INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

B-cell lymphopoiesis occurs in the bone marrow of adult mammals and involves both secreted factors and cell-cell interactions (12, 13, 20). A variety of tissue culture methods have been used to study the molecular requirements for B-cell development (11, 39, 49). These methods have shown that interleukin-7 (IL-7) is required for in vitro pre-B-cell growth (28, 30), although other secreted factors that reduce or eliminate IL-7 requirements have been recently described (29, 36). These methods have also demonstrated the importance of cell-cell interactions between B-cell precursors and stromal cells that cannot be replaced by soluble factors (11, 49, 50). The molecular basis of this stromal-cell-pre-B-cell interaction is not well characterized. Several cellular or extracellular matrix proteins are involved in these interactions, including Pgp-1/CD44 (26), VLA-4/CD49d, VLA-5/CD49e (24), and VCAM-1/CD106 (25). Despite these recent advances, a complete understanding of the factors and mechanisms regulating B lymphopoiesis is lacking (33).

Long-term bone marrow cultures have facilitated the study of the biological properties of stromal cells, including the observation that stromal cells could undergo differentiation toward the adipocyte or osteoblast phenotypes (16). Adipocytes are the prevalent stromal cell type in adult bone marrow, and they have been shown to play an important role in the hematopoietic environment (14). For example, the ability of bone marrow adipocytes to support lymphopoiesis or myelopoiesis was different than that of their nondifferentiated precursors (16). The pattern of secreted cytokines also differs between adipocytes and their precursors (31). Although no correlation between the profile of cytokine production and hematopoietic supportive ability of stromal cells appears to exist (47), a positive correlation between the ability to undergo adipocyte differentiation and the ability to support in vitro pre-B-cell growth has been documented repeatedly (10, 11) and has been recently confirmed (14).

These observations suggested a close relationship between adipocyte differentiation and hematopoiesis in the bone marrow. Adipogenesis in the bone marrow stromal cells appears to occur by the same mechanisms, and it is under the control of the same molecules that regulate adipogenesis of other cells (16, 17). Since cell-to-cell interactions are necessary for both in vitro adipogenesis (9, 23) and lymphopoiesis, we hypothesized that membrane molecules involved in one of these processes could influence or modulate the other. One of the molecules involved in the cell contact interactions controlling adipocyte differentiation is dlk. dlk belongs to the epidermal growth factor (EGF)-like homeotic family and was named due to its homology with the Drosophila neurogenic protein Delta (dlk = Delta-like). Subsequent to its initial characterization by our laboratory (21), dlk was shown to be involved in several differentiation processes, including adipogenesis (44, 45) and fetal liver hematopoiesis (27). Various forms of dlk have been isolated (22), including Pref-1 (preadipocyte factor 1) (45), FA-1 (fetal antigen 1) (19), and SCP-1 (stromal cell protein 1; Genbank accession no. D16847). The analysis of all of these variants indicates that dlk is a transmembrane molecule that contains six cysteine-rich EGF repeats in the extracellular region, a single transmembrane domain, and a short intracellular tail. Downregulation of dlk expression is complete in differentiated adipocytes, and its overexpression has been shown to inhibit adipocyte differentiation of 3T3-L1 preadipocytes (44). Inhibition of adipogenesis can be obtained either by transmembrane dlk or by a soluble dlk molecule containing the six EGF repeats (43), suggesting that dlk may function as a cell-cell contact or paracrine molecule. It has been suggested that alternately spliced dlk species regulate these two functions.

It was recently reported that dlk participates in cell-to-cell interactions between fetal liver stromal cells and hematopoietic precursors (27). This molecule, either added in soluble form or expressed by transfection of the stromal cells, promoted "cobblestone area" colony formation in Dexter-type stromal cocultures. dlk appears, therefore, to differ from its homologs, Delta and Serrate, in that the latter molecules are not released to the extracellular medium. This property may increase the range of action of dlk. Alternately, released, soluble dlk may be a regulator of the cell-to-cell interactions in which transmembrane dlk may participate.

