Abstract
Despite their critical function as APCs for primary immune responses, dendritic cells (DC) and Langerhans cells (LC) have been rarely used as targets of gene-based manipulation because well-defined regulatory elements controlling LC/DC-specific expression have not been identified. Previously, we identified dectin-2, a C-type lectin receptor expressed selectively by LC-like XS cell lines and by LC within mouse epidermis. Because these characteristics raised the possibility that dectin-2 promoter may direct LC/DC-specific gene expression, we isolated a 3.2-kb nucleotide fragment from the 5′-flanking region of the dectin-2 gene (Dec2FR) and characterized its regulatory elements and the transcriptional activity using a luciferase (Luc) reporter system. The Dec2FR contains a putative TATA box and cis-acting elements, such as the IFN-stimulated response element, that drive gene expression specifically in XS cells. Dec2FR comprises repressor, enhancer, and promoter regions, and the latter two regions coregulate XS cell-specific gene expression. In transgenic mice bearing a Dec2FR-regulated Luc gene, the skin was the predominant site of Luc activity and LC were the exclusive source of such activity within epidermis. By contrast, other APCs (DC, macrophages, and B cells) and T cells expressed Luc activity close to background levels. We conclude that epidermal LC are targeted selectively for high-level constitutive gene expression by Dec2FR in vitro and in vivo. Our findings lay the foundation for use of the dectin-2 promoter in LC-targeted gene expression systems that may enhance vaccination efficacy and regulate immune responses.
Dendritic cells (DC)3 are professional APCs that are critical to immunity because of their unsurpassed potency in presenting Ags to naive T cells (1). Although widely distributed in various tissues, DC comprise a minuscule segment (<5%) of the cell population in these organs. Langerhans cells (LC) are considered to be an immature form of DC that reside in the epidermis and in mucosal tissues (2, 3), where they can endocytose a wide variety of Ags including contact hapten sensitizers, tumor-associated Ags, and microbial Ags. After capturing Ags, LC migrate into the T cell zone of regional lymph nodes, during which process they become mature DC characterized by reduced endocytic ability and increased immunostimulatory capacity (3). Finally, LC activate naive T cells in lymph nodes and thereafter are probably subjected to apoptosis induced by the activated T cells (4).
DC are derived from the bone marrow (2) and express surface molecules (e.g., MHC class II, CD80, and CD86) required for T cell activation. LC are thought to undergo differentiation pathways that distinguish them from other DC (5, 6, 7). Unlike DC, LC express E-cadherin, an adhesion molecule that connects them to keratinocytes (8), and CD1a (in the case of human LC) (9), a molecule important for presenting lipid and glycolipid Ags (10). LC are also unique in expressing Birbeck granules (11), an organelle system consisting of superimposed and zippered membranes that may account for distinct Ag processing properties (12). It would appear that some of these LC-specific features could be used in the development of gene-based manipulation of LC function. However, none of them to date have been so applied.
Our goal has been to identify and characterize genes that distinguish LC from other APCs and to use their regulatory elements to selectively target gene expression to LC. By subtractive cDNA cloning of the LC-like XS cell line derived from the epidermis of newborn BALB/c mice (13), we have previously isolated the dectin-2 gene that encodes for a novel C-type lectin receptor (14) with costimulatory function. Because dectin-2 mRNA is expressed selectively by XS cell line and exclusively by epidermal LC in vivo, we hypothesized that its promoter may be a LC-specific regulatory element. To prove this, we isolated the 5′-flanking region of the dectin-2 gene (Dec2FR) and examined its regulatory role in controlling LC-specific expression. Our findings indicate that Dec2FR possesses unique promoter sequences that drive gene expression specifically in XS cell lines in vitro and at differentially high levels in epidermal LC in transgenic mice.
Materials and Methods
Cell lines
XS106 and XS52 cells were obtained from A. Takashima (University of Texas Southwestern Medical Center, Dallas, TX). These cells are long-term DC lines established from the epidermis of newborn BALB/c mice maintained and expanded in complete RPMI 1640 supplemented with mouse rGM-CSF (1 ng/ml) and NS47 fibroblast culture supernatant (10% v/v) as source of CSF-1 (13). The NS47 line of dermal fibroblasts was maintained in complete RPMI 1640 supplemented with 10% FBS (13).
Ab and reagents
The mAb specific for CD8, Mac-1, MHC class II I-Ad/I-Ed (2G9), CD45R/B220, CD3ε, CD11c, and FcR (CD16/CD32, FcγIII/IIR) were purchased from BD PharMingen (San Diego, CA). All chemical reagents used were obtained from Sigma-Aldrich (St. Louis, MO).
