A transforming growth factor–β–induced protein stimulates endocytosis and is up-regulated in immature dendritic cells

Dendritic cells (DCs) are potent antigen-presenting cells (APCs) that play critical roles in the activation of naïve T cells and cross-presentation of antigens to CD8 T cells.^1-3  The potency of DCs as APCs is, to a certain extent, mediated by the ability of these cells to perform antigen uptake and T-cell stimulation at distinct maturation stages.^4,5  Immature DCs (imDCs), present mainly in peripheral tissues, are potent in antigen uptake, using mechanisms such as receptor-dependent endocytosis and receptor-independent macropinocytosis.^1-3,6  Many microbial structures and inflammatory stimuli can activate imDCs, causing DCs to undergo a complex maturation program when the cells increase their expression of surface major histocompatibility complex (MHC) and costimulatory molecules, such as CD80/CD86, CD54, and CD40, and produce proinflammatory cytokines.^1-3  These maturing DCs also migrate to the T-cell area of draining lymph nodes, guided by their surface chemokine receptor CCR7.^7-11  In contrast to imDCs, mature DCs (mDCs) exhibit reduced endocytic activities and are poor in antigen uptake.^1-5  imDCs can expand regulatory T cells.¹²  Although mDCs can directly activate CD8⁺ T cells, imDCs can expand antigen-specific CD8 regulatory T cells.^13,14  imDCs also expand CD4⁺CD25⁺ regulatory T cells.¹⁵  Therefore, besides antigen uptake, imDCs can also actively regulate adaptive immunity.

The molecular mechanisms underlying the functional properties of imDCs are poorly understood. For example, although imDCs are potent in macropinocytosis,^4,5  how these cells acquire this macropinocytotic activity is unclear. Macropinocytosis requires membrane ruffling to form sealed macropinosomes,¹⁶  and 2 independent studies have shown that activation of the small GTPase Rac is required for imDC macropinocytosis.^17,18  One of these studies also implicated constitutively active cdc42 in DC macropinocytosis.¹⁸  Rac and cdc42 are members of the Rho subfamily of small GTPases that are activated on GTP binding.¹⁹  It remains to be determined how imDCs acquire the ability to sustain Rac or cdc42 activation. In addition, although imDCs are known to expand regulatory T cells, the underlying mechanisms regulating this activity have not yet been determined. In the present study, we focus on the identification of genes that are associated with immature DCs.

LPS (Escherichia coli O55:B5) was purchased from Sigma-Aldrich (St Louis, MO). Human interleukin 4 (IL-4), granulocyte-macrophage colonystimulating factor (GM-CSF), and macrophage colony-stimulating factor (M-CSF) were obtained from R&D Systems (Minneapolis, MN). The following mouse monoclonal antibodies were obtained from Ancell (Bayport, MN): CD1a (FITC), CD14 (PE), CD40 (FITC), and CD54 (FITC). Antibodies for CD80 (PE), CD86 (PE), and CD83 (FITC) were purchased from BD PharMingen (San Diego, CA).

Monocytes were isolated from peripheral-blood buffy coat as previously described,²⁰  and isolated monocytes were confirmed to be approximately 95% pure based on the expression of surface CD14. The cells were cultured for 6 days at 37°C in the presence of IL-4 and GM-CSF (each at 20 ng/mL) to generate imDCs⁴  and cultured with M-CSF (20 ng/mL) to generate macrophages.

To activate imDCs and macrophages, the cells were harvested and cultured for 48 hours at 1 × 10⁶/mL with LPS (0.5 μg/mL). The activation of DCs and macrophages was judged by surface expression of CD83 and CD54, respectively.²⁰  Monocytes and macrophages also express surface CD83, but this is transient and lost within 24 hours.²⁰  Briefly, cells were washed in PBS and then incubated for 30 minutes with specific antibodies or, as controls, isotype mouse IgG. The cells were washed 3 times in FACS wash (PBS containing 2.5% [vol/vol] bovine calf serum [BCS] and 0.05% [wt/vol] NaN₃) and fixed in PBS containing 1% (wt/vol) paraformaldehyde (pH 7.6) before analysis on a FACSCalibur using CellQuest software (Becton Dickinson Immunocytometry Systems, San Jose, CA).

