If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Recent studies traced inflammatory bowel disease in some patients to deficiency of CD55 [decay-accelerating factor (DAF)], but the mechanism underlying the linkage remained unclear. Herein, we studied the importance of DAF in enabling processes that program tolerance in the gut and the eye, two immune-privileged sites where immunosuppressive responses are continuously elicited. Unlike oral feeding or ocular injection of ovalbumin in wild-type (WT) mice, which induced dominant immune tolerance, identical treatment of DAF–/– mice or DAF–/– to WT bone marrow chimeras did not. While 10% to 30% of mesenteric and submandibular lymph node CD4+ cells became robust T-regulatory cells (Tregs) in WT forkhead box P3 (Foxp3)–green fluorescent protein mice, few in either site became Tregs with little suppressor activity in DAF–/– Foxp3–green fluorescent protein mice. Phenotyping of CD103+ dendritic cells (DCs) from the ovalbumin-fed DAF–/– mice showed impaired expression of inducer of costimulation (ICOS) ligand, programmed death receptor 1-ligand 1 (PD1-L1), CxxxC chemokine receptor 1 (Cx3CR1), CCR7, and CCR9. Analyses of elicited DAF–/– Foxp3+ Tregs showed reduced expression of interferon regulatory factor 8 (IRF-8)/aldehyde dehydrogenase 1 family member A2 (Aldh1a2) and glycoprotein A repetitions predominant/latency-associated protein associated with Treg transforming growth factor-β production and presentation, as well as integrin β6/integrin β8 associated with Treg and CD103+ DC transforming growth factor-β release. Thus, DAF is required for the properties of CD103+ DCs and their naïve CD4+ cell partners that together program tolerance.
Peripheral tolerogenic responses interdict T-effector (Teff) cell activation against nonpathogenic environmental antigens at mucosal surfaces. In clinical settings, the tolerogenic responses prevent the development of inflammatory bowel disease (IBD), asthma, and other allergic and/or autoimmune conditions. In transplantation, their induction against donor tissue promotes successful organ engraftment. In normal health, tolerogenic responses operate continuously to maintain homeostasis in sites where immune activation must be suppressed. Among these sites are the anterior chamber (a.c.) of the eye and the gastrointestinal (GI) tract, in each setting in coordination with the respective lymphoid organ(s). Determining how acquisition of the tolerogenic properties of antigen-presenting cells (APCs) and CD4+ T cells in these sites mechanistically occurs would provide new insights regarding physiological programming of tolerogenic outcomes.
With respect to homeostasis in the eye, early work showed that foreign tissue survived longer when placed in the a.c. as opposed to the peritoneum or the skin,
showing that in contrast to s.c. injection of ovalbumin (OVA), which evoked a cytotoxic T-lymphocyte response and robust delayed-type hypersensitivity (DTH), a.c. preinjection prevented these and other Teff responses on subsequent s.c. OVA challenge. This phenomenon was designated anterior chamber–associated immune deviation in that it conferred long-lasting dominant tolerance. With respect to the GI tract, early work documented an analogous immune altered state. The immune altered phenomenon there, first termed immunologic unresponsiveness,
The Collaborative Corneal Transplantation Studies Research Group The collaborative corneal transplantation studies (CCTS): effectiveness of histocompatibility matching in high-risk corneal transplantation.
that cognate DC-CD4+ cell partners jointly produce C3a and C5a from endogenously synthesized complement. The two anaphylatoxins establish autocrine signaling loops with C3a and C5a receptors (C3ar1 and C5ar1, respectively) in each partner, which provide costimulatory and survival signals to CD4+ cells.
have highlighted the requisite role of local complement production and its activation in human CD4+ cell responses and provided evidence that the signaling operates intracellularly.
Work building on our finding that autocrine C3ar1/C5ar1 transduction promotes Teff activation led to the insight that the absence of transduced C3ar1/C5ar1 signals into CD4+ cells conversely favors the differentiation of naïve CD4+ cells into pTregs. The extent, however, to which DAF control of immune cell C3ar1/C5ar1 signaling is obligatory or modulatory to the physiological imprinting of Treg commitment has remained unclarified. Moreover, the linkage of DAF's activity with the phenotypic properties of DCs and CD4+ cells that play requisite roles in immune homeostasis also has remained unknown. An important example is the physiological induction of de novo peripheral tolerance, an issue of broad immunologic significance.
Autoimmune uveitis results from broken tolerance in the a.c. Both anterior and posterior disease can occur, and both are connected with CD4+ cell production of tumor necrosis factor-α, interferon (IFN)-γ, and IL-17
identified 11 patients with early-onset (aged <2 years) protein losing enteropathy. The patients (six males and five females) belonged to eight families. Their disease was characterized by malnutrition, vomiting, diarrhea, hypoalbuminemia, hypogammaglobulinemia, lymphangiectasia, and inflammation with lymphatic infiltrates. Three patients had thromboses. Family histories included two patients who died. Immunologic analyses showed increased C3 antigen on CD4+ cell blasts (presumed to derive from plasma C3 protein) and increased CD4+ cell production of tumor necrosis factor-α and IFN-γ. Although increased CD4+ cell production of C5a was noted, the pathogenic mechanism underlying the enteropathy was not characterized.
Past work by several investigators on mucosal tolerance has implicated discrete subsets of DCs
in playing critical roles. Important among these subsets is CD103+ DCs. After antigen processing, these DCs migrate to the regional lymph node (LN) and convert CD4+ cells into pTregs. In view of the critical role of this DC subset in the induction of tolerance, we studied whether the loss of DAF function on CD103+ DCs and/or CD4+ cells could underlie the pathogenesis of the above described cases of enteropathy and predispose to IBD and uveitis. The interplay between DAF function, the properties of CD103+ DCs that enable their migration to the LN, and the tailoring of the properties of CD4+ cells that together evoke tolerogenic responses were studied. It was found that DAF control of C3ar1/C5ar1 signaling accounts for many DC and CD4+ cell phenotypic properties reported to be integrally involved in induction of tolerance.
