Effects of Delta1 and Jagged1 on Early Human Hematopoiesis: Correlation with Expression of Notch Signaling-Related Genes in CD34+ Cells



It has been shown that Notch signaling mediated by ligands of both Jagged and Delta families expands the hematopoietic stem cell compartment while blocking or delaying terminal myeloid differentiation. Here we show that Delta1- and Jagged1-expressing stromal cells have distinct effects on the clonogenic and differentiation capacities of human CD34+ CD38+ cells. Jagged1 increases the number of bipotent [colony-forming unit-granulocyte macrophage (CFU-GM) and unipotent progenitors (CFU-granulocytes and CFU-macrophages), without quantitatively affecting terminal cell differentiation, whereas Delta1 reduces the number of CFU-GM and differentiated monocytic cells. Expression analysis of genes coding for Notch receptors, Notch targets, and Notch signaling modulators in supernatant CD34+ cells arising upon contact with Jagged1 and Delta1 shows dynamic and differential gene expression profiles over time. At early time points, modest upregulation of Notch1, Notch3, and Hes1 was observed in Jagged1-CD34+ cells, whereas those in contact with Delta1 strikingly upregulated Notch3 and Hes1. Later, myeloid progenitors with strong clonogenic potential emerging upon contact with Jagged1 upregulated Notch1 and Deltex and downregulated Notch signaling modulators, whereas T/NK progenitors originated by Delta1 strikingly upregulated Notch3 and Deltex and, to a lesser extent, Hes1, Lunatic Fringe, and Numb. Together, the data unravel previously unrecognized expression patterns of Notch signaling-related genes in CD34+ CD38+ cells as they develop in Jagged1- or Delta1-stromal cell environments, which appear to reflect sequential maturational stages of CD34+ cells into distinct cell lineages.


Notch receptors and their ligands are evolutionary conserved trans-membrane proteins that regulate cell-fate decisions in many developmental processes [1]. Recent evidence indicates that different Notch receptors and ligands have differential effects in distinct hematopoietic microenvironments and cell differentiation stages. For example, Notch1 commits the common lymphoid progenitor into the T/NK cell lineage, at the expense of the B-cell lineage [reviewed in ref. 2] and is essential for the generation of definitive hematopoietic stem cells in early mouse embryos [3], whereas Notch2 promotes the development of marginal zone B-cells in the spleen, while not affecting T-cell development in the thymus [4]. Nonredundant actions of different Notch ligands in lymphopoiesis have also been reported. Delta1, but not by Jagged1, completely inhibits the differentiation of human CD34+ into the B-cell lineage, while promoting their development into T/NK precursor cells [5], which are capable of terminal T-cell differentiation in vitro and in vivo [6, 7].

The role of Notch signaling in myeloid differentiation is still controversial. Conditional deletion of Notch1 [8, 9] and RBP-J [10] (a transcription factor that mediates the Notch signal in the nucleus) in bone marrow cells in vivo, did not apparently cause abnormalities in myeloid lineages, nor did the in vivo overexpression of active forms of Notch1 [11, 12]. However, mice transplanted with bone marrow cells transduced with the active intracellular domain of Notch1 show increased terminal differentiation of nontransduced cells, suggesting that Notch1 may affect hematopoiesis in a non-cell-autonomous manner [13].

Most in vitro experiments (but not all [14, 15]) in which activated Notch1 was expressed in cell lines [1619] and/or primary cells [12, 1921] have shown that active Notch1 expands the stem cell compartment but blocks or delays terminal myeloid differentiation. Studies in which Notch signaling has been induced by members of the Jagged family [20, 2226] or Delta family [6, 2729] corroborated these findings. Thus, both Jagged and Delta Notch ligands appear to act as growth- and self-renewal factors for hematopoietic stem cells. Whether such a redundancy of effects also occurs at later stages of hematopoiesis remains to be determined. The restricted and developmentally regulated patterns of expression of Jagged and Delta proteins (and Notch receptors) in bone marrow stromal cells and developing hematopoietic cells strongly suggest that they may have specific effects in myeloid cell development.