RNA was isolated from fresh monocytes, macrophages, imDCs, and mDCs using TRIzol (Life Technologies, Bethesda, MD) followed by a repeated extraction with TRIzol to yield RNA with an A₂₆₀/A₂₈₀ ratio more than 1.8. For each cell type, RNA samples were pooled equally from 2 to 3 donors to prepare microarray probes. For each microarray hybridization, 2 to 3 μg total RNA was used following a single round of linear mRNA amplification.²¹  All samples were compared against a standard commercially available mRNA reference pool (Stratagene, La Jolla, CA) that had been similarly amplified. cDNA microarrays were fabricated using an SDDC-2 microarrayer (Virtek, Waterloo, CA) following standard procedures,²²  using cDNA clones obtained from various commercial vendors (Incyte, Research Genetics). Samples were fluorescently labeled using Cy3 dye; the reference was labeled with Cy5. After hybridization, microarray images were captured using a CCD-based microarray scanner (Applied Precision, Issaquah, WA). Genes listed in Table 1 were further verified by sequencing.

Fluorescence intensities corresponding to individual microarrays were uploaded into a centralized Oracle 8i database, accessible at www.omniarray.com. Establishment of various data sets and gene retrievals were performed using standard SQL queries, and are available from www.omniarray.com/dendritic_cells. All genes and arrays were mean-normalized to facilitate interarray comparisons. Hierarchic clustering was performed using the program Xcluster (Stanford, Palo Alto, CA) and visualized using the program Treeview.²³  To identify genes that were differentially regulated across all 4 cellular populations, array elements were chosen that consistently exhibited more than 3-fold regulation across at least 3 arrays.

RNA was isolated from monocytes, macrophages, imDCs, and mDCs. RNA was also isolated from cultured monocytes at different time points of differentiation into imDCs, that is, at 0, 1, 2, 4, and 6 days. RNA from 19 different human tissues was purchased from BD Clontech (Palo Alto, CA). cDNA was synthesized from these RNA samples using the reverse transcription-for-polymerase chain reaction (RT-for-PCR) kit (BD Clontech). βig-h3 mRNA was detected by 35 cycles of RT-PCR using βig-h3–specific primers (5′-gctctagacgcgtcgtcccgctccat/cccttcgaagttggtgatggtggagatga3′). As a control, the β-actin mRNA was similarly amplified using specific primers (ggaaggaaggctggaaga/ggcgtgatggtgggcatg). The PCR products were separated on 1% (wt/vol) agarose gels and visualized using ethidium bromide. For real-time PCR, the following primers were used (5′-3′): βig-h3 (gactttgaaccgtatcctg/gagtagctcatcaatgtagt) and β-actin (aaatcgtgcgtgacattaagg/agcactgtgttggcgtacag). PCR was performed on a Light Cycler (Roche Diagnostics, Basel, Switzerland) using the DNA Master SYBR Green I kit. The reactions were carried out for 45 cycles in 20 μL volume containing 4 mM MgCl₂, 0.5 μM of each primer, 2 μL DNA Master SYBR Green I, and 2 μL cDNA template. After a starting 10 minutes at 95°C, each following cycle consisted of 10 seconds at 95°C, 10 seconds at 56°C, and 10 seconds at 72°C, and finished by gradual product melting up to 95°C. The specificity of the PCR product was determined by melting curve analysis.

