Neurokinin-B transcription in erythroid cells: direct activation by the hematopoietic transcription factor GATA-1.

The GATA family of transcription factors establishes genetic networks that control developmental processes including hematopoiesis, vasculogenesis, and cardiogenesis. We found that GATA-1 strongly activates transcription of the Tac-2 gene, which encodes proneurokinin-B, a precursor of neurokinin-B (NK-B). Neurokinins function through G protein-coupled transmembrane receptors to mediate diverse physiological responses including pain perception and the control of vascular tone. Whereas an elevated level of NK-B was implicated in pregnancy-associated pre-eclampsia (Page, N. M., Woods, R. J., Gardiner, S. M., Lomthaisong, K., Gladwell, R. T., Butlin, D. J., Manyonda, I. T., and Lowry, P. J. (2000) Nature 405, 797-800), the regulation of NK-B synthesis and function are poorly understood. Tac-2 was expressed in normal murine erythroid cells and was induced upon ex vivo erythropoiesis. An estrogen receptor fusion to GATA-1 (ER-GATA-1) and endogenous GATA-1 both occupied a region of Tac-2 intron-7, which contains two conserved GATA motifs. Genetic complementation analysis in GATA-1-null G1E cells revealed that endogenous GATA-2 occupied the same region of intron-7, and expression of ER-GATA-1 displaced GATA-2 and activated Tac-2 transcription. Erythroid cells did not express neurokinin receptors, whereas aortic and yolk sac endothelial cells differentially expressed neurokinin receptor subtypes. Since NK-B induced cAMP accumulation in yolk sac endothelial cells, these results suggest a new mode of vascular regulation in which GATA-1 controls NK-B synthesis in erythroid cells.

The development of blood cells from hematopoietic stem cells is critically dependent upon integration of the activities of diverse transcription factors. Of particular importance are three members of the GATA family of transcription factors that regulate hematopoiesis via both unique and overlapping activities (1). GATA-1 is required for definitive or adult erythropoiesis, for megakaryocyte maturation, and for eosinophil production (2-7), whereas GATA-2 is important for the survival and function of multipotent hematopoietic precursor cells (8 -10). GATA-1 and GATA-2 have redundant activities to control primitive or embryonic erythropoiesis (11). GATA-3 uniquely regulates aspects of lymphopoiesis and central nervous system development (12)(13)(14)(15)(16). Understanding how GATA factors regulate development requires the elucidation of genetic circuits that control expression of GATA factor target genes and the GATA factors themselves.
The C-terminal zinc finger of dual zinc finger GATA factors recognizes a simple DNA motif (WGATAR) (17)(18)(19)(20) that is distributed abundantly throughout genomes. In contrast, the N-terminal zinc finger mediates interactions with the friend of GATA-1 (FOG-1) 1 coregulator (21,22) and stabilizes DNA binding at certain GATA motifs (17,23). Despite the high frequency of WGATAR in genomes, analyses of GATA-1 and GATA-2 binding to chromatin by quantitative ChIP analysis have revealed an exquisite specificity of chromatin occupancy. For example, the ␤-globin locus contains greater than 250 GATA motifs, but GATA-1 only occupies motifs within hypersensitive sites 1-4 of the ␤-globin LCR and within a subset of the ␤-globin promoters (24,25). Despite the more than 80 GATA motifs within the GATA-2 locus, GATA-1 only occupies motifs within a restricted upstream region (26,27). Thus, despite there being abundant GATA motifs within chromatin, including motifs that are conserved from mouse to humans, the vast majority of these motifs are not occupied by GATA factors in cells. ChIP analysis (27) has also revealed GATA-1 occupancy at sites previously implicated as being GATA-1 targets: the aminolevulinate synthase-2 promoter, hypersensitive site 1 of the GATA-1 locus, and hypersensitive site Ϫ26 of the ␣-globin locus (28 -32).
The determinants of GATA factor recognition of chromatin are unknown, but presumably neighboring cis-elements, intrinsic features of GATA motifs, and the chromatin environment are crucial. Considering the lack of knowledge of these determinants, one cannot predict whether a given GATA motif will be occupied in vivo. Thus, establishing the ensemble of GATA factor target genes that control hematopoiesis cannot be accomplished solely by delineating the distribution of conserved GATA motifs.
