Heregulin-induced apoptosis is mediated by down-regulation of Bcl-2 and activation of caspase-7 and is potentiated by impairment of protein kinase C alpha activity

Heregulins are a group of growth factors that play diverse and critical roles in the signaling network of the human epidermal growth factor receptor (HER or EGFR) superfamily. Our earlier studies have shown that recombinant heregulinbeta1 (HRG) induces apoptosis in SKBr3 breast cancer cells that overexpress HER2. Here we report molecular mechanisms of HRG-induced apoptosis. HRG treatment of SKBr3 cells for 72 h decreased the level of Bcl-2 protein. HRG treatment led to degradation of poly (ADP-ribose) polymerase (PARP) and activated both caspase-9 and caspase-7. No significant activation of caspase-3, -6, or -8 was detected. Expression of exogenous caspase-7 by adenovirus-caspase-7 (Ad-casp-7) in SKBr3 cells resulted in apoptosis, which mimicked the effect of HRG treatment. Expression of exogenous caspase-7 had no impact on Bcl-2 expression, but promoted PARP degradation. Two highly selective inhibitors of protein kinase C (PKC), GF109203X (GF) and Ro318425 (Ro), significantly enhanced HRG-induced apoptosis as determined by flow cytometric analysis and DNA fragmentation assay. Accordingly, the PKC inhibitor GF further decreased the level of Bcl-2 protein and further degraded PARP in HRG-treated cells. Assay of PKC activity indicated that HRG activated PKC in SKBr3 cells, predominantly affecting the PKCalpha isoform. To confirm which PKC isoform(s) mediated potentiation of HRG-induced apoptosis, the profile of PKC isoforms was measured in SKBr3 cells. Five PKC isoforms, PKCalpha, PKCiota, PKCzeta, PKClambda, and PKCdelta as well as their receptors (RACK1) were expressed in this cell line. Treatment with PKC inhibitors GF and Ro decreased protein levels of both PKCalpha and PKCdelta at 24 h. PKCalpha levels were still depressed at 72 h. GF and Ro had little effect on the expression of other PKC isoforms. An inhibitor of classical PKC isoforms (Go6976) enhanced HRG-induced apoptosis, whereas the PKCdelta selective inhibitor rottlerin did not. As PKCalpha was the only classical isoform expressed in SKBr3 cells, the effect of Go6976 on HRG-induced apoptosis largely related to inhibition of PKCalpha. Constitutive expression of wild-type PKCalpha attenuated the apoptosis produced by HRG and GF. Consequently, HRG-induced apoptosis in SKBr3 cells appeared to involve down-regulation of Bcl-2 protein, activation of caspase-9 and caspase-7, and degradation of PARP. Inhibition of PKC function enhanced HRG-induced apoptosis, leading to synergistic down-regulation of Bcl-2 expression. Impairment of the PKCalpha isoform alone was sufficient to potentiate HRG-induced apoptosis.

Differential signaling by an anti-p185(HER2) antibody and heregulin

To understand the molecular mechanisms by which anti-p185HER2 antibody and the ligand heregulin inhibit tumor growth, we have investigated several signaling proteins and pathways. We report here that anti-p185HER2 monoclonal antibody ID5 induced tyrosine phosphorylation of HER2 in SKBr3 breast cancer cells that overexpress p185HER2. Heregulin beta1 induced phosphorylation of both HER3 and HER2. ID5 produced a greater association of phospholipase C (PLC)-gamma1 with HER2 than did heregulin. Concordantly, ID5, but not heregulin, increased PLC-gamma1 activity. However, the G1 cell cycle arrest and induction of p27Kip1 produced by ID5 were not affected by the inhibition of PLC-gamma. ID5 preferentially induced binding of the Mr 46,000 isoform of SHC to HER2, whereas heregulin preferentially induced binding of the Mr 52,00 isoform of SHC to HER3. Heregulin, but not ID5, induced the p85 subunit of phosphatidylinositol 3'-kinase (PI3-K) to interact with HER3. Heregulin induced sustained activation of P13-K signaling, whereas ID5 had only a transient effect. Heregulin, but not ID5, activated the c-Jun-NH2-terminal kinase cascade. Pretreatment of SKBr3 cells with ID5 decreased heregulin-induced association of HER2 with HER3 as well as the activation of c-Jun-NH2-terminal kinase and PI3-K activities. Inhibition of the mitogen-activated protein kinase pathway in SKBr3 cells did not affect heregulin-induced G2-M-phase arrest, apoptosis, and differentiation. Heregulin-induced apoptosis could be blocked by inhibition of p70s6k, but not by inhibition of PI3-K. Heregulin-induced differentiation could be eliminated by inhibition of PI3-K. We conclude that ID5 and heregulin signal via different pathways, although both agents can inhibit the clonogenic growth of cells that overexpress HER2.

