Harwood Academic Publishers imprint, part of the Gordon and Breach Publishing Group Printed in Malaysia The Avian Transcription Factor c-Rel Is Expressed in Lymphocyte Precursor Cells and Antigen-Presenting Cells During Thymus Development

Transcription factors of the Rel/NF-κB family are widely involved in the immune system. In this study, we investigate the in vivo expression of the avian protein c-Rel in the T-cell lineage during thymus development. The majority of thymocytes do not express the c-Rel protein. However, lymphocyte precursor cells that colonize the thymus express the c-Rel protein shortly after their homing in the organ and before they begin to differentiate, c-Rel is also detected in different subsets of,antigen-presenting cells such as epithelial cells, dendritic cells, and macrophages. In vitro studies have shown that Rel/NF-κB proteins are sequestered in an inactive form in the cytoplasm by interaction with the IκBα inhibitory protein. By immunocytochemistry, we show that in vivo c-Rel is localized in the cytoplasm of antigen-presenting cells but in both the cytoplasm and nucleus of lymphocyte precursor cells. The cytoplasmic localization of c-Rel in antigen-presenting cells correlates with a high expression of IκBα, whereas the nuclear localization of c-Rel in lymphocyte precursor cells correlates with a much lower expression of IκBα. These results suggest that c-Rel might be constitutively activated in lymphocyte precursor cells.


INTRODUCTION
The proto-oncogene c-rel encodes a transcription factor of the Rel/NF-tcB family. The members of this family share a highly conserved 300 amino acid Nterminal domain, the Rel Homology Domain (RHD), involved in DNA binding, dimerization, nuclear localization, and interaction with inhibitory proteins *Corresponding author. of the ItcB family. Rel/NF-tcB proteins associate as homoor heterodimers and bind a specific DNA sequence, the tcB motif, in the promoter of their target genes. Inside the Rel/NF-tcB family, two subfamilies of proteins can be defined based on the function of their C-terminal domains. The first one includes c-Rel, RelA, RelB, Dorsal, and Dif, which contain a transactivating domain in their C terminus. The second family is composed of the precursor proteins NF-tB 1 and NF-tB2, which contain ItB-like ankyrin repeats in their C terminus. NF-tB1 and NF-tB2 undergo a proteolytical processing that eliminates their ankyrin motifs and generates proteins p50 and p52, respectively. These mature p50 and p52 are able to homoor heterodimerize with other members of the Rel/NF-tB family, thus becoming active transcription factors (reviewed in Miyamoto and Verma, 1995;Verma et al., 1995).
In most cell types, Rel/NF-tB dimers are found in an inactive form sequestered in the cytoplasm by an ankyrin repeat protein of the ItB family. The IKB family is composed of ItBo, ItB/3, and Bcl3, as well as NF-tB1 and NF-tB2 precusors in their unprocessed form. The activation of the Rel/NF-tB proteins is regulated by a posttranslational mechanism, which includes the phosphorylation of the ItB inhibitor and its subsequent degradation, thus liberating the Rel/NF-B dimer, which can translocate into the nucleus (reviewed in Verma et al., 1995;Miyamoto and Verma, 1995). The inhibitor degradation occurs within minutes after stimulation, making the Rel/NF-tB transcription factors very efficient when rapid responses are needed. The immune system takes advantage of this property and extensively uses the Rel/NF-tB proteins for regulating cytokine, growth factor, acute phase protein, and immunoreceptor gene expression (review in Bauerle and Henkel, 1994;Kopp and Ghosh, 1995). The Rel/NF-B proteins are also involved in pathogenesis of the immune system. In humans, alterations at the rel locus have been found in follicular and diffuse large cell lymphomas (Lu et al., 1991) and amplifications of the c-rel gene have been detected in primary mediastinal thymic Bcell lymphoma (Joos et al., 1996) and in extranodal diffuse large cell lymphoma (Houldsworth et al., 1996).
The expression of c-Rel has been widely described in the hematopoietic organs during mouse development. In the fetal liver, the c-Rel protein expression is restricted to hematopoietic precursor cells of the erythroid and B-cell lineage. In the spleen, c-Rel expression is also found in B cells. In the embryonic thymus, c-Rel is essentially detected in medullary epithelial cells and in some B cells, whereas thymocytes do not express any c-Rel protein. In the T-cell lineage, only mature helper and cytotoxic T cells of the lymph nodes express c-Rel . A better understanding of the mouse c-Rel functions in hematopoiesis and immune system was obtained with the establishment of null mice (K6ntgen et al., 1995). In these mice, no developmental defect in any hematopoietic lineage could be evidenced, indicating that c-Rel is not essential for embryonic hematopoiesis. However, c-rel-deficient mice displayed defects in humoral immunity and unresponsiveness of mature B and T cells to most mitogenic stimuli.
In the avian embryo, such exhaustive studies have not yet been performed. Nevertheless previous in vitro analysis suggested a potential involvement of some Rel/NF-tB members in avian hematopoiesis (reviewed in Huguet et al., 1994). Northern blot analysis revealed that c-rel is predominantly expressed in chick hematopoietic organs such as the bursa of Fabricius, the thymus, the spleen, and the bone marrow (Moore and Bose, 1989). Moreover, the avian retrovirus Rev-T, bearing the v-Rel oncogene, causes fatal lymphomatosis in chicken (Sevoian et al., 1964;Olson, 1967). The in vivo preferential target cells for Rev-T are immature and mature lymphoid cells expressing Band T-cell determinants as well as myeloid cells (Lewis et al., 1981;Barth et al., 1990;Zhang et al., 1991). In vitro studies also showed that the majority of v-Rel transformed hematopoietic cells bore lymphoid markers (Morrison et al., 1991). More recently, a v-RelER fusion protein has been shown to transform an early progenitor of dendritic and neutrophil cells (Boehmelt et al., 1995).
Altogether, these results prompted us to further investigate the relationships between c-Rel and hematopoiesis in chicken. We focused our study on the embryonic thymus that constitutes the primary site of T-cell development. Since the activity of the avian c-Rel is tightly dependent on interactions with the IBc inhibitor (Morrison et al., 1989;Davis et al., 1990;Kerr et al., 1991;Kochel et al., 1991;Hrdlickova et al., 1995;Schatzle et al., 1995), we studied in parallel the expression of both genes and searched for the nucleocytoplasmic localization of the c-Rel protein, which gives evidence about its activation.