Genetic complementation analysis with a GATA-1 fusion to the estrogen receptor hormone binding domain in G1E cells has proven to be a powerful approach for defining GATA-1 target genes (27,33). The G1E cell line was derived from GATA-1-null embryonic stem cells, and G1E cells recapitulate the phenotype of adult bone marrow proerythroblasts (34). Oligonucleotide array analysis in this system revealed that a similar number of genes is activated and repressed by ER-GATA-1 (33). Although GATA-1 was known to confer activation and repression in a context-dependent manner, it was surprising that a similar number of genes was activated and repressed, since few GATA-1-repressed genes were known. However, since gene-profiling analysis identifies both direct and indirect target genes, it is crucial to use other approaches, such as ChIP, to determine whether a given gene is a direct target.
Insights into the mechanism of GATA-1-mediated repression have arisen from the observation that GATA-1 and GATA-2 have reciprocal expression patterns during hematopoiesis (1). Consistent with this expression pattern, targeted deletion of GATA-1 resulted in up-regulation of GATA-2 (3,34). These results suggested that GATA-1 represses GATA-2 transcription. Since a mutant of GATA-1 impaired in binding FOG-1 was inefficient in repressing GATA-2 (22), FOG-1 functions in the repression mechanism. We provided evidence for a direct repression mechanism in which GATA-1 displaces GATA-2 in a FOG-1-dependent manner from an upstream region of the GATA-2 locus (26,27). This GATA switch was temporally coupled with expulsion of the histone acetyltransferase CREBbinding protein from GATA-2 regulatory elements and broad deacetylation throughout the GATA-2 locus. We hypothesized that GATA-1 binding instigates a bimodal repression mechanism consisting of the GATA switch coupled with domain-wide deacetylation.
To assess whether the bimodal repression mechanism is used by GATA-1 to repress other loci, we asked whether GATA-2 occupies the Nab2 (nerve growth factor-activated factor-binding protein 2) locus and whether GATA-1 displaces GATA-2. Nab2 was identified by gene profiling analysis as a GATA-1repressed gene (33). Nab2 encodes a corepressor for the early growth regulator family of transcription factors, which mediate mitogen-induced cell proliferation (35,36). Whereas extensive ChIP analysis throughout the Nab2 locus failed to identify sites occupied by GATA-1 and GATA-2, GATA-1 and GATA-2 occupancy was detected at intron-7 of the nearby Tac-2 gene, which encodes a NK-B proprotein (37,38). These results are discussed with respect to the underlying transcriptional mechanism and the consequences of GATA factor-mediated control of Tac-2 transcription.
Primary Erythroid Cell Isolation-Bone marrow cells were flushed from the hind limbs of young female C57BL/6J (Jackson Laboratory, Bar Harbor, ME) mice into PBS (BioWhittaker, Walkersville, MD) containing 2% heat-inactivated FBS (Hyclone, Logan, UT). The cells were collected by centrifugation and resuspended at a concentration of 2 ϫ 10 6 cells/ml in PBS containing 2% FBS. The cells were incubated with phycoerythrin-conjugated TER119 or the appropriate isotype control antibodies (Pharmingen, San Diego, CA) for 20 min at 4°C with regular mixing at the concentrations recommended by the supplier. The cells were washed twice in PBS prior to fluorescence-activated cell sorting on a Vantage SE instrument (BD Biosciences) with a 488-nm argon laser. After sorting, TER119ϩ cells (ϳ20% of bone marrow cells) and TER119Ϫ cells were collected in PBS and concentrated by centrifugation. RNA was isolated with Trizol (Invitrogen).
For ex vivo human erythropoiesis experiments, adult human erythroblasts were cultured from peripheral blood CD34ϩ cells as previously described (42). Total RNA was isolated from these cells on culture days 0, 2, 4, 6, 8, 10, 12, and 14 using Trizol. The isolated RNA was further processed by DNase I treatment (Qiagen, Valencia, CA).