Materials and Methods
Reagents and Antibodies
C5a was from Cell Sciences (Newburyport, MA). C3ar1-A was purchased from Calbiochem (now EMD Millipore, Burlington MA). C5ar1-A synthetic peptide was kindly provided by John Lambris (University of Pennsylvania, Philadelphia, PA). Anti-mouse C3a (I87-1162) and C5a (I52-1486) monoclonal antibodies (mAbs) were from BD Biosciences (San Jose, CA). Mouse IL-2, IL-10, and IFN-γ and human IL-2, IL-10, IFN-γ, and transforming growth factor (TGF)-β1 were from Prospec Bio (Ness-Ziona, Israel). Anti-murine C5ar1 (20/70) was purchased from Abcam (Cambridge, UK). Abs against mouse CD11c, CD11b, F4/80, CD103, B7-1 (16-10A1), B7-2 (GL1), PD-L1, and ICOS ligand (ICOS-L) were from BD Biosciences. Anti-CD40 ligand mAb was from BioExpress (Kaysville, UT). Anti-murine C3ar1 (sc-20138) was purchased from Santa Cruz Biotechnology (Dallas, TX). Anti-murine C5ar2 (AKA C5L2; HP8015) was purchased from Hycult Biotech (Wayne, PA). Mouse anti-Foxp3 mAbs and Treg staining kits were from eBiosciences (now Thermo Fisher Scientific, Waltham, MA). Anti-CD4 (RM-4-5; fluorescein isothiocyanate), anti-Thy1.1 [HIS51; phosphatidylethanolamine-cyanine 7 (Cy7)], anti-Thy1.2 (53-2.1; APC), anti-CD45.1 (A20; phosphatidylethanolamine), and anti-CD45.2 (104; APC–eFluor 780) were purchased from eBiosciences. Carboxyfluorescein succinimidyl ester (CFSE), CellTracker Red, and CellTracker Violet were from Invitrogen (Thermo Fisher Scientific). Grade VI chicken egg OVA and human IgG were purchased from Sigma (St. Louis, MO). Complete Freund adjuvant (CFA) containing heat-killed Mycobacterium tuberculosis strain H37Ra and incomplete Freund adjuvant (IFA) were from Difco Labs (Detroit, MI).
Foxp3–green fluorescent protein (GFP) mice were a generous gift from Vijay Kuchroo (Harvard University, Boston, MA). Foxp3-GFP and OT-II mice were bred with Daf1–/– and C3ar1–/–C5ar1–/– mice in our animal facility. Bone marrow transplants were prepared, as previously described.
All studies were approved by the Institutional Animal Care and Use Center (protocol number 2015-0015) in our Assessment and Accreditation of Laboratory Animal Care–accredited facility and conducted according to American Veterinary Medical Association guidelines. Mice were anesthetized with ketamine/xylazine rodent cocktail or isoflurane to reduce suffering. Proparacaine was used as an analgesic to reduce pain caused by intraocular injections.
a.c. Injections and s.c. Challenge
OVA (2 mg/mL) in CFA and OVA (0.5 mg/mL) in IFA were prepared by emulsifying equal volumes of OVA in phosphate-buffered saline (PBS) with each adjuvant. Mice were anesthetized by i.p. injection with 0.6 mL of 1.5 mL ketamine HCl (100 mg/mL), 1.5 mL xylazine HCl (20 mg/mL), and 0.5 mL acepromazine (10 mg/mL). For a.c. injection, the eye was anesthetized topically with proparacaine HCl drops (Alcon Labs, Humacao, Puerto Rico). A paracentesis was performed by inserting a 30-gauge needle at the limbus parallel to the iris. An air bubble then was introduced into the a.c. by injecting 2 μL of air with a Hamilton microsyringe (Hamilton, Reno, NV) fitted with a 33-gauge beveled needle. After this, 50 μg of OVA in 2 μL of PBS or 2 μL of PBS alone was injected. The air bubble sealed the paracentesis, thereby preventing leakage.
Wild-type (WT) OT-II, Daf1–/– OT-II, and C3ar1–/–C5ar1–/– OT-II mice were fed 20 mg/mL OVA through their drinking water for 5 days, after which the mice were immunized with 100 μg of OVA in IFA, as described by others.
CD4+ T cells were isolated from spleens and LNs using the CD4+ negative selection cocktail from Miltenyi Biotec (Bergisch Gladbach, Germany), per the manufacturer's instructions. CD11c+ cells were isolated from spleens, LNs, and eyes (20 mice were used from each genotype for resident eye DC isolation) using the positive selection cocktail from Miltenyi, per the manufacturer's instructions. Both were purified using the Automacs Pro (Miltenyi Biotec). For corneal DCs, 40 mice of each genotype were used.
DTH and Alamar Blue Cytotoxicity Assays
In studies of ocular tolerance, 7 days after a.c. injections, mice were injected subcutaneously with 12.5 μg OVA in IFA in the footpad of the right leg and with PBS in IFA in the footpad of the left leg; and swelling was measured 24 hours later. In studies of oral tolerance, 7 days after oral feeding, mice were injected subcutaneously with 100 μg OVA in CFA at the base of the tail. Seven days later, they were injected subcutaneously with 12.5 μg OVA in IFA in the right leg and with PBS in IFA in the left leg; and swelling was measured 24 hours later.
For cytotoxicity assays, isolated T cells were incubated for 3 days at 37°C with 20 μg/mL OVA. They were then incubated with OVA-transfected EL-4 cells in decreasing ratios and Alamar blue for 24 hours. Cytotoxicity was calculated using the following formula: 100 × ([(AF of targets alone)] − [(AF of mix) − (AF effectors alone)]) × (AF of targets alone), where AF represents the mean of the absolute fluorescence units for the wells minus the average fluorescence units of the media alone.