To address this possibility, the effects of bone-marrow-derived stromal cells expressing human Delta1 or Jagged1 [5] on cord blood CD34+ CD38+ cells were investigated. We observed that Delta1 and Jagged1 have different effects on myeloid bipotent and unipotent progenitors and differentially regulate the development of granulocytic and monocytic cell lineages. Dynamic changes in the transcriptional activity of genes coding for Notch receptors, Notch targets, and Notch signaling modulators were observed, which differed according to whether CD34+ were cultured in the absence or presence of Jagged1 or Delta1.

Materials and Methods

CD34+ Cells

Umbilical cord blood (CB) was collected according to the guidelines approved by the Ethical Committee of the Lisbon Medical School, mononuclear cells were isolated by density gradient centrifugation (Ficoll-Paque; Plus, Amersham Biosciences, Lisbon, Portugal, http://www.amersham.com), and CD34+ cells were obtained using the VarioMACS separation column (Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com). Expression of CD34+ on collected cells was 96.5% ± 1.5% (n = 10). Cells were immediately processed for coculture experiments. CD34+ cells differentiating upon contact with Delta1 or Jagged1 were purified using the same procedure (purity >97%). Sorting of Delta1-derived CD7 and CD7+ subpopulations was further performed with purity higher than 98%.

Retroviruses and Producer Cell Lines

Full-length cDNAs encoding either the human Delta1 or Jagged1 (provided by G. Artavanis-Tsakonas, Harvard Medical School, Charlestown, MA) were cloned into the LZRS linker internal ribosome entry site (IRES)-enhanced green fluorescent protein (GFP) retrovirus (provided by H. Spits, The Netherlands Cancer Institute, Amsterdam, Holland; Garry Nolan, Stanford University, Stanford, CA), as previously described in detail [5].

Transduction of S17 Cells with Retroviruses Containing Delta1 or Jagged1 c-DNA

S17 bone marrow-derived stromal cells (provided by K. Dorshkind, University of California, Los Angeles, CA) were transduced and functionally assayed as previously described in detail [5] (except for the use of a “nonlymphoid” batch of calf serum, to permit the development of myeloid cells in this system). Transduction efficiency was identical for both retroviral vectors, and Jagged1 and Delta1 proteins were expressed by stromal cells, as assessed by immunocytochemistry and Western blot as previously described in detail [5].

Functional Assays for Jagged1 and Delta1 Proteins

Efficiency of transduced Delta1 and Jagged1 proteins in the activation of the Notch pathway was assessed as described [5]. The intensity of Notch signals triggered by the two ligands in C2 cells was compared by using increasing doses (0.3 μM, 1 μM, and 5 μM) of the γ-secretase inhibitor [N-(3,5-difluorophenylacetyl-l-alanyl)]-S-phenylglycine t-ButylEster (DAPT [Calbiochem, Darmstadt, Germany, http://www.emdbiosciences.com]), as recently described by Lehar et al. [30]. Myotube formation was assessed by phase contrast analysis and expression of Myosin (clone My-32; Sigma-Aldrich, Sintra, Portugal, http://www.sigmaaldrich.com) (immunocytochemical detection).

Coculture Assays

Twenty-four hours before their use in coculture, 2 × 104 S17 cells in 1 ml of medium were plated in 24-well flat-bottomed plates. Cocultures were initiated by seeding 4 × 104 CD34+ cells to the wells precoated with parental or transduced S17 stroma and maintained at 37°C, 5% CO2. Half of the culture medium (RPMI [Sigma-Aldrich], 5% fetal bovine serum [Life Technologies, Barcelona, Spain, http://www.lifetech.com], 50 μM 2-mercaptoethanol [Life Technologies], 2 mM l-glutamine [Life Technologies], 100 U/ml penicillin-streptavidin [Life Technologies]) was replaced by fresh medium once a week. No exogenous cytokines were added to the cultures in any time point. After 48 hours, 1 week, 2 weeks, and 4 weeks of incubation, cells were harvested, counted, stained, and used for subsequent analyses. The expression of Notch signaling-related genes was analyzed in purified CD34+ cells by quantitative reverse transcription polymerase chain reaction (qRT-PCR) at the different time points. CD34+ cells persisting at 4 weeks of culture were further used for clonogenic assays.