Adult human reticulocytes were isolated from 20 ml of peripheral blood in heparin following a process of leukodepletion. Leukocytes were concentrated by centrifugation at 3500 rpm for 30 min at 4°C, forming a disc (buffy coat) at the plasma-red cell interface. Both the plasma and buffy coat were removed, and the cells were washed three times with 10 ml of ice-cold PBS. At the third wash, reticulocytes were concentrated by centrifugation at 3500 rpm for 30 min at 4°C. The supernatant was removed, and the top 700-l reticulocyte-rich fraction was added to 4 ml of ice-cold PBS. This red cell suspension was layered on top of a column consisting of ϳ4.5 ml of 2 parts cellulose (C-8002; Sigma) and 1 part sigmacell type 50 microcrystallin cellulose (S-5504; Sigma) in normal saline. The red cells were eluted by gently centrifuging the column at 1000 rpm at 4°C for 1 min. This leukodepleted eluate (ϳ2.5 ml) was washed three times in 10 ml of ice-cold PBS. Finally, the reticulocyteenriched eluted cells were recovered by centrifugation and resuspended in 2 ml of ice-cold PBS. RNA was extracted with TRI reagent (T-9424; Sigma).
Quantitative Chromatin Immunoprecipitation Assay-Cells were grown in media containing 15% FBS with or without 1 M 4-hydroxytamoxifen (Sigma) for various times as specified in the figure legends. Protein-DNA cross-linking was conducted by treating cells with formaldehyde at a final concentration of 0.4% (1% for FOG-1) for 10 min at room temperature with gentle agitation. Glycine (0.125 M) was added to quench the reaction, and real time PCR-based quantitative ChIP was conducted as described (24,26,39,43). Measurements were made under conditions in which signals were in the linear range.
Quantitative RT-PCR-Total RNA was purified with Trizol. cDNA was prepared by annealing RNA (1 g) with 250 ng of a 1:5 mixture of random and oligo(dT) primers at 68°C for 10 min. This was followed by incubation with Moloney murine leukemia virus reverse transcriptase (Invitrogen) combined with 10 mM dithiothreitol, RNasin (Promega), and 0.5 mM dNTPs at 42°C for 1 h. cDNA synthesis reactions were diluted to a final volume of 100 l and heat-inactivated at 98°C for 10 min. Reactions (25 l) contained 2.5 l of cDNA, 12.5 l of SYBR Green Master Mix (Applied Biosystems, CA), and the appropriate primers. Product accumulation was monitored by SYBR Green fluorescence. Control reactions lacking RT yielded very low signals. Relative expression levels were determined from a standard curve of serial dilutions of G1E cDNA samples, and measurements were made under conditions in which signals were in the linear range.
Quantitation of cAMP Levels-MAE and YSEC cells (1 ϫ 10 6 ) were resuspended in growth medium containing either 250 M IBMX (Calbiochem), a phosphodiesterase inhibitor, IBMX ϩ 10 M forskolin (Sigma), or IBMX ϩ 50 M NK-B (Sigma) for 1 h at 37°C. The reaction was terminated by aspiration of the medium. Cells were washed with PBS and were lysed in 500 l of HCl (0.1 N, 10 min). Cell lysates (100 l) were assayed for cAMP using a Direct cAMP enzyme-linked immunosorbent assay kit (Assay Designs, Inc., Ann Arbor, MI) after acetyla-tion as recommended by the manufacturer. In brief, the cAMP assay utilizes a 96-well microtiter plate in which anti-cAMP antibody is immobilized. Cell lysates or cAMP standards are added in the presence of alkaline phosphatase-conjugated cAMP, and after a colorimetric reaction with phosphatase substrate, absorbance is measured at 405 nm under conditions of linearity.