The βig-h3 cDNA was amplified, by RT-PCR, from imDC RNA with βig-h3–specific primers (5′-gctctagacgcgtcgtcccgctccat/cccttcgaaatgcttcatcctctctaataac-3′) and cloned into the XbaI and SfuI sites of the pcDNA3.1 myc-His plasmid (Invitrogen, Carlsbad, CA) in frame with 3′myc/His-coding sequences (pβig-h3-MH). The human embryonic kidney 293T cells (American Type Culture Collection, Manassas, VA) were cultured in DMEM supplemented with 10% (vol/vol) BCS (HyClone, Logan, UT), 100 U/mL penicillin, and 100 μg/mL streptomycin and transfected with the pβig-h3-MH vector (50 μg/T75 flask) using calcium phosphate.²⁴  After 6 hours, the media were changed to serum-free DMEM containing bovine serum albumin (BSA; 100 μg/mL). The media were harvested in 2 days and incubated overnight at 4°C with 0.5 mL Ni-NTA-agarose (Qiagen, Valencia, CA). In some experiments, iodoacetamide was added to 2.5 mM before incubation with the beads. The beads were packed in a column and washed extensively with a wash buffer (20 mM Tris, pH 8.0, 500 mM NaCl, and 10 mM imidazole). Bound βig-h3 was eluted in 0.15-mL fractions using an elution buffer (20 mM Tris, pH 8.0, 500 mM NaCl, and 250 mM imidazole). βig-h3, eluted in fractions 2 to 5, was examined on 10% (wt/vol) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels by Coomassie blue staining or Western blotting using a mouse monoclonal anti-myc or polyclonal anti–βig-h3 antibody. The latter was prepared by immunization of 6-week-old female Balb/c mice, through the peritoneal route, with purified βig-h3 (30 μg/mouse) in Freund complete adjuvant. After 2 boosters at 2-week intervals, sera were collected. Purified βig-h3 was quantified using the Bio-Rad Protein Assay kit (Bio-Rad, Hercules, CA) and dialyzed against PBS using the Pierce Microdialyzer System 100 (Pierce, Rockford, IL). Soluble CD14 was similarly expressed and purified.

Monocytes, macrophages, imDCs, and mDCs were cultured for 24 hours in the serum-free CD hybridoma medium (Gibco, Grand Island, NY) and media were collected. 293T cells were transfected in 6-well plates with pβig-h3-MH or pcDNA3.1 (4 μg/well) as described in “Cloning and expression of βig-h3.”After 6 hours, the cells were washed and cultured for another 24 hours in serum-free DMEM containing BSA (100 μg/mL) to collect media. The media were subjected to SDS-PAGE on 10% (wt/vol) gels under reducing conditions (1% [vol/vol] 2-mercaptoethanol). On electroblotting, βig-h3 was detected with a mouse polyclonal anti–βig-h3 antibody (1:1000 dilution). The blots were developed with alkaline phosphatase-conjugated goat anti–mouse IgG (1:5000 dilution; Sigma-Aldrich). Signals were visualized using the Immune-Star substrate pack (Bio-Rad).

Endocytosis was examined as previously described.⁴  Cells were washed and resuspended in fresh medium at 1 × 10⁶/mL. Purified βig-h3 or CD14 was added to the cell culture to 10 μg/mL and incubated for 30 minutes at 37°C before addition of FITC-dextran (FITC-DX) or lucifer yellow (LY; 1 mg/mL; Molecular Probes, Eugene, OR). In some experiments, purified βig-h3 was preincubated for 30 minutes with polymyxin B (20 μg/mL) before incubation with macrophages that were then incubated with the fluorescent dye molecules. imDCs were also preincubated with affinity-purified anti–βig-h3 IgG (2 μg/mL) for 1 hour before incubation for another 10, 20, or 30 minutes with the dye molecules (20 μg/mL). The antibody was purified by affinity chromatography using immobilized βig-h3 on cyanogen bromide-activated agarose (Sigma-Aldrich). After washing 4 times, the cells were analyzed by flow cytometry. In all experiments, cells incubated with the probes at 4°C were used as negative controls.