ELISPOT Assays, Cytokine Enzyme-Linked Immunosorbent Assays, and Ab Isotype Enzyme-Linked Immunosorbent Assays
ELISPOT assays (ImmunoSpot, Shaker Hts, OH) and cytokine assays were performed as previously described.
For quantitation of cytokines produced in the eye, explants of eyes were incubated for 48 hours at 4°C in RPMI 1640 medium in the absence of antigen. Supernatants were assayed by the enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN), per the manufacturer's instructions. Plasma samples were collected 10 days after s.c. OVA injection.
Levels of anti-OVA antibodies were measured using the mouse monoclonal Ab ID/SP kit (Zymed, San Francisco, CA).
Foxp3+ Treg Induction and Foxp3+ Treg Suppressor Assays
Foxp3+ or Foxp3– Cell Isolations and Treg Induction
CD4+ cells were isolated using negative selection beads (Miltenyi Biotec). Foxp3+ or Foxp3– cells from Foxp3-GFP transgenic mice were sorted using APC-labeled anti-CD4 mAb (Thermo Fisher Scientific) and BD Biosciences Aria Facsorter. Foxp3+ or Foxp3– cells (1 × 105) were incubated for 24 hours at 37°C with anti-CD3/28 Dynabeads (Invitrogen, Carlsbad, CA), per the manufacturer's instructions, and 2 μg/mL IL-2 (Peprotech, Rocky Hill, NJ) in the presence and absence of WT DCs in RPMI 1640 medium.
Foxp3+ Staining Assays
For both human and murine Treg staining assays, the relevant Foxp3+ T-regulatory staining kits (236A/E7 for human and FJK-16s for mouse) were purchased from Thermo Fisher Scientific and used per the manufacturer's instructions.
Murine Treg Suppressor Assays
Sorted GFP− Foxp3-GFP OT-II CD4+T cells (1 × 106), labeled with CellTracker Red (Thermo Fisher Scientific), were stimulated with 1 × 105 autologous CD11c+ DCs and OVA323-339 alone or with various ratios of sorted GFP+ Foxp3-GFP CD4+ cells. The cells were cultured for 3 days in 96-well flat-bottom plates, and CellTracker Red dilution was analyzed by fluorescence-activated cell sorting.
Phenotyping and Flow Cytometry
Spleens, mesenteric LNs (mLNs), and submandibular LNs (smLNs) were isolated from mice, incubated for 30 minutes in spleen dissociation buffer (Miltenyi Biotec), passed through a 70-μm cell strainer, and incubated for 2 minutes in isotonic red blood cell lysis buffer. After Fc receptor blocking with anti-CD16/32 Ab (BD Biosciences), cells were stained with anti-CD4 and anti-CD11c+ anti-CD103+ or anti-F4/80+ anti-CD11b mAbs. Cells were then further stained for costimulatory molecules, coinhibitory molecules, major histocompatibility complex (MHC)-II, chemokine receptors, or complement fragment receptors, as described in each experiment, and assayed using an Aria, LSR I, or LSR II flow cytometer (BD Biosciences). Flow data were initially analyzed via CellQuest (BD Biosciences) or FACSDiva (BD Biosciences) and then subsequently with FlowJo version 10 (FlowJo LLC, Ashland, OR). Data are presented as mean fluorescent intensity.
All experiments were repeated at least twice. Statistical analyses were performed using a two-tailed t-test using Microsoft Excel (Microsoft Corp., Redmond, WA) or GraphPad Prism 6.0 (GraphPad Software, San Diego, CA). All data are presented as means with SD, unless otherwise noted.
Because published studies of ocular and oral tolerance in mice
in large part have used OVA as antigen, OVA was used in the experimental systems.
DAF Is Required for Ocular Tolerance
To study the extent to which DAF function is required for ocular tolerance, the outcome of injecting OVA into the a.c. of Daf1–/– (murine equivalent of human DAF) mice versus the a.c. of WT (Daf1+/+) littermates was compared. After 7 days, the mice were challenged subcutaneously with OVA + IFA in the right leg and with PBS + IFA in the left leg (control); and 24 hours later, leg swelling and lesion histology were measured. WT mice that received OVA in the a.c showed approximately threefold reduced DTH compared with PBS-treated controls (Figure 1A). In contrast, identically treated Daf1–/– mice developed lesions (Figure 1A) with even more robust leukocyte infiltration than those of WT recipients that received PBS alone in the a.c. (ie, were susceptible to characteristic DTH) (Figure 1B). The data in WT, but not Daf1−/−, mice are thus consistent with the anterior chamber–associated immune deviation phenomenon in which antigen that is initially seen in the a.c. elicits a tolerogenic response that inhibits OVA-specific DTH. This is opposed to a.c.-injected PBS in which antigen initially seen systemically elicits a Teff response and DTH consequently is not inhibited. Consistent with the effect being antigen specific, DTH in response to s.c. injection of the nonrelevant peptide was not altered in either WT or Daf1–/– mice a.c. preinjected with OVA (data not shown).
The findings that DTH was abrogated in WT mice but heightened in Daf1–/– mice argued that a.c. injection in DAF's presence and absence had opposing extraocular effects. In accordance with DAF playing an obligate role in conferring dominate systemic tolerance, splenocytes from OVA a.c. preinjected WT littermates showed <2% of IFN-γ–producing CD4+ cells in response to whole OVA, whereas OVA a.c. preinjected Daf1–/– mice conversely showed approximately twofold heightened IFN-γ–producing CD4+ cells (Figure 1C). A comparable difference was observed with the immune dominant peptide of OVA (ie, OVA323-333) (Figure 1C). Induction of OVA-specific CD8+ T cells similarly was diminished. Although all WT littermates injected in the a.c. with OVA exhibited 75% decreased cytotoxic activity against OVA-transfected (syngeneic H-2b) EL-4 targets,
on corneal transplants in the context of ocular tolerance implicated shifts away from complement-fixing Ig isotypes. Anti-OVA antibodies produced in Daf1–/– mice that were a.c. preinjected with OVA showed significant increases (P < 0.05) in (complement-fixing) anti-OVA IgG2a and IgG2b antibodies as opposed to other isotypes (Figure 1E). No alterations in anti-OVA Abs occurred in response to a nonrelevant antigen (data not shown), consistent with specificity.