To determine whether Tac-2 is expressed in other erythroid cells, quantitative RT-PCR was used to measure Tac-2 mRNA in erythroid cell lines and in primary murine and human erythroid cells. Tac-2 expression was induced by tamoxifen in G1E-ER-GATA-1 but not G1E cells ( Fig. 2A), indicating that ER-GATA-1 mediated the induction. Tac-2 was expressed in uninduced MEL cells, and expression was induced ϳ2 fold upon Me 2 SO-induced erythroid maturation ( Fig. 2A). GATA-1 expression is similar in uninduced and induced MEL cells. 2 Tac-2 was also expressed in primary murine Ter119ϩ erythroid cells isolated from bone marrow (Fig. 2A). The major adult ␤-globin gene, ␤major, and Tac-2 shared a similar pattern of expression, being repressed in G1E cells, induced by tamoxifen in G1E-ER-GATA-1 cells, induced upon MEL cell maturation, and expressed in Ter119ϩ cells ( Fig. 2A).
Tamoxifen-mediated differentiation of G1E-ER-GATA-1 cells and MEL cell maturation mimic the later stages of definitive erythropoiesis (34,52). To assess whether Tac-2 was differentially expressed over a broader developmental continuum, we analyzed expression of the human Tac-2 ortholog, TAC-3, during erythropoiesis ex vivo. Primary human erythroid progenitors from peripheral blood were differentiated for 14 days under conditions that recapitulate erythropoiesis in vivo (42). GATA-2 mRNA decreased as GATA-1 mRNA increased (Fig.  2B). GATA-1 mRNA was induced maximally by day 10 (Fig.  2B). Human TAC-3 mRNA increased by day 8 and was highly 2 S. Pal and E. H. Bresnick, unpublished data.

FIG. 2. Murine Tac-2 and the orthologous human gene TAC-3 are expressed in primary mouse and human erythroid cells and are induced upon erythroid differentiation.
A, quantitative RT-PCR analysis of Tac-2 and ␤major mRNA levels in mouse erythroid cell lines (tamoxifen-treated (24 h) G1E and G1E-ER-GATA-1 and untreated and Me 2 SO-treated (1.5% Me 2 SO, 4 days) MEL) and in primary mouse bone marrow erythroid cells (Ter119ϩ). Relative mRNA levels were normalized by the levels of GAPDH transcripts. The plots depict the mRNA/GAPDH mRNA ratios (mean, two independent experiments) in which the ratio obtained from analysis of G1E cells was designated as 1. B, reciprocal regulation of TAC-3 and Nab2 during ex vivo human erythropoiesis. CD34ϩ peripheral blood cells from healthy donors were isolated and cultured for 14 consecutive days. Total RNA was isolated from the cells, and mRNA levels were quantitated by real time RT-PCR. The relative levels of specific mRNAs were normalized by the levels of HPRT transcripts. The plots depict the mRNA/HPRT mRNA ratios (mean, two or three independent experiments) in which the ratio obtained from analysis of the day 0 culture in the presence of reverse transcriptase was designated as 1. C, TAC-3 mRNA in human peripheral reticulocytes. Reticulocytes were isolated from seven healthy donors. Total RNA was purified and quantitated by real-time RT-PCR, and the relative level of mRNA for sample 1 was designated as 1.

FIG. 3. Coregulation of Tac-2 and
Nab2 by tamoxifen-activated ER-GATA-1. Total RNA was isolated from G1E-ER-GATA-1 cells treated with tamoxifen for various times, and Tac-2 (A), Nab2 (B), and ␤major (C) mRNA levels were quantitated by real-time RT-PCR. The relative levels of Tac-2, Nab2, and ␤major mRNAs were normalized by the levels of GAPDH transcripts. The plots depict the mRNA/GAPDH mRNA ratios (mean Ϯ S.E., three independent experiments) in which the ratio obtained from the time 0 culture was designated as 1. D, correlation between the magnitude of Tac-2 transcriptional activation and Nab2 repression upon tamoxifen-mediated activation of ER-GATA-1.

GATA-1-mediated Activation of Neurokinin-B Transcription
induced by day 14 (Fig. 2B). Similar to the G1E-ER-GATA-1 results, Nab2 mRNA levels declined as TAC-3 mRNA levels increased. The patterns of TAC-3 and ␤-globin expression were similar, albeit not identical, since ␤-globin mRNA was maximal by day 12, whereas TAC-3 mRNA was maximal by day 14.
