ICSBP/IRF-8 differentially regulates antigen uptake during dendritic-cell development and affects antigen presentation to CD4⁺ T cells

Dendritic cells (DCs) are the most powerful antigen-presenting cells (APCs), with unique features in initiating antigen (Ag)–specific immune responses, activating both naive CD4⁺ and CD8⁺ T lymphocytes.¹  To do so, DCs undergo a series of developmental changes in a process known as “maturation.” Immature DCs are endowed with potent ability to capture Ags but are inefficient at presenting the processed Ags on their MHC class II (MHC-II) molecules, and thus are poor T-cell stimulators. Upon contact with microbial products or inflammatory cytokines, DCs are induced to mature and become highly efficient at presenting Ags and activating T cells while exhibiting reduced capacity for uptake and processing of newly encountered antigens.²  Immature DCs are tuned to sample their environment and to capture any foreign particle. To accomplish this task, DCs use at least 2 mechanisms to internalize soluble proteins, fluid-phase macropinocytosis or receptor-mediated endocytosis by members of different receptor families.³  The most prominent member of these membrane-bound proteins is the mannose receptor (MR). The MR acts as a pattern recognition molecule that enables efficient capture of mannosylated or fucosylated antigens, like dextran (DX), and is the main effector of receptor-mediated endocytosis.^4-6  In this regard, MR-mediated uptake of Ags has been shown to potently enhance MHC-II–restricted Ag presentation in both human and mouse DCs.^7,8 

In mouse spleen, distinct DC subtypes, including CD8α⁺ and CD8α^– DCs, can be identified with diverse localization, ontogeny, and nature of immune response induced, but with a common ability to capture, process, and present soluble antigens.⁹  Lymphoid organ–resident DCs have been generally considered as mature DCs based on high-level expression of costimulatory molecules as well as for their exceptional ability to stimulate both allogeneic T cells and Ag-specific naive T cells.^9,10  However, a number of reports have shown that spleen DC (s-DCs) can efficiently endocytose Ags, both in vitro and in vivo, suggesting that these cells may also exhibit features of immature DCs.^11-15  A recent study has clarified this controversy by showing that most subsets of lymphoid organ DCs, including CD8α⁺ and CD8α^– s-DCs, are phenotypically and functionally immature in the steady state and are induced to mature only upon stimulation in vitro or in vivo.^16,17  This immature feature of s-DCs is particularly relevant for the maintenance of peripheral tolerance and allows DCs to promptly capture blood-circulating antigens and to mature in situ.¹⁸ 

Interferon consensus sequence-binding protein (ICSBP), also known as interferon regulatory factor 8 (IRF-8), is a transcription factor that plays a prominent role in myeloid cell development as well as in the immune response to various infectious agents.^19,20  We and others have recently demonstrated that ICSBP is critically required for the normal development of defined DC populations. ICSBP^–/– mice lack plasmacytoid DCs and display a selective reduction of CD8α⁺ DCs and epidermal Langerhans cells as well as dermal DCs.^21-25  In addition to controlling the differentiation of some DC subsets, it appears that ICSBP also affects the functional maturation of DCs. In fact, the few CD8α⁺ DCs arising in ICSBP^–/– mice exhibit very low levels of costimulatory molecules and the CD8α^– DCs display reduced expression of MHC-II.^21,23  In addition, bone marrow (BM)–derived DCs from ICSBP^–/– mice display an immature phenotype, altered chemotactic behavior, and defective responses to inflammatory stimuli.^24,25  Finally, ICSBP^–/– mice show impaired trafficking of skin DCs in vivo that results in failure to initiate T-cell–mediated responses.²⁵  Together, these findings indicate that ICSBP affects both the differentiation and the functional maturation of DCs and suggest that this factor may regulate many aspects of DC-mediated T-cell immunity.

In the present work, we provide evidence that ICSBP plays a role on the ability of DCs to capture and to present exogenous Ags to MHC-II–restricted CD4⁺ T lymphocytes. We show that s-DCs in ICSBP^–/– mice, especially the CD8α⁺ subset, display a marked defect in Ag internalization as well as a poor capacity to stimulate Ag-specific naive CD4⁺ T cells. Moreover, although BM-DCs from ICSBP^–/– mice exhibit high endocytic activity, they are much less able to stimulate CD4⁺ T cells than their wild-type (WT) counterparts. Altogether, our data demonstrate that ICSBP affects the capacity of DCs for both Ag uptake and MHC-II presentation and suggest that this factor may control these 2 processes independently and at different stages of DC development.