In this study, we used an in vitro system to explore whether dlk could affect the adipogenesis of stromal cells and whether this effect could modulate B lymphopoiesis. We found that constitutive downregulation of dlk in BALB/c 3T3 cells increases their adipocyte differentiation in response to insulin. BALB/c 3T3 are cells of mesenchymal origin with a differentiation potential similar to that of bone marrow stromal cells (4, 35). We therefore used these cells as stroma for in vitro growth of pre-B cells. In this system, normal pre-B cells can be maintained indefinitely in culture in the presence of exogenous IL-7 and suitable stromal cells (38, 39). Removal of either IL-7 or stromal cells causes pre-B cells to die from apoptosis or to differentiate to surface immunoglobulin-positive B cells which subsequently also die from apoptosis (38, 40). We found that when cells with downregulated dlk are used as stroma, IL-7 is no longer required to support the growth of pre-B-cell lines and removal of IL-7 does not trigger apoptosis or differentiation to mature cells. Our experiments show that this effect is not likely due to the release of a soluble factor, either by the stromal or the pre-B cells, that would compensate for the lack of IL-7 in the medium. These results suggest that dlk participates in the cell-to-cell interactions that occur in the hematopoietic environment of the bone marrow and regulates differentiation signals directed both to the stromal cells and to B-lymphocyte precursors.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Establishment of transfectant cell lines. BALB/c 3T3 cells, clone A31 (ATCC CCL-163), or S10 stromal cells (8) were transfected with control plasmid pCD2 (a generous gift from J. Battey) containing no insert or the same plasmid containing full-length dlk cDNA either in the sense or antisense orientation under the control of the cytomegalovirus promoter. Transfectants were selected by G418 (Gibco-BRL, Bethesda, Md.) treatment, and resistant clones were pooled to give the sense and antisense dlk cell lines used for these experiments. Several individual clones from antisense dlk were selected and called Tr1, Tr2, and Tr3. Since Tr3 expressed the least cell surface dlk, it was used in the experiments presented here.

Adipocyte differentiation studies. Differentiation of BALB/c 3T3 cells was achieved by treatment with 1 µM insulin for 7 to 10 days. At the end of this period, cells were stained with Oil-Red O to detect lipid accumulation indicative of adipocyte differentiation. The extent of differentiation was estimated by counting adipocytes among nondifferentiated cells in three randomly selected regions of a plate in a microscope field of 3 mm². The total number of adipocytes and undifferentiated cells counted was greater than 1,000.

Derivation and maintenance of normal pre-B-cell lines. Pre-B-cell cultures were initiated by using fetal livers from BALB/c or DBA/2 mice at day 13 to 15 of embryonic development as described previously (38). Livers were sterilely resected, a single cell suspension was made in phosphate-buffered saline (PBS), and nucleated cells were counted. The cells were then centrifuged for 10 min at 1,000 × g. The cell pellet was resuspended in Iscove's medium (Gibco-BRL) containing 2% fetal calf serum (FCS), 5 × 10⁵ M 2-mercaptoethanol, penicillin-streptomycin, and 10% of a conditioned medium from IL-7-producing cells (see below) containing 2,000 U of mouse IL-7 per ml (complete Iscove's medium). Serial dilutions containing from 20 to 6,000 liver cells in 100 µl were plated on a 96-well culture dish containing 10⁴ irradiated (1,200 rads) adherent S10 stromal cells (8). Cells were incubated for 5 to 7 days at 37°C until colonies of round, lymphoid cells appeared. Tissue culture plates containing colonies in fewer than 33% of the wells were examined for wells containing single colonies. These single colonies were then propagated in larger tissue culture flasks to establish cell lines from each strain. Cell lines from three separate BALB/c fetal livers and two separate DBA/2 fetal livers were established. Based on analysis of cell surface markers and expression of 5 and VpreB mRNAs, the phenotypes of these cells match the phenotype of normal pre-B cells described previously (38).