Cloning of Dec2FR and nucleotide sequencing
Genomic nucleotide sequences for dectin-2 were isolated from a genomic library prepared from BALB/c mouse (Clontech Laboratories, Palo Alto, CA) by hybridization with the cDNA for dectin-2 (GenBank accession number AF240357) (14). One of the isolated phage clones contained a 3.2-kb fragment of the 5′-flanking region that was amplified by PCR using Expand Long Template PCR System (Roche Diagnostics, Indianapolis, IN) with primers complementary to a 5′ end cloning site and a 3′ end sequence of 5′-untranslated region. The amplified nucleotide was subcloned into a plasmid vector, pGEM-7zf(−) (pGEM-Dec2FR; Promega, Madison, WI), and its nucleotide sequence was determined at both sense and antisense strands by the automated sequencing of deletion mutants produced by Erase-a-base system (Promega).
Primer extension assay
This assay was used to map mRNA initiation sites on the 5′-flanking region. Fifty femtomoles of the synthetic oligonucleotide, 5′-CCAGAGTTCAGAATCAACTTCCACACACACTT-3′, was 5′ end-labeled with [γ-32P]ATP and hybridized with 30 μg of total RNA isolated from XS52 cells. In the presence of the reverse transcriptase (10 U), Superscript II (Life Technologies, Grand Island, NY), the cDNA strand was extended from the primer toward the 5′ end of dectin-2 mRNA. This extended strand was separated on 8% urea-PAGE in parallel with sequencing ladders (C, T, A, and G) synthesized from genomic DNA by the same primer. The location of the 5′ end (mRNA initiation site) was determined by the position of the nucleotide in the ladders showing the same size of the extended strand.
Construction of Luc expression vectors
The area containing a 5′-flanking region (nt −3176 to +1, designated the mRNA initiation site as +1) and the 5′-untranslated region was excised by digestion of pGEM-Dec2FR with MluI and XhoI restriction enzymes and introduced upstream a luciferase (Luc) coding sequence in pGL3-basic (Promega) without transcriptional control elements (pDec2FR-Luc). A second control was pGL3-control (or pSV40-Luc) (Promega), which contains SV40 promoter (SVP) and enhancer upstream and downstream, respectively, of the Luc gene. A set of deletion mutants lacking 5′-flanking sequences from its 5′ end was constructed by PCR-based mutagenesis. Briefly, a nucleotide fragment spanning the indicated 5′ end (Fig. 3⇓) to the +126 was PCR-amplified and cloned into a pGL3-basic Luc vector using MluI and XhoI sites. For deletion mutants designed to determine repressor and minimum promoter regions, a putative enhancer fragment (nt −2741 to −1850) was also PCR-amplified and inserted into some of the previous deletion mutants through SacI and MluI sites. For experiments examining whether the dectin-2 enhancer and a minimum promoter region (nt −123 to +126) control LC/DC specificity, we constructed a third set of Luc expression vectors. Dectin-2 enhancer and minimum promoter fragments were inserted separately or together into a pGL3-basic vector (pDec2E-Luc, pDec2P-Luc, and pDec2E/Dec2P-Luc, respectively). The enhancer fragment was also inserted into a pGL3-promoter (Promega) containing a SVP sequence, using MluI and XhoI sites (pSVP/Dec2E-Luc).
DNA transfection and Luc assays
Luc vector DNA was delivered into various cell lines using liposome-mediated transfection (15). XS106 or NS47 cells, seeded on a 60-mm dish at a density of 1 × 106 1 day before, were cultured in the presence of 1 μg of DNA and 3 μl of Fugene (Roche Diagnostics). Following a 24-h incubation, whole cell extracts were prepared from cells by lysis in 1× Reporter lysis buffer (Promega). For experiments in which tissue Luc activity was examined, proteins were extracted from excised tissues by homogenization in the lysis buffer (1 ml/100 mg of wet weight). An aliquot (10–20%) was used to measure Luc activities determined by light emitted for 30 s with an Optocomp I Luminometer (MGM Instruments, Hamden, CT) using Luciferase Assay System (Promega). Protein concentration was determined by the Bradford method, and their values were used to normalize Luc activity (16).