Flow Cytometry and Cell Sorting

Flow cytometry analysis was performed using a FACSCalibur (BD Biosciences, Madrid, Spain, http://www.bdbiosciences.com). Fresh CB progenitors or supernatant cells harvested at different time points were stained with the following monoclonal antibodies as described [5]: CD10-FITC, CD15-FITC, CD14-PE, CD19-PE, CD34-PE, CD38-PE, CD3-PerCP, CD10-PE-Cy5, CD34-PerCP-Cy5.5, CD45-APC (all from BD Biosciences), and CD7-PE (Beckman Coulter, Buckinghamshire, U.K., http://www.beckmancoulter.com). To sort CD34+CD7+ and CD34+CD7 cells from Delta1-derived CD34+ cells, cells were stained with PE-conjugated anti-CD7 (BD Biosciences) and sorted using triple laser (488 nm argon laser, 599 nm dye laser, and UV laser) fluorescence-activated cell sorting (FACS DiVa flow cytometer and cell sorter; BD Biosciences).

Methylcellulose Assays

At day 0 (control for clonogenic assays) and 4 weeks of coculture (pool of cells from 2–3 different donors per sample), supernatant-purified CD34+ cells were plated in quadruplicate of two dilutions in methylcellulose medium (Methocult GF+ H4435; Stem Cell Technologies, Vancouver, BC, Canada, http://www.stemcell.com) according to the manufacturer's recommendations. Cultures were maintained at 37°C in 5% CO2, and colonies were evaluated by phase microscopy, stained by May-Grünwald Giemsa, and classified according to morphologic criteria after 14–16 days.

Real-Time qRT-PCR

The expression of selected Notch signaling-related genes was analyzed by qRT-PCR of purified CD34+ cells freshly isolated and after 48 hours, 1 week, 2 weeks, and 4 weeks of contact with control, Jagged1, and Delta1 stromas (further sorted for CD7 and CD7+ subpopulations) (a pool of cells from 2–3 different donors was used for each sample). Three different samples were analyzed in triplicate for each culture condition. All procedures were done as described [5]. Expression of Notch2, Notch3, Manic Fringe, and Numb was analyzed using the Gene Expression Assays-on-Demand (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com) and of Radical Fringe and Lunatic Fringe by the Gene Expression Assays-by-Design (Applied BioSystems). Primers and probes for Hes1 and Deltex have been published [5]. Primers and probe for Notch1 were designed using the Primer Express software (Applied BioSystems): Notch1 forward primer, 5′-TGC CTC TTC GAC GGC TTT-3′, reverse primer, 5′-GGG TAC ATG CTC CGC ACA GT-3′; Notch1 probe, 5′-ACT CGC ACT CCG CGC TGT TGC A-3′ (GenBank accession number AF308602). Expression of each target gene was normalized for the endogenous gene huGAPDH (TaqMan PDAR; Applied BioSystems). PCRs were performed in a TaqMan 7900 machine (Applied BioSystems), according to the manufacturer's instructions. The Universal Human Reference RNA (Stratagene, La Jolla, CA, http://www.stratagene.com) was used as standard to allow relative quantification and expression values of control cells used as calibrators. All tests were performed in triplicate, and the mean cycle threshold (Ct) value was taken as the final result (mean Ct ± SD, if SD < 0.5). Intra-assay variation for each gene was studied through amplification of 10 replicates of the 10−2 standard cDNA dilution. For each gene, the Ct values of the replicates had SD < 0.5.

Statistical Analysis

Statistical analysis was performed using the nonpaired Student's t test.


Stromal Cells Expressing Delta1 or Jagged1 Differentially Affect the Expansion of CD34+ and CD14+ Cells

The experimental design of the present study is depicted in the diagram of Figure 1A. Figure 1B shows that S17 stromal cells transduced with Jagged1 and Delta1 are efficient in inhibiting the fusion of C2 myoblasts into multinucleated myotubes, as previously described [5]. This indicates that exogenous Jagged1 and Delta1 expressed by S17 cells are capable of activating Notch signaling in Notch-expressing myoblasts [31, 32]. Furthermore, myotube formation occurred to similar degrees, for equal doses of DAPT, in the presence of Jagged1 and Delta1 (Fig. 1B, results obtained with the lowest concentration of DAPT), indirectly showing that Jagged1 and Delta1 induce Notch signals of similar intensity in these cells [30].