ICSBP-deficient mice were generated and bred as described.^20,21  ovalbumin (OVA)–specific MHC-II–restricted, T-cell receptor (TCR)–transgenic mice (OT-II mice)²⁶  were kindly provided by Dr D. F. Tough (The Edward Jenner Institute, Compton, United Kingdom). All mice were kept under specific pathogen–free conditions. Animal procedures were performed in conformance with institutional guidelines.

Dextran-FITC (70 000 kDa) was purchased from Molecular Probes (Leiden, The Netherlands). OVA-FITC was obtained by conjugation of OVA–Chicken Egg (Sigma-Aldrich, St Louis, MO) with FITC, using the FluoroTag FITC-conjugation kit (Sigma-Aldrich). Horseradish peroxidase (HRP) was purchased from Amersham Biosciences (Little Chalfont, United Kingdom). Granulocyte-macrophage colony-stimulating factor (GM-CSF) was obtained from Peprotech (London, United Kingdom). All monoclonal antibodies (mAbs) used for flow cytometry were from Pharmingen (BD Biosciences, San Jose, CA). Biotinylated mAbs were detected with streptavidin-Tricolor (Caltag Laboratories, Burlingame, CA). Stained cells were analyzed on a FACSsort flow cytometer (BD Biosciences).

The procedure of DC isolation from spleens has been described in detail elsewhere.^21,27  Briefly, spleen fragments were digested in collagenase/DNase, followed by density-gradient centrifugation. The low-density fraction was positively selected using anti-CD11c Microbeads (Miltenyi Biotech, Bergisch Gladbach, Germany). The cells obtained were routinely more than 97% CD11c⁺. In some experiments, DCs were cultured overnight in Iscove modified Dulbecco medium (IMDM; Cambrex, East Rutherford, NJ) containing 10% fetal calf serum (FCS; Euroclone, Pero, Italy), with or without 0.5 μg/mL lipopolysaccharide (LPS; Sigma-Aldrich). BM-DCs were generated with GM-CSF.²⁵  Cells were used for experiments at day 7 of culture. Where indicated, BM-DCs were replated with or without LPS (0.5 μg/mL) for 1 day before use.

For flow cytometric analysis, 10⁵ DCs were incubated for various times with 1 mg/mL DX-FITC or with 0.2 mg/mL OVA-FITC at 37°C. To calculate background fluorescence, control cells were incubated with Ag at 4°C. Cells were washed in cold phosphate-buffered saline (PBS), stained with anti-CD8a–PE mAb, and analyzed by fluorescence-activated cell sorting (FACS). For confocal laser scanning microscopy (CLSM), 3 × 10⁵ DCs were incubated for 40 minutes with 2 mg/mL DX-FITC or with 0.3 mg/mL OVA-FITC at 37°C. Cells were then washed in cold PBS and left to adhere on Superfrost Plus glass slides (Menzel Glasser, Braunschweig, Germany) for 1 hour at 4°C. Slides were stained with biotin anti-CD8a for 20 minutes at 4°C, followed by Streptavidin–Alexa 594 (Molecular Probes), washed and fixed in acetone for 20 minutes at –20°C. Cover glasses were then mounted on microscope slides with antifade Vectashield reagent (Vector Laboratories, Burlingame, CA). CLSM observations were performed using a Leica TCS SP2 AOBS apparatus (Leica Lasertechnik, Heidelberg, Germany), using excitation spectral laser lines 488 and 594 nm properly tuned by Acousto-optic tunable filters (AOTF; Leica LaserFechnik). Image acquisition and processing were conducted by using the Leica confocal software (Leica Lasertechnik). Signals from different fluorescent probes were taken in parallel. For immunohistochemistry, 2 × 10⁵ DCs were incubated in RPMI in presence of 1 mg/mL HRP at 37°C for 10 minutes, washed in cold PBS, and cytospun for 5 minutes at 500g. Slides were fixed in 4% paraformaldehyde and washed in Tris-NaCl (TN) buffer; the enzymatic reaction was developed by diaminobenzidine (DAB; SigmaAldrich) reaction. Slides were counterstained with hematoxylin. Slides were observed by using a Zeiss Axioplan microscope (Zeiss, Göttingen, Germany). Images were captured with a Canon Powershot G5 camera (Canon, Tokyo, Japan) and were processed with Photoshop software (Adobe Systems, Mountain View, CA).

Mice were injected intravenously with 1 mg OVA-FITC or PBS. Two hours later, mice were killed and spleens were removed. Splenic DCs were isolated, double-stained with anti-CD11c and anti-CD8, and analyzed by FACS to assess the uptake of OVA-FITC.