Abrogated Ocular Tolerance in Daf1–/– Mice Is due to Loss of DAF's Restraint of Immune Cell C3ar1/C5ar1 Signaling
Because DAF function can impact multiple processes, each process was examined in the context of the ocular tolerogenic response. DAF's interaction with CD97, a seven-membrane spanning surface protein expressed on leukocytes and other cell types,
The tolerogenic response to OVA was intact in CD97–/– mice preinjected with OVA into the a.c. (Figure 1F), excluding DAF-CD97 interactions as being involved.
To test whether increased sensitivity of DAF–/– ocular cells to systemic complement attack accounted for the disabled tolerogenic response, lethally irradiated WT recipients were reconstituted with bone marrow cells from Daf1–/– mice and vice versa. Eight weeks later [after confirming >90% donor bone marrow was comparable in both groups (data not shown)], the ocular OVA challenge experiments were repeated. In WT → Daf1–/– chimeras, both OVA-specific CD4+ cell IFN-γ (Figure 1, G and H) and OVA-specific CD8+ cell cytotoxic responses (Figure 1, G and H) were markedly suppressed, as in WTs. In contrast, potent and specific OVA-reactive CD4+ cell and CD8+ cell responses occurred in the Daf1–/– → WT chimeras, pointing to effects mediated by immune cells. Because DAF on immune cells suppresses C3ar1/C5ar1 signaling and affects cytokine production as well as cooperates with other regulators in circumventing systemic complement attack,
the ocular tolerance protocol was repeated in Daf1–/–C3ar1–/–C5ar1–/– mice. The tolerogenic response was completely restored in Daf1–/–C3ar1–/–C5ar1–/– mice (Figure 1I), indicating that the failed tolerogenic response in Daf1–/– mice resulted from the loss of DAF restraint of C3ar1/C5ar1 signaling in immune cells. Prior findings
that addition of C5a to WT CD4+ cells simulated the effect of DAF deficiency supported the interpretation that DAF's function in the context of tolerance is repression of autocrine C3ar1/C5ar1 signaling in immune cells.
DAF Regulatory Activity Is Required for Induction of Tregs against a.c.-Injected Protein
To decipher mechanisms accounting for the requirement of DAF for the induction of ocular tolerance, the a.c. tolerance protocol was repeated in Daf1−/− and WT mice, after which the mice were challenged subcutaneously with OVA in CFA. Seven days later, ocular explants were isolated and cultured with OVA for 48 hours. Supernatants of explants from OVA a.c.-injected WTs showed approximately 1.8-fold more TGF-β1 compared with those from PBS-injected controls (Figure 2A). Explants of identically treated Daf1–/– mice conversely showed approximately 1.5-fold less TGF-β than those from WT mice (Figure 2A). Consistent with these findings, splenic cells from OVA a.c.-injected WTs produced increased TGF-β1, IL-10, and little IFN-γ in response to OVA, whereas splenic cells from OVA a.c.-injected Daf1–/– mice produced approximately fourfold less TGF-β1, more than ninefold less IL-10, and approximately twofold more IFN-γ (Figure 2B) than WTs.
Because TGF-β1 and IL-10 are connected with Treg induction and function, WT and Daf1–/– Foxp3-GFP reporter mice were prepared and pretreated with OVA in the a.c., after which Treg induction was assessed. The gating strategy is shown in Supplemental Figure S1A. Basal percentages of Foxp3+ cells among CD4+ cells in the draining smLN and spleen of unimmunized WT mice were 5.7% ± 0.7% and 8.0% ± 0.5%, respectively. This compared with only 1.4% ± 0.4% and 3.5% ± 0.4%, respectively, in unimmunized Daf1–/– mice (P < 0.05 for both sites). After a.c. preinjection of OVA or PBS in WT Foxp3-GFP mice and subsequent tail immunization (Figure 1), Foxp3+ cell percentages in the spleen and smLN increased from 5.7% to 6.7% ± 0.8% and from 8.0% to 10.2% ± 0.6%, respectively (P < 0.05 in each site). In contrast, after identical treatment of Daf1–/– Foxp3-GFP mice, no increases from the lower basal Foxp3+ cell levels occurred (ie, from 1.4% ± 0.4% to 1.3% ± 0.5% in the spleen and from 3.5% ± 0.4% to 3.0% ± 0.3% in the smLN; P < 0.0002 and P < 0.0001, respectively, compared with WTs). Flow-sorted Foxp3+ cells from the smLN or spleen (Figure 2C) of the WTs efficiently suppressed OT-II cell proliferation, whereas the Daf1–/– Foxp3+ cells showed minimal, if any, suppressor activity. Thus, in the context of this in vivo a.c. antigen initial encounter system, DAF function is required for Treg induction and the ocular tolerance response.
To distinguish whether the Treg differences reflected the generation of de novo pTregs versus the expansion of preexisting (thymic Tregs and/or pTregs), sorted Foxp3– (GFP–) CD4+ cells were adoptively transferred from WT or Daf1–/– Foxp3-GFP mice (both Thy-1.2) into WT Thy-1.1 recipients. Repeat studies using the above protocol and gating on Thy-1.2 CD4+ cells showed an increase of 3.8% ± 0.04% Thy-1.2 GFP+ cells in recipients of WT (Thy-1.2) Foxp3– cells, compared with 0.8% ± 0.02% in recipients of Daf1–/– (Thy-1.2) Foxp3– cells (P < 0.008) (Figure 2D). This argued that the Tregs induced by the a.c. OVA injection derived from uncommitted CD4+ cells. Although only small numbers of Thy-1.2 Daf1−/− Foxp3+ cells could be retrieved from the recipients of GFP–Daf1–/– CD4+ cells, their inclusion at a 4:1 Teff/Treg ratio in Treg suppression assays confirmed that they possessed little suppressive activity (Figure 2E). This contrasted with Thy-1.2 Foxp3+ WT cells from the recipients of Foxp3– WT cells (Figure 2E). These data argued that failed ocular tolerance in Daf1–/– mice is connected with impaired induction as well as impaired function of pTregs. The data, thus, argued that DAF control of DC and CD4+ cell C3ar1/C5ar1 signaling is obligatory for the in vivo induction of tolerance in this site rather than constituting a redundant or modulatory process.