Electromobility shift assay
The nucleotide sequence −123 to −34 was segmented into three duplexed oligonucleotides: DM1 prober (35-mer), 5′-CCACATTAGGAACTGAGAAAGTAATGAGAACATTC-3′ (nt −123 to 89); DM2 probe (30-mer), ACATTCTTGACAGAGTTTTTAGGAACAAAT (nt −94 to −65); and DM3 probe (33-mer), AAATTTAGGTATGTTTCTCAATTTCCTCTTTCC (nt −66 to −34). The oligonucleotide probe for SP1 (22-mer; ATTCGATCGGGGCGGGGCGAGC) and a control competitor (GTGTTGGGCGCGTTATTTATCGGAGTTGCA) derived from the coding sequence for Luc version 3 (Promega) were also synthesized. These oligonucleotides were 5′ end-labeled with T4 DNA kinase and mixed with 10 μg of nuclear extracts. The preparation of nuclear extracts and electromobility shift assays were performed as described previously (17, 18). For competition experiments, nuclear extracts prepared from XS106 DC were incubated with unlabeled DM2 or control competitor oligonucleotide at various concentrations before addition of radiolabeled DM2 probe.
Generation of transgenic mice
pDec2FR-Luc DNA was depleted of its plasmid sequence by digestion with MluI and BamHI restriction enzymes and was highly purified by agarose electrophoresis and DNA extraction from gel plugs using a Nucleo Spin Extraction kit (Clontech Laboratories). The purified DNA was microinjected into fertilized oozytes obtained from ICR mice (The Jackson Laboratory, Bar Harbor, ME) (performed by the Transgenic Mouse Facility at University of Texas Southwestern Medical Center). Transgenic founders were identified by PCR analysis of genomic DNA (1 μg) extracted from tail biopsies using primers for Luc gene: 5′-GAAGTCGGGGAAGCGGTTGC-3′ and 5′-CGGCGTCATCGTCGGGAAGA-3′. After 30-cycle amplification, PCR products were separated on 1.5% agarose gel, transferred onto a membrane, and followed by Southern hybridization with the corresponding DNA probe. The founders whose tail DNA showed expected size of the PCR and specific hybridization were selected and examined for Luc activity in their tails. Consequently, three transgenic founders were generated, and two of them, showing the highest Luc activity, were selected as breeding partners. Their F2 offspring mice were produced by mating with ICR wild-type mice, and the mice aged 4–20 wk were used for experiments addressing the LC/DC specificity for Dec2FR promoter activity. Animals were housed in the pathogen-free facility of the Animal Resource Center at the University of Texas Southwestern Medical Center. All the experiments were conducted according to guidelines of the National Institutes of Health.
Depletion and purification of epidermal LC
Depletion and purification were performed using immunomagnetic beads. Epidermal cells were prepared from ear skins of eight transgenic mice, as described previously (19). Cells were processed in three ways: one batch (5 × 105 cells) was not treated, whereas the other batches (1.5 × 106 cells per batch) were incubated for 30 min on ice with 10 μg/ml rat mAb specific for CD8 (control) or MHC class II molecule (Ia). After extensive washing, cells were incubated with magnetic beads (2 × 107) coated with anti-rat IgG (Dynal Biotech, Lake Success, NY). Unbound (depleted of Ia+ cells) and bound (purified for Ia+ cells) fractions were resuspended in 200 μl of PBS, and an aliquot (20 μl) was used for the cell counting. Protein was extracted from the rest of cells and assayed for Luc activity using 40% of the total extract. The extent of depletion of Ia+ cells was estimated by reduction in frequency of Ia+ cells after treatment with the magnetic beads. Small aliquots of unfractionated epidermal cells (just before use of magnetic beads) and of a fraction not bound to anti-Ia were stained with 10 μg/ml FITC-conjugated anti-rat IgG. Untreated epidermal cells and cells of anti-CD8-unbound fraction were also stained with anti-Ia mAb. These stained cells were examined for frequency of Ia+ cells by flow cytometry.
Purification of leukocytes
DC, T cells, and B cells were purified from spleen. For DC, spleen cells were prepared from the same pool of transgenic mice as was used for preparing epidermal cell suspensions (Fig. 6⇓). After pretreatment with anti-FcR Ab (10 μg/ml), the spleen cells (5 × 106 cells) were incubated with 10 μg/ml biotin-conjugated rat anti-CD11c mAb, followed by treatment with streptavidin-coated magnetic beads (5 × 107; Dynal Biotech). For T and B cells, after staining of spleen cells (2 × 106) with 10 μg/ml FITC-conjugated anti-CD3ε or anti-B220, T cells (CD3+) and B cells (B220+) were purified by flow cytometric sorting. For peritoneal macrophages, at 4 day after i.p. injection of 3% thioglycolate (in PBS), peritoneal cells were collected from the transgenic mice and cultured for 3 days in 10% FCS-RPMI 1640 (20). Following extensive washing of the culture dishes to remove floating cells and weakly adherent cells, peritoneal macrophages were harvested by scraping and examined for expression of Mac-1 and B220+ to estimate contamination of peritoneal B cells (B220+) using flow cytometric analysis. Typically, the adherent cells comprise <10% of B cells. Cell counting and Luc assays were performed as described above. Selective isolation of DC, macrophages, T cells, and B cells was confirmed by flow cytometric analyses and by parallel setting with control Abs.