As CD34+ CD38+ cells derived from cord blood, when cultured with stromal cells, do not generate colony-forming units (CFU) beyond day 40 (reaching the maximum of expansion and clonogenic potential around day 30) [33], coculture experiments were ended at 4 weeks. The phenotypic analysis of supernatant cells obtained by culturing freshly collected CD34+ cells with parental (nontransduced) S17 cells over time is shown in Figure 2A. After an initial decline, a progressive expansion of CD34+ cells, together with expansion of early B and myeloid cells, was observed.

Phenotypic analysis of supernatant cells grown on transduced S17 cells (Fig. 2B and Table 1) was then performed. The following was observed: no significant differences could be detected, either in cell numbers (not shown) or expression of differentiation markers, in cells grown with S17 cells transduced with the vector containing only IRES-GFP sequences (Vector), compared with the parental conditions. Significant proportions of CD34+ cells were detected at 4 weeks in all conditions; Jagged1 did not influence the percentages of CD15+ (granulocytic) and CD14+ (monocytic) cells; Delta1 increased the percentage of CD15+ cells, whereas CD14+ cells were maintained at percentages similar to control. As expected [5], an increase in CD10+CD19+ cells was observed during the 4 weeks of culture both in control and Jagged1 stromas, and no CD10+CD19+ cells were detected at any time point in the presence of Delta1; in the latter condition, almost half of the supernatant CD34+ cells co-expressed CD7+ at 4 weeks (not observed in the other conditions). No TdT-positive, CytCD3-positive, or CytCμ-positive cells were detected in any condition.

The absolute numbers of the different cell subpopulations at 4 weeks, normalized for the control values, show that the increase in total cell numbers was significantly lower in Delta1 than in Jagged1 and control conditions (Fig. 2C). CD15+ cell numbers were similar in Jagged1 and Delta1, CD14+ and CD34+ cells were lower in Delta1 conditions, and CD10+ CD19+ cell numbers were similar in Jagged1 and control conditions (Fig. 2C).

Thus, after 4 weeks of culture of CD34+ cells with Delta1-or Jagged1-S17 cells, Delta1 caused, in addition to a block of B-cell development and increase in CD7+ cells, a reduction in the absolute number of monocytic cells (CD14+ cells), whereas the absolute number of CD10+CD19+ and CD14+ cells in Jagged1 conditions did not differ from control. In addition, Delta1 induces a lower expansion of CD34+ cells at 4 weeks, compared with Jagged1 and control stromas.

Stromal Cells Expressing Delta1 or Jagged1 Differentially Affect the Myeloid Potential of CD34+ Cells

CD34+ cells purified after 4 weeks were used for CFU assays (Table 2). The total number of CFU originated by fresh CD34+ cells and CD34+ cells grown with the control stroma (Vector) were similar for the two dilutions (results from three independent experiments are shown in Table 2). In contrast, the total number of CFU originating from Jagged1-CD34+ cells was higher than in control (p < .05), and those from Delta1-derived CD34+ cells were lower than in control (p < .05). However, 42% of purified Delta1-CD34+ cells at 4 weeks coexpressed CD7 (Table 1), a phenotype of early T-cell precursors [12], likely to be devoid of myeloid potential. Methylcellulose assays of sorted Delta1-CD34+CD7+ cells confirmed this assumption, as they only originated 0.5 CFU per 200 plated cells (3 CFU per 400 cells; 17.5 CFU per 2,000 cells). The myeloid potential of CD34+ cells grown in contact with Delta1 is thus restricted to CD34+CD7 cells (60% of CD34+ cells), indicating that the actual clonogenicity of Delta-derived myeloid progenitors does not differ from that in control condition (last column of Table 2).

In summary, Jagged1, as opposed to Delta1, increases the overall myeloid clonogenic potential of CD34+ cells.

Stromal Cells Expressing Delta1 or Jagged1 Have Different Effects on the Number, Size, and Cell Morphology of Different Colony Types Generated by CD34+ Cells

Inspection of colonies emerging in the various conditions revealed that, compared with control conditions, all types of colonies (CFU-GM, CFU-G, and CFU-M) increased in Jagged1 conditions, whereas Delta1-CD34+ cells (of which only 60% had myeloid potential) originated less CFU-GM and CFU-M and more CFU-G (see Table 2, last column); that all types of Jagged1 colonies were usually larger (more cells) than in control (Fig. 3A, 3B), the opposite being observed for all types of Delta1 colonies (not shown); and that all colonies included all expected cell differentiation stages, with the exception of Delta1-CFU-G, in which most cells were morphologically promyelocytes (Fig. 3B). In some of the largest Jagged1-derived CFU-GM, some erythroid cells could also be observed.

