Notch4 inhibits endothelial apoptosis via RBP-Jkappa-dependent and -independent pathways.

Notch4, a member of the Notch family of transmembrane receptors, is expressed primarily on endothelial cells. Activation of Notch in various cell systems has been shown to regulate cell fate decisions, partly by regulating the propensity of cells to live or die. Various studies have demonstrated a role for Notch1 in modulating apoptosis, either in a positive or negative manner. In this study, we determined that constitutively active Notch4 (Notch4 intracellular domain) inhibited endothelial apoptosis triggered by lipopolysaccharide. Notch signals are transmitted by derepression and coactivation of the transcriptional repressor, RBP-Jkappa, as well as by less well defined mechanisms that are independent of RBP-Jkappa. A Notch mutant lacking the N-terminal RAM domain showed only partial antiapoptotic activity relative to Notch4 intracellular domain but stimulated equivalent RBP-Jkappa-dependent transcriptional activity. Similarly, constitutively active RBP-Jkappa activated a full transcriptional response but only demonstrated partial antiapoptotic activity. Additional studies suggest that Notch4 provides endothelial protection in two ways: inhibition of the JNK-dependent proapoptotic pathway in an RBP-Jkappa-dependent manner and induction of an antiapoptotic pathway through an RBP-Jkappa-independent up-regulation of Bcl-2. Our findings demonstrate that Notch4 activation inhibits apoptosis through multiple pathways and provides one mechanism to explain the remarkable capacity of endothelial cells to withstand apoptosis.

Notch4, a member of the Notch family of transmembrane receptors, is expressed primarily on endothelial cells. Activation of Notch in various cell systems has been shown to regulate cell fate decisions, partly by regulating the propensity of cells to live or die. Various studies have demonstrated a role for Notch1 in modulating apoptosis, either in a positive or negative manner. In this study, we determined that constitutively active Notch4 (Notch4 intracellular domain) inhibited endothelial apoptosis triggered by lipopolysaccharide. Notch signals are transmitted by derepression and coactivation of the transcriptional repressor, RBP-J, as well as by less well defined mechanisms that are independent of RBP-J. A Notch mutant lacking the N-terminal RAM domain showed only partial antiapoptotic activity relative to Notch4 intracellular domain but stimulated equivalent RBP-J-dependent transcriptional activity. Similarly, constitutively active RBP-J activated a full transcriptional response but only demonstrated partial antiapoptotic activity. Additional studies suggest that Notch4 provides endothelial protection in two ways: inhibition of the JNK-dependent proapoptotic pathway in an RBP-J-dependent manner and induction of an antiapoptotic pathway through an RBP-J-independent upregulation of Bcl-2. Our findings demonstrate that Notch4 activation inhibits apoptosis through multiple pathways and provides one mechanism to explain the remarkable capacity of endothelial cells to withstand apoptosis.
The Notch proteins comprise a family of transmembrane receptors that have been highly conserved through evolution as mediators of cell fate and are comprised of four members in mammals (Notch1 to -4) (1). Following intracellular processing of the full-length protein by a furin-like convertase, Notch is expressed at the cell surface as a heterodimeric receptor (2,3). Engagement by ligand results in a two-step cleavage of the Notch heterodimer. These cleavage events release the intracellular domain of Notch (NotchIC) 1 from its membrane tether, whereupon NotchIC translocates to the nucleus and interacts with the DNA-binding factor, RBP-J (CBF1). RBP-J is a DNA binding protein that has dual function: it represses transcription in the absence of NotchIC and activates transcription in its presence (2). In the nucleus, NotchIC and RBP-J associate with other factors to form a multimeric complex that results in transcriptional activation of various basic helix-loophelix factors of the HES (Hairy and enhancer of Split) and HRT (Hairy-related transcription factor, also called HEY, HESR, Gridlock, and HERP) families, through the release of a corepressor complex from RBP-J, and recruitment of a co-activator complex (2). There is cell type-dependent activation of the HRTs, but all three HRTs have been shown to be expressed in vascular cells (4 -7). RBP-J-independent Notch activity has also been described, but Notch signaling through this mechanism is less well elucidated.