Anti-HER2 antibody and heregulin suppress growth of HER2-overexpressing human breast cancer cells through different mechanisms

Previous reports have shown that certain anti-HER2 antibodies and heregulin can inhibit clonogenic growth of breast and ovarian cancers that overexpress HER2. Anti-HER2 antibodies bind to HER2 directly, whereas heregulin does not bind to HER2 alone, but rather interacts with HER2 through the formation of heterodimers with HER3 or HER4. The purpose of the present study was to elucidate the mechanisms by which anti-HER2 antibody and heregulin inhibit tumor growth. The anti-HER2 monoclonal antibody (mAb) ID5 was found to block G1-S progression of the cell cycle, whereas heregulin inhibited passage through G2-M. Compatible with the effects on the cell cycle, treatment with mAb ID5 decreased levels of cyclin-dependent kinase (CDK) 2, cyclin E, and CDK6 proteins and reduced cyclin E-CDK2-associated kinase activity; mAb HD5-treated cells had increased p27Kip1 expression and an increased association of p27Kip1 with CDK2. In contrast, treatment with heregulin increased protein levels of CDK2, CDK6, CDC2, and cyclin B1. More Retinoblastoma protein was found in the hypophosphorylated state in the cells treated with mAb ID5, whereas more retinoblastoma protein was in the hyperphosphorylated state in heregulin-treated cells. Heregulin was able to induce cell differentiation as assessed by Oil Red O staining and apoptosis as assessed by sub-G1 peak on flow cytometry and the presence of DNA fragmentation in ApopTag histochemistry staining. Neither differentiation nor apoptosis was observed in the cells treated with mAb ID5. We conclude that anti-HER-2 mAb ID5 and heregulin exert growth inhibition through different mechanisms. In mammary cells overexpressing HER2, anti-HER2 mAb ID5 induces G1 arrest, whereas heregulin induces G2-M arrest, cell differentiation, and apoptosis.

The outcome of heregulin-induced activation of ovarian cancer cells depends on the relative levels of HER-2 and HER-3 expression