DAF Is Required for Oral Tolerance
Unlike the a.c., the GI tract is in constant contact with commensal bacteria and is exposed to numerous dietary antigens, which past studies
as being critical for the oral tolerogenic response.
To determine the importance of DAF in the induction of oral tolerance, OVA or PBS was added to the drinking water of naïve WT and Daf1–/– (OVA-specific) OT-II mice for 5 days, after which both groups were immunized subcutaneously with OVA in IFA. Seven days later, CD4+ cells were harvested from the mLNs. Intracellular staining for Foxp3 showed that in WT mice fed PBS, the percentage of OT-II cells that were Foxp3+ was 9.0% ± 0.5%; and in those fed OVA, the percentage increased to 30% ± 2% (Figure 3A). In contrast, in Daf1–/– OT-II mice fed PBS, <5.4% ± 4% of the OT-II cells were Foxp3+; and OVA feeding did not significantly increase the percentage (Figure 3A). Culture supernatants of mLN CD4+ cells and OVA-primed DCs from WT OT-II mice fed OVA produced approximately 2.5-fold more TGF-β1 and approximately 5-fold less IFN-γ compared with those of PBS-fed controls (Figure 3B). In contrast, culture supernatants from Daf1–/– OT-II mice fed OVA showed a slightly smaller increase in TGF-β1 production compared with those of PBS-fed controls, but no decrease in IFN-γ production (Figure 3C). As occurred in the a.c., DTH in OVA-fed WTs decreased from 1.5 to 0.9 mm, whereas DTH did not decrease in OVA-fed Daf1–/– mice (Figure 3D). Likewise, sorted CD25+ cells (>95% Foxp3+ in other aliquots) harvested from mLNs of WT OT-II mice showed robust suppressor activity, whereas those harvested from mLNs of Daf1–/– OT-II mice exerted little suppressor activity (Figure 3E).
To determine whether the differences in Treg induction and function reflected differences in the induction of de novo pTregs, Thy-1.1 WT recipients of sorted Thy-1.2 Foxp3– WT and Daf1–/– Foxp3– cells were studied, as in ocular tolerance. In view of the finding that ocular tolerance in Daf1–/– → WT chimeras was restored in Daf1–/–C3ar1–/–C5ar1–/– mice, linking DAF's repression of C3ar1/C5ar1 signaling with tolerance (Figure 1I), C3ar1–/–C5ar1–/– mice were included in the comparisons of oral tolerance. OT-II Foxp3-GFP mice were prepared on each genetic background, and sorted WT, Daf1–/–, or C3ar1–/–C5ar1–/– Foxp3– OT-II cells (all Thy-1.2) were transferred to Thy-1.1 recipients, and the protocol used with the manipulated mice on each background was followed. Thy-1.2 Foxp3+ cells increased from 10% to 30% in OVA-fed WTs, from 9% to 64% Foxp3+ cells in OVA-fed C3ar1–/–C5ar1–/– mice, and from <0.5% to only 12% in OVA-fed Daf1–/– mice (Figure 3F). Cultures of CD4+ and CD11c+ cells plus OVA from the mLN of Daf1–/– mice generated approximately eightfold less TGF-β and threefold more IFN-γ than those of recipients of WT cells (Figure 3, G and H). Conversely, cultures from C3ar1–/–C5ar1–/– mice generated 140-fold more TGF-β1 and approximately threefold less IFN-γ than WTs (Figure 3, G and H). As observed in ocular tolerance, Thy-1.2 Foxp3+Daf1−/− OT-II cells, recovered from the mLN, showed little, if any, suppressor activity, whereas Foxp3+C3ar1–/–C5ar1–/– OT-II cells showed more than fourfold more suppressor activity than those recovered Foxp3+C3ar1–/–C5ar1–/– OT-II cells (Figure 3I). Returning to the comparisons of DTH in Daf1−/− versus WT mice, now including C3ar1–/–C5ar1−/− mice (Figure 3D), C3ar1–/–C5ar1–/– OT-II mice fed OVA showed a more profound reduction in DTH than in WTs (Figure 3D). Consistent with the failed immunosuppressive activity of Daf1−/− Tregs, intracellular staining of CD4+ cell precursors on each genetic background showed that Daf1−/− CD4+ cells produced markedly lower amounts of TGF-β and IL-10, whereas C3ar1–/–C5ar1–/– cells produced markedly more of both immunosuppressive cytokines (Supplemental Figure S2, A and B). These findings indicated that repression of C3ar1/C5ar1 signaling by DAF is required for oral tolerance. They argued that this control is an obligate process.
MHC Class II Is Up-Regulated and PD-L1/ICOS-L Are Constitutively Down-Regulated on APCs in the smLN and mLN of Naïve Daf1–/– Mice
To decipher mechanism(s) accounting for the abrogated versus enhanced tolerogenic responses in Daf1–/– versus C3ar1–/–C5ar1–/– mice, DCs and CD4+ cells were first compared for their C3a/C5a production and C3ar1/C5ar1 levels in both cases, assaying levels both on the cell surface and intracellularly. These data were then correlated with immune cell processes connected in the literature with tolerance. Tolerogenic responses occur as a result of the following: i) uptake of antigen by resident APCs, which are immature; ii) transit of the primed APCs to the draining LN; and iii) stimulation of (naïve) cognate CD4+ cells by the immature entrant APCs.