Therefore, S17 cells expressing Delta1 and Jagged1 differentially affect the differentiation of CD34+ progenitors into the myeloid lineages. Jagged1 increases the overall clonogenicity of CD34+ cells; Jagged1 and Delta1 have opposite effects on the emergence of CFU-GM, the former increasing the latter decreasing their number. A consistent, yet nonsignificant, reduction of CFU-M and increase of CFU-G were observed in Delta1 conditions.

CD34+ Cells Arising in Contact with Stromal Cells Expressing Delta1 or Jagged1 Have Different Transcription Profiles of Notch Signaling-Related Genes

In an attempt to investigate the expression dynamics, over time, of genes related to Notch signaling, qRT-PCR was performed at 48 hours, 1 week, 2 weeks, and 4 weeks in purified CD34+ cells emerging in the different stromal conditions. In the case of Delta1-CD34+ cells, gene expression analysis at 1, 2, and 4 weeks was done in sorted CD7 and CD7+ subpopulations (the latter appearing at significant numbers in the supernatant at 2 weeks). Three groups of gene-transcripts were studied: 1) Notch receptors (Notch1, Notch2, and Notch3); 2) Notch targets (Hes1 and Deltex, all time points); and 3) modulators of Notch signaling (Lunatic Fringe at all time points; Manic Fringe, Radical Fringe, and Numb at 4 weeks).

An overview of the expression dynamics of these transcripts over time is shown in Figure 4. Figure 4A shows the expression levels of all genes (relative to those in fresh CD34+ cells) after 48 hours of cell contact with parental- and vector-S17 cells. No upregulation of any transcript was observed. This is in contrast with, at the same time point, Jagged1- and Delta1-CD34+ cells: Jagged1-cells upregulated Notch1, Notch3, and Hes1 at modest levels (two- to threefold) (Fig. 4B), whereas Delta1-CD34+ cells (whose phenotype is identical to that of control and Jagged1 CD34+ cells at this point) strikingly upregulated Notch3 (36-fold) and Hes1 (13-fold) (Fig. 4C).

At 1 and 2 weeks, levels of Notch1, Notch2, and Hes1 were maintained, with mild fluctuations, in both Jagged1- (Fig. 4B) and Delta1-CD34+ cells (Fig. 4C), with the exception of Notch3 and Deltex, which increased sharply from 1–2 weeks in Delta1-CD34+ cells (Fig. 4C, 4D). A sustained upregulation (fourfold) of Lunatic Fringe, starting at week 1, was also observed in Delta1-CD34+ (Fig. 4C).

At 4 weeks, the quantitative and qualitative differences between Jagged1- and Delta1-CD34+ cells were as follows: most genes were transcribed at much higher levels in Delta1-than in Jagged1-CD34+ cells, and within Delta1-CD34+ cells, the same genes were transcribed at higher levels in CD7+ than in CD7 cells. One further observation was that genes of each functional group were expressed in a differential manner within each subset. Jagged1-CD34+ cells upregulated Notch1 and Notch3 genes (the former at higher levels than the latter) and Deltex and Hes1 (the former at much higher levels than the latter), and all modulators of Notch signaling were downregulated (Fig. 4B). As to Delta1-CD34+ cells, Notch3 was the most upregulated gene, followed by Notch1 and Notch2, a pattern that was common to both Delta1-CD34+ cell subpopulations (Fig. 4C, 4D). Hes1 was markedly upregulated in Delta1-cells, especially in CD7 cells (Fig. 4C), whereas Deltex transcripts predominated in CD7+ cells (Fig. 4D). All Notch signaling modulators (with the exception of Radical Fringe) were up regulated in Delta1-CD34+ cells, more so in CD7+ than in CD7 cells (Fig. 4C, 4D).

Therefore, dynamic changes in the transcriptional activity of genes coding for Notch receptors, Notch targets, and Notch signaling modulators were observed over time, which differed according to whether CD34+ were cultured in the absence or presence of Jagged1 or Delta1.