Because Notch function requires ligand-dependent cleavage of the intracellular domain, enforced expression of NotchIC results in a constitutively active, signaling form of the receptor, which results in altered cell fate decisions in several models (2,3). The intracellular domain can be divided into two major subdomains; C-terminal to the transmembrane domain, there is a RAM domain, which is followed by an ankyrin domain composed of six Cdc10/ankyrin repeats. The region C-terminal to the ankyrin repeats acts as a putative transactivation domain in Notch1 but not Notch4 (8).
Notch1, -2, and -4 have been reported to be expressed in endothelial cells in vivo, and similar results have been reported in cultured endothelial cells (9). Several studies point to a role for Notch and its ligands in influencing vascular development (10). Notch signaling is required for arterial-venous differentiation in zebrafish (11). Mutant mice that are null for Notch1 show defects in the vasculature, and the severity of these vascular defects is enhanced in mice that are null for both Notch4 and Notch1 (12). A homozygous Notch2 hypomorphic allele disrupts development of vasculature of the glomerulus, heart, and eye (13). Interestingly, constitutive activation of Notch4 also causes defects in vascular remodeling (14,15).
The regulation of vascular cell survival and death is critical during vascular development and homeostasis as well as in diverse pathological processes including inflammation (16 -18). Despite continual exposure to various inflammatory cytokines and exogenous toxins, endothelial cells have a remarkable capacity to resist apoptosis, and this may be a mechanism to preserve vascular integrity in pathological situations (19,20). However, we have shown that in certain circumstances inflammatory mediators, such as tumor necrosis factor and bacterial lipopolysaccharide (LPS), are able to induce endothelial apoptosis. In particular, we have shown that LPS induces endothelial apoptosis by activating the mitogen-activated protein kinase family member, c-Jun NH 2 -terminal kinase (JNK) (20). Although Notch1 activation appears to promote endothelial viability when cells are starved of serum, little is known about the mechanism (9). Since Notch and its ligands are expressed at high levels in vascular endothelium, we postulated that Notch activation may play a protective role in maintaining endothelial survival in inflammatory situations (21)(22)(23).
In this paper, we examined the functional activity of Notch4, a Notch member that is expressed selectively in the endothelium. We demonstrate that activated Notch4 is able to inhibit endothelial apoptosis in response to the inflammatory mediator, LPS, in at least two different ways. First, Notch activation is able to inhibit LPS-mediated JNK activation, through an RBP-J-dependent pathway. Notch also provides antiapoptotic activity by up-regulating Bcl-2 via an RBP-J-independent mechanism. This dual antiapoptotic mechanism makes the activation of Notch a particularly potent inhibitor of the intrinsic apoptotic pathway.

Cell Culture
Transformed human microvascular endothelial cells (HMEC-1, hereafter referred to as HMEC) were provided by the Centers for Disease Control and Prevention (Atlanta, GA) and cultured as previously described (14,24). Human umbilical vein endothelial cells (HUVEC) were isolated and cultured as previously described (25). Cells were maintained at 37°C in 5% CO 2 .

Plasmid Constructs and Gene Transfer
The Notch4 intracellular region (Notch4IC) construct, described previously (14), contains a C-terminal hemagglutinin epitope tag (HA) and includes amino acids 1476 -2003 of the 2003 residue full-length Notch4. The Notch4IC deletion mutants were constructed by PCR, using Notch4IC as a template, and inserted into the LNCX retroviral vector. The Notch4IC mutants (see Fig. 3A) include constructs (i) lacking the entire RAM domain (⌬RAM; encodes amino acids 1518 -2003); (ii) lacking the RAM and N-terminally fused with an SV40-derived NLS (NLS-⌬RAM; the NLS tag codes for the amino acid sequence DPKKKRKV); and (iii) lacking all six ankyrin repeats (⌬Ank encodes amino acids 1476 -1578 and 1801-2003). Notch4IC was also cloned into the MSCV-IRES-YFP (MIY) retroviral vector, as was RBP-VP16, a constitutively active RBP-J. RBP-VP16 was constructed by PCR amplification of the 3Ј region of the murine RBP-VP16 cDNA (gift of E. Manet) containing the coding region for the VP16 transactivation domain (26). The product was digested with AflII and ligated to the corresponding AflII site of the cDNA for FLAG-RBP-J derived from the RBP-2N isoform of human RBP-J (gift of R. Schmid) (27). The 4ϫRBP-J luciferase plasmid (gift of S. D. Hayward) includes four copies of an RBP-J binding element cloned into the pGL2pro (Promega) firefly luciferase plasmid (28).