Pre-B-cell colony assay. Pre-B cells were serially diluted and added to duplicate wells of 24-well tissue culture dishes previously seeded with 2 × 10⁴ irradiated stromal cells. Pre-B cells were plated on the different stromal cells in medium with or without IL-7. Medium with IL-7 consisted of complete Iscove's medium as described above. Medium without IL-7 was similar except that it contained 4% FCS to match the concentration of FCS in the IL-7-containing conditioned medium. Some experiments included 1:1 mixtures of fresh Iscove's medium-4% FCS and Iscove's medium-4% FCS medium conditioned by 72 h of incubation over the various stromal cell lines. Conditioned media without IL-7 were generated from Tr3 cells, irradiated Tr3 cells, and a culture of irradiated Tr3 cells with D-1-3 pre-B cells. After 5 days, pre-B-cell colonies were counted in each well. Pre-B-cell colonies appeared as clusters of round, uniform lymphoid cells growing on the surface of the adherent fibroblastoid stromal cell layer. It was no longer possible to distinguish individual colonies when the number of colonies in a well exceeded 40.

Pre-B-cell growth curves. Fifty thousand pre-B cells were added to wells of six-well tissue culture dishes previously seeded with 5 × 10⁴ irradiated stromal cells. Pre-B cells were plated on the different stromal cells in medium with or without IL-7 as described above. Viable pre-B cells from duplicate wells were harvested and counted by trypan blue exclusion after various intervals. The total number of viable cells was plotted against number of days in culture to assess the growth under different culture conditions.

Cell surface marker analysis. Aliquots of 10⁶ cells were incubated for 30 min at 4°C with FcR blocking antibody 2.4G2 (Pharmingen) and the following fluorescein isothiocyanate (FITC)- or phycoerythrin (PE)-labeled antibodies: anti-BP-1-FITC (Pharmingen), anti-Mac-1-PE, anti-major histocompatibility complex class II, anti-CD43-PE, anti-CD44-FITC, anti-gamma interferon (IFN-) receptor-FITC, anti-CD45RB-FITC, anti-CD45 (B220)-PE, anti-ThB-FITC, and anti-immunoglobulin M (IgM)-FITC. Staining for Notch protein was performed by incubation with an anti-mouse Notch-1 rabbit antisera developed as previously described (15), followed by incubation with anti-rabbit IgG-FITC.

Western blotting. Western blotting was performed by standard methodology. Cell extracts were obtained, and their protein concentrations were evaluated by use of the bicinchoninic acid protein assay kit (Pierce, Rockford, Ill.). Thirty micrograms of soluble cell protein extract was run on a commercial 10% polyacrylamide gel (Novex) and blotted onto nitrocellulose filters. The filters were incubated with an affinity-purified rabbit anti-dlk antiserum (generously provided by Bronek Pytowski) directed against the entire extracellular part of the dlk molecule. After the filters were washed, 10 µCi of ¹²⁵I-protein A was added to each and the filters were incubated for 1 h. Following extensive washing, the filters were analyzed by autoradiography.

Analysis of apoptosis. The kinetics of apoptosis in pre-B cells was studied by staining the cells with Annexin V-FITC (Pharmingen). Apoptosis was induced by transferring the cells to several conditioned or normal media without IL-7 in the presence of different types of stromal cells. Annexin V-positive cells were analyzed by flow cytometry as explained above.

Gene expression analysis. Total RNA was isolated from cells as described before (5). cDNAs were made from 1 µg of total RNA as described previously (41). Following synthesis, the cDNA was diluted to a total volume of 100 µl. PCR analysis utilized 1 µl of cDNA in reaction mixtures containing 100 µM nucleotide triphosphates, 0.25 mM oligonucleotide primers, 0.04 U of Taq polymerase per µl, and 1× PCR buffer (Perkin-Elmer Cetus Co.). Thirty PCR cycles of the following steps were done: denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s. We used PCR primers specific for 5, -actin, recombinase-activating gene 1 (RAG-1), RAG-2, VpreB, IL-7, pre-B-cell stimulatory factor (PBSF), and dlk. PCR products were analyzed by electrophoresis through 2% agarose gels in Tris-acetate-EDTA buffer. The sizes of the PCR products were determined by comparison to a 123-bp ladder or molecular size standards (Gibco-BRL). Following electrophoresis, PCR products were detected by using either 0.5 µg of ethidium bromide per ml or SYBER Green II (Molecular Probes) and scanning with a FluorImager (Molecular Dynamics).