OVA-specific CD4⁺ T cells were magnetically sorted from lymph nodes (LNs) and spleens of OT-II mice by using anti-CD4 Microbeads (Miltenyi Biotech). T cells (50 000) were cultured with graded numbers of OVA-loaded s-DCs or BM-DCs from ICSBP^–/– and WT mice. OVA loading of BM-DCs was performed by incubating cells with 0.2 mg/mL OVA for 40 minutes at 37°C. OVA loading of s-DCs was performed in vivo by intravenous injection with 1 mg OVA or PBS; DCs were isolated 2 hours later. In some experiments, OVA-loaded DCs were cultured for 18 hours with or without LPS (0.5 μg/mL) and then assayed for their ability to prime OVA-specific CD4⁺ T cells. Cell cultures were incubated in RPMI containing 10% FCS for 3 days at 37°C in 5% CO₂ and then pulsed with ³H-thymidine (1 μCi/well [0.037 MBq/well]) for 16 hours. Incorporation of ³H-thymidine was analyzed by liquid scintillation counting.

CD4⁺ T cells purified from OT-II mice were incubated with 5 μM carboxyfluorescein diacetate succinimidyl ester (CFSE; Molecular Probes) for 10 minutes at 37°C, then washed in cold PBS.²⁸  5 × 10⁶ CFSE-labeled OT-II cells were transferred intravenously into WT recipients. One day later, recipients were injected in the hind footpads with 5 × 10⁵ OVA-loaded s-DCs or BM-DCs from ICSBP^–/– and WT mice. PBS-injected mice served as controls. Three days after immunization, popliteal LNs were harvested and cell suspensions were stained with PE-labeled anti-CD4. OVA-specific proliferation of OT-II cells was measured by cytofluorometric analysis of CFSE dilution.²⁹ 

To assess the capacity of DCs to induce Ag-specific T-cell sensitization in vivo, BM-DCs from ICSBP^–/– or WT mice were incubated with 12.5 μg/mL FITC for 20 minutes at 37°C.³⁰  FITC-loaded DCs (5 × 10⁵) were then injected in 1 footpad of WT recipients. Six days later, mice were challenged by painting 1 ear with 10 μL FITC. Contact hypersensitivity (CHS) response was determined 24 to 72 hours after challenge by measuring ear thickness using a dial thickness micrometer. Ear swelling was calculated by subtracting the thickness of the ear not treated with FITC.

Total RNA was extracted from 2-5 × 10⁶ splenic DC or BM-DC and reverse-transcribed as previously described.²⁵  Quantitative real-time reverse transcriptase–polymerase chain reaction (qRT-PCR) was performed using QuantiTect SYBR kit (Applied Biosystems, Epson, United Kingdom). Forward and reverse primers at a final concentration of 0.3 mM and 2 μL of each cDNA sample were also used. The conditions of PCR reaction were as follows: 15 seconds at 95°C, 30 seconds at 60°C, and 30 seconds at 72°C (50 cycles). PCR products were continuously measured by means of an ABI PRISM 7700 detection system (PerkinElmer/Cetus, Norwalk, CT). Triplicates were performed for each experimental point. Threshold cycle (C_T) values were determined automatically by the Sequence Detection System software (PerkinElmer/Cetus). These values were used to calculate relative expression of mannose receptor (Mrc1) mRNA expression with respect to β-actin expression using the ΔΔC_T method.³¹  The primer pairs used were as follows: Mrc1: 5′-ATGCCAAGTGGGAAAATCTG-3′ (forward), 5′-TGTAGCAGTGGCCTGCATAG-3′ (reverse); β-actin: 5′-AGAGGGAAATCGTGCGTGAC-3′ (forward), 5′-CAATAGTGATGACCTGGCCGT-3′ (reverse).³² 