Results
Dec2FR contains essential promoter and cis-acting elements
To isolate the transcriptional regulatory region of the dectin-2 gene, we screened a BALB/c mouse genomic library. Five independent phage clones were isolated, which cover ∼40 kb of the dectin-2 gene containing six exons and a long stretch (3.2 kb) of Dec2FR (14). Primer extension assay and nucleotide sequence analyses revealed that Dec2FR has one major site for mRNA initiation at 126 bp from the initiation codon and a putative TATA box at 24 bp upstream of the initiation site (Fig. 1⇓), both of which are essential elements for de novo mRNA synthesis.
Characterization of a proximal region of Dec2FR. A, Primer extension assay was used to map initiation sites for dectin-2 mRNA. Total RNA isolated from XS52 DC was primed for cDNA synthesis with 32P-labeled oligonucleotide, and the extended cDNA strand (cDNA) was separated on denatured PAGE in parallel with sequencing ladders (C, T, A, and G). An arrow indicates the 5′ end position of cDNA strand (i.e., mRNA initiation site). A putative TATA box and the initiation site (designated as +1) are shown on the nucleotide sequence of Dec2FR. B, A proximal region (nt −550 to +1) of Dec2FR was searched for NF binding sites using a transcription factor database, revealing the presence of the following sites: AP-1 (12-O-tetra-decanoylphorbol-13-acetate-inducible element), adenovirus-conserved (inverted terminal repeats in adenovirus promoter), BHLH-CS (consensus binding site for helix-loop-helix NF family), CCAAT/enhancer binding protein-TIRS2 (multiple elements in a promoter region of transthyretin), γIRE-CS (IFN-γ-responsive element in MHC class II gene promoters), GM-CSF-CS (responsible for GM-CSF expression by Ag- or mitogen-stimulated T cells), H2A-conserved (conserved site in multiple histone H2A promoter), H-2RIIBP (retinoic responsive element in MHC class I gene), IE1.2 (NF binding site in human CMV enhancer), NF-Y-MHCII (conserved sequence for class II gene promoters), and TATA box (essential promoter element).
The entire nucleotide sequence of Dec2FR was determined and a proximal region (nt −550 to +1) of the mRNA initiation site was searched for putative NF binding sites using a transcription factor II database. AP-1 binding sites (AP-1) (21), IFN-γ-responsible elements (22), retinoic responsive element (H-2RIIBP) (23), and an element (GM-CSF-consensus sequence (CS)) responsible for gene expression of GM-CSF by T cells in response to Ag stimulation (24) were found (Fig. 1⇑B), suggesting that Dec2FR transcriptional activity may be up-regulated by a wide variety of stimuli, including phorbol esters, IFN-γ, retinoic acids, and Ag stimulation. We also noted the presence of an NF-Y-MHCII site, which controls MHC class II gene expression (25), suggesting that the Dec2FR may have a relationship with class II expression.
Dec2FR drives gene expression specifically in XS cell lines
Dec2FR or SV40 transcription units were linked separately to the Luc gene (pDec2FR-Luc and pSV40-Luc, respectively) (Fig. 2⇓A) and introduced into various cell lines. pDec2FR-Luc produced 18-fold higher activity compared with pGL3-basic (promoterless Luc vector) in XS106 DC (Fig. 2⇓B), a line that resembles epidermal LC phenotypically and functionally (13, 26). By contrast, the promoter activity was close to the baseline level in NS47 fibroblasts (Fig. 2⇓B) and considerably lower or minimal in other cell lines including macrophage, B cell, T cell, and keratinocyte lines (48). The control SV40 transcription unit induced strong Luc activity in both XS106 DC (65-fold) and NS47 fibroblasts (153-fold higher than the baseline). This XS106 DC-selective expression of Dec2FR activity is consistent with our previous finding of exclusive dectin-2 mRNA expression in the same cell line (14).
Promoter activity of the Dec2FR fragment. A, Map of Dec2FR-driven Luc expression vector (pDec2FR-Luc). A 3.2-kb fragment isolated from Dec2FR contains a putative TATA box, a major mRNA initiation site (+1), and a full length of 5′-untranslated region (+2 to +126, shown by an open box). The filled box represents an SV40 late poly(A) signal. B, XS106 DC and NS47 fibroblasts were transfected in triplicate with pDec2FR-Luc, SV40 (enhancer and promoter)-driven Luc (pSV40-Luc), or Luc alone (pGL3-basic) and assayed for protein concentration and Luc activity (RLU). The Luc activity (1760 average RLU/μg) expressed by Dec2FR in NS47 cells was just above the baseline expression shown by pGL3-basic (1496 RLU).