Members of the epidermal growth factor receptor family of tyrosine kinases, including epidermal growth factor receptor, c-erbB-2 (HER-2), c-erbB-3 (HER-3), and c-erbB-4 (HER-4), can be coexpressed at different levels in nonhematopoietic tissues. Amplification and overexpression of HER-2 is found in approximately one-third of cancers that arise in the breast and ovary. In our previous studies, heregulin (HRG) and anti-HER-2 antibodies inhibited proliferation, increased invasiveness, and enhanced tyrosine autophosphorylation of SKBr3 breast cancer cells that overexpressed HER-2. In the present report, the effects of HRG and anti-HER-2 antibody have been compared in six ovarian cancer cell lines. HRG inhibited anchorage-independent growth of SKOv3 cells that overexpressed HER-2 (10(5) receptors/cell) but stimulated the growth of OVCA420, OVCA429, OVCA432, OVCA433, and OVCAR-3 cells that expressed lower levels of the receptor (10(4) receptors/cell). Thus, cell lines with a high level of HER-2 relative to HER-3 or HER-4 were growth inhibited, whereas cell lines with lower levels of HER-2 were growth stimulated by HRG. Stimulation or inhibition of clonogenic growth did not correlate with endogenous expression of HRG or with the impact of exogenous HRG on phosphorylation of HER-2, HER-3, or HER-4. Anti-HER-2 antibodies inhibited the growth of SKOv3 cells but failed to affect the growth of the other cell lines. In OVCAR-3 cells that had been transfected with HER-2 cDNA to increase expression to 10(5) receptors/cell, HRG inhibited rather than stimulated growth. Conversely, when HER-2 expression by SKOv3 cells was downregulated by transfection of the viral E1A gene, HRG stimulated rather than inhibited growth. To evaluate the relative importance of HER-3 and HER-4, NIH 3T3 cells were cotransfected with HER-2 and HER-3 or with HER-2 and HER-4. HRG inhibited the growth of cells with a high ratio of HER-2:HER-3, whereas HRG stimulated the growth of cells with low levels of the two receptors. In cells that express only HER-2 and HER-4, HRG stimulated the growth of cells that expressed HER-4 independent of HER-2 levels. Anti-HER-2 antibodies inhibited the growth of transfectants with high levels of HER-2 expression independent of HER-3 or HER-4 expression. In ovarian cancer cells that express all three receptors, the relative levels of HER-2 and HER-3 appear to determine the response to HRG. Taken together, these studies support the concept that the level of HER-2 expression can modulate response to HRG, determining whether the response is stimulatory or inhibitory. In contrast, agonistic antibodies that bind to HER-2 alone inhibit anchorage-independent growth but fail to mimic HRG's ability to stimulate growth of cells with low HER-2: HER-3 ratios.

Regulation of AML2/CBFA3 in hematopoietic cells through the retinoic acid receptor alpha-dependent signaling pathway

AML2 is a member of the acute myelogenous leukemia, AML family of transcription factors. The biologic functions of AML1 and AML3 have been well characterized; however, the functional role of AML2 remains unknown. In this study, we found that AML2 protein expressed predominantly in cells of hematopoietic origin is a nuclear serine phosphoprotein associated with the nuclear matrix, and its expression is not cell cycle-related. In HL-60 cells AML2 expression can be induced by all three natural retinoids, all-trans-retinoic acid (RA), 13-cis-RA, and 9-cis-RA in a dose-dependent manner. A synthetic retinoic acid derivative, 4HPR, which neither activates RA receptor (RAR) alpha nor retinoic X receptor alpha was unable to induce the expression of AML2. A RAR-selective activator, TTNPB, induced AML2 expression similar to RA. Our study further showed that AGN193109, a potent RARalpha antagonist, suppressed AML2 expression induced by RA and that a retinoic X receptor pan agonist AGN194204 had no effect on its expression. Taken together, these studies conclusively demonstrated that the expression of AML2 in HL-60 cells is regulated through the RARalpha-specific signaling pathway. Our study further showed that after all-trans-retinoic acid priming, AML2 expression could be augmented by vitamin D(3). Based on these studies we hypothesize that AML2 expression is normally regulated by retinoid/vitamin D nuclear receptors mainly through the RARalpha-dependent signaling pathway and that it may play a role in hematopoietic cell differentiation.