Analysis of IL-7 production. Supernatants from stromal cell lines were tested for the presence of IL-7 by ELISA using microtiter wells coated with anti-IL-7 monoclonal antibody (Genzyme) and secondary biotin-labeled polyclonal goat anti-mouse IL-7 (Research and Diagnostic Systems, Inc.). Streptavidin-horseradish peroxidase was then added to the wells, and the wells were incubated and washed. Paranitrophenyl phosphate substrate was added to the wells, and the wells were incubated until a yellow color developed in positive control wells. The reaction was stopped with 0.3 M NaOH, and then absorbance at 405 nm was determined. A standard dilution series of mouse recombinant IL-7 (Immunex) was used to determine the concentration of IL-7 in supernatants. The limit of detection in this assay was 10 pg.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Transfection of BALB/c 3T3 cells with antisense dlk constructs increases their adipogenic potential. BALB/c 3T3 cells were established from BALB/c mouse embryos (1) and are considered to be a model of normal fibroblasts. Subsequent studies indicated, however, that these cells are multipotent mesenchymal cells capable of differentiating into a variety of cell types, including chondrocytes, myocytes, and adipocytes (4). A similar multipotent differentiation pattern is displayed by many stromal cell-derived clones (35). In contrast to 3T3-L1 cells, which readily downregulate dlk during adipogenesis and show a high potential for adipocyte differentiation, only a small percent of BALB/c 3T3 cells undergo adipocyte differentiation when maintained under confluence for several days or when treated with differentiating agents (Fig. 1A). Northern blot analysis of the expression of dlk in BALB/c 3T3 A31 cells upon treatment with differentiating agents fails to detect decreased dlk levels (data not shown), consistent with the fact that the majority of cells do not differentiate. Since transfection with dlk expression constructs inhibits adipocyte differentiation of 3T3-L1 cells, we explored whether enforced downregulation of dlk expression, by means of antisense dlk mRNA expression constructs, could modify the adipogenic response to insulin of BALB/c 3T3 cells. We found that antisense-dlk-transfected BALB/c 3T3 cells showed a dramatically increased differentiation response upon treatment with insulin (Fig. 1A). Compared to less than 0.1% of control cells differentiating to adipocytes, 5 to 10% of the antisense dlk cells underwent differentiation. Sense-dlk- or mock-transfected BALB/c 3T3 cells showed differentiation responses to insulin similar to those of parental cells (data not shown).

View larger version (95K): [in this window] [in a new window]   FIG. 1.   Effects of antisense dlk expression on BALB/c 3T3 cells. BALB/c 3T3 cells were transfected with a vector expressing full-length dlk cDNA in antisense orientation. (A) Untransfected (control) cells or pooled antisense transfectants (AS dlk) were treated with 1 µM insulin for 7 to 10 days and then assessed for adipocyte differentiation by Oil-Red O staining as described in Materials and Methods. Dark areas indicate lipid accumulation. (B) Cell surface expression of dlk was examined by flow cytometry by using dlk-specific rabbit antiserum as described in Materials and Methods. Mean fluorescence intensity was plotted against cell numbers of untransfected BALB/c 3T3 controls, pooled antisense dlk transfectants (AS dlk), and three individual antisense transfectant clones (Tr1, Tr2, and Tr3). (C) Western blot analysis of dlk expression in cellular extracts from BALB/c 3T3 cells transfected with antisense dlk expression constructs. Lanes: M, molecular size markers; E, Escherichia coli glutathione S-transferase-dlk fusion protein (73-kDa molecular mass); B, nontransfected BALB/c 3T3 cells; T, clone Tr3; A, pool of antisense-dlk-transfected BALB/c 3T3 cells. The multiple dlk bands result from alternate splicing or differences in glycosylation.