RESULTS
High Expression of c-rel in the Thymus Correlates with the Differentiation of the Medulla The analysis of c-rel expression was performed at three stages representative of the avian thymus development: (1) at E9.5, the thymus is an epitheliomesenchymal rudiment filled up with very immature precursor cells; (2) at El5, small medullary areas are separating from the cortex, figuring the maturation of the thymic stroma; (3) at El9, the avian thymic structure is almost completed with well-defined cortical lobules and large medullary areas. The cortex is essentially composed of double-positive CD4+CD8 immature thymocytes enclosed in an epithelial meshwork, whereas the medulla is mainly composed of single-positive mature thymocytes among antigen-presenting cells and epithelial cells.
In both El5 and El9 stages, high magnifications show that inside the medulla the areas of highest expression correspond to cells with large cytoplasm and nucleus that could be clusters of dendritic or epithelial cells and macrophages, according to their morphology and localization. In contrast, cortical and medullary thymocytes, recognizable by a small round nucleus and a low cytoplasmic volume, express lower levels of c-rel mRNAs ( Figure 2). Thus, during the thymus development, the high expression of c-rel correlates with the establishment of mature medullary areas, whereas the expression in developing thymocytes is continuous and at low levels.
To investigate the c-rel expression at the protein level, Western blot experiments were performed with an immuno-purified anti-c-Rel antibody on hematopoietic organs at different developmental stages. One major band at 68 kD, corresponding to the avian c-Rel protein molecular mass (Simek and Rice, 1988), is detected in the thymus at El5 and El9, although less intensely than in the spleen and the bursa of Fabricius (Figure 3). Immunocytochemistry experiments were performed with this same anti-c-Rel antibody on serial thymus sections to identify the c-Rel-expressing cells. At El5 and El9, consistent with its mRNA distribution, the c-Rel protein is predominantly expressed inside the medulla in large clusters of cells ( Figure 4). Some isolated cells scattered in the cortex also express c-Rel. Cortical as well as medullary thymocytes, identified with anti-CD3, -CD4, and -CD8 antibodies (data not shown), do not express any detectable amounts of c-Rel. At least four different types of cells were found to express c-Rel. These cells displayed the violet color characteristic of myeloid cells when stained by the panoptic method of Pappenheim (data not shown). According to their morphology, localization and staining properties, they could be dendritic cells. At the corticomedullary junction, inside the medulla and scattered inside the cortex, isolated large round cells also express c-Rel in their cytoplasm [ Figure 5(E)]. These cells possess a characteristic eccentric nucleus that is clearly negative. We assume that these cells are macrophages because of their morphology, pink staining by the described in a mature mouse B-cell line and was explained by a high and continuous degradation of It<Bc in the absence of external signals (Miyamoto et al., 1994). Here, we show that avian LPCs express low levels of ItBo, suggesting that the constitutive nuclear localization of c-Rel might be due to a low expression of the inhibitor. This hypothesis requires further investigation of It<Bo expression at the protein level.
The involvement of c-Rel in proliferation has already been suggested by in vitro experiments showing that overexpression of c-Rel in chicken embryo fibroblasts led to morphological transformation, life-span extension, and proliferation (Abbadie et al., 1993;Kralova et al., 1994). The constitutive nuclear localization of c-Rel in LPCs documented in this paper represents the first in vivo argument in favor of a participation of this transcription factor in the proliferation of very immature progenitor of T cells.

