the sole IL12Rβ2 chain was not sufficient to confer full IL12 responsiveness via JAKmediated pathway, T cells from these individuals could still develop into TH1 cells and respond to IL12 (22). In this case, IL12 signaling occurred via an IL12Rβ1independent pathway that involved MAPK pathways. In agreement with these observa tions, we reported that IL12/IL12Rβ2 ligation on freshly isolated IL12Rβ1 − /β2 + IECs induced activation of phophoMAPKp38 in a manner similar to that obtained by a non-specific MAPK activa tor. At present the possibility that other surface molecules might be associated with IL12Rβ2 on IECs to form an additional variant of the IL12R that operates via MAPK cannot be ruled out; however, our results, strongly suggested that the presence of IL12Rβ2 alone on IECs is sufficient to confer responsiveness to IL12.
The presence of a functional, albeit incomplete IL12R on IECs suggested that the IL12/IL12Rβ2 ligation could be important in controlling the production by IECderived cytokines. We tested this hypothesis in a mouse model of food allergy. Its utilization was prompted by the notions that in this model the allergic sen sitization to food components (peanuts) was characterized by the lack of IL12 in the gut but not in the systemic immune system (10); in addition, using the same model others have previously shown that oral delivery of liposome encapsulated recombinant IL12 suppressed allergic reaction (11). We report here that in chimeric mice the lack of IECassociated IL12Rβ2 significantly reduced but not completely abolished the antiallergic capability of a L. lactis strain genetically engineered to secrete bioactive IL12. This result suggested that IL12 delivered orally via the recombinant bacterial vector produced the antiallergic activity by engaging a variety of IL12 responding cells either locally or systemically that included IECs. This result could not be attrib uted to an unsuccessful BM engraftment in IL12Rβ2 −/− recipi ents; in these mice the production of IFNγ following systemic administration of IL12 was completely restored at 8 weeks after transplant of BM from wt mice.
Oral delivery of LcIL12 had a significant impact on the levels of TSLP produced by IECs in sensitized mice. The production of TSLP by IECs is an important example of the importance of the tight regulation of the finely tuned cytokine network in the gut. On one hand, TSLP plays a critical role in maintaining the intestinal immune homeostasis by shaping the antiinflammatory features of dendritic cells (2) and Treg (30); on the other, strong evidence suggested that its overexpression is instrumental for the development of allergic reactions (23,24). We found, in agreement with another report (25) that mice undergoing the sensitization protocol employed in our study displayed a significant increase of TSLP production by IECs during and following the sensitizing procedure. However, oral administration of rLcIL12 significantly downregulated TSLP production by IECs while at the same time supressing allergic symptoms. The action of IL12 on TSLP pro duction however did not appear to be directly linked to the IL12/ IL12R ligation on IECs. Naive IL12Rβ2 −/− mice showed only a slight increase of IECderived TSLP, thus suggesting that the pro duction of TSLP by IECs is under the control of multiple factors. It is known that the production of certain cytokines, such as IL13 required the simultaneous presence of two cytokines (i.e., IL33 and IL7) (31) and this notion lends support to the hypothesis that a dual signal might be required also to control the production of TSLP by IECs. This would suggest that IL12 mediated regula tion of TSLP production required the synergy with factor(s) that FigUre 5 | Effects of oral delivery of rLc-IL12 on IgE levels and body temperature in sensitized wild-type (wt) and chimeric mice. Administration of rLc-IL12 by oral gavage in a single dose to wt mice 24 and 2 h before administration of the sensitizing mixture significantly reduced the levels of allergen (peanut)-specific IgE antibody (a) and prevented the drop of the body temperature as assessed by rectal probe (B) showing that oral delivery of IL-12 via a genetically engineered strain of Lactococcu lactis could suppress food allergic reaction by restoring appropriate levels of IL-12. By contrast, administration of IL-12 in chimeric IL12Rβ2 −/− mice (BM transplant from IL-12Rβ2 −/− →IL-12Rβ2 −/− ) had no effect. Indeed, neither plasma levels of IgE nor body temperature were affected by treatment with rLc-IL12 (c,D). The pattern was different in chimeric IL-12Rβ2 ∆IEC mice (BM transplant from WT→IL-12Rβ2 −/− ) that lacked the IL-12Rβ2 exclusively on intestinal epithelial cells (IECs) (e,F). In these mice, oral delivery of rLc-IL12 afforded a reduced protection against food allergy. Administration of rLc-IL-12 significantly reduced plasma levels of IgE compared to rLc-IL12 untreated sensitized mice; however, the extent of the reduction was much less pronounced compared to what was observed in fully competent IL-12Rβ2 mice. Similarly, oral administration of rLc-IL12 did not completely abolished the allergy-associated drop in body temperature although the drop was significantly (**) less pronounced compared to sensitized rLc-IL12 untreated mice. Statistical analysis (8-10 mice group) was carried out by (ANOVA) Wilcoxon rank test. (*, **) indicates statistical differences. p < 0.05. FigUre 6 | Efficacy of rLc-IL12 on additional parameters of food allergy in sensitized wild-type (wt) and chimeric mice. Oral administration of IL-12 reduced levels of plasma histamine by 75% in sensitized wt mice compared to IL-12 untreated syngeneic mice (a). By contrast, no effects were seen in chimeric IL-12Rβ2 −/− mice; instead the IL-12 mediated reduction of plasma histamine in chimeric IL-12Rβ2 ∆IEC totaled only to 45%. Also, levels of IL-4 produced by splenocytes following in vitro recall allergen challenge showed a similar pattern (B). Production of IL-4 in wt mice treated with rLc-IL12 dropped 67% compared to sensitized untreated mice; the same treatment was not effective in chimeric IL-12Rβ2 −/− mice and only partially effective in IL-12Rβ2 ∆IEC mice (54%). By contrast, levels of IL-5 following in vitro recall allergen challenge were reduced by rLc-IL12 only in wt but not in either chimeric mouse variants (c). Hematocrit was evaluated as a measure of vascular leakage (D). Administration of rLc-IL12 restored near physiological levels of hematocrit in wt mice; no effects were seen in chimeric IL-12Rβ2 −/− mice and only partial restoration of the hematocrit was seen in chimeric IL-12Rβ2 ∆IEC . Finally, representative histograms showed that levels of IL-12Rβ2 on intestinal epithelial cells (IECs) did not vary between sensitized (F) and naïve mice (e). Statistical analysis (8-10 mice group) was carried out by (ANOVA) Wilcoxon rank test. *p < 0.05.
are suppressed/downregulated in sensitized mice but that are still retained in nonsensitized IL12Rβ2 −/− mice. Overall, our study suggested that IL12 and TSLP are part of an important regulatory cytokine loop operating at the interface between the intestinal epithelium and the underlying immune system that ultimately plays a role on the host's ability to mount appropriate immune responses. Indeed, it has been shown previously that TSLP directly suppressed the production of IL12 in DCs (32); here, we report that in turn IL12, in combination with other factors plays a role in controlling the production of TSLP by IECs.
The utilization of a mouse model of food allergy to understand the biological role of IL12Rβ2 in controlling the production of IECderived cytokines led to the parallel finding that the suppres sion of food allergy by IL12 was indirectly linked to the downreg ulation of TSLP but did not affect levels of another allergycritical epitheliumderived cytokine, IL33. Instead, at the same time oral delivery of IL12 partially suppressed allergyassociated expres sion of OX40L. This was rather surprising considering that IL33 but not TSLP was previously indicated as having a critical, non redundant role in allergy (25). Thus, our work in agreement with a recent report (33) argued against the concept that the alteration of a single immunoregulatory pathway (i.e., IL33) is at the core of allergic reactions. Our data point to a synergistic action of TSLP and IL33 on OX40L expression, as also more recently shown in a rhinovirusmediated allergic reaction to aeroallergen (34), rather than an exclusive IL33mediated effect as previously suggested (25). In summary, taken together these data show that IL12Rβ2 chain functions as an incomplete yet functional IL12R on IECs and overall further highlight the importance of a finely tuned lymphoepithelial cross talk in the intestine in maintaining health.

sUPPleMenTarY MaTerial
The Supplementary Material for this article can be found online at https://www.frontiersin.org/articles/10.3389/fimmu.2018.01177/ full#supplementarymaterial. FigUre 7 | IL-12-mediated downregulation of IEC-derived TSLP does not occur directly via IEC-associated IL-12Rβ2. RT-PCR analysis of the gut epithelium (a) showed that oral delivery of the adjuvant cholera toxin (CT) induced a significant increase of TSLP in the gut epithelium similar to that observed following delivery of the complete sensitizing mixture (s-untreated) compared to control mice (naive). Allergy-associated levels of TSLP were abrogated by treatment with rLc-IL12. This notion was further confirmed by immunohistochemistry. Immunofluorescence analysis with anti-TSLP antibody did not detect levels of TSLP in the IECs of both control (B) and rLc-IL12-treated mice (c). In contrast, specific anti-TSLP staining (green) was observed in the epithelium of sensitized rLc-IL-untreated mice (D), detail in (e). Sections were counter-stained with antiactin antibody (red). Finally, expression of TSLP gene in the intestinal epithelium of naïve IL-12Rβ2 −/− mice (F) showed only a minor increase that did not reach statistical significance compared to wt syngeneic mice. In this case, the thymus from wt mice was used as positive control. Transcript analysis was carried out in replicate mice (four to six mice/group) in two independent assays. Data were analyzed using a Student's t-test followed by application of the Bonferroni correction the level of which was set at p ≤ 0.0125. (*) indicate statistical difference. p < 0.05.