At day 14, the majority of erythroid cells generated in the ex vivo system are orthochromatic normoblasts, which have not yet undergone enucleation (42). Upon extravasation through the marrow sinusoids in vivo, the normoblasts enucleate, yielding peripheral reticulocytes. Since reticulocytes are competent to carry out protein synthesis, we asked whether TAC-3 transcripts are present in purified human reticulocytes. TAC-3, ␤-globin, and HPRT transcripts were detected in reticulocytes from seven normal patients, whereas Nab2 transcripts were undetectable (Fig. 2C). Thus, similar to the Tac-2 results in the murine system, TAC-3 is expressed in human erythroid cells and is highly regulated during human erythropoiesis.
Since ER-GATA-1 regulates Nab2 and Tac-2 within the same chromosomal region, it is possible that Nab2 and Tac-2 are coordinately regulated. To test this hypothesis, G1E-ER-GATA-1 cells were treated with tamoxifen for increasing amounts of time, and Nab2 and Tac-2 mRNA levels were quantitated. Tac-2 activation (Fig. 3A) and Nab2 repression (Fig.  3B) were apparent by 4 h and were maximal by 24 h. These changes in mRNA levels were more rapid than the induction of ␤major transcripts (Fig. 3C), which is known to be a direct transcriptional target of GATA-1. The correlation between Tac-2 activation and Nab2 repression (Fig. 3D) supports a mechanism in which these two genes are coordinately regulated in a reciprocal manner in the G1E-ER-GATA-1 system. Both Nab2 and Tac-2 are members of protein families in which the individual members have both unique and shared activities. Like Nab2, Nab1 is an early growth regulator corepressor (53) but is expressed differently from Nab2 in certain scenarios. For example, nerve growth factor treatment of PC12 pheochromocytoma cells and serum stimulation of fibroblasts can induce Nab2 but not Nab1 expression (35). In other scenarios, Nab1 and Nab2 are co-regulated. 3 Tac-2 is highly related to Tac-1, which encodes preprotachykinin-1, the precursor to Substance P that mediates pain perception (47). To assess the specificity of GATA-1-mediated repression of Nab2 and activation of Tac-2, we tested whether ER-GATA-1 regulates Nab1 and Tac-1 expression. Nab1 mRNA was detected in G1E-ER-GATA-1 cells but was unaffected by tamoxifen treatment (Fig. 4A). Tac-1 mRNA was undetectable in G1E-ER- GATA-1 cells, and tamoxifen did not induce Tac-1 expression (Fig. 4B). In contrast, Tac-1 was highly expressed in the mouse brain (Fig. 4B).
GATA-1 Occupancy of a Conserved Region within Intron-7 of Tac-2-Based on the abundance of high affinity GATA motifs within genomes, the mere presence of a GATA motif does not imply functional significance. Functional insights can be derived from evaluating the evolutionary conservation of GATA motifs, but even a conserved motif does not equate to a functional motif. Fourteen conserved GATA motifs exist within the region upstream of Stat-6 to downstream of Tac-2 (Fig. 5A). Each of these motifs (Fig. 5B) would bind GATA-1 with high affinity in vitro, based upon the established DNA binding specificity of GATA-1 (18,19). However, analyses of GATA-1 occupancy at the ␤-globin (24,26) and GATA-2 (24,26) loci have revealed that the majority of GATA motifs are not occupied by GATA-1 in cells.
Quantitative ChIP analysis was conducted to assess whether ER-GATA-1 occupies the Nab2-Tac-2 chromosomal region in G1E-ER-GATA-1 cells. Analysis of 15 regions spanning all of the conserved GATA motifs and the Nab2 and Tac-2 promoters revealed that ER-GATA-1 occupies a single region within intron-7 of Tac-2 (Fig. 5C). Endogenous GATA-1 in Me 2 SO-induced MEL cells also occupies this intron-7 region, but not the Tac-2 promoter, which contains a GATA motif (Fig. 5D). Thus, GATA-1 selects exquisitely among the GATA motifs within the Nab2-Tac-2 chromosomal region. The highly selective occupancy of Tac-2 intron-7 strongly suggests that GATA-1 activates Tac-2 transcription via interaction with this site.