Here we show that stromal cell environments expressing Jagged1 or Delta1, to which no exogenous cytokines were added, have differential effects on the phenotype, clonogenic potential and expression of Notch signaling-related genes in CD34+CD38+ hematopoietic progenitors.

Both Jagged1 and Delta1 stromas affected myeloid progenitors, not by forcing them to undergo cell fate decision processes, as is the case with Delta1 effects on lymphopoiesis [5], but by differentially regulating the balance between their granulocytic and monocytic cell progenies. As expected [34], CFU-M derived from cord blood CD34+ cells were more abundant than CFU-G (M/G = 2) in control conditions. This imbalance disappears, however, upon contact with Jagged1, by an increase in both the granulocytic and monocytic compartments and with Delta1, which decreases the monocytic compartment and increases granulocytic cells. Although the observations in Delta1 cells could, in principle, be a consequence of part of progenitors having followed a T-cell development pathway, the morphological evidence for abnormal dynamics of granulocytic differentiation, as shown by the accumulation of promyelocytes in Delta1 CFU-G colonies, points to a direct effect of Delta1 on myeloid progenitors. In support of this assumption is the recent observation that strong Delta1 signals induce apoptosis on CD34+ cells and decrease myeloid differentiation, while not affecting the emergence of T-cell development [35]. Previous reports have shown that Delta1 has proliferative and antidifferentiative effects on hematopoietic stem cells [6, 27, 28]. Here we observed that, when acting on more mature progenitors (CD34+CD38+), Delta1 decreases the generation of CFU-GM. This, together with the previously shown pro-apoptotic effect of Delta1 on monocytic cells [36], indicates that the negative effects of this ligand on the monocytic lineage must start beyond the stem cell stage, possibly at the level of the pluripotent myeloid progenitor.

In contrast, Jagged1-expressing stromal cells increase the proliferation of bipotent and unipotent myeloid progenitors. Jagged1 was shown to induce proliferative and antidifferentiative effects in hematopoietic stem cells [2224, 37], although differentiation could occur under certain cytokine conditions [29]. Here we show that when acting in more advanced progenitors, Jagged1 maintains its proliferative effect in all cell progenies, but the block of terminal differentiation is no longer present.

In an attempt to investigate whether these phenotypic and functional differences were related to changes in transcription levels of Notch signaling-related genes, qRT-PCR analysis of supernatant CD34+ cells, emerging in the different stromal conditions, was done at different time points.

Despite the functional heterogeneity of the cell subpopulations under analysis, a few insights emerge when gene expression data are interpreted together with phenotypic changes occurring in supernatant cells over time. At the earliest time point investigated (48 hours), at which no evidence for cell differentiation was detectable, the data indicate that Delta1 and Jagged1 do indeed induce distinct responses in CD34+CD38+ progenitors. In fact, whereas upregulation of Hes1 was present in both culture conditions, indicating that both ligands were activating Notch signaling through this CBF1-dependent pathway [1], an impressive differential regulation of Notch genes themselves was observed. Jagged1-CD34+ cells had a very modest, yet similar, upregulation of Notch1 and Notch3, whereas Delta1-CD34+ cells exhibited a striking upregulation of Notch3 in relation to Notch1 and Notch2.

When looking at subsequent time points, abrupt changes were observed at 2 weeks in Delta1-CD34+CD7 cells, consisting of further increase in Notch3 levels, modest upregulation of Notch1, downregulation of Hes1, and upregulation of Deltex. This time point coincides with the appearance of substantial numbers of CD34+CD7+ cells in the supernatant (40% of all CD34+ cells expressed CD7 at this point). Indeed, a similar, but quantitatively more robust, expression pattern was observed in CD34+CD7+ cells (pure T/NK progenitors) purified at 2 weeks. Therefore, the composite expression pattern of Notch3, Notch1, Hes1, and Deltex in Delta1-CD34+ cells at 2 weeks appears to reflect a transcription profile of T/NK progenitors. If so, the presence of the same profile, albeit at much lower quantitative levels in the CD7 subpopulation, suggests that very early T/NK progenitors might co-exist with myeloid progenitors in Delta1-CD34+CD7 cells.