To determine how disabled DAF affects APC immaturity in the eye and gastrointestinal tract, MHC class II and costimulatory molecule expression levels were compared on unstimulated resident DCs in the cornea, smLN, and mLN of naïve WT and Daf1–/– mice. The gating strategy is shown in Supplemental Figure S1B. While unstimulated DCs in the cornea of naïve WT mice showed constitutive PD-L1 (B7-H1)
and ICOS-L (B7-H2) expression in conjunction with low MHC class II expression and essentially no B7-1,2 or CD40 expression (Figure 4A), those in naive Daf1–/– mice showed virtually no PD-L1 and ICOS-L expression, but 450% to 1200% (P < 0.01) increased constitutive expression levels of MHC class II, 560% (P < 0.01) higher levels of B7-1, 450% (P < 0.01) higher levels of B7-2, and 1200% (P < 0.01) higher levels of CD40. Resident corneal CD11b+ F4/80+ macrophages (Figure 4D), smLN DCs and macrophages (Figure 4, B and E), as well as mLN DCs and macrophages (Figure 4, C and F), in WT and Daf1–/– mice exhibited the same differences. A similar pattern was observed for inguinal LN (iLN) DCs (Supplemental Figure 1C). The increases in MHC class II and costimulatory molecule expression levels were completely reversed in Daf1–/–C3ar1–/–C5ar1–/– mice or C3ar1–/–C5ar1–/– mice (Figure 4A). As in Figure 1I, Daf1–/–C3ar1–/–C5ar1−/− cells were included to show that DAF's effect in the context of tolerance was solely related to its repression of autocrine C3ar1/C5ar1 signaling. Consequently, the effects of its deficiency were completely overcome by disabled C3ar1/C5ar1 signaling in immune cells. More important, selective studies showed the increases in PD-L1 and ICOS-L in WT and C3ar1–/–C5ar1−/− DCs were accompanied by increases in C5L2 (ie, C5ar2), a C5ar G protein–coupled receptor (GPCR) devoid of a G protein, which is implicated in scavenging C5a so as to prevent C5ar1 signaling. Also more important, studies with human plasmacytoid DCs and CD4+ cells silenced in either DAF or C3ar1/C5ar1 with siRNA(s) showed comparable changes (Figure 4, G and H), indicating that these findings apply in human immune cells.
Although the previous studies examined immune cell C3ar1/C5ar1 signaling in the context of proinflammatory experimental autoimmune encephalitis, they did not study it in the context of peripheral tolerance.
They also did not study the properties of immune cells that program the de novo induction of tolerance. DCs in the smLN of Daf1–/– mice contained less TGF-β1 and IL-10 than DCs in the smLN of WT mice (Figure 5A), whereas those of Daf1–/–C3ar1–/–C5ar1–/– mice contained more of both immunosuppressive cytokines (P < 0.01 versus WTs). There was greater up-regulation of TGF-β and IL-10 production in Daf1−/−C3ar1−/−C5ar1−/− T cells compared with WT T cells. This increase is because of the fact that disabled C3ar1/C5ar1 signaling overcomes the effect of DAF deficiency because augmented C3a/C5a production, resulting from DAF's absence, is unable to signal through C3ar1/C5ar1. Moreover, the C3ar1/C5ar1 deficiency induces immunosuppressive TGF-β and IL-10 that would overcome any effect of DAF. Comparable differences were obtained for DCs in the mLN (data not shown). Co-cultures of WT DCs and Daf1–/– CD4+ cells produced less active TGF-β1 and IL-10 than co-cultures of WT DCs and WT CD4+ cells. In contrast, co-cultures of WT DCs and Daf1–/–C3ar1–/–C5ar1–/– CD4+ cells produced more active TGF-β1 and IL-10 (Figure 5B), indicative of CD4+ cells modifying APC partners to become tolerogenic. Accordingly, WT DCs in co-cultures containing stimulated Daf1–/– CD4+ cells expressed markedly up-regulated B7-1,2 and CD40 levels compared with DCs in co-cultures with stimulated WT CD4+ cells, whereas WT DCs in co-cultures with stimulated Daf1–/–C3ar1–/–C5ar1–/–CD4+ cells expressed more coinhibitory PD-L1 and ICOS-L (Figure 5C). Adding C3a/C5a or C3ar1/C5ar1 antagonists to WT DCs induced differences corresponding to those in the different genotypes and anti-C3a/anti-C5a treatment could not overcome the maturity of WT Daf1–/– cells (Supplemental Figure S2, C and D)
interactions are centrally implicated in Treg lineage commitment. To determine whether these linkages are mechanistically interconnected with absent C3ar1/C5ar1 signaling in DC-CD4+ cell partners, sorted CD25– WT, PD-1–/–, or ICOS–/– CD4+ cells were incubated with WT DCs, or WT CD4+ cells were incubated with PD-L1–/– DCs, in all cases with IL-2 in the presence of C3ar1 and C5ar1 pharmaceutical antagonists (C3ar1-A and C5ar1-A). After 5 days, the CD4+ cells were assayed intracellularly for Foxp3. While >60% of the CD4+ cells became Foxp3+ in the presence of C3ar1-A/C5ar1-A when both the WT CD4+ cells and DCs possessed PD-1-PD-L1 and ICOS-ICOS-L couplings in PD-1–PD-L1 and ICOS–ICOS-L couplings, <20% became Foxp3+ when PD-1 or ICOS was deficient in the CD4+ cells (Figure 5D) or PD-L1 was deficient in the DCs (Figure 5E). These data showed that the participation of PD-1–PD-L1 and ICOS–ICOS-L couplings in Treg induction is dependent on C3ar1/C5a1 antagonism. To gain mechanistic insight, WT and C3ar1–/–C5ar1–/– CD4+ cells were examined for Src homology region 2 domain-containing phosphatase-2 (Shp-2), a phosphatase tied to PD1 expression.