HMEC and HUVEC were transduced with the various constructs as described previously (19). Polyclonal HMEC lines were isolated by selection in 300 g/ml of G418 (Invitrogen) for the LNCX constructs and by sorting for yellow fluorescent protein (YFP) expression using a FACS 440 (BD Biosciences) for the MIY constructs. Polyclonal HMEC lines were used in order to avoid artifacts due to the retroviral integration site.
Loss of ⌬⌿ m -To measure mitochondrial transmembrane potential (⌬⌿ m ), 5 ϫ 10 5 cells were incubated with tetramethyl rhodamine ethyl ester (TMRE) (Molecular Probes, Inc., Eugene, OR) and analyzed for fluorescence on a flow cytometer. The mitochondrial uncoupler, carbonyl cyanide m-chlorophenylhydrazone (Sigma) was used as a positive control for the detection of loss of ⌬⌿ m .
Caspase Activity Assay-DEVD-p-nitroaniline (pNA) (caspase 3/7) cleavage activity was quantitated with a colorimetric assay kit according to the manufacturer's instructions (R & D Systems). Briefly, 200 g of whole cell lysates from HMEC cells exposed to LPS (100 ng/ml) was combined with DEVD-pNA (200 M) in a 96-well plate and incubated at 37°C. The release of the chromophore by active caspases was quantitated at 405 nm and normalized to untreated cell lysates.

Immunofluorescence
Transduced HMEC lines were cultured overnight on chamber slides, fixed with 4% paraformaldehyde for 15 min, and then permeabilized with cold methanol for 3 min. Nonspecific binding was blocked by incubation with 5% goat serum. Cells were stained with the mouse anti-HA monoclonal primary antibody (1:100 dilution) for 1 h and then for 30 min with an AlexaFluor 488-conjugated goat anti-mouse IgG secondary antibody (1:500 dilution). Nuclei were counterstained with 4Ј,6-diamidino-2-phenylindole for 5 min, and coverslips were mounted with 50% glycerol. Slides were viewed using a Zeiss Axioplan II Imaging inverted microscope (Carl Zeiss Canada), and images were captured with a 1350EX cooled CCD digital camera (QImaging).

Transient Transfection and Luciferase Assays
Transient transfection of luciferase reporter plasmids was carried out by electroporation as described (29). Transduced HMEC lines were grown until ϳ80% confluence and then trypsinized and resuspended in HMEC medium. Cells (1.5 ϫ 10 6 /transfection) were pelleted at 1000 rpm for 5 min, washed with PBS, pelleted as previous, and then resuspended in 0.4 ml of electroporation buffer (20 mM HEPES, 137 mM sodium chloride, 5 mM potassium chloride, 0.7 mM sodium phosphate, 6 mM D-glucose, pH 7.0) containing luciferase reporter plasmid DNA. The cell/DNA mixture was transferred to a 4-mm gap electroporation cuvette (Bio-Rad), left for 10 min at room temperature, and then electroporated at a fixed capacitance of 900 microfarads and 200 V using a Bio-Rad Gene Pulser II instrument. For each transfection, 2.5 g of 4ϫRBP-J-binding promoter luciferase and 1 g of RL-CMV was used. The RL-CMV reporter contains the Renilla luciferase cDNA expressed under control of the CMV immediate early enhancer/promoter and serves as a normalization control for transfection efficiency. After electroporation, the cells were left for 10 min at room temperature before plating in prewarmed HMEC medium. The medium was changed 24 h later, and cells were harvested for assay 48 h after transfection. Lysis and dual luciferase reporter assays were performed according to the manufacturer's recommendations (Promega) with luminescence measured on a Tropix tube luminometer (BIO/CAN Scientific). Luminescence values of mock transfections were subtracted from sample luminescence readings to give the net firefly and net Renilla luciferase units. The net firefly units divided by the net Renilla units determined the relative luciferase units.