Diminished stromal-cell dlk expression modulates pre-B-cell response to IL-7 deprivation. Previously published data showed that preadipocyte cell lines, and stromal cells from fetal liver or thymus, can support pre-B-cell growth (38). Consistent with these observations, when used as stromal cells in our in vitro pre-B-cell growth system, BALB/c 3T3 cells supported pre-B-cell growth in the presence, but not in the absence, of exogenous IL-7. It was recently reported that the ability of stromal cells to support hematopoiesis in vitro correlates with their ability to undergo adipogenesis (14). Therefore, we tested whether the ability to support the in vitro growth of pre-B cells was modified in our antisense dlk transfectants. As a preliminary experiment, we utilized control or antisense-dlk-transfected BALB/c 3T3 cells in a pre-B-cell colony-forming assay (see Materials and Methods). Despite different levels of dlk cell surface expression, no differences in colony formation were observed when IL-7 was present (Fig. 2A). As expected, no colonies formed on the control BALB/c 3T3 cells when IL-7 was omitted. Surprisingly, however, pre-B-cell colonies also formed on the antisense-dlk-transfected cells seeded in the absence of IL-7. A representative experiment is shown in Fig. 2A. The colony assay, using the pre-B-cell line D-1-3, was repeated four times with similar results. In two further experiments, three additional pre-B-cell lines also displayed similar colony formation patterns. In total, four pre-B-cell lines, two of BALB/c and two of DBA/2 origin, were able to grow in the absence of IL-7 on stromal cells with diminished dlk expression.

View larger version (31K): [in this window] [in a new window]   FIG. 2.   Effects of diminished stromal cell dlk expression on pre-B-cell growth. (A) Pre-B-cell colony assay. D-1-3 pre-B cells were seeded into 24-well dishes on irradiated BALB/c 3T3 (control) or dlk antisense transfectants (AS dlk) in the presence (left panel) or absence (right panel) of IL-7. Pre-B cells were seeded in 10-fold dilutions (x axis), and the number of colonies that formed at each pre-B-cell dilution (y axis) was plotted. The greatest number of individually discernible colonies was 40, so bars reaching 40+ indicate wells that were virtually confluent. Bar height represents the mean of duplicate wells. (B) Pre-B-cell growth analysis. Pre-B-cell line D-1-3 was seeded into wells of six-well culture dishes with irradiated stromal cell layers in the absence of IL-7. The mean number of viable pre-B cells in duplicate wells harvested on the indicated days is shown for nontransfected BALB/c 3T3 cells or for BALB/c 3T3 cell lines made by transfection with control plasmid containing no insert or the same plasmid containing full-length dlk cDNA either in the sense (Sense dlk) or antisense (AS dlk) orientation. The control, sense, and antisense lines were pooled following selection. Tr3 is a cloned derivative of the AS dlk line.

View larger version (31K): [in this window] [in a new window]   FIG. 3.   Flow cytometry analysis of transfected BALB/c 3T3 cells. (A) Cells transfected with sense or antisense dlk expression constructs. Expression levels of CD44 and IFN- receptor (IFN-) are compared. (B) Cells transfected with antisense Notch constructs. Expression levels of Notch, CD44, and CD45RB are indicated. The thicker lines represent specific staining of the corresponding markers, whereas dotted or thinner lines represent the unstained controls.

Downregulation of dlk of S10 stromal cells also affects IL-7 requirements. Although we found that BALB/c 3T3 could supply stromal cell support for pre-B cells, we studied whether modulating the expression of dlk on the membrane of S10 cells, widely used as pre-B-supporting stromal cells in vitro, could also have an effect on the IL-7 requirements of pre-B cells growing in contact with them. S10 cells express dlk on the membrane, and its expression can be substantially decreased by transfection with antisense dlk expression constructs, leaving unaffected the expression levels of CD44 (Fig. 4A), suggesting that the effect of the antisense dlk transfection remains limited to dlk expression. When transfected S10 cells were used as stroma in a cell colony assay, no differences in pre-B-cell support were observed in the presence of IL-7. Interestingly, however, only antisense-dlk-transfected cells allowed the growth of pre-B-cell colonies in the absence of the cytokine (Fig. 4B). These results confirm that modulation of dlk expression on stromal cells influences pre-B-cell IL-7 requirements.