Dec2FR consists of three functionally independent regions
To identify nucleotide sequences controlling Dec2FR promoter activity, we conducted deletion mutant analyses. Various nucleotide sequences were deleted from the 5′ end of Dec2FR (nt −3176 to +126), ligated to the pGL3-basic vector, and transfected into XS106 DC or NS47 fibroblasts (Fig. 3⇓A). Dec2FR produced high levels of Luc activity (29 ± 6% of Luc activity by pSV40-Luc) in XS106 DC. Deletion of the region between −2741 and −1852 dramatically reduced transcriptional activity, suggesting that it is required for high activity, probably functioning as an enhancer. Progressive deletion within region −1851 to −507 partially restored activity, whereas further deletion of region −506 to +1 completely abrogated activity. To more precisely characterize region −1851 to −36, segments were progressively deleted and inserted into a Luc vector containing the putative enhancer fragment −2741 to −1852. Deletion of region −1851 to −507 markedly enhanced promoter activity (11-fold; 1681 vs 149 relative Luc activity of wild type), suggesting that it serves as a repressor (Fig. 3⇓B). By contrast, further deletion of region −506 to −277 led to the lowest level of reduced activity shown by mutant −276. Surprisingly, mutant −123 produced increased activity, indicating the presence of a second repressor region between −276 and −124. Mutant −35, which contains a TATA box and its downstream region, lost almost all activity. When transfected into NS47 fibroblasts, all mutants produced considerably low activity ranging between 1.0 and 6.4 relative Luc activity to pGL3-basic. Because even mutant −123 retained selective expression in XS106 cells, we surmised that region −123 to +1 and the putative enhancer contain the minimal sequences required for the XS cell selectivity. Removal of a 5′-untranslated region (5′UTR), spanning nt +1 to +126, from the wild type and a mutant −123 reduced transcriptional activity by 72 and 20%, respectively. However, it should be noted that the 5UTR itself did not produce any detectable Luc activity even in the presence of an enhancer (data not shown). These analyses indicate that Dec2FR consists of an enhancer (−2741 to −1852), two repressors (−1851 to −507 and −276 to −124), and a minimum promoter (−123 to +1) including a TATA box and an mRNA initiation site (Fig. 3⇓C).
Dec2FR contains three functionally distinct regions. A, Deletion mutants of Dec2FR were generated by PCR-based mutagenesis. 5′ end positions of mutants are shown by distance (bp) from the mRNA initiation site. Mutants were transfected into XS106 DC and examined for Luc activity. Transcriptional activity was indicated by the percentage of relative Luc activity to that of pSV40-Luc vector. B, Putative repressor and minimal promoter regions were mapped. A putative enhancer fragment (−2741 to −1852) was inserted downstream the Luc gene in deletion mutants of the region −1851 to −35 and measured for Luc activity in XS106 and NS47 cells. Each transcriptional activity was calculated as RLU of mutant relative to pGL3-basic. C, Data from two sets of deletion mutant analyses are summarized. Dec 2FR consists of, from the 5′ end, an enhancer, two repressors, and a minimal promoter including a TATA box and an mRNA initiation site. D, Enhancer and promoter components in the indicated Luc vectors are shown at the left; dectin-2 enhancer fragment (Dec2E), the minimal promoter (Dec2P), and an SVP. These Luc vectors were transfected into XS106 and NS47 cells and their transcriptional activities were compared with that of pSV40-Luc. These data (A, B, and D) were derived from triplicate assays, representative of three independent experiments.
We next determined the role of the putative enhancer (Dec2E) and/or minimal promoter (Dec2P) in regulating the XS DC-specific expression. Dec2E fragment was inserted into a Luc vector with Dec2P or with the SVP and examined for Luc activity in XS106 DC or NS47 fibroblasts (Fig. 3⇑D). Neither Dec2E nor Dec2P alone displayed strong activity in the XS DC, whereas activity was greatly enhanced when the two fragments were linked, with the level of enhancement much higher in XS cells than in fibroblasts. Moreover, Dec2E also enhanced the activity of the heterologous promoter, SV40, more intensely in XS cells than in fibroblasts (Fig. 3⇑D). These data verify the presence of enhancer function in this region and also indicate that enhancer and minimal promoter coregulate XS DC specificity of dectin-2 gene expression.