Activated Notch4
Inhibits Endothelial Cell Apoptosis-Members of the Notch family have previously been shown to have Notch4 Activation Inhibits Endothelial Apoptosis either anti-or proapoptotic effects depending on the Notch member, cell type, or apoptotic stimulus (30,31). Because Notch4 is structurally distinct from the other Notch members and exhibits endothelium-selective expression, we tested whether Notch4 was able to regulate endothelial apoptosis elicited by inflammatory mediators. We have previously shown that LPS can induce endothelial apoptosis; thus, HMEC expressing an activated form of Notch4 were exposed to LPS to induce cell death (20). As seen in Fig. 1A, Notch4IC inhibited HMEC death in response to LPS, as measured by MTT assays.
LPS utilizes a mitochondria-dependent death pathway to induce apoptosis (20). We thus confirmed that endothelial cells were able to maintain mitochondrial integrity by confirming the ability of Notch4IC-expressing cells to retain their transmembrane potential. The cationic fluorophore, TMRE, partitions preferentially to the mitochondria. Loss of ⌬⌿ m results in a mitochondrial permeability transition, which can be detected by loss of TMRE fluorescence by flow cytometry. Fig. 1B demonstrates that activated Notch4 is able to maintain ⌬⌿ m over time in response to LPS stimulation. LPS interacts with Tolllike receptors to activate caspase 3/7 downstream of mitochondrial disruption (20). Determination of caspase 3/7 activity using the caspase 3/7-specific substrate, DEVD-pNA, showed that Notch4IC inhibited activation of caspase 3/7 in HMEC exposed to LPS (Fig. 1C).
Because HMEC are a transformed microvascular endothelial cell line, we tested whether Notch4IC also protected primary endothelial cells from apoptosis. HUVEC were transduced either with activated Notch4 linked by an internal ribosome entry site to YFP (MIY-Notch4IC), or with the empty vector as a control. The percentage of cells demonstrating phosphatidylserine exposure, as determined by annexin V binding, was used to quantitate apoptosis by flow cytometry (Fig. 2, right upper  quadrant). The proportion of apoptotic cells was determined by gating only on the YFP-positive HUVEC (Fig. 2, right upper  and lower quadrants). Fig. 2 shows the results of three such experiments, demonstrating that activated Notch4 inhibits apoptosis of primary endothelial cells in response to LPS stimulation.
The Ankyrin Repeats Are Required for the Notch4 Antiapoptotic Function-The Notch intracellular domain contains two major subdomains that have been implicated in binding to the downstream effector, RBP-J. These are the RAM domain at the NH 2 terminus and the ankyrin repeat region. To determine which of these domains are important in the antiapoptotic activity of Notch, deletion mutants were generated lacking each of these subdomains (Fig. 3A, ⌬RAM and ⌬Ank). Expression of these constructs was confirmed by immunoblotting (Fig.  3B) and immunofluorescent staining (Fig. 3C). Whereas the intact Notch4 intracellular domain and the ⌬Ank mutant localized mainly to the nucleus, the ⌬RAM mutant was expressed mainly in the cytoplasm, although a small amount of nuclear expression was also seen. The localization pattern of the ⌬RAM mutant was confirmed by subcellular fractionation studies (data not shown). Although there is some nuclear expression of this construct, the mainly cytoplasmic localization of the ⌬RAM mutant is consistent with the presence of an important nuclear localization signal in the region NH 2 -terminal to the ankyrin repeats as has previously been suggested (32). In order to ensure nuclear localization of the ⌬RAM mutant, an SV40 nuclear localization signal (NLS) was fused to the NH 2 terminus of the ⌬RAM construct (NLS-⌬RAM), and this construct showed mainly nuclear expression (Fig. 3C).
The various Notch4 mutants were tested for their ability to protect HMEC against LPS-initiated apoptosis. As seen in Fig.  3D, deletion of the ankyrin repeats abrogated the cytoprotective effect of Notch4. In contrast, the ⌬RAM mutant only partially lost cytoprotective activity. Because the ⌬RAM mutant exhibited a defect in nuclear entry, we tested whether the ⌬RAM mutant targeted to the nucleus would provide similar antiapoptotic activity to Notch4IC. Interestingly, the NLS-⌬RAM mutant was not able to protect endothelial cells to the same extent as Notch4IC, but rather this mutant showed a similar degree of protection as the nontargeted ⌬RAM mutant (Fig. 3E). Thus, increased nuclear expression is not sufficient for the ⌬RAM mutant to provide full cytoprotective activity. These findings indicate that the ankyrin repeats are essential for antiapoptotic function, and the partial protection conferred by the ⌬RAM mutant suggests that the RAM motif may signal one of multiple cytoprotective pathways induced by Notch4. Alternatively, The RAM domain may be required for "full" derepression and activation of its downstream effector RBP-J.