Spleen-resident DCs in the steady state display a relatively immature phenotype and therefore retain considerable capacity to internalize soluble Ags both in vitro and in vivo.¹⁶  Previous work has shown that s-DCs from ICSBP^–/– mice tested in vitro display reduced expression of costimulatory as well as MHC molecules, indicative of a poorly differentiated phenotype with respect to their WT counterparts.^21,23  Therefore, we asked whether the more immature features of s-DCs in ICSBP^–/– mice could be associated with a higher potential for Ag uptake. To this end, we used different kinds of Ag internalized either through the MR-mediated pathway or by macropinocytosis, namely DX and OVA.^5,33,34  As shown in Figure 1A, s-DCs from WT mice exhibited considerable endocytic activity for both DX-FITC and OVA-FITC, as revealed by 3-fold and 5-fold increases, respectively, of FITC positivity in cells incubated with Ag at 37°C over those incubated at 4°C. Surprisingly, CD11c⁺ s-DCs from ICSBP^–/– mice exhibited almost undetectable uptake of DX-FITC and low uptake of OVA (3-fold increase of FITC-positive cells over control; Figure 1A). As an alternative approach to visualizing Ag internalization, we used HRP as the Ag. Like DX-FITC, HRP is known to be internalized through the MR-mediated pathway.^5,35  Morphologic analysis revealed that while macropinosomes containing HRP could be detected in many WT CD11c⁺ DCs, few particles were found in ICSBP^–/– DCs (Figure 1B). Interestingly, in ICSBP^–/– s-DCs, the HRP spots were observed to be primarily on the surface membrane rather than in macropinosomal vesicles, suggesting a defective ability to internalize the Ag. To better understand the role of ICSBP in the MR-dependent pathway of Ag internalization, we analyzed the expression of Mrc1 mRNA in ICSBP^–/– and WT s-DCs. As shown by qRT-PCR, Mrc1 transcripts were 5.5-fold higher in s-DCs of WT mice than those of ICSBP^–/– mice (Figure 1C).

We previously reported that ICSBP^–/– mice are characterized by an abnormally reduced frequency of the CD8α⁺ DC subset, which display a phenotype less mature than that of CD8α^– DCs.²¹  It was therefore important to dissect the Ag uptake features of these 2 DC subsets. To this end, we analyzed the uptake of DX-FITC and OVA-FITC in CD8α⁺ and CD8α^–\ DCs. As shown in Figure 2A and 2C, CD8α⁺ DCs from ICSBP^–/– mice failed to exhibit any endocytic activity, whereas the same cells from WT mice displayed significant levels of DX uptake. Similarly, ICSBP^–/– CD8α⁺ DCs exhibited almost undetectable endocytosis of OVA-FITC when compared with their WT counterparts (Figure 2B-C). CD8α^– DCs from ICSBP^–/– mice displayed very small ability to capture DX-FITC, whereas their WT counterparts were more efficient (Figure 2D). In addition, ICSBP^–/– CD8α^– DCs displayed a moderate reduction of OVA-FITC uptake with respect to WT cells (Figure 2D). We also examined the kinetics of uptake by CD8α⁺ and CD8α^– DCs from WT and ICSBP^–/– mice (Figure 2E). The mean fluorescence intensity (MFI) of WT CD8α⁺ DCs incorporating DX-FITC was increased 2-fold over control cells in as short a time as 5 minutes and increased over time to 3.5-fold higher at 40 minutes. Strikingly, no increase in the MFI was detected in ICSBP^–/– CD8α⁺ DCs throughout the time of incubation, thus confirming the lack of DX-FITC uptake in these cells (Figure 2E). Furthermore, the MFI of WT CD8α⁺ DCs capturing OVA-FITC increased significantly over time, peaking at 4.5-fold higher compared with that of control cells. In contrast, the MFI of CD8α⁺ DCs from ICSBP^–/– mice increased only to 2-fold higher compared with that of control cells. In WT CD8α^– DCs the MFI increased up to 3.3-fold more than the incubation time with DX-FITC and up to 4.8-fold during incubation with OVA-FITC compared with that of control cells. Conversely, ICSBP^–/– CD8α^– DCs showed almost no increase in MFI during incubation with DX-FITC (1.5-fold more than control cells), and up to 3.4-fold MFI increase during incubation with OVA-FITC (Figure 2E).

To better explore Ag endocytosis in CD8α⁺ s-DCs we used confocal laser-scanning microscopy. Strikingly, no evidence for uptake of FITC-labeled DX (Figure 3A) or OVA (Figure 3B) was found in any of the CD8α⁺ s-DCs from ICSBP^–/– mice. In contrast, both DX-FITC (Figure 3A) and OVA-FITC (Figure 3B) were internalized by WT CD8α⁺ s-DCs.