Antigen-Presenting Cells
In avian thymuses ,from El5 to El9, c-Rel is expressed in epithelial cells, dendritic cells, and macrophages. Avian thymic epithelial cells, dendritic cells, and macrophages share common surface antigens and particularly MHC molecules (Peck et al., 1982;Guillemot et al., 1984;Boyd et al., 1992); they constitute the three main types of antigen-presenting cells (APCs) of the thymus. These cells govern positive and negative selection of T cells in the thymus, via interactions between their MHC and the T-cell receptor (TCR) (reviewed in von Boehmer, 1994;Nossal, 1994).
In vitro studies have shown that avian MHC class I and II genes are induced when c-Rel or v-Rel are overexpressed (Hrdlickova et al., 1994). Therefore, a potential function of c-Rel in APCs could be to regulate the expression of MHC genes. However, in macrophages, besides genes involved in T-cell maturation, c-Rel could regulate genes involved in basic antimicrobial functions, since in HDll cells, a chicken myelomonocytic cell line, RelA, and c-Rel are able to activate the transcription of the lysozyme gene (Phi Van, 1996).
In vitro experiments have demonstrated that in chicken, ItBc is able to inactivate c-Rel by sequestering it in the cytoplasm (Davis et al., 1990;Kerr et al., 1991). In accordance with these studies, our in situ hybridization analysis revealed that in vivo, high expression of ikba is detected in all of the stromal cells that express the c-Rel protein in the cytoplasm, that is, in endothelial cells, epithelial cells, dendritic cells, and macrophages.
The expression pattern of c-Rel is more widespread in chick than in mouse APCs, where c-Rel appears restricted to medullary epithelial cells. Moreover, in mouse, the expression of c-Rel in progenitor cells is restricted to the erythroid and B-cell lineages . Our results on ikba are also different from those obtained in mouse. In the mouse thymus, ikba expression is higher in the cortex than in the medulla .
In conclusion, both avian and murine c-Rel proteins might be indirectly involved in T-cell maturation via their expression in antigen-presenting cells, but this expression seems restricted to epithelial cells in mouse, whereas it also extends to dendritic cells and macrophages in chicken. In contrast, the sole avian c-Rel protein might be also involved in early lymphopoiesis via its expression in proliferating lymphocyte precursor cells.

Animals
Fertilized White Leghorn chicken eggs were incubated at 39C in a humidified chamber. The age of embryos is indicated as El, E2 E1 corresponding to 24 hr of incubation.

Histological Sections
Chick embryo thymuses at different developmental stages were dissected in PBS and fixed at 4C for 18 hr in 4% paraformaldehyde in PBS (NazHPO4-NaHzPO4, 0.1 M, pH 7.4), then dehydrated in ethanol and toluene, embedded in paraffin, and processed for 5-6-#m histological sections.
fragments of approximately 150 bases as recommended by (Cox et al., 1984).

Isolation and Primary Culture of Macrophages and Dendritic Cells
The protocol was adapted from Oliver and Le . El9 thymuses were dissected, minced with scissors, and crushed through a fine nylon gauze, which retains connective tissue and epithelial cells. The collected cell suspension was rinsed in culture medium (RPMI 1640, 10% FCS, 1000 IU/ml penicillin, 1000 #g/ml streptomycin) and resuspended in 10% BSA (Sigma) in RPMI 1640 buffered with Hepes pH 7.3. A discontinuous gradient of 22%, 27%, and 29% BSA in RPMI 1640, pH 7.3, was prepared in a 5-ml polyallomer tube and the cell suspension was layered on top of the gradient. After centrifugation at 35,000 g for 30 min, the fraction located above the 22% layer was collected. Cells enriched in macrophages and dendritic cells were rinsed in culture medium and plated on glass coverslips for 4 hr at 37C in 5% CO2 humidified incubator. Contaminating thymocytes were eliminated by hard pipettilg. Adherent cells were fixed in 4% paraformaldehyde in PBS overnight at 4C and processed for in situ hybridization or immunocytochemistry.