Only the Ankyrin Repeats Are Required for Notch4-activated RBP-J-dependent Signaling-Little is known about the down-

FIG. 2. Activated Notch4 inhibits apoptosis of primary endothelial cells.
Notch4IC-or vector-transduced HUVEC were exposed to LPS for 8 h followed by incubation with annexin V-PE. The proportion of YFP-positive, annexin V-positive cells was determined as a proportion of the total YFP population. Results represent the mean Ϯ S.E. of three experiments. stream pathways required for the antiapoptotic activity of Notch, and nothing is known regarding this function with respect to Notch4. Further, because of cell-specific signaling events, the critical downstream pathways activated by Notch members in endothelial cells remain to be defined. A major signaling pathway utilized by Notch homologues in other cell types involves derepression/activation of the transcriptional repressor RBP-J (2). Although the RAM domain was initially identified as a crucial region required for RBP-J-dependent Notch activity, others have not found this to be true (33,34). To test the ability of the various constructs to derepress/activate RBP-J, an RBP-J-dependent promoter construct fused to a luciferase reporter was used to assay for Notch4 activity. Transient transfections of the various cell lines with an RBP-J-dependent promoter-luciferase construct (Fig. 4A) demonstrate that the RBP-J-dependent promoter was activated equally in response to the wild-type Notch4 construct, the ⌬RAM mutant, or the NLS-⌬RAM mutant. This finding suggests that the Notch4 RAM motif is not required for RBP-J derepression/ activation in endothelial cells. However, the mutant lacking the ankyrin repeats was not able to activate the RBP-J promoter at all, confirming the essential requirement of this domain for Notch4 function (Fig. 4A).
RBP-J has been shown to promote transcription of downstream basic helix-loop-helix factors of the HES and HRT families (4). In particular, HRT2 has been shown to be expressed in endothelial cells and appears to play an important role in endothelial function (4). Thus, to confirm the ability of the ⌬RAM mutant Notch4 construct to activate an endogenous RBP-J-dependent promoter, RT-PCR of HRT2 from RNA of the mutant Notch4 cell lines was performed. These results confirmed that the ⌬Ank mutant was not able to activate the HRT2 promoter, whereas both the ⌬RAM and the NLS-⌬RAM mutants increased the expression of HRT2 mRNA to a similar extent as wild-type Notch4IC (Fig. 4B). Thus, our findings suggest that only the ankyrin repeats are necessary for Notch4 to signal through RBP-J, whereas the RAM domain is dispensable for this activity. Further, the finding that the ⌬RAM and the NLS-⌬RAM mutants showed similar protective activity is in keeping with the idea that minimal nuclear localization is sufficient for functional RBP-J activation by Notch.
Notch4 Inhibits Apoptosis through RBP-J-dependent and -independent Pathways-The NLS-⌬RAM mutant activates RBP-J-dependent promoters to a similar extent as the wildtype Notch4IC construct, yet this mutant only provides partial protection against apoptosis. Given these findings, we posited that Notch4 is able to protect endothelial cells against apoptosis through RBP-J-dependent and -independent pathways. To test whether blocking RBP-J activation would only partially inhibit Notch-driven antiapoptotic activity, we co-transduced a dominant negative RBP-J construct into Notch4IC endothelial cells. Although this mutant only partially blocked the Notch4 antiapoptotic function, we were also not able to completely block RBP-J-promoter activity using this approach (data not shown). Thus, this approach was not robust enough to determine whether complete inhibition of RBP-J would reveal a Notch-dependent, RBP-J-independent cytoprotective activity. We therefore attempted an alternative approach using a constitutively active RBP-J mutant. To generate a constitutively active RBP-J mutant, the

Notch4 Activation Inhibits Endothelial Apoptosis
VP16 activation domain was fused to the COOH terminus of RBP-J, and this chimeric construct (RBP-VP16) was transduced into HMEC. RBP-VP16 was able to activate the RBP-J-dependent promoter to the same extent as Notch4IC (Fig.  5A) but only partially protected HMEC against LPS-induced apoptosis (Fig. 5B). Taken together with the function of the ⌬RAM mutant, our findings demonstrate that activated Notch4 is able to inhibit endothelial apoptosis induced by LPS and that RBP-J-dependent and -independent signals are both required for full cytoprotective activity.