View larger version (36K): [in this window] [in a new window]   FIG. 4.   Effects of enforced downregulation of dlk expression on S10 stromal cells on pre-B-cell growth. (A) S10 cells stably transfected with sense or antisense dlk expression constructs were analyzed for their levels of dlk and CD44 expression. The thicker lines represent specific staining of the corresponding markers, whereas the thinner lines represent the unstained controls. (B) Transfected S10 cells were also used as stroma in cell colony assays using the D-1-3 pre-B-cell line. The bar graph represents the number of colonies that develop in relation to the number of pre-B cells seeded per well. The key shows what construct was transfected into S10 cells.

Pre-B cells grown over antisense-dlk-transfected cells in the absence of IL-7 retain a normal phenotype. Since removal of IL-7 from pre-B cells results in either differentiation or apoptotic death (39), the observations described above raised the possibility that pre-B cells grown without IL-7 may have differentiated. Therefore, analyses to assess this possibility were performed. Three lines of evidence demonstrate that pre-B cells growing on antisense-dlk-transfected BALB/c 3T3 cells in the absence of IL-7 did not display any of the phenotypical modifications that occur during B-cell maturation. First, analysis of several cell surface phenotypic markers, including B220, CD43, BP-1, ThB, major histocompatibility complex class II, and surface IgM, showed no difference between pre-B cells grown under normal conditions and those grown over antisense-dlk-transfected BALB/c 3T3 cells in the absence of IL-7 (Fig. 5A). Second, genes whose expression changes during B-cell maturation, such as RAG-1, RAG-2, VpreB, and 5, showed no changes in their expression levels when these cells were cultured with antisense-dlk-transfected BALB/c 3T3 cells in either the presence or the absence of IL-7 (Fig. 5B). Finally, pre-B cells maintained in culture over antisense dlk cells in the absence of IL-7 remained responsive to IL-7 when restored to normal culture conditions, as evidenced by a normal growth response and growth kinetics similar to those of cells grown continuously in the presence of IL-7 (data not shown). This suggests that the ability of the pre-B cells to grow without IL-7 in the presence of antisense dlk BALB/c 3T3 cells is not associated with a lack of response to this cytokine and argues against adaptation to new culture conditions as an explanation for the observations described here. To date, pre-B cells have been maintained in culture on antisense dlk cells in the absence of IL-7 for over 6 months.

View larger version (36K): [in this window] [in a new window]   FIG. 5.   Effects of diminished stromal cell dlk expression on pre-B-cell differentiation. (A) Cell surface marker analysis using the indicated antibodies was done on D-1-3 pre-B cells propagated either under normal culture conditions with S10 stromal cells and IL-7 or without IL-7 on low-dlk-expressing antisense transfectants (as-dlk or Tr3). Each dot plot panel displays fluorescence intensity for PE-labeled antibodies (FL-2; y axis) plotted against fluorescence for FITC-labeled antibodies (FL-1; x axis). (B) Gene expression analysis was done by using reverse transcription-PCR on RNAs isolated from D-1-3 pre-B cells propagated either under normal culture conditions with S10 stromal cells and IL-7 (indicated by +) or without IL-7 on low-dlk-expressing antisense transfectant Tr3 cells (indicated by ). The genes studied are indicated at the top of the panel. A 123-bp ladder (lane M) was included for determination of PCR product size. Lengths in base pairs (MW) of standards are shown on the left.

Pre-B cells can be isolated from fetal liver in the absence of IL-7 by using antisense-dlk-transfected cells as stroma. The colony assay and growth kinetic experiments were consistent and suggested that diminished dlk expression on stromal cells modulates the pre-B-cell requirement for IL-7. However, these experiments utilized normal pre-B-cell lines established by using S10 stromal cells. Therefore, we attempted to initiate primary pre-B-cell cultures directly from fetal mouse livers. Dilutions of fetal liver cells were seeded in 96-well plates with irradiated normal BALB/c 3T3 cells, antisense dlk transfectant Tr3 cells, or the pooled antisense-dlk-transfected BALB/c 3T3 cells by using medium without IL-7. Colonies were found exclusively on low-dlk-expressing BALB/c 3T3 cells transfected with antisense dlk. Five of these colonies were further expanded by using Tr3 cells as stroma in the presence of IL-7, and three were expanded by using the pooled antisense-dlk-transfected BALB/c 3T3 cells with IL-7. In addition, three colonies were expanded over Tr3, and another three colonies were expanded over pooled antisense-dlk-transfected BALB/c 3T3 cells in the absence of this cytokine. Figure 6 shows representative flow cytometry analysis of several cell surface markers after expansion, confirming that the cells displayed a normal pre-B-cell phenotype. These results show that pre-B-cell lines can be directly isolated from fetal liver cells without IL-7 if cells with diminished levels of dlk expression are used as stroma.