In Situ Hybridization
Synthesis of 3sS RNA Probes The chicken full-length c-rel c-DNA was cloned in Bluescript SK-(Stratagene) (Abbadie et al., 1993). A 0.9-kb EcoRI fragment from the chicken ikba cDNA, kindly provided by H. Bose, was cloned in Bluescript SKphagemid (Stratagene). 35S RNA probes were transcribed from 2 #g of linearized plasmids by 20 U of either T7, T3, or SP6 RNA polymerase for sense and antisense probes in a 20-/xl reaction mixture containing 200 /zCi 35S-CTP (1300 Ci/mMol), 200 /zM UTP, ATP, and GTP for hr at 39C. To facilitate their penetration into cells, probes were submitted to a limited alkaline hydrolysis generating In Situ Hybridization In situ hybridization was adapted from the method of Cox et al., (1984), as described by Qudva et al. (1992). Briefly, after being deparaffinized and rehydrated, sections were incubated in 0.1 M glycine, 0.2 M Tris-HC1, pH 7.4, for 10 min at 20C, treated with /zg/ml proteinase K (Boehringer Mannheim) for 15 min at 37C, postfixed in 4% paraformaldehyde, washed in PBS, acetylated 10 min with 0.25% acetic anhydride in 0.1 M triethanolamine, washed in 2X SSC (0.3 M NaC1, 30 mM sodium citrate), and dehydrated in ethanol. 3S RNA probes were denatured at 80C and diluted in the hybridization buffer at a concentration of 50 pg/#l. Hybridization was performed at 60C for 18 hr. After a wash in 4X SSC at 20C, slides were treated with 10 #g/ml of RNase A (type III A, Sigma) for 30 min at 37C, subsequently washed in 0.1X SSC at 60C, dehydrated by ethanol, and dipped in nuclear track emulsion (Kodak NTB2). The sections were exposed at 4C for 2 weeks. After developing, sections were stained with a DNA intercalating fluorescent dye (Hoechst 33258), and mounted and observed under dark-field and UV illumination. Both antisense and sense probes were used, and the sense probes never gave any signal.

Immunocytochemistry and Western Blot
Antibodies A rabbit immuno-purified anti-c-Rel serum (SB 146) raised against the 15 carboxy terminal amino acids specific of the chicken c-Rel protein (Abbadie et al., 1993) was used both in Western blot and immunocy.tochemistry experiments. A normal rabbit immunoglobulin fraction (Dako) was used as negative control. A rabbit polyclonal anti-keratine serum (Dako) was used in immunocytochemistry to identify epithelial cells. Mouse monoclonal anti-chicken-CD3,-CD4, and -CD8 antibodies (Southern Biotechnology Associates) were used in immunocytochemistry to identify the different types of thymocytes.

Immunocytochemistry
After being deparaffinized and rehydrated, sections were incubated in 0.5 mg/ml Saponine (Sigma) in PBS for 30 min; then endogenous peroxidase activity was quenched by incubation in 80% methanol, 20% PBS, 0.6% H202 for 30 min followed by 30 min saturation in 5% dry defatted milk in PBS. Sections were incubated with the primary antibody overnight at 4C. After three washes in PBS, the EXTRA-3 kit (Sigma Immunochemicals) and the AEC substrate system (Dako) were used to detect the primary antibody. Sections were counterstained with hematoxylin (Sigma) and mounted in Glycergel (Dako).

Western Blot of Hematopoietic Organs
Hematopoietic organs at various developmental stages were dissected in PBS, transferred into a solution containing 0.06 M tris-HC1, pH6.8, 5% /3mercaptoethanol, 2%, SDS, 10% glycerol, 0.002% bromophenol blue, homogenized in a Dounce, sonicated and centrifuged at 10,000 g. In order to measure protein concentration in each sample, 20-/xl aliquots were separated on a 10% SDS-PAGE; proteins were then stained with Coomassie-blue and the gel scanned lane by lane. Equicharged 10% SDS-PAGE were run and transferred onto nitrocellulose sheets. Blots were first incubated 30 min with 5% dry defatted milk in PBS, then 2 hr at room temperature with SB 146. After three washes in PBS, the sheets were incubated with horseradish-peroxidase-conjugated swine anti-rabbit immunoglobulin (Dako) for 2 hr at room temperature. After washing, the peroxidase activity was revealed by incubation in diaminobenzidine (30 mg in 100 ml 0.05 M tris-HC1, pH 7.4, 0.01% H202).