Notch4 Inhibits JNK Activation and Up-regulates Bcl-2 Expression via RBP-J-dependent and -independent Signals, Respectively-Notch has been shown to modulate JNK activity, and we have previously shown that inhibition or delay of JNK activation protects endothelial cells against LPS-induced apoptosis (20,35). We therefore tested whether activated Notch4 was able to inhibit JNK activation in endothelial cells, as one potential mechanism of its antiapoptotic activity. Fig. 6A demonstrates that Notch4 significantly attenuates JNK activation induced by LPS and that deletion of the RAM domain does not abrogate this function. In contrast, deletion of the ankyrin repeats does not inhibit LPS-stimulated JNK activation. Since the ⌬RAM mutant is able to activate RBP-J-dependent signaling, the above finding suggested that the constitutively active RBP-VP16 mutant would also potentially inhibit JNK activation. Fig. 6B demonstrates that RBP-VP16 does indeed block JNK activation by LPS, thus providing one mechanism for the Notch4-induced, RBP-J-dependent antiapoptotic activity.
Others have demonstrated that Notch1 induces the antiapoptotic protein Bcl-2 in some but not all T cell lines (36). We have seen that various Bcl-2 family members are able to protect endothelial cells against apoptosis induced by LPS or serum starvation (25,37). We thus tested whether Notch4 and the deletion mutants were able to induce Bcl-2 expression. Interestingly, only wild-type Notch4IC was able to up-regulate Bcl-2 expression, whereas neither mutant had any effect on Bcl-2 levels (Fig. 7A). Bcl-X L and Bax levels were unchanged by Notch4 activation (Fig.  7A). Given that deletion of the RAM domain did not affect Notch4-induced RBP-J activation, we theorized that Notch4 probably induces Bcl-2 expression via an RBP-J-independent pathway. As seen in Fig. 7B, constitutive activation of RBP-J was not sufficient to induce Bcl-2 in endothelial cells, thus confirming that Bcl-2 up-regulation by activated Notch4 occurs through an RBP-J-independent pathway. DISCUSSION Our findings show that activated Notch4 is able to inhibit endothelial apoptosis through multiple mechanisms and suggest that Notch activation plays a role in maintaining vascular stability. Indeed, studies by Taylor et al. (38) raise the notion that Notch activation may be required for the establishment of a mature, quiescent endothelial phenotype by down-regulating vascular endothelial growth factor receptor-2. Further, the prominent vascular defects observed in Notch-and Notch ligand-deficient mice suggest that inappropriate apoptosis may play a role in the observed phenotypes. Interestingly, the lack of Notch1 and/or Notch4 does not prevent endothelial differentiation in mice (10). The primary vascular plexus is laid down, but remodeling of this initial endothelial network does not take place (10). Thus, it is possible that Notch activation is required to maintain endothelial viability only in reorganizing or mature vasculature.
Notch plays a critical role in the determination of cell fate, and in many cases this is related to the regulation of apoptosis. Notch1 activation has been implicated both in promoting and in inhibiting cell death. In T lymphocytes and tumor cells, Notch1 shows antiapoptotic activity (30,39). Another study has suggested that activated Notch1 synergizes with papilloma virus oncogenes and inhibits apoptosis through the activation of phosphatidylinositol-3-kinase (40). In contrast, in chicken B lymphocytes and human monocytes, Notch1 activation has been reported to promote apoptosis (31,41).