Anti-CD3/28 activation of flow-sorted Daf1–/–, WT, or C3ar1–/–C5ar1–/– Foxp3– CD4+ cells with IL-2 showed that basal and stimulated Shp-2 as well as Shp-1 mRNA levels in Daf1−/− cells were lower than in WT cells and in C3ar1−/−C5ar1−/− cells. Both Shp-1 and Shp-2 were downregulated in unactivated and activated Daf1−/− CD4+ cells (Figure 5F). These data indicated that the linkages of Shp-2–dependent PD-1–PD-L1 and ICOS–L-ICOS couplings with Treg commitment are tied to absent C3ar1/C5ar1, which our past studies have shown represses phosphoinositide-3 kinase-γ (PI-3Kγ) activation.
but have not been studied in the smLN draining the a.c. CD103+ DCs were analyzed in the mLN. The gating strategy used for examining CD103+ DCs is shown in Supplemental Figure S3. Compared with CD11c+CD11b–CD103+ DCs from WT mice, CD103+ DCs from Daf1–/– mice showed reduced ICOS-L and PD-L1 (Figure 6A), paralleling reduced levels in whole DCs (Figure 4, A and B). They also showed reduced expression of Cx3CR1 involved in priming CD8+ cells,
and CCR9 involved in DC transit to the lamina propria (Figure 6B). CD11c+CD103+ DCs were identified in the smLN in proportions corresponding to those in the mLN (Figure 6, C and D). These cells as well as macrophages in both sites (Supplemental Figure S1D) showed the same alterations as those measured earlier in Daf1–/– mice.
provided evidence that CD103+ DCs are indispensable for oral tolerance. Phenotyping distinguished CD8a+ CD11b− DCs, which are IRF-8 positive, as the relevant CD103+ DC subset. High transcription of the Aldh1a2 and TGFB2 genes (which encode retinaldehyde dehydrogenase 2 and TGF-β2, respectively) in both this CD103+ DC subset and CD4+ cells was linked with high-efficiency Treg induction. To determine how DAF deficiency impacts the expression of these markers, their expression was compared in WT, Daf1−/−, and C3ar1−/−C5ar1−/− CD4+ cells in Treg induction cultures of WT, Daf1−/−, and C3ar1−/−C5ar1−/− CD4+ cells supplemented with C3ar1-A/C5ar1-A. Unlike WT CD4+ cells activated under Treg conditions, Daf1–/– CD4+ cells showed reduced IRF-8, Aldh1a2, and TGF-β2 mRNAs, whereas C3ar1–/–C5ar1–/– CD4+ cells showed increases in each (Figure 6E). Cx3CR1, CCR7, and CCR9 as well as IRF-8, Aldh1a2, and TGF-β2 were not measured in other APC sets.
DAF Expression on Tregs Is Essential for Processes Implicated in Their Tolerogenic Function
Tregs possess inactive (latent) TGF-β1 on their surface in association with latency-associated protein (LAP), which is bound to glycoprotein A repetitions predominant.
The interaction of these complexes with integrin β6 (Itg-β6; pulmonary epithelial cells) and Itg-β8 on either the Tregs themselves or CD103+ DCs (see Discussion) is implicated in the release of active TGF-β1 and consequent maintenance of gut homeostasis.
Phenotyping of Foxp3+ cells, isolated from the mLN of OVA-fed WT, Daf1–/–, and C3ar1–/–C5ar1–/– OT-II Foxp3-GFP mice, showed that although glycoprotein A repetitions predominant and LAP expression increased on WT and C3ar1–/–C5ar1–/– pTregs, neither protein did so on Daf1–/– Tregs (Figure 6, F and G). Unlike WT or C3ar1–/–C5ar1–/– pTregs, which showed up-regulated Itg-B6 and Itg-B8 mRNA (Figure 6H), Daf1–/– pTregs showed little mRNA up-regulation of either integrin. Although Itg-β8 on CD103+ DCs could not be measured because of insufficient cell recovery for real-time quantitative PCR and the lack of a suitable anti-mouse Itg-β8 antibody, as reported by others,
the same difference between the three genotypes was found for Itg-β6 protein on CD103+ DCs (Figure 6I). Thus, the normal expression of multiple factors on both CD103+ DCs and pTregs, which play requisite roles in ocular and oral tolerance, depends on DAF.
The maturation status of APCs that process antigens and migrate to draining LNs, in large part, determines whether priming of cognate naïve CD4+ cells in the LN produces effector or tolerogenic responses.
In this study, two independent sites where tolerogenic responses are continually and physiologically mounted were examined: the a.c. of the eye and the GI tract. Our prior studies showed absence of C3ar1/C5ar1 signaling in CD4+ cells enables autocrine TGF-β production by DC and CD4+ cell partners and programs pTreg induction
; but it was unclear whether this is a requisite, redundant, or modulatory process in T-lineage commitment. In addition, they did not address physiological processes that underlie T-cell homeostasis versus activation, such as tolerance, or interrogate the immune cell properties of APCs and Tregs traditionally linked with tolerance. The experiments herein, together with the recent clinical findings that IBD-like enteropathy develops in DAF's absence, show that DAF expression on immune cells is obligatory for many of these properties and is needed for CD103+ DC induction of peripheral gut and ocular tolerance. They show that physiological DAF expression does the following: i) represses DC MHC class II expression, ii) induces coinhibitory PD-L1/ICOS-L expression and PD1/ICOS expression on the interactive partners while repressing costimulatory molecule expression, iii) enables CD103+ DC IRF-8+Aldh1a2+TGF-β2+ expression connected with CD103+ DC migration to the regional LN, iv) increases Shp-2 expression needed for PD1 expression, v) increases glycoprotein A repetitions predominant and increases surface LAP and Itg-β6/Itg-β8 connected with Treg TGF-β1 secretion, and vi) enables CCR7 and CCR9 expression needed for transit of CD4+ cells to and from tolerogenic LNs, where Tregs are generated. More important, the studies show that DAF expression on CD4+ cells as well as on APCs is required for tolerogenic APC processes in the GI tract and in the a.c.