Membrane ruffling was determined using rhodamine-labeled phalloidin as previously reported.²⁵  Macrophages were cultured for 24 hours on glass coverslips. After stimulation for 30 minutes with purified βig-h3 or CD14 (10 μg/mL) or left unstimulated, the cells were fixed for 15 minutes in 3.7% (wt/vol) formaldehyde in PBS and permeabilized for 5 minutes in 0.2% (vol/vol) Triton X-100. The cells were then incubated for 30 minutes with rhodamine-phalloidin (5 U/mL) and were, after washing, viewed using a LSM510 laser-scanning microscope equipped with a 63×/1.4 oil-immersion objective lens and using Zeiss LSM Image Browser software (Carl Zeiss, Jena, Germany).

Macrophages cultured in 6-well plates were stimulated for 30 minutes with βig-h3 or CD14 at 10 μg/mL or left unstimulated. The cells were lysed in a buffer consisting of 20 mM Tris (pH 7.5), 300 mM NaCl, 2 mM EDTA, 2 mM EGTA, 1% (vol/vol) Triton X-100, 10 μg/mL leupeptin, 10 μg/mL aprotinin, and 2 mM PMSF and, after centrifugation, the lysates were incubated for 1 hour at 4°C with the p21-binding domain (PBD) of p21-activated kinase 1 (PAK-1) immobilized on agarose (Upstate Biotechnology, Lake Placid, NY).²⁶  As a positive control, lysate of unstimulated macrophages was incubated for 15 minutes with GTPγS (0.1 mM) before incubation with PAK-1 PBD-agarose. The agarose beads were washed in the lysis buffer and boiled in SDS-PAGE sample buffer. The supernatants were subjected to Western blotting using Rac and cdc42 antibodies. As controls, equal volumes of the different cell lysates (20% of total lysate) were similarly analyzed by Western blotting.

Cells were washed and resuspended at 2 × 10⁷/mL in cold RPMI containing 20 mM Tris (pH 7.4), 1 mM CaCl₂, and 1 mM MgCl₂ (RPMI-Tris). The cells (50 μL) were incubated with purified βig-h3 (50 μg/mL) for 2 hours at 4°C with gentle rocking. On washing twice in the binding buffer, cells were blocked for 30 minutes with 20% (vol/vol) goat serum in RPMI-Tris and then incubated for 30 minutes on ice with a mouse anti-myc antibody (10 μg/mL) or an isotype IgG. The cells were washed twice and then incubated for 30 minutes with PE-labeled goat anti–mouse IgG. After 2 washes, the cells were fixed in 1% (wt/vol) paraformaldehyde in PBS (pH 7.6) and analyzed by flow cytometry. The cells were also spotted on glass coverslips and examined using a BX51 Olympus fluorescence microscope equipped with a UPFL600I 63× oil-immersion objective lens (NA, 0.65-1.25). Images were captured using Olympus Micro Image software (Olympus, Tokyo, Japan).

The differentiation of CD14⁺ monocytes into macrophages and DCs in vitro has been well documented.⁴  We isolated monocytes using an adhesion method, which yields monocytes of about 95% purity (Figure 1), and cultured these cells into macrophages and DCs. Although both monocytes and macrophages are CD14^hiCD1a^– (Figure 1A), the cultured imDCs exhibited little surface CD14 and acquired CD1a (Figure 1A). On activation, DCs but not macrophages sustained surface CD83 expression as previously reported (Figure 1A).²⁰  Compared with macrophages, DCs also exhibited more significant increase in CD40, CD80, and CD86 expression on activation. These 4 types of cell were used for microarray studies.