Priming of naive CD4⁺ T cells in draining lymphoid organs is largely the function of DCs that capture Ags in peripheral tissues.^16,18  It was therefore important to evaluate the ability of s-DCs of ICSBP^–/– mice to present Ags and to prime CD4⁺ T cells in vivo. To this end, we injected WT and ICSBP^–/– mice intravenously with 1 mg OVA-FITC and analyzed OVA uptake in the CD8α⁺ and CD8α^– DC subsets. As shown in Figure 4A, CD8α⁺ s-DCs from ICSBP^–/– mice were unable to capture OVA-FITC in vivo. In contrast, OVA was efficiently taken up by the CD8α⁺ DC subset of WT mice. In keeping with the in vitro studies, CD8α^– DCs from ICSBP^–/– mice showed a moderately reduced endocytic activity when compared with their WT counterparts. Splenic DCs from ICSBP^–/– and WT mice injected with OVA were also tested for their capacity to stimulate OT-II CD4⁺ T cells in vitro. As expected, the ability of ICSBP^–/– s-DCs to stimulate OVA-specific CD4⁺ T-cell proliferation was greatly impaired when compared with the function of WT s-DCs (Figure 4B). To determine whether the defective Ag-presenting function of s-DCs from ICSBP^–/– mice could be overcome by treatment with stimuli capable of activating DCs, s-DCs loaded with OVA in vivo were cultured overnight in presence or absence of LPS before testing their ability to stimulate the proliferation of OT-II CD4⁺ T cells. Notably, treatment of ICSBP^–/– s-DC with LPS not only failed to increase their stimulatory capacity for CD4⁺ T cells, but seemed to actually diminish it when compared with the activity of untreated ICSBP^–/– s-DCs (Figure 4C). In contrast, treatment of WT s-DCs with LPS markedly enhanced the proliferative response of OT-II CD4⁺ T lymphocytes over that observed using unstimulated WT s-DCs (Figure 4C). In line with these findings and with a previous report, treatment of ICSBP^–/– s-DCs with LPS resulted in minor activation of MHC-II and costimulatory molecules with respect to LPS-exposed WT cells (Figure 4D).

Finally, we assessed how the reduced efficiency of Ag uptake and the altered T-cell stimulatory capacity of ICSBP^–/– s-DCs affected Ag presentation to MHC-II–restricted T lymphocytes in vivo. To this end, s-DCs from ICSBP^–/– and WT mice injected with OVA were inoculated into the footpads of WT recipients that had been adoptively transferred with CFSE-labeled OT-II CD4⁺ T cells. As shown in Figure 4E, s-DCs from ICSBP^–/– mice induced only weak proliferation of OT-II cells, as revealed by a reduced number of division cycles (4 cycles) compared with those observed in mice immunized with WT OVA-DC (6 cycles). Of note, the absolute number of CFSE-positive cells recovered in LNs of mice primed with OVA-DCs from ICSBP^–/– mice (0.87%) was significantly lower than the levels observed with recipients of OVA-DCs from WT counterparts (1.51%; Figure 4E). Thus, s-DCs in ICSBP^–/– mice display marked defects in both capturing and presenting soluble Ags in a MHC-II–restricted context in vivo, and Ag presentation could not be enhanced by activation signals.

We have previously shown that ICSBP^–/– BM-DCs display immature features resulting in defective chemotactic responses compared with BM-DCs from WT mice.²⁵  We then asked whether ICSBP^–/– BM-DCs, by virtue of their immature phenotype, exhibited a higher endocytic potential with respect to WT cells. For this purpose, we assessed the ability of BM-DCs from ICSBP^–/– and WT mice to capture Ags. To define the maturational stage of the DCs, cells were stained for CD86 expression. As illustrated in Figure 5A, the frequency of ICSBP^–/– BM-DCs taking up DX-FITC was 3-fold higher than for WT cells. Of interest, the population displaying endocytic activity in both WT and ICSBP^–/– BM-DCs expressed CD86 at low levels or were negative (CD86^–/low), indicating that the efficiency of DX-FITC uptake was associated with an immature state of DCs (Figure 5A). Previous studies have shown that the MR is expressed at high levels by immature DCs and that expression is down-regulated during the process of maturation.^4,5  To test whether the efficiency of DX uptake correlated with the expression of its receptor, we analyzed the levels of Mrc1 transcripts in WT and ICSBP^–/– BM-DCs at various stages of development. Interestingly, expression of Mrc1 in ICSBP^–/– DCs increased 22-fold after 7 days in culture, whereas Mrc1 transcripts in WT cells increased only 2-fold during the same time frame (Figure 5B). To investigate receptor-independent pathways of endocytosis in ICSBP^–/– BM-DCs, we tested the uptake of OVA. The frequency of ICSBP^–/– BM-DCs taking up OVA-FITC was 3-fold higher than for BM cells of WT mice (Figure 6A). Again, the population exhibiting endocytic activity in both WT and ICSBP^–/– BM-DCs displayed an immature phenotype (CD86^–/low; Figure 6A). These findings suggest that the higher endocytic potential of BM-DCs from ICSBP^–/– as compared to WT mice may be the result of a higher frequency of immature DCs.