Recombinant PML adenovirus suppresses growth and tumorigenicity of human breast cancer cells by inducing G1 cell cycle arrest and apoptosis

Our previous studies demonstrated that the promyelocytic leukemia gene, PML which involved in the 15;17 translocation in acute promyelocytic leukemia (APL) is a growth and transformation suppressor. In this study, recombinant PML adenovirus, Ad-PML was constructed and used to infect human breast cancer cells in vitro and in vivo, the anti-oncogenic function of PML and its mechanism of growth suppressing effect in breast cancer cells were examined. We showed that Ad-PML effectively infected the MCF-7 and SK-BR-3 cells. A high level of PML protein was expressed within 24 h post-infection and a detectable level remained at day 16. Ad-PML significantly suppressed the growth rate, clonogenicity, and tumorigenicity of breast cancer cells. Intratumoral injections of MCF-7-induced tumors by high titer Ad-PML suppressed tumor growth in nude mice by about 80%. The injection sites expressed high level of PML and associated with a massive apoptotic cell death. To elucidate the molecular mechanism of PML's growth suppressing function, we examined the effect of Ad-PML on cell cycle distribution in MCF-7 and SK-BR-3 cells. We found that Ad-PML infection caused a cell cycle arrest at the G1 phase. We further showed that G1 arrest of MCF-7 cells is associated with a significant decrease in cyclin D1 and CDK2. An increased expression of p53, p21 and cyclin E was found. The Rb protein became predominantly hypophosphorylated 48 h post-infection. These findings indicate that PML exerts its growth suppressing effects by modulating several key G1 regulatory proteins. Our study provides important insight into the mechanism of tumor suppressing function of PML and suggests a potential application of Ad-PML in human cancer gene therapy.

Characterization of the ETO and AML1-ETO proteins involved in 8;21 translocation in acute myelogenous leukemia

The AML1 and ETO genes are disrupted by the nonrandom chromosomal translocation t(8;21) in acute myelogenous leukemia (AML). While the AML1 gene encodes a transcription factor indispensable for definitive hematopoiesis, the biological function of ETO is unknown. To understand the role of ETO and AML1-ETO in the pathogenesis of AML, the full length cDNAs of ETO and AML1-ETO were cloned and antibodies against AML1 and ETO proteins have been developed in our laboratory. Western blot analysis showed that ETO and AML1-ETO were identified as 70 kDa and 94 kDa proteins, respectively, and that both proteins, like AML1, were associated with the nuclear matrix. To examine whether the t(8;21)-positive AMLs expressed a 94-kDa AML1-ETO, protein fractions isolated from leukemia blasts of 10 patients with t(8;21)-positive AML and the Kasumi-1 cells were analyzed by Western blotting. The 94 kDa AML1-ETO fusion protein was detected in all samples. However, this fusion protein was not detectable in all 40 patients with t(8;21)-negative AMLs. The biological significance of AML1-ETO was examined in K562 cells, which stably overexpress AML1-ETO. We found that AML1-ETO blocked the erythroid differentiation of K562 cells induced by low doses of Ara-C. Thus, t(8;21)-positive AMLs appear to overexpress the AML1-ETO fusion protein, which may be responsible for differentiation block and leukemogenesis in AML.

Stable overexpression of PML alters regulation of cell cycle progression in HeLa cells

Our previous studies demonstrated that PML is a growth suppressor that suppresses oncogenic transformation of NIH/3T3 cells and rat embryo fibroblasts. PML is a nuclear matrix-associated phosphoprotein whose expression is regulated during the cell cycle. Disruption of PML function by t(15;17) in acute promyelocytic leukemia (APL) plays a critical role in leukemogenesis. To further study the role of PML in the control of cell growth, we have stably overexpressed PML protein in the HeLa cell line. This overexpression of PML significantly reduced the growth rate of HeLa cells and suppressed anchorage-independent growth in soft agar. We consequently investigated several parameters correlated with cell growth and cell cycle progression. We found that, in comparison with the parental HeLa cells, HeLa/PML stable clones showed proportionally more cells in G1 phase, fewer cells in S phase and about the same number in G2/M phase. The HeLa/PML clones showed a significantly longer doubling time as a result of a lengthening of the G1 phase. No effect on apoptosis was found in HeLa cells overexpressing PML. This observation indicates that PML suppresses cell growth by increasing cell cycle duration as a result of G1 elongation. To further understand the mechanism of the effect of PML on HeLa cells, expression of cell cycle-related proteins in HeLa/PML and parental HeLa cells was analyzed. We found that Rb phosphorylation was significantly reduced in PML stable clones. Expression of cyclin E, Cdk2 and p27 proteins was also significantly reduced. These studies indicate that PML affects cell cycle progression by mediating expression of several key proteins that normally control cell cycle progression. These results further extend our current understanding of PML function in human cells and its important role in cell cycle regulation.