To identify genes that are preferentially expressed in DCs, especially imDCs, the gene-expression profiles of monocytes, macrophages, imDCs, and mDCs were serially compared using cDNA microarrays containing approximately 13 000 array targets. The hybridization scheme is summarized in Figure 1B. To control for possible experimental variability, the hybridizations for each cell type were performed using 3 independent batches of pooled RNA, leading to a total of 12 samples (3 for each cell type), where each batch was equally pooled from 2 to 3 healthy donors to minimize individual donor variability.

After hybridization and scanning, approximately 8000 array elements were found to exhibit fluorescence signals significantly above background levels. To identify genes in this data set that were specifically regulated across all 4 cell types, we selected array elements that exhibited a minimum of 3-fold regulation across at least 3 of the arrays. This secondary filter resulted in a truncated data set of 578 genes. To identify global patterns of regulation in this reduced data set, a 2-way hierarchic clustering algorithm²³  was applied to cluster both the genes and samples in order of their similarity to one another (Figure 1C). The 12 samples were found segregated into 4 broad branches, each branch representing a different cell type. Consistent with their lineages, imDCs and mDCs occupy a separate branch from monocytes and macrophages, suggesting that imDCs and mDCs are more related to one another than to the latter 2 cell types.

A closer examination of the 578 genes revealed 121 genes that displayed highly consistent signal values across the 3 independent serial hybridizations. Of this 121-gene set, the expression of 55 genes was found to be preferentially increased or decreased in imDCs or mDCs or both (Table 1), whereas another 38 genes were globally increased or decreased in macrophages and DCs compared with the monocyte precursors (Table S1, available on the Blood website; see the Supplemental Tables link at the top of the online article). In addition, 28 genes were preferentially up-regulated in macrophages (Table S2). In the present study, we focus on genes that are preferentially up-regulated in imDCs.

Among the 15 genes that are preferentially up-regulated in imDCs, the βig-h3 gene (TGFBI) is most prominent (Table 1). Using RT-PCR, we confirmed that βig-h3 mRNA is low in monocytes and macrophages and is abundant in imDCs, suggesting its selective up-regulation in imDCs (Figure 2A). Expression of βig-h3 was diminished when DCs were activated with LPS (Figure 2A), but activated macrophages showed no significant change in their low basal level of βig-h3 mRNA. βig-h3 expression is therefore associated with the immature state of DCs. The expression of βig-h3 mRNA was also examined by RT-PCR following a 6-day course of imDC culturing (0, 1, 2, 4, and 6 days). βig-h3 mRNA was transiently up-regulated within the first 24 hours of culturing but returned to the low basal level at day 2 (Figure 2B). It increased again at day 4 and became highly prominent on day 6 when the cells have acquired imDC properties (Figure 2B). Real-time PCR was performed with these RNA samples and the results showed similar patterns of βig-h3 expression (Figure 2C-D). This result shows that high βig-h3 expression is associated with the immature state of DCs.

To our knowledge, this is the first report of selective βig-h3 up-regulation in imDCs. In previous studies, βig-h3 was detected mainly in cell types other than leukocytes.²⁷  We therefore examined βig-h3 expression in 19 different human tissues by RT-PCR and observed that it was highly expressed in the spleen and bone marrow and also abundant in colon and small intestines, which are rich in lymphoid tissues (Figure 2E). Its expression in the lung, liver, placenta, testis, trachea, and uterus was also significant. Although the cellular origin of the detected tissue βig-h3 mRNA is not clear, its association with lymphoid-rich tissues implies a role in immunologic mechanisms.