By week 4, when CD34+ cells have reached their maximal myeloid clonogenic potential [33], new changes occur in Delta1-CD34+ cells, which are characterized by a trend toward opposite patterns of Hes1 and Deltex in CD7 and CD7+ subsets. In the former, Hes1 predominates over Deltex, whereas in the latter, Deltex is more expressed than Hes1. One possible explanation for this observation is that it reflects sequential steps of cell differentiation occurring after commitment into the T/NK cell lineage. For example, levels of Notch3 and Deltex (higher than Notch1 and Hes1, respectively) in CD7+ cells (which are pure NK/T precursors), is similar to that of immature thymocytes [38, reviewed in ref. 39], the putative immediate progeny of those cells. Also, the dynamics of Deltex expression (decreasing from 2–4 weeks) in CD7+ cells may reflect sequential stages of early T cell development, considering that Deltex1 is expressed at different levels during early thymocyte development in the mouse [40, 41].

As to changes in gene transcription observed at 48 hours (striking upregulation of Notch3 and, to a much lesser extent, of Notch1), at least two possibilities should be considered. One is that they could be a direct consequence of Delta1-mediated Notch stimulation on CD34+ cells. In fact, when the latter were switched, after 48 hours of contact with Delta1, to control stroma, no T-cell development (or block of B-cell differentiation) was observed (data not shown), in agreement with recently reported findings for mouse fetal liver precursors [42]. This indicates that irreversible Delta1-induced T-cell commitment [5] has not yet occurred by this early time point. However, a small proportion (6%–7%) of starting CD34+ cells co-expressed CD7, which may contain the early T/NK progenitor recently described in human cord blood (1–2.5/1000 CD34+ cord blood cells) [43]. The possibility cannot, therefore, be formally excluded that the observed early expression patterns reflect the effect of Delta1 on this very rare, yet already lineage-committed, cell population.

In contrast with Delta1-CD34+ cells, no substantial changes in gene expression were observed in Jagged1-CD34+ cells until week 4, when a marked increase in Notch1 and Deltex transcripts occurred. Although no phenotypic differences, in relation to control cells, could be observed at any time point, Jagged1-CD34+ cells were, by this time, more immature and proliferative than those in other stromal conditions, as shown by colony assays. How these differences relate to the concomitant Notch1 and Deltex transcription profiles is uncertain, as no clonogenic assays were performed at earlier time points. However, the absence in Jagged1 cells of the above-described T/NK transcription profile, together with the knowledge that murine immature B-cells (immature B-cells were also present in Jagged1 cultures; Table 1) express very low levels of Notch1 and Deltex [4, 40], makes it conceivable that the upregulation of these genes at 4 weeks might be myeloid-specific and related to the increased myeloid clonogenic potential of Jagged1-CD34+ cells.

As to the expression of genes coding for Notch signaling modulators, Lunatic Fringe, Manic Fringe, and Numb were upregulated in Delta1-CD34+ cells (CD7+ cells at 4 weeks showing the highest levels for all transcripts) and downregulated in Jagged1-CD34+ cells. Given that Lunatic Fringe suppresses Jagged1 signaling while enhancing that mediated by Delta1 [44], whereas Numb inhibits Notch signaling [45], these changes are likely to be related to the regulation of Notch signaling intensity mediated by the two ligands.

In summary, the present study shows that CD34+CD38+ progenitors arising on Delta1- or Jagged1-expressing stromal cells differ as to their phenotype and myeloid clonogenic capacities. The observed differences correlate with distinct and dynamic gene transcription patterns for Notch signaling-related genes. As to the latter, the results reveal that myeloid progenitors with strong clonogenic potential emerging in contact with Jagged1 upregulate Notch1 and Deltex and downregulate Notch signaling modulators, whereas the T/NK progenitors originated by Delta1 [5] are characterized by a striking upregulation of Notch3 and Deltex and, to a lesser extent, of Hes1, Lunatic Fringe, and Numb. The quest for the mechanisms underlying these responses and their putative physiological role is certainly worth pursuing.

Table Table 1.. CD34+ cells cultured in the presence of Delta1 or Jagged1: phenotype of supernatant cells at different time points
original image
Table Table 2.. Myeloid colonies originated by CD34+ cells grown in contact with Delta1 or Jagged1 for 4 weeks
original image
Figure Figure 1..