The experiments on ocular and oral tolerance yielded essentially identical results. They showed that resident DCs and CD4+ cells in the smLN and mLN of unmanipulated Daf1–/– mice tonically are more mature than in WTs, reflecting their acquisition of an M1 rather than a tolerogenic M2 phenotype. Consequently, lineage commitment to pTregs in Daf1–/– mice, needed for a tolerogenic response, was virtually abolished. In accordance with the effect being due to restraint of immune cell C3ar1/C5ar1 signaling, this contrasted with Daf1–/–C3ar1–/–C5ar1–/– or C3ar1–/–C5ar1–/– mice in which the M1 maturation status of DCs possessing Daf1–/– deficiency alone was reversed and pTreg commitment was augmented above WTs. The superphysiological tolerogenic properties when C3ar1/C5ar1 are deficient are due to the fact that DAF in WT cells does not fully block C3ar1/C5ar1 signaling, allowing constitutive signaling through these GPCRs, which supports the viability of conventional Teff cells.
The data herein, thus, argue that the default state of APCs is connected, at least in part, with low-level APC C3a/C5a production, which, when increased, would serve as an APC maturation signal and render the APCs able to convert naïve CD4+ cells to a Teff lineage. In support of this characteristic being DAF specific rather site specific and alterations of DAF control of C3ar1/C5ar1 signaling being site specific via elaboration of cytokines, the same connections of DAF with immature DCs and its absence with mature DCs were found in the inguinal LN absent a toll-like receptor or maturation signal. In the presence of restraint of autocrine C3ar1/C5ar1 signaling in C3ar1−/−C5ar1−/− DCs, TGF-β1/IL-10 production by smLN and mLN CD4+ cells was increased above WTs and IFN-γ production reciprocally decreased, whereas the opposite pattern was observed in the presence of potentiated C3ar1/C5ar1 signaling in Daf1−/− DCs.
Collectively, the data herein on DAF control of autocrine C3ar1/C5ar1 signaling tie together many individual APC and CD4+ cell processes associated in the literature with tolerance.
interpreted in the context of serum complement that dextran sodium sulfate–induced colitis is markedly more severe in Daf1–/– mice. They likewise provide a mechanism for reports that i) the absence of C5ar1 signaling is important for tolerance in the dextran sodium sulfate model,
As discussed, the same presence or absence of DAF control of autocrine C3ar1/C5ar1 signaling in macrophages, which biases between PD1-L1/ICOS-L expression and TGF-β1/IL-10 production, versus B7/CD40 expression and IFN-γ production is likely relevant to M1 versus M2 macrophage polarization. This macrophage polarization is linked with type 2 diabetes, atherosclerosis, and cirrhosis,
Several different immune processes have been proposed as playing a principal role in ocular tolerance. A common denominator has not been identified. Among the processes proposed as being central have been the following: i) local production of TGF-β1 in the a.c.
The data herein provide a common mechanism for processes i) through iv), in part reconciling this previous literature. Further work will be needed to determine whether there are mechanistic linkages with processes (v) α-MSH, (vi) corneal and endothelial cell FasL expression, and (vii) natural killer T cells.
Although tolerance in the gastrointestinal tract and airways is more complex than that in the a.c. in that it involves IL-10–producing Tr1 cells,
implicated initial antigen interaction with macrophages, followed by transfer to DCs, both of which, according to our data, would depend on DAF expression to maintain their immaturity. More important, a more recent study distinguished that the critical APC is the CD103+ CD8a+ CD11b– IRF-8+Aldh1a2+TGF-β2+ DC.
In addition to linking expression of coinhibitory PD-L1, ICOS-L, IRF-8+Aldh1a2+TGF-β2+ CCR7, and CCR9 on CD103+ DCs with DAF, the data herein show that surface expression of glycoprotein A repetitions predominant, which anchors LAP–TFG-β1 complexes to the surface of Tregs,
depends on DAF. Most important, they show that expression levels of CD103, the integrin mechanistically linked with conversion of surface-associated LAP–TGF-β1 complexes into active TGF-β1 on CD103+ DCs,
are decreased. They likewise show that up-regulation of Itg-β6 and Itg-β8 on pTregs is dependent on DAF. CD103+ DCs in the smLN having the same properties essential to ocular tolerance are identified for the first time.
DAF's modulatory effect on Teff responses is mechanistically due to the fact that Gβγ signaling of the G proteins of the C3ar1/C5ar1 GPCRs evokes PI-3Kγ activation.
Because C3ar1 and C5ar1 are Gi-coupled GPCRs, the Gα subunits of their G proteins repress adenylyl cyclase and consequently suppress protein kinase A activation. Increased AKT phosphorylation and decreased protein kinase A activation prevent the nuclear translocation of Foxo1
The physiological and pathogenic relevance of C3ar1/C5ar1 signaling in DCs and CD4+ cells has remained unclear, and whether this pathway operates redundantly as one of several pTreg lineage commitment mechanisms in the context of tolerance has remained unstudied. The data herein argue that, although other processes operate to provide immunosuppression, DAF is indispensable in the context of physiological tolerance–associated pTreg induction that protects against IBD. They show that immune cell DAF is essential for expression of CD103+ cell and CD4+ cell markers that play requisite roles in the induction of tolerance.
We thank Ethan Shevach for helpful discussion, Michael Sramkoski for help with flow cytometry, John Lambris (University of Pennsylvania, Philadelphia, PA) for C5ar1-A synthetic peptide, J. Hamann for CD97–/–, and Vijay Kuchroo (Harvard University) for the Foxp3–green fluorescent protein mice.
Transplantation immunology of the anterior chamber of the eye.