To examine the functional properties of βig-h3, we expressed this protein in 293T cells with C-terminal myc and His tags and purified this protein using Ni-NTA-agarose. βig-h3 was eluted in 6 stepwise fractions, with βig-h3 usually eluting in fractions 2 to 5 as a protein of about 70 kDa (Figure 3A). This 70-kDa protein reacted with a polyclonal anti–βig-h3 antibody (Figure 3B) and a monoclonal anti-myc antibody (Figure 3C). Based on its sequence, a mature βig-h3 polypeptide is expected to contain 668 amino acids (GenBank accession no. NM_000358), which predicts a polypeptide of about 75 kDa, and it is at present not clear why the observed molecular mass of the purified βig-h3 is smaller than this. The lack of potential Asn-linked glycosylation site in βig-h3 and neutral PI (7.6) cannot explain this apparent molecular weight reduction. Under nonreducing conditions, βig-h3 also migrated mainly as a 70-kDa protein irrespective of the presence or absence of iodoacetamide during its purification (Figure 3D). This result shows that under these experimental conditions βig-h3 may not exist as a disulfide-linked oligomer although native βig-h3 purified from tissues was previously observed as a fibrillike structure involving disulfide bonding.²⁸ 

To examine whether βig-h3 is secreted by imDCs, serum-free media from imDCs and, as comparison, monocytes, macrophages, and mDCs were examined by Western blotting. As controls, media were also collected from 293T cells transfected with pβig-h3-MH or pcDNA3.1. βig-h3 was prominently detected in the media of imDCs and 293T cells transfected with pβig-h3-MH but not the other cells (Figure 3E). These results show that βig-h3 is secreted by imDCs.

We then examined whether βig-h3 binds to imDCs by flow cytometry. imDCs were washed and then incubated with βig-h3. Bound βig-h3 was detected using an anti-myc antibody followed by fluorescent goat anti–mouse IgG. βig-h3 bound prominently to imDCs and to macrophages (Figure 4A), but showed no detectable binding to mDCs (Figure 4A). When these cells were examined by fluorescence microscopy, βig-h3 was also shown to bind to imDCs and macrophages but not mDCs (Figure 4B). These results raise the possibility that βig-h3 may regulate imDC activities in an autocrine manner and may also regulate macrophage function through paracrine stimulation. Interestingly, we note that the inability of βig-h3 to bind to mDCs may be consistent with its diminished expression in these matured DCs.

The high βig-h3 expression by imDCs and its ability to bind to imDCs prompted us to examine whether it regulates imDC-related functions. We first examined whether it stimulates imDC endocytosis or macropinocytosis using 2 soluble fluorescent probes, FITC-DX and LY. imDCs are constitutively potent in FITC-DX uptake and also exhibit significant uptake of LY (Figure 5A). The mean fluorescence index (MFI) for FITC-DX and LY was 810.93 and 72.07, respectively (Figure 5B). Prestimulation of imDCs for 30 minutes with purified βig-h3 did not change imDC endocytosis (Figure 5A). However, preincubation of imDCs with affinity-purified anti–βig-h3 antibodies could inhibit subsequent imDC uptake of FITC-DX by approximately 50% (Figure 5C). Nonimmune mouse IgG showed no significant inhibition. Indirectly, this suggests a possible autocrine regulation of imDC endocytosis through secreted βig-h3.

Because βig-h3 binds to macrophages (Figure 4A), we then examined whether it regulates macrophage endocytosis. Macrophages exhibit intermediate levels of FITC-DX and LY endocytosis without stimulation (Figure 5D). Stimulation of macrophages with βig-h3 increased FITC-DX uptake 3-fold (MFI = 737.79; Figure 5E). The uptake of LY was also increased 75% (MFI = 89.2; Figure 5E). In fact, βig-h3–stimulated macrophages exhibited a rate of FITC-DX and LY uptake comparable to imDCs. These results suggest a possible paracrine regulation of macrophage endocytosis by imDCs through βig-h3. In contrast, βig-h3 was unable to stimulate mDC endocytosis. mDCs exhibited low endocytosis that could not be stimulated by βig-h3 (Figure 5D). This is consistent with the lack of βig-h3 binding to mDCs.