Experimental design and functional assay of S17 cells transduced with Delta1 or Jagged1. (A): Diagram depicting the sequential steps of the experimental approach used in the present study. (B): Differentiation of C2 myoblasts cells after 4 days of coculture on S17-vector, S17-Delta1, or S17-Jagged1 in the presence of carrier alone (0.5% dimethyl sulfoxide [DMSO]) or with 0.3 μM γ-secretase inhibitor [N-(3,5-difluorophenylacetyl-L-alanyl)]-S-phenylglycine t-butylester (DAPT). Original magnification ×20. Abbreviations: CFU, colony-forming unit; DAPI, 4,6-diamidino-2-phenylindole; GFP, green fluorescent protein; qRT-PCR, quantitative reverse transcription polymerase chain reaction; S17-Delta1, S17 cells transduced with Delta1 cDNA; S17-Jagged1, S17 cells transduced with Jagged1 cDNA; S17-vector, S17 cells transduced with GFP-vector; wk, week.

Figure Figure 2..

CD34+ cells cultured in the presence of parental and transduced S17 stromal cells for 4 weeks. (A): Expansion of supernatant cells grown on parental stroma. (B): Phenotypic analysis of supernatant cells grown on transduced and parental S17 stromal cells. Cells were analyzed for CD45 (hematopoietic cells), CD34 (progenitor cells), CD14/CD15 (myeloid cells), CD10/CD19 (pre-B cells) and CD7, within gated CD34+ cells, for detection of T/NK cell precursors. (C): Expansion of supernatant cells grown in the presence of Jagged1 and Delta1 for 4 weeks (normalized to vector). Absolute cell numbers were as follows. At 2 weeks: (9.6 ± 4.1) × 104 (Vector), (1.1 ± 0.5) × 105 (Jagged1), (7.8 ± 5.1) × 104 (Delta1); at 4 weeks: (4.0 ± 1.8) × 105 (Vector), (4.0 ± 1.5) × 105 (Jagged1), (1.7 ± 0.6) × 105 (Delta1). Abbreviations: Delta1, S17 cells transduced with Delta1 cDNA; Jagged1, S17 cells transduced with Jagged1 cDNA; Vector, S17 cells transduced with green fluorescent protein vector; Wk, week; WT, parental S17 stromal cells.

Figure Figure 3..

Myeloid clonogenic assays. Phase contrast images and May-Grunwald-Giemsa staining of cytospins obtained of colonies from vector-derived CD34+ cells (A) or ligand-derived CD34+ cells (B). Bar = 500 μm. Abbreviations: CFU-GM, colony-forming unit-granulocyte macrophage; CFU-G, colony-forming unit-granulocyte; CFU-M, colony-forming unit macrophage.

Figure Figure 4..

Expression of Notch signaling-related genes in CD34+ cells emerging at different time points of culture (quantitative reverse transcription polymerase chain reaction). (A): CD34+ cells after 48 hours of contact with parental (WT) or vector-transduced (Vector) stromal cells. (B): CD34+ cells emerging, at different time points, in contact with Jagged1. Phenotype of CD34+ cells at 2 and 4 weeks is shown in boxes (mean values; see also Table 1). (C, D): CD34+ cells emerging, at different time points, in contact with Delta1. (C): CD7 cells. (D): CD7 cells. Expression of each transcript was measured as a ratio with the glyceraldehyde-3-phosphate dehydrogenase transcript and expressed in arbitrary units normalized to control (fresh CD34+ cells − 1 [A]; vector-derived CD34+ cells = 1 [B–D]). Three different samples were analyzed in triplicate for each culture condition. Abbreviations: F, Fringe; Vector, vector-transduced stromal cells; Wk, week; WT, parental stromal cells.


The authors indicate no potential conflicts of interest.


We are grateful to Prof. António Coutinho (Instituto Gulbenkian de Ciência, Oeiras, Portugal) for helpful discussions, to Dr. Domingos Henrique (Instituto de Medicina Molecular, Faculdade de Medicina de Lisboa) for cDNA clones, and to Dr. Isabel Alcobia for helping with culture experiments. This work was supported by grants from Fundação para a Ciência e Tecnologia (POCTI/37953/2001) and Fundo de Investigação Oncológica.