Next, we studied the ability of ICSBP^–/– BM-DCs to prime CD4⁺ T cells in vitro. Notably, OVA-loaded ICSBP^–/– BM-DCs were much less effective than WT cells in stimulating the proliferation of OVA-specific CD4⁺ T lymphocytes (Figure 6B). To investigate whether Ag presentation by ICSBP^–/– BM-DCs could be triggered by activation stimuli, OVA-loaded BM-DCs were treated with LPS and then tested for their ability to stimulate OT-II CD4⁺ T cell proliferation. Exposure to LPS enhanced the ability of both WT and ICSBP^–/– BM-DCs to activate CD4⁺ T cells, although ICSBP^–/– DCs were largely less stimulatory than WT DCs (Figure 6B). Notably, the magnitude of the proliferative response elicited by the BM-DC preparations correlated with their stage of maturation, as the frequency of ICSBP^–/– BM-DCs expressing high levels of CD86 and MHC-II molecules was only about half that of WT DCs (Figure 6C). Following stimulation with LPS, the frequencies of mature cells increased to a similar extent for cells from both mice, leaving I-A⁺ and CD86⁺ cells in WT cultures present at frequencies that increased 1.6-fold and 1.8-fold, respectively, over stimulated cultures from ICSBP^–/– mice (Figure 6C).

Last, we examined the ability of ICSBP^–/– BM-DCs to present OVA to OT-II CD4⁺ T lymphocytes in vivo. BM-DCs were loaded with OVA and then injected in the hind footpads of WT recipients reconstituted with CFSE-labeled OT-II CD4⁺ T cells. We found that the ability of OVA-loaded ICSBP^–/– BM-DCs to stimulate OT-II–cell proliferation was significantly reduced when compared with the activity of WT DCs (Figure 6D). In addition, recovery of CFSE-positive cells in mice primed with OVA-loaded BM-DCs from ICSBP^–/– mice (0.79% of LN cells) was more than 4-fold less than the number of CFSE⁺ cells found in mice injected with WT DCs (3.35%; Figure 6D). Altogether, these data indicate that BM-DCs from ICSBP^–/– mice are characterized by a high endocytic activity but a reduced ability to activate MHC-II–restricted T cells in vitro and in vivo, 2 hallmarks of immature DCs.

To further assess the ability of BM-DCs to prime a T-cell–mediated response in vivo, we examined their ability to sensitize mice for a CHS response to the hapten, FITC.³⁶  BM-DCs from ICSBP^–/– and WT mice were pulsed with FITC and injected in the footpad of WT hosts. Strikingly, only mice immunized with FITC-pulsed WT-DCs mounted a CHS response, as detected by specific ear swelling. No ear swelling was detected with recipients of FITC-pulsed DCs from ICSBP^–/– mice (Figure 7).

It is becoming increasingly clear that DCs comprise subsets of highly specialized cells primed to capture and to process Ags in distinct manners from other APCs. These distinctions are regulated by different activation stimuli in the developmental process termed “maturation.” Maturation links innate immunity with adaptive responses and plays a pivotal role in the discrimination of self-nonself differences that are orchestrated by DCs.²  It thus becomes extremely important to identify the factors and downstream signals that allow the maturation program to execute these distinctions. During the last few years, ICSBP has emerged as a transcription factor with a series of critical roles in regulating the differentiation and maturation of distinct subpopulations of DCs. Mice with a nonconditional knockout of ICSBP have essentially no plasmacytoid DCs and exhibit a selective reduction in the number and functionality of CD8α⁺ DCs, dermal DCs, and epidermal Langerhans cells.^21-25  In addition, even DCs that develop in normal numbers in ICSBP^–/– mice display a poorly differentiated phenotype, indicating that the developmental and maturation program of DCs may be controlled by ICSBP at multiple stages. In the present paper, we demonstrate that during the course of DC development, ICSBP is a crucial determinant for the ability of these cells to capture and to present exogenous Ags to MHC-II–restricted T lymphocytes.