LPS stimulates endocytosis.²⁹  This was also observed in this study and it was inhibited by polymyxin B (Figure 5F). However, polymyxin B showed no inhibition of βig-h3–induced endocytosis (Figure 5F). Instead, it slightly enhanced βig-h3–stimulated macrophage endocytosis. Similarly purified CD14 could not stimulate macrophage endocytosis showing insignificant LPS contamination in proteins purified with this protocol. Therefore, βig-h3 stimulates macrophage endocytosis.

Because βig-h3 stimulates macrophage endocytosis or macropinocytosis, we examined whether it involves the activation of membrane ruffling. Macrophages, stimulated with purified βig-h3, were fixed and stained using phalloidin-rhodamine.²⁵  As shown in Figure 6A, untreated macrophages had few membrane ruffles. Stimulation of macrophages with CD14 did not induce significant membrane ruffles (Figure 6B). However, βig-h3–stimulated macrophages displayed extensive membrane ruffling (Figure 6C). Therefore, βig-h3 probably induces macrophage endocytosis or macropinocytosis through activation of membrane ruffles.

To examine whether βig-h3 activates Rac and cdc42, macrophages were stimulated with βig-h3 and then examined for Rac or cdc42 activation. As positive controls, both Rac-GTPγS and cdc42-GTPγS were precipitated by PAK-1 PBD-agarose (Figure 6D). Although activated Rac was also detected in the lysate of βig-h3–stimulated macrophages, no activated cdc42 was detected (Figure 6D). As negative controls, neither Rac nor cdc42 was precipitated from unstimulated or CD14-stimulated macrophages (Figure 6E). This result shows that βig-h3 stimulation of macrophages activates Rac but not cdc42, implying that Rac activation may be required in βig-h3–elicited macrophage membrane ruffling and endocytosis.

Microarray and other approaches have been used in previous studies to identify genes that are preferentially expressed in DCs and related cell types as summarized by Tang and Saltzman.³⁰  In this report, we compared the gene-expression profiles of monocytes, macrophages, imDCs, and mDCs serially with particular interests in identifying genes that are preferentially regulated in imDCs (Figure 1B). The reliability of the expression data identified for 121 differentially expressed genes is supported by reports of similar expression profiles of many of these genes.³⁰ 

One subset of genes is related to cell migration or motility. CCR1 and CCR7 were found to be preferentially expressed by imDCs and mDCs, respectively (Table 1). This switch from CCR1 to CCR7 is consistent with previous reports and is in line with DC migration from peripheral to lymphoid tissues on activation.^7-11  RGS1 regulates chemokine receptor signaling and it has been shown to be up-regulated during DC maturation (Table 1).^31,32  The Wiskott-Aldrich Syndrome protein (WASp) and the subunit 4 of actin-related protein 2/3 (Arp2/3), which are required for the podosome formation at the front edges of imDCs or DC migration,³³  are up-regulated in DCs (Table 1). Rho GDP dissociation inhibitor β (RhoGDIβ), which regulates Rho GTPase activation, was selectively up-regulated in imDCs (Table 1).³³ 

The M-CSF receptor was highly up-regulated in imDCs, but down-regulated on activation (Table 1). Indeed, M-CSF was shown to revert imDC, but not mDC, differentiation into macrophages.³⁴  Compared with monocytes, imDCs and mDCs have reduced G-CSF receptor 1 expression (Table 1), which is generally consistent with monocytes typically ceasing to proliferate after 2 days in culture.³⁵  The elevated expression of IFNγ receptor 2 in DCs (Table 1) is consistent with potent IFNγ costimulation of IL-12 production in DCs.³⁶ 

Surface CD83 is a hallmark of human DC activation,³⁷  but differential CD83 mRNA expression was not observed in the microarray results. In fact, CD83 mRNA was detected in all 4 cell types although, at the protein level, it was only stably expressed on mDCs.²⁰  Monocytes and macrophages only transiently express surface CD83.²⁰ 