The biologic function of PML and its role in acute promyelocytic leukemia

Patients with acute promyelocytic leukemia (APL) are characterized by the presence of a t(15;17) chromosomal translocation. The fusion protein PML-RAR alpha encoded from the breakpoint can form a heterodimer and acts as a dominant negative inhibitor against the normal function of PML. Recently we demonstrated that PML is a growth suppressor and transcription suppressor expressed in all cell lines tested. We also found that PML suppresses the clonogenicity and tumorigenicity of APL-derived NB4 cells, as well as the transformation of rat embryo fibroblasts by cooperative oncogenes and NIH/3T3 by neu. Overexpression of PML in human tumor cell lines induces a remarkable reduction in growth rate in vitro and in vivo. More recently, we have shown that PML is a phosphoprotein associated with the nuclear matrix and that its expression is cell cycle related. PML expression is altered during human oncogenesis, implying that PML may be an anti-oncogene involved not only in APL but also in other oncogenic events. Mutation analysis of the functional domains of PML demonstrated that its ability to form PML nuclear bodies or PODs (PML oncogenic domains) is essential for suppressing growth and transformation. In light of the above studies it appears that disruption of the normal function of PML plays a critical role in the pathogenesis of APL.

Analysis of the growth and transformation suppressor domains of promyelocytic leukemia gene, PML

The promyelocytic leukemia gene (PML) involved in the t(15;17) (q22;q12) translocation in acute promyelocytic leukemia is a growth suppressor. To elucidate the functional domains of PML, several mutants lacking the nuclear localization signal (PMLnls-), the dimerization domain (PMLdim-), the proline-rich domain at the N-terminal (PMLpro-), the proline-rich RING finger motif (PMLpr-), the proline-rich RING finger B-box-1 (PML-prb-), the serine-proline-rich domain at the C-terminal (PMLsp-), and the double mutant (PMLprb-nls-) have been constructed. Immunofluorescence staining of transiently transfected NIH3T3 cells demonstrated that the RING finger motif, dimerization domain, and nuclear localization signal are all required for the formation of PML oncogenic domains (PODs). Immunofluorescence staining of transiently transfected GM637D human fibroblasts indicated that expression of PMLprb-, PM-Lnls-, and PMLprb-nls- led to a significant reduction or, in some cases, complete elimination of PODs. PMLdim-, PMLnls-, PMLpr-, PMLprb-, and PMLprb-nls- mutants were found to lose their ability to suppress transformation of NIH3T3 cells by activated neu, while PMLpro- and PMLsp- mutants did not. These results suggest that the ability of PML to form a POD is essential for suppression of growth and transformation. Furthermore, since PMLprb-, PMLnls-, and PMLprb-nls- mutants could block the suppression effect of wild-type PML on transformation of NIH3T3 cells by the neu oncogene, these PML mutants are potential dominant negative inhibitors of PML. Our study also suggests that the RING finger motif may interact with other nuclear proteins.

Characteristics of paroxysmal nocturnal hemoglobinuria in China. Clinical analysis of 476 cases

476 cases of paroxysmal nocturnal hemoglobinuria (PNH) were analysed and compared with PNH patients in foreign countries and its classification was discussed. The clinical characteristics in Chinese PNH patients were: males were affected more; the age of onset was young; more common in north China; hemoglobinuria as an initial symptom was less common; thrombosis as a complication was relatively rare; the common causes of death were heart failure, infections and hemorrhage; and bone marrow picture might show mild dyshematopoiesis.