Internalization of soluble Ags by DCs can occur by 3 major routes: micropinocytosis, involving fluid-phase uptake of small molecules (0.1 μm) via clathrin-coated pits; macropinocytosis, involving fluid-phase uptake of larger vesicles (0.5-3 μm) through cytoskeleton-dependent membrane ruffling; and receptor-dependent phagocytosis, mediated primarily by the MR. Both macropinocytosis and phagocytosis convey Ags to the endocytic route that leads to the presentation of derivative peptides by MHC-II molecules.^4,37  In our studies, OVA-FITC, DX-FITC, and HRP were used as useful markers to monitor endocytosis in DCs of ICSBP^–/– mice. OVA-FITC is known to be taken up exclusively by macropinocytosis, whereas DX-FITC and HRP use both the MR-dependent pathway and the macropinocytic route.^2,4,5,33,34  We show that the ability of s-DCs of mutant mice to capture these antigens is greatly impaired. In particular, ICSBP^–/– CD8α⁺ DCs were completely unable to capture DX and OVA, as revealed by both FACS and confocal microscopic analyses. Moreover, CD8α^– s-DCs of ICSBP mice displayed impaired uptake of DX and, to a lesser extent, of OVA. These results were somehow unexpected, since s-DCs of ICSBP^–/– mice, especially the CD8α⁺ subset, are characterized by a prominent immature phenotype,^21,23  a feature that normally correlates with high endocytic potential. On the other hand, and consistent with their immature traits, ICSBP^–/– BM-DCs were fully competent at capturing Ags. Precisely, uptake of DX or OVA was markedly increased in BM-DCs of ICSBP^–/– mice compared with WT cells. Of note, the endocytic activity was exclusively observed in immature CD86^–/low DCs, which represented the largest population in ICSBP^–/– cultures. Thus, it appears that the enhanced level of endocytosis in ICSBP^–/– BM-DCs is a result of a more immature phenotype of these cells. These findings suggest that ICSBP regulates Ag uptake only at late stages of DC development and is critical in maintaining efficient endocytic activity in s-DCs.

Although the mechanisms by which ICSBP controls the endocytic activity of s-DCs remain to be determined, we hypothesize that this function may be switched off at some point during maturation of these cells through the control of different processes. For instance, the strong impairment in the internalization of mannosylated Ags, such as DX and HRP, observed with ICSBP^–/– s-DCs, may be correlated, at least in part, with reduced expression of MR. The idea that MR expression is regulated by ICSBP is worth considering since transcriptional regulation of MR gene is known to be under the control of PU.1, a transcription factor which interacts with ICSBP to bind target elements.^19,38  In this regard, the high efficiency of DX uptake by ICSBP^–/– BM-DCs correlated with elevated MR mRNA levels. On the other hand, since receptor-independent endocytosis of OVA was also impaired in s-DCs of ICSBP^–/– mice, other mechanisms have to be taken into account. Thus, ICSBP may also be involved in structural mechanisms that allow macropinosomes to be adequately transported from the cell surface to the cytosol. This concept is supported by the strong reduction of HRP-containing macropinosomal vesicles in the cytosol of ICSBP^–/– s-DCs, which indicates a defect in the internalization process used by these cells. In this regard, polymerization of cytoskeleton-bound actin is a key event regulating the selective mobilization of macropinosomes into DCs and affects both macropinocytosis and receptor-mediated endocytosis.^39,40  The organization of actin cytoskeleton is regulated by the Rho family of guanosine triphosphatases (GTPases), whose activity is finely regulated during DC maturation and controls both the rates of endocytosis and the interaction of DCs with T cells during T-cell priming.^41,42  Among these small GTPases, Rac1 and Cdc42 are the best-characterized members implicated in DC endocytosis.^43-45  Interestingly, selective inhibition of Rac1 in DCs results in reduced endocytic activity and impaired development of murine CD8α⁺ DCs, 2 features shared with ICSBP^–/– mice.⁴⁶  Therefore, it may be hypothesized that ICSBP controls the endocytic potential of DCs through both the regulation of MR expression and possibly by affecting processes involved in the organization of acting cytoskeleton.