View larger version (28K): [in this window] [in a new window]   FIG. 6.   Comparison of cell markers expressed by pre-B-cell lines freshly isolated from mouse fetal liver cells cultured over Tr3 stromal cells in the presence or absence of IL-7 and markers expressed by the pre-B-cell line D-1-3 cultured over S10 cells in the presence of IL-7. Developing colonies of fetal liver cells were expanded over Tr3 cells in the presence or absence of IL-7. Cell type abbreviations: FL, fetal liver origin; Tr3, stromal cell used; Isc, Iscove's medium-4% FCS; IL-7, IL-7 added to the medium. The expression of cellular markers characteristic of the pre-B phenotype, including B220, ThB, CD43, and BP-1, was analyzed by flow cytometry.

Production of IL-7 or other cytokines is not responsible for the modified lymphopoiesis-supportive abilities of antisense dlk cells. One possibility to explain the growth of pre-B cells in the absence of exogenous IL-7 is that low dlk expression on the stromal cells could induce production of this cytokine in either the stromal cells or the pre-B cells. We found, however, no IL-7 by ELISA in supernatants from BALB/c 3T3 or antisense transfectants after 2 weeks of culture. Also, we found no IL-7 mRNA expression in the normal or transfected BALB/c 3T3 cells. In pre-B cells, we detected a small amount of IL-7 mRNA in 40-cycle PCR experiments. There was no difference between the levels of the message, however, in pre-B cells grown in the presence and those grown in the absence of exogenous IL-7 (data not shown). We also examined by reverse transcription-PCR the expression of PBSF, a recently described pre-B-cell growth factor (29). PBSF was equally expressed in all transfected BALB/c 3T3 cell lines and all pre-B-cell lines whether grown with or without IL-7. This suggests that PBSF expression does not explain the maintenance of pre-B-cell growth on antisense-dlk-transfected cells in the absence of IL-7.

View larger version (33K): [in this window] [in a new window]   FIG. 7.   Effects of several stromal-cell-conditioned media on pre-B-cell growth. (A) Kinetics of apoptosis of D-1-3 pre-B cells cultured over S10 stromal cells in the presence of the indicated conditioned media or in the presence of normal medium with or without IL-7; (B) cell colony growth assay of D-1-3 pre-B cells cultured over S10 stromal cells in the presence of the indicated conditioned media.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Our results suggest that dlk is involved in the cell-cell interactions that modulate the B-lymphopoietic and adipocytic differentiation processes that take place in the bone marrow. A decrease in dlk expression on the membrane of stromal cells increases their adipogenic potential and modifies the requirements of soluble factors, namely, IL-7, for the maintenance of pre-B cells in vitro. Since, in the absence of IL-7, pre-B cells grew faster over the stromal cells expressing the lowest levels of membrane dlk, our data also suggest a dose-dependent relationship between lower levels of stromal cell surface dlk expression and the ability of the stromal cells to support pre-B-cell growth in the absence of IL-7. Decreasing dlk expression levels by the stromal cells allows the pre-B cells to grow and escape the programmed cell death signal that the lack of IL-7 would normally stimulate (39).