Our previous analyses of GATA factor interactions with chromatin indicated that a subset of the sites occupied by GATA-1 in erythroid cells were occupied by GATA-2 in GATA-1-null G1E cells (26,27) and in a FOG-1-null hematopoietic precursor cell line that expresses both GATA-1 and GATA-2 (27,55). A distinct subset of sites, including regions of the ␤-globin locus was occupied by GATA-1, but GATA-2 occupancy was either not detected or was very low. 2 GATA-2 and FOG-1 co-localized at chromatin sites, and only in the presence of FOG-1 was GATA-1 able to efficiently induce a GATA switch in which GATA-2 was displaced (27). Occupancy of the ␤major promoter by a GATA-1 mutant exhibiting reduced FOG-1 binding was
Quantitative ChIP analysis was used to test whether GATA-2 associates with intron-7 of Tac-2 in untreated G1E-ER-GATA-1 cells. GATA-2 occupancy was detected under conditions of very little ER-GATA-1 occupancy (Fig. 6A). Tamoxifen-mediated activation of ER-GATA-1 resulted in the loss of GATA-2 occupancy concomitant with increased ER-GATA-1 occupancy. Reduced GATA-2 occupancy was detected 30 min after tamoxifen treatment, and the switch was nearly complete by 4 h. Based on the previous analysis of the kinetics of GATA-1-mediated repression of GATA-2 (26), these data strongly suggest that the loss of GATA-2 occupancy from intron-7 results from GATA-1-mediated displacement of GATA-2 from the intron-7 region rather than repression of GATA-2, which would lead to insufficient levels of GATA-2 for intron-7 occupancy.
G1E-ER-GATA-1 cells express FOG-1, and, as noted above, FOG-1 co-localizes with GATA-2 at a subset of the chromatin sites bound by GATA-1 later in erythropoiesis (27). To test whether FOG-1 is a component of the endogenous Tac-2 intron-7 complex when either GATA-1 or GATA-2 is bound, quan-titative ChIP analysis was conducted with uninduced and tamoxifen-treated G1E-ER-GATA-1 cells. FOG-1 occupancy was detected at intron-7 of the locus in both the transcriptionally inactive and active states (Fig. 6B), similar to our previous findings with the GATA-2 locus (27). Thus, the intron-7 region is occupied by a GATA-2-FOG-1 complex in the inactive state, and the GATA switch, which correlates with transcriptional activation, results in either the replacement of GATA-1 by GATA-2 with FOG-1 retention or the substitution of a GATA-2-FOG-1 complex with a GATA-1-FOG-1 complex.
An Erythroid Cell-derived Neurokinin: Implications for Vascular Regulation-Whereas neurokinins were classically studied in the nervous system, recent studies have expanded the repertoire of sites of neurokinin synthesis and function (47,50). However, whether erythroid cells are competent to generate neurokinins had not been investigated. We describe herein that erythroid cell lines and primary erythroid cells express Tac-2 in a highly regulated manner during erythropoiesis. Furthermore, GATA-1 and GATA-2 bind directly to intron-7 of Tac-2 in living cells, and GATA-1 rapidly activates Tac-2 transcription. The mechanistic analyses in the G1E system provided strong evidence that Tac-2 is a direct target of GATA-1 and GATA-2. These findings raise the question of what might be the function of erythroid cell-derived NK-B.