Ags that access the endocytic route of DCs are processed and loaded onto MHC-II molecules for presentation to naive CD4⁺ T cells.⁴⁷  In our studies, we found that defective Ag uptake by s-DCs of ICSBP^–/– mice reflects a functional impairment in presenting the Ag to MHC-II–restricted T cells. In fact, ICSBP^–/– s-DCs were remarkably ineffective at presenting injected soluble OVA to OT-II CD4⁺ T cells both in vitro and in vivo. Although the impaired ability of ICSBP^–/– s-DCs to stimulate CD4⁺ T cells strongly correlates with the incapacity of these cells to effectively capture Ags, we observed a complete lack of endocytic activity only in ICSBP^–/– CD8α⁺ DCs, a subset that comprises a minor fraction of total s-DCs (3% in ICSBP^–/– mice, 25%-30% in WT mice).²¹  Several studies indicate that in the MHC-II context, CD8α^– s-DCs exhibit higher potential for the presentation of exogenous Ags, whereas CD8α⁺ s-DCs are specialized for presentation of cell-associated Ags.^12,14,16,48  Thus, it can be envisaged that ICSBP^–/– CD8α^– s-DCs may convey additional functional alterations resulting in defective stimulation of CD4⁺ T cells. In this regard, previous work showed that CD8α^– DCs of ICSBP^–/– mice express reduced levels of MHC-II following activation in vitro or in vivo.^21,23  Consistently, we found that exposure to LPS resulted in defective activation of ICSBP^–/– s-DCs. On the other hand, ICSBP^–/– s-DCs loaded in vitro with OVA peptide are able to stimulate OT-II CD4⁺ T cells as effectively as WT s-DCs⁴⁹  (Figure S1, available on the Blood website; see the Supplemental Figure link at the top of the online article). As peptide loading does not involve Ag internalization and processing, but direct recruitment to surface MHC molecules, these data suggest that when Ag is appropriately loaded onto MHC-II, ICSBP^–/– s-DCs are competent at stimulating CD4⁺ T cells. In this view, the defects in Ag uptake in ICSBP^–/– s-DCs appear to affect the fate of MHC-II Ag presentation critically. Deficiencies during the maturation process may ultimately contribute to the defective T-cell stimulatory ability of ICSBP critically s-DCs. Nonetheless, other pathways may be disrupted in ICSBP critically s-DCs, namely Ag processing into peptides, peptide targeting to MHC-II molecules in endocytic compartments, and transport of MHC-II peptide complexes to the cell surface. Of interest, a recent report showed that ICSBP directly regulates 3 cytoskeleton-related genes and 4 endosomal/lysozomal enzyme–related genes.⁵⁰ 

Regardless of the mechanisms involved, it appears clear that ICSBP regulates endocytosis and Ag presentation independently and at distinct stages of DC development. This concept is supported by our finding that while ICSBP critically BM-DCs exhibit increased endocytic activity, these cells are significantly impaired at presenting OVA to CD4⁺ T lymphocytes both in culture and after in vivo injection into competent WT hosts. The impaired Ag-presenting function by ICSBP critically BM-DCs in vivo was revealed by the markedly reduced proliferation of OVA-specific CFSE-labeled OT-II cells as well as by the failure to induce a CHS response to FITC. From an immunologic perspective, CHS can be considered as a prototype of T-cell–mediated delayed-type hypersensitivity and is a reliable in vivo approach to follow the whole process of immune responses to some haptens where CD4⁺ T cells are thought to play a pivotal role.^51,52  One mechanism that may account for defective MHC-II–restricted stimulation of T cells in vivo by Ag-loaded ICSBP^–/– BM-DCs may stem from a defective ability of these cells to migrate from the injection site to the draining LN. This view is based on the observations that ICSBP^–/– BM-DCs express low levels of CCR7 as well as impaired chemotactic response to the CCR7 ligands Mip-3β and 6CKine.²⁵  However, the finding that ICSBP^–/– BM-DCs were also impaired at stimulating CD4⁺ T cells in vitro clearly suggests that ICSBP^–/– BM-DCs may have an intrinsic defect to undergo full maturation and to effect T-cell stimulation. This concept is supported by the finding that treatment with LPS did not activate ICSBP^–/– BM-DCs as effectively as WT cells in terms of CD86 and MHC-II expression as well as of CD4⁺ T-cell stimulatory capacity, thus confirming the previously noted defective responses by BM-DCs of ICSBP^–/– mice to maturation signals.²⁴ 

In summary, our study demonstrates that in addition to its known roles in driving DC development, ICSBP critically affects key DC functions, including endocytosis of exogenous Ags and Ag-specific stimulation of CD4⁺ T cells. We suggest that ICSBP affects the capacity of lymphoid organ DCs to present blood-borne Ags to naive CD4⁺ T cells by controlling DC endocytosis and maturation. In parallel, ICSBP influences MHC-II presentation in BM-DCs by regulating the maturation process of these cells. Of interest, the concept that ICSBP is central in preserving the Ag uptake function of CD8α⁺ s-DCs may be particularly relevant since these cells are endowed with the unique ability to cross-present cell-associated antigens to CD8⁺ T cells, a crucial mechanism underlying immune responses to viral diseases and cancer.⁵³  Given the potential of DCs to promote antitumor immunity due to their unique ability to prime T-cell responses, DC-based therapies hold promise for developing new effective immunoadjuvant therapies for cancer. Our studies clearly suggest that ICSBP may represent a crucial factor to be considered in efforts to enhance the immunopotency of therapeutic DCs.

We are grateful to Annamaria Fattapposta and Angela Fresolone for excellent secretarial assistance. Special thanks also go to Maria Teresa D' Urso and Anna Maria Pacca for their excellent technical assistance with breeding.