Our results demonstrate that pre-B cells growing in the absence of IL-7 on BALB/c 3T3 or S10 cells transfected with antisense dlk are able to expand while maintaining their differentiation state. Cell surface markers, including BP-1 and CD43, associated with an IL-7-dependent phenotype (39), remained similar to those found on normal pre-B cells cultured in the presence of IL-7. Maturation from pre-B- to mature B-cell stages would be accompanied by transcriptional downregulation of the surrogate light chain genes VpreB and 5, as well as loss of RAG gene expression (3, 32). The lack of changes in the expression of these genes was another indication that differentiation was not occurring. The pre-B cells could multiply for over 6 months without exogenous IL-7, ruling out a short-term adaptation to these culture conditions. The ability of the antisense-dlk-transfected cells to support the growth of phenotypically normal pre-B cells from freshly isolated fetal mouse liver cells in the absence of IL-7 suggests that these results apply not only to cell lines adapted to grow in vitro but also to cells growing in vivo. Recent reports describe stromal cell clones capable of maintaining the expansion of human pre-B cells in vitro in the absence of IL-7 (14, 34), but an understanding of the conditions that allow this IL-7-independent growth is lacking. Our results suggest that dlk expression may play a role in this phenomenon and that dlk could have a substantial influence on normal pre-B-cell differentiation requirements.

By which mechanisms could dlk function? In the culture system we used, the majority of pre-B cells removed from either IL-7 or stromal cell contacts die from apoptosis within 3 days, whereas a minority of cells undergo maturation to surface immunoglobulin-positive B cells and die soon thereafter (38, 39). The data presented here argue, therefore, in favor of dlk as playing a role in modulating the apoptotic signals in pre-B cells that are normally inhibited by the presence of IL-7. dlk-dependent interactions between pre-B cells and stromal cells could increase the level of expression of genes protecting pre-B cells from apoptosis. Alternately, changes in the levels of dlk expression in the stromal cells may affect the response of B-cell precursors to growth or apoptotic signals, including the absence of IL-7 in the extracellular medium. The fact that dlk affects both insulin effects on BALB/c 3T3 cells and the requirement for soluble factors for pre-B cells suggests that dlk may alter the interpretation of, or need for, some external signals. Our results invite speculation that dlk could modulate signaling events common to both IL-7 and insulin pathways. Some molecules have been shown to participate in the signaling triggered by many cytokines and also by insulin (6, 46). Insulin and IL-7 signaling have been shown, for instance, to activate MAP kinases (28, 51), and IL-7 signaling also results in phosphorylation of IRS-1 and IRS-2 in human thymocytes (42).

Cell differentiation, however, involves more than the control of apoptotic signals. The homology between dlk and other EGF-like homeotic proteins, such as Notch and its ligands Delta and Serrate, suggests that dlk may function through mechanisms of differentiation control similar to those in which these molecules participate. EGF-like homeotic proteins are involved in cell-to-cell interactions that regulate the choice between two possible differentiation fates through a mechanism called lateral specification (2, 7). In this mechanism, neighboring cells expressing both Notch and Delta send signals to each other through ligand-receptor interactions. Random variations in the expression of the receptor or ligand are amplified in such a way that cells expressing a greater amount of receptor upregulate its expression and downregulate the expression of the ligand, decreasing the signal delivered to neighboring cells. Reciprocally, cells receiving less signal downregulate the receptor and upregulate the ligand, increasing the signal delivered to neighboring cells. A particular pattern of ligand or receptor cells is obtained; this will determine the spacial distribution of differentiated cells in the adult animal. The importance of dosage of the EGF-like genes that participate in a lateral specification mechanism has been extensively documented for both invertebrate and mammalian systems. Stoichiometric relations between Notch and Delta play an important role in the control of ectodermal differentiation in Drosophila. Notch expression levels have been shown to influence cell fate determination between CD4 and CD8 and between / and / during mouse T-cell development, although changes in Notch-1 expression do not seem to affect B-cell differentiation (37, 48), as our own data obtained with antisense Notch-1 transfectants also suggest. Determination of whether dlk participates in mechanisms similar to those described for Notch and its ligands must await characterization of the molecules that interact with it on the membrane of pre-B or other cells.

Although previous reports indicated that dlk expression was restricted to neuroendocrine, preadipose, placental, and fetal liver stromal tissue, our results show that at least two distinct cell lineages that develop in the bone marrow microenvironment are influenced by the cell surface level of dlk expression. We have demonstrated that dlk has a role in cell-cell interactions that take place between stromal cells and B-cell precursors and that control their differentiation. The establishment of possible roles of dlk in the differentiation of other hematopoietic lineages requires continued investigation into the nature of the interactions and signals mediated by dlk and related cell surface molecules.