In T cells, different studies suggest multiple modes of apoptosis inhibition by Notch1. On the one hand, Jehn et al. (39) isolated Notch1 in a yeast two-hybrid screen using Nur77 as bait. Notch1IC inhibited Nur77-dependent transcription and prevented T cell receptor-mediated but not glucocorticoid-dependent apoptosis in D011.10 cells (39). In contrast, Deftos et al. (36) have demonstrated that in the AKR1010 and 2B4.11 T cell lines, Notch1IC protected against glucocorticoid-triggered apoptosis. Interestingly, these investigators demonstrated upregulation of Bcl-2 in AKR1010, but not 2B4.11, cells, although Notch1IC protected both cell lines against death (36). The above findings demonstrate the cell type specificity of Notch signaling activity and highlight the multiple potential pathways of cytoprotection.
We used a Notch4IC mutant lacking the RAM domain to attempt to separate distinct pathways of Notch-mediated protection of endothelial cells. The ⌬RAM mutant showed partial cytoprotective activity despite the ability to activate RBP-Jdependent transcriptional activity to the same level as the intact Notch4IC. Nevertheless, the RAM domain alone is not sufficient to provide cytoprotective activity, since deletion of the ankyrin repeats abrogates all of the Notch-mediated antiapoptotic effect. Thus, the functionality provided by the RAM domain must require the cooperation of the ankyrin repeats. Interestingly, a constitutively active RBP-J mutant also displayed robust transcriptional activity but only partial antiapoptotic function. Taken together, the above findings would suggest that Notch4IC signals antiapoptotic activity that is both dependent and independent of RBP-J. Indeed, neither the ⌬RAM mutant nor the constitutively active RBP-J constructs were able to induce Bcl-2 expression, whereas both mutants inhibited JNK activation. One recognized Notch-mediated, RBP-J-independent signal is transmitted via Deltex (42,43). Interestingly, in the study by Deftos et al. (36), AKR1010 cells that up-regulated Bcl-2 in response to Notch1IC also induced Deltex expression. Although it was not reported whether the T cell line that lacked Bcl-2 induction (2B4.11) also lacked Deltex up-regulation, it is conceivable that the RBP-Jindependent induction of Bcl-2 is Deltex-dependent. In contrast, our data indicate that inhibition of JNK activation by Notch4 activation in endothelial cells is RBP-J-dependent.
Our data implicate Notch in inhibition of the intrinsic, mitochondria-directed apoptotic pathway. We have previously seen that enforced expression of Bcl-2 family members is sufficient to inhibit endothelial cell death in response to LPS as well as other apoptotic triggers (25,37). Similarly, inhibition of JNK activation using a dominant negative JNK mutant inhibited endothelial apoptosis in response to LPS stimulation or serum starvation (20,35). Both mechanisms of cytoprotection identified in this study are mitochondrially mediated. Bcl-2 is known to localize to the mitochondrial membrane and inhibit loss of mitochondrial transmembrane potential and release of cytochrome c (44). JNK promotes apoptosis by activating proapoptotic Bcl-2 members of the BH3 family and potentiating Bax-induced cell death (45)(46)(47). Thus, by inhibiting JNK and up-regulating Bcl-2, Notch signals a two-pronged antiapoptotic pathway that limits the intrinsic mitochondria-dependent death pathway. Although we did not specifically test the effect of Notch1IC in these studies, a previous study showed that Notch1IC delays endothelial death induced by serum starvation (9). Given that serum starvation-induced endothelial death is also dependent on JNK activation and is inhibited by Bcl-2, our findings suggest that the cytoprotective mechanisms described herein would explain the protection seen in the previous study (25,35). Our preliminary studies with tumor necrosis factor-induced endothelial apoptosis demonstrate that there is only a slight protective effect of Notch4IC. 2 The greater effect of Notch cytoprotective activity on LPS-induced apoptosis is consistent with the intrinsic pathway being critical for LPSbut not tumor necrosis factor-induced endothelial cell death, which is mainly mediated through the extrinsic death receptor pathway (20,37,48).
Our study thus demonstrates that activated Notch4 is able to inhibit multiple apoptotic pathways converging on the mitochondria. The Notch4 activity is separable into a pathway that is signaled by RBP-J and one that is independent of RBP-J. Endothelial cells are remarkably resistant to various apoptotic triggers in vivo, despite continual exposure to various toxins and proapoptotic cytokines (49). The findings reported in this paper implicate Notch activation in providing one mechanism to explain this capacity for survival.