Since nucleated definitive erythroid cell precursors exist pre- In the inactive state, a GATA-2-FOG-1 complex occupies the Tac-2 intron-7 region. GATA-1 induces a "GATA switch" in which GATA-2 is displaced, but FOG-1 is retained. For simplicity, the model depicts free GATA-1 displacing GATA-2, but equally likely is that a GATA-1-FOG-1 complex displaces the GATA-2-FOG-1 complex. The GATA switch would instigate the assembly of activating complexes that induce Tac-2 transcription. Since YSEC express NK1 and NK3 and are activated by exogenously added NK-B, the model assumes that erythroid cell-derived NK-B would have the capacity to activate NK receptors on certain endothelial cell subtypes. dominantly in the bone marrow, NK-B released from erythroid progenitors might function within the marrow compartment. To assess whether NK-B functions via autocrine or paracrine regulation of erythroid progenitors, we tested whether the G protein-coupled NK receptors (NK1, NK2, and NK3) that mediate NK-B actions (57) are expressed in uninduced or induced G1E-ER-GATA-1 cells. NK-B has the highest affinity for NK3 but can also bind and activate NK1 and NK2 (47). NK receptor activation results in stimulation of adenylyl cyclase and cyclic AMP generation, as well as activation of phospholipase C and the accumulation of intracellular calcium (58,59). NK receptor expression was not detected in uninduced or induced G1E-ER-GATA-1 cells using a sensitive RT-PCR-based assay (Fig. 7). In addition, NK receptor expression was not detected in murine Ter119ϩ bone marrow erythroid cells. 2 However, NK1 (Fig. 7A) and NK2 (Fig. 7B) were expressed in MAE cells, and NK1 and NK3 (Fig. 7C) were expressed in YSEC. MAE and YSEC cells did not express Tac-2, 2 indicating that endothelial NK receptors must be activated by NK-B derived from distinct cell types. Taken together with the findings that NK1 and NK2 receptors are expressed in human umbilical vein endothelial cells (60,61) and that modulators of NK receptors are vasoactive in vivo (49,(61)(62)(63)(64), these results suggest that NK-B regulates the function of certain endothelial cells.
To test the functionality of MAE and YSEC NK receptors, we assessed whether these cells were competent to mount a signaling response to NK-B. NK-B has not been shown to activate signaling pathways in endothelial cells, despite the report of NK1 and NK2 receptor expression in human umbilical vein endothelial cells (60,61). Whereas the direct activator of adenylyl cyclase forskolin stimulated cAMP levels in MAE and YSEC cells, NK-B only stimulated cAMP production in YSEC (Fig. 7D). Intriguingly, YSEC, but not MAE cells, expressed NK3, which binds NK-B with higher affinity than other NK receptors (47). Thus, YSEC express functional NK receptors, which can mediate activation of adenylyl cylase in response to NK-B. These results support a model in which erythroid cellderived NK-B has the capacity to activate NK receptors on certain endothelial cell subtypes (Fig. 8).
Human bone marrow stromal cells express NK1 and NK2, and treatment of these cells with stem cell factor increased and decreased NK1 and NK2 mRNA levels, respectively (65). Thus, the bone marrow stroma also represents a prospective target for erythroid cell-derived NK-B. However, given the low affinity of NK-B for NK1 and NK2 and the failure of NK-B to induce signaling in MAE cells that express NK1 and NK2, it is more likely that another member of the neurokinin family, such as substance P, acts through the stromal NK receptors. A cleavage product of Tac-1-encoded Substance P (SP-(1-4)) induced transforming growth factor-␤ and tumor necrosis factor-␣ production in bone marrow stromal cells (65), indicating that substance P and derivative peptides might regulate hematopoiesis. Since Tac-1 expression was not detected in untreated and tamoxifen-treated G1E-ER-GATA-1 cells (Fig. 4B) and GATA-1 specifically regulates Tac-2 transcription, erythroid cells produce NK-B rather than substance P.
The bone marrow compartment is separated from the circulation by a single layer of endothelium, and blood cell precursors traverse this endothelium upon mobilization to the periphery. It is attractive to speculate that the binding of GATA-1 to intron-7 of Tac-2 establishes and/or maintains NK-B transcription, resulting in NK-B regulation of endothelial cell components of the bone marrow vascular network. Whereas functional properties of the bone marrow endothelium are not well understood, especially relative to other endothelial cell subtypes (66 -68), the bone marrow endothelium has a central role in regulating the mobilization of blood cell precursors from the marrow into the periphery. Since this crucial function is likely to be highly regulated via signaling mechanisms, it will be important to develop immunological reagents to quantitate murine pro-NK-B, NK-B, and NK receptor levels and to test whether the NK-B/NK receptor system controls vascular function in the bone marrow compartment.