Nuclear proteins that bind to the minimum promoter sequence are expressed abundantly by XS106 cells
Because a nucleotide sequence spanning −123 to −36 (88 bp) in the minimum promoter was shown to regulate transcriptional activity and selectivity of cell expression of the dectin-2 promoter, we performed electromobility shift assays to detect nuclear proteins that bind to the 88 bp. The sequence was divided into three regions: DM1 (nt −123 to 89), DM2 (nt −94 to −65), and DM3 (nt −66 to −34). DM1 probe detected several species of DNA-protein complexes in nuclear extracts isolated from XS106 cells. Three of these species were also expressed in other cell lines (e.g., Raw macrophage, J558 B cell, BW5147 thymocyte, and S105 fibroblast lines) (Fig. 4⇓A), whereas two minor species were expressed only by XS106 cells. However, the respective binding activities of these two latter species were not blocked with unlabeled DM1 oligonucleotide, indicating that binding was not specific (data not shown). DM3 probe detected six species, all of which were also widely expressed. Finally, DM2 probe detected four species in complexes formed with XS106 nuclear extracts, two of which (third and fourth species) were highly expressed in XS106 nuclear extracts, only minimally in macrophage extracts, and not at all in other cell extracts. The abundant expression in XS106 cells was not due to a higher concentration of nuclear proteins in the cells, because expression levels of complexes with DM1 probe and Sp1 transcription factor (Ref. 27 and data not shown) were similar for XS106 and Raw macrophage cells. Moreover, binding activity of XS106 nuclear extracts was almost completely inhibited by 50 ng of unlabeled DM2 (corresponding to a molecular excess of 1:300), but not by control oligonucleotide (Fig. 4⇓B). These results indicate that the Dec2FR minimum promoter contains cis-acting element(s) that may play a role in regulating preferential expression of dectin-2 gene in XS cells.
XS cells express NFs that bind to minimum promoter sequence. A, Double-stranded DM1 (35-mer), DM2 (30-mer), or DM3 (33-mer) oligonucleotides corresponding to the minimum promoter sequence were 32P-labeled and mixed with nuclear extracts (NE) prepared from XS106 DC, Raw macrophages (Mφ), J558 B cells, BW5147 thymocytes, and S105 fibroblast cells. DNA-protein complexes were electrophoretically separated from free probes on native PAGE. B, Sequence specificity of DNA binding activity was examined. Before binding reaction, XS106 nuclear extracts were pretreated with 50 or 100 ng of unlabeled DM2 or control competitor (Ctrl) containing no NF binding sites. Four DNA-protein complexes are shown by arrows with number in decreasing order of m.w. The third and fourth species marked with ∗ are preferentially expressed in XS106 DC.
Tissue distribution of Luc activity in Dec2FR-Luc-transgenic mice
To evaluate the cell-type specificity of Dec2FR activity in vivo, we generated transgenic mice bearing the Dec2FR-driven Luc gene. Various organs were excised and examined for Luc-specific activity (relative light units (RLU) per microgram of protein) (Fig. 5⇓). Luc activity was highest in skin (from ear), with lower levels in lymphoid organs (spleen, lymph node, and thymus) known to contain relatively large numbers of DC. Nonlymphoid organs (e.g., adipose tissue, heart, and kidney) showed only background levels, although organs (lung, intestine, and testis) reported to harbor modest LC/DC densities showed similar Luc activity to lymphoid organ (Fig. 5⇓) (28, 29, 30). The expression in skin becomes even more marked when Luc activity was normalized by DNA concentration. Because the ear is made up of skin (epidermis and dermis) and cartilage, we determined what tissue is a major source for the high activity. Luc activity was traced almost entirely to skin, with 80% in epidermis, 20% in dermis, and negligible activity in cartilage. Therefore, the highest Dec2FR activity appears linked to epidermal LC.
Tissue distribution of Luc activity in transgenic mice with Dec2FR-Luc. Proteins and DNA were prepared from various organs of two individual transgenic mice with Dec2FR-Luc. Protein, DNA concentrations, and Luc activity were measured, and Luc activity in tissues was calculated as RLU per microgram of protein extract and also did as RLU per microgram of DNA (for only one transgenic mouse). Filled and shaded bars represent activity in skin and lymphoid tissues, respectively. Organs from control littermates did not produce detectable activity.
Epidermal Luc activity is expressed exclusively by LC
Epidermal cells prepared from ear skin (pool of eight transgenic mice) expressed activity (8,023 RLU/104 cells). Fractionation into Ia+ cells (LC) and Ia− cells (mostly keratinocytes) by immunomagnetic beads revealed markedly high activity in the Ia+ fraction (119,922 RLU/104 cells), 15-fold higher than unfractionated epidermal cells, whereas the Ia− fraction possessed only background activity (Fig. 6⇓A). Selectivity and efficacy of LC depletion was confirmed by marked reduction in frequency of Ia+ cells (from 4.8 to 0.05%) following depletion with anti-Ia treatment, but not with anti-CD8-coated beads (Fig. 6⇓B). Therefore, Dec2FR activity is targeted to epidermal LC in the skin.