Among genes that are preferentially up-regulated in imDCs, a gene encoding βig-h3 was most prominent (Table 1). Previous studies have described βig-h3 as an extracellular matrix (ECM) protein secreted by nonleukocytes, which regulates cell proliferation, differentiation, and adhesion.^27,28  We observed that βig-h3 mRNA was more abundant in lymphoid-rich tissues. At the cellular level, its expression is selectively up-regulated during monocyte differentiation into imDCs. It was also transiently induced after monocytes were cultured for 24 hours with GM-CSF and IL-4 but this did not sustain beyond 2 days (Figure 2B). It is not clear whether this was simply due to transient GM-CSF and IL-4 induction. TGF-β1 could not induce βig-h3 mRNA expression in monocytes (data not shown). Its expression became stabilized from day 4 of the culture when the cells also acquired DC properties. By differential display and microarray, βig-h3 expression in DCs has also been reported by others but not further investigated.^38,39  The specific role of βig-h3 expression with regards to DC functions in vivo remains to be investigated.

Our demonstration of imDC-associated βig-h3 expression and its ability to stimulate endocytosis represents a novel observation for both βig-h3 and imDCs/macrophages. The ability of βig-h3 to bind to macrophages, to activate Rac, and to induce membrane ruffles are all in line with its ability to stimulate macrophage endocytosis. In vivo, how βig-h3–stimulated macrophage endocytosis affects macrophage functions remains to be investigated. imDCs are likely to lose βig-h3 production under inflammatory conditions when these cells become mDCs. Therefore, it is possible that imDC stimulates macrophage and its own endocytosis of self-antigens under steady state to promote self-tolerance.

Addition of βig-h3 to the culture did not affect the constitutively potent endocytic activities of imDCs. It is possible that βig-h3 secreted by imDCs binds to imDCs rapidly with high affinity to maintain potent endocytosis. To assess the role of βig-h3 in imDC endocytosis, βig-h3–specific siRNA was used to transfect imDCs. But the transfection experiments were not successful. However, we have demonstrated that imDC endocytosis was inhibited by affinity-purified anti–βig-h3 antibodies. The βig-h3–stimulated macrophage endocytosis was not due to contaminating LPS in the purified βig-h3 but the protein itself. This was demonstrated by the lack of inhibition of βig-h3–stimulated macrophage endocytosis by polymyxin B, which inhibited LPS-stimulated macrophage endocytosis. Similarly purified CD14 exhibited no macrophage endocytosis, enhancement of showing that this protocol of protein expression and purification does not usually result in significant LPS contamination.

The βig-h3 polypeptide is dominated by 4 tandem fas-1 domains, defined based on homology with the Drosophila protein fasciclin-I.⁴⁰  Although our results suggest a role for βig-h3 in the regulation of macrophage endocytosis, previous studies suggest a role of βig-h3 as an ECM protein that regulates cell adhesion, differentiation, and proliferation.^27,41-44  Our demonstration of the lack of disulfide-linked βig-h3 oligomerization is also in contrast with previous detection of native tissue βig-h3 as highly oligomerized fibrils. It is possible that βig-h3 is found in ECM-associated and soluble forms, depending on the tissue compartments or cell types where it is produced, and the 2 forms may exhibit distinct functional properties. An ECM-related βig-h3 function is also supported by the fact that mutations in βig-h3, for example, the residue Arg124His mutation, are associated with lattice cornea dystrophy or granular dystrophy in which βig-h3 was found degraded and abnormally located.^45,46  We introduced the Arg124His mutation in the recombinant βig-h3 but it remained potent in stimulating macrophage endocytosis (data not shown). Indeed, immunologic malfunctions have not been reported in these patients.

Although ECM-related functions of βig-h3 have been reported, the abundant βig-h3 mRNA detected in lymphoid tissues and its secretion by imDCs strongly suggests a role for βig-h3 in immune regulation. Our demonstration of βig-h3 stimulation of macrophage endocytosis is consistent with such a role.

P.T. thanks Prof Hui Kam Man for support and encouragement.