Epidermal Luc activity is expressed exclusively by LC. A, Epidermal cells prepared from ear skin of transgenic mice (a pool of eight offspring) were divided into three batches; one batch was not treated and the other batches were incubated with magnetic beads coated with rat anti-Ia or with anti-CD8 mAb (control). Unbound and bound fractions were separately collected and examined for cell number and Luc activity. These data were derived from two independent experiments. Aa, Luc activity is expressed as RLU/104 cells. Ab, The anti-CD8-bound fraction was not counted. Ac, Values obtained in A were compared with those of epidermal cells (1.0). B, In parallel, small aliquots of unfractionated epidermal cells and of fractions not bound to anti-Ia or anti-CD8 were collected and stained with goat anti-rat IgG or rat anti-Ia Ab (for CD8-depleted). These samples were assayed for the frequency of Ia+ cells by flow cytometric analysis.
Epidermal LC is targeted for high expression of Luc among leukocytes
We next examined Luc expression levels in splenic DC and in non-DC leukocytes. Spleen cells prepared from the previous pool of transgenic mice showed 101 RLU/104 cells, 80-fold lower than in epidermal cells, and this preparation was used to isolate splenic DC, T cells, and B cells. Splenic DC, purified using anti-CD11c Ab-magnetic beads, expressed 150-fold lower activity than epidermal LC (Fig. 7⇓A). T and B cells, purified by flow cytometric sorting of CD3+ cells and CD45R/B220+ cells, respectively, both expressed mere background activity, documenting lack of Dec2FR influence in these cells. Finally, macrophages (Mac-1+) isolated from peritoneal cells elicited with thioglycolate displayed similar Luc activity as splenic DC. These findings lead us to conclude that epidermal LC are targeted selectively for high-level constitutive gene expression by Dec2FR in transgenic mice.
Dec2FR activity is targeted to epidermal LC in transgenic mice. A, DC (CD11c+), B cells (B220+), and T cells (CD3+) were purified from spleen cells prepared from the same pool of transgenic mice as was used in Expt. 1 of Fig. 6⇑, using immunomagnetic beads or flow cytometric sorting. Peritoneal macrophages (90% Mac-1high) were isolated from peritoneal cells of a different pool of five transgenic mice stimulated with thioglycolate. Luc activity was calculated as RLU/104 cells and compared with that of epidermal LC (set at a value of 100%). Note that Luc activity for epidermal cells in both pools of mice was similar (8,023 and 11,873 RLU/104 cells). Selective isolation of DC, macrophages, T cells, and B cells was confirmed by flow cytometric analysis as performed in Fig. 6⇑. B, Data are shown in a graphic form.
Discussion
Current consensus acknowledges DC to encompass different subsets that reside in various tissues, each subset bearing overlapping as well as distinct features. In this context, two possibilities exist with respect to the Dec2FR transcriptional regulatory unit. Dec2FR activity may be a feature common to all DC. Alternatively, it may be a distinguishing feature of only some DC subsets (like LC). We have shown Dec2FR activity (on a per-cell basis) to be expressed in epidermal Ia+ LC at differentially high levels compared with splenic DC. Dec2FR activity may be targeted to a certain subset of splenic DC which includes at least two distinct subpopulations with different phenotype and function (31, 32). In support of this possibility is our previous finding that dectin-2 molecule is expressed only by a subset (20%) of CD11c+ spleen DC (data not shown). Furthermore, we have not excluded the possibility that dermal DC express Dec2FR activity, although most certainly at very low levels if at all. Finally, LC reside not only in the epidermis but also in lung and mucosal tissues (e.g., intestine) (33). However, in our transgenic mice, Luc activity was markedly expressed in skin, with 10-fold lower activity per DNA base (Fig. 5⇑) in lung and intestine, suggesting that Dec2FR transcriptional activity may also be targeted to a certain LC subset (epidermal but not lung or intestine).
Dec2FR contains cis-acting regulatory elements in the proximal region of mRNA initiation site, which are responsible for transcriptional activation by IFN-γ (IFN-stimulated response element), retinoic acid (H-2RIIBP), phorbol esters (AP-1), and Ag stimulation (GM-CSF-CS). Their presence raises the possibility that Dec2FR activity can be enhanced by these stimuli. However, we have found each of these stimuli to have null effects on constitutive Dec2FR activity in epidermal LC. By contrast, splenic DC, peritoneal macrophages, and splenic T cells (but not B cells) showed some up-regulated Luc activity following treatment with specific stimuli, although the resultant levels were still much lower than in resting LC (data not shown). Thus, the cis-acting elements may not function in LC, or Dec2FR activity in resting LC may be optimally expressed, with no effects detectable from further stimulation. In contrast, in searching for NFs that bind to the minimum promoter region, we found that a nucleotide sequence of nt −94 to −65 detected two species of DNA-protein complexes abundantly expressed in XS106 cells (Fig. 4⇑), suggesting that the sequence contains at least one cis-acting element recognized by the putative transcription factors. Interestingly, the sequence is AT-rich and does not contain previously identified NF binding sites. We have not yet determined the exact binding sites recognized by XS106 nuclear proteins nor have we functionally characterized the putative sites. However, we are testing the possibility that nuclear proteins forming the third and fourth complexes may be responsible for the preferential selectivity of dectin-2 promoter expression in XS cells.
Transcriptional units of other genes have been used or are potential candidates for DC/LC-targeted gene expression systems. A promoter region of the CD11c gene was used to target DC expression in transgenic mice (34), but this gene is expressed by non-DC such as NK cells (35), intraepithelial cells (36), and eosinophils (37). DEC-205 is expressed by epidermal LC and splenic DC, but also by B cells (20). DC-SIGN is expressed by DC in dermis and other tissues, but not by LC (38). DC-LAMP is expressed by interdigitating DC, but not by LC or dermal DC (39). Finally, Langerin was reported to be expressed selectively by LC (40), but its promoter region has not been isolated or characterized for LC specificity. Thus, limitations related to DC/LC specificity of these markers make Dec2FR the best promoter to date for genetically manipulating LC function in vivo.
We have begun to use a LC-targeted gene expression system to study the in vivo function of LC. DC including LC play critical roles: certainly in initiating primary immune responses through activation of naive T cells; most likely in regulating the nature of T cell responses (e.g., Th1 vs Th2, immunogenic vs tolerogenic); and possibly in inducing innate immunity (e.g., against microorganisms) (41). This diversity of DC function may be reflected in heterogeneity of DC subsets (42). Our epidermal LC-targeted system may allow us to better distinguish between LC and DC (and even among functional subsets of LC), because constitutive Dec2FR activity is markedly higher in LC than in DC. We will use strategies involving gain of function achieved by LC-targeted gene transduction and loss of function through LC-targeted gene disruption via Cre-loxP site-specific recombination (43). Dec2FR can be coupled with inducible gene expression systems (e.g., Tet-ON and Tet-OFF) that will permit conditional regulation of gene expression. Furthermore, we will link our system to a gene producing a readily accessible visual marker (e.g., enhanced green fluorescent protein) that will allow more precise tracking of LC migration in and out of skin, because conventional methods for detecting LC, such as immunostaining using Abs and labeling of epidermal LC with FITC or rhodamine B (44), are limited by loss of the markers during LC migration to lymph nodes and leakage of the fluorescent reagents resulting in nonspecific labeling of DC other than the emigrated LC.
Our LC-targeted system will be used to develop new DNA vaccination strategies. DC (or APC)-targeting delivery of Ags induces rapid Ab responses following single step immunization (45) at 100 to 10,000-fold higher efficiency compared with nontargeted conventional protocols (46, 47). Indeed, we have demonstrated recently that gene gun-mediated DNA delivery of the Dec2FR genetic element leads to efficient induction of immunity against a surrogate Ag (48).
In conclusion, Dec2FR-based LC-targeted gene expression represents a new opportunity to manipulate gene expression in LC, to better understand LC function, and to develop more effective ways of activating and regulating immune responses for preventing and treating diseases.
Acknowledgments
We thank Dr. Akira Takashima for providing XS52 and XS106 DC and NS47 fibroblast lines and Dr. Dorothy Yuan for critical reading of this manuscript. We are also grateful to Alok Das for excellent technical assistance and Susan Milberger for secretarial assistance for preparation of this manuscript.
Footnotes
-
↵1 This study was supported by National Institutes of Health Grant RO1 AR44189-01.
-
↵2 Address correspondence and reprint requests to Dr. Kiyoshi Ariizumi, Department of Dermatology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-9069. E-mail address: Kiyoshi.Ariizumi{at}UTSouthwestern.edu
-
↵3 Abbreviations used in this paper: DC, dendritic cell; CS, consensus sequence; Dec2FR, 5′-flanking region of the dectin-2 gene; LC, Langerhans cell; Luc, luciferase; RLU, relative light unit; H-2RIIBP, retinoic responsive element in MHC class I gene; SVP, SV40 promoter.
- Received July 6, 2001.
- Accepted October 15, 2001.
- Copyright © 2001 by The American Association of Immunologists