Catecholamine stimulation induces cleavage and phosphorylation of Her2
When we treated human breast cancer cell line BT474 with β-AR agonist isoproterenol (ISO), an extra ~ 80 kDa molecular mass was accumulated in a time-dependent manner as detected by Western blot with the antibody against the C-terminus of Her2 (Fig. 1a). Similar result was also obtained in human ovarian cancer cell line SKOV3 (Fig. 1b), suggesting that the 80 kDa fragment may represent Her2 CTF and appearance of the 80 kDa Her2 fragment (p80Her2) was associated with ISO stimulation. To determine whether the generation of p80Her2 is mediated by activation of β-AR, we used β-AR activators including ISO (5 µM) and naturally occurring catecholamines, epinephrine (10 µM) and norepinephrine (10 µM), to treat human breast cancer cell line MCF-7. As shown in Fig. 1c, both full-length Her2 and p80Her2 were remarkably increased. Pretreatment with the specific inhibitor of β2-AR ICI 118,551 (1 µM) strikingly impaired the effect of ISO on the formation of p80Her2, whereas the specific inhibitor of β1-AR ATEN (1 µM) had only marginal effect (Fig. 1d), indicating that activation of β2-AR is a prerequisite for the generation of p80Her2.
P95Her2 was assumed to contain the transmembrane and cytoplasmic domains (24). To clarify whether p80Her2 derives from proteolysis or alternative initiation of translation, we constructed MCF-7 cells expressing Her2-GFP fusion protein (MCF-7/Her2-GFP). Western blot analysis with the antibody against GFP showed that a new product, whose size perfectly fits the molecular weight of p80Her2-GFP fusion protein (~ 100 kDa), appeared after exposure of the cells to ISO (Fig. 1e), testifying that the p80Her2 fragment is a product cleaved from full-length Her2. We examined the soluble Her2 ECD in the cultural supernatant of SKOV3 cells by using the antibody against the N-terminus of Her2. Concomitant with markedly increased p80Her2 in the whole cell lysates, an about 100 kDa protein was clearly detected in the cultural supernatant of the cells (Fig. 1f). Notably, p80Her2 was apparently smaller than the membrane-associated p95Her2. We hypothesized that p95Her2 may undergo the secondary cleavage. Moreover, the phosphorylation of both full-length Her2 and p80Her2 was prominently enhanced in a time-dependent manner, indicating a rapid posttranslational modification of Her2 following ISO stimulation (Fig. 1g).
Catecholamine modulates the cleavage of Her2 ECD by promoting ADAM10 expression through downregulation of miR-199a-5p and upregulation of SIRT1
The ectodomain cleavage of the transmembrane proteins is generally mediated by membrane-associated metalloproteases under the regulation of multiple signaling pathways such as the activation of PKC (30, 31). Earlier studies showed that Her2 ECD shedding could be suppressed by the broad-spectrum metalloprotease inhibitors TAPI, batimastat and the tissue inhibitor of metalloproteases-1 (24). A previous study by utilizing the siRNAs selectively inhibiting ADAM10 expression suggested that ADAM10 may be one of the proteases responsible for Her2 cleavage (32). However, shedding of Her2 is inefficient in contrast to the majority of shedding events. In addition, how the sheddase is controlled under physiological conditions is unclear. To test whether ADAM10 is engaged in catecholamine-induced Her2 ECD cleavage, we examined the expression of ADAM10 in MCF-7 cells stably transfected with Her2 expression plasmid (MCF-7/Her2). We found that ISO stimulation induced a significant upregulation of ADAM10 expression, which was obviously coherent with the accumulation of p80Her2 (Fig. 2a). A recent study indicates that the NAD-dependent deacetylase SIRT1 regulates the transcription of the gene encoding ADAM10 by direct interaction with ADAM10 promoter (33). The data in Fig. 2b showed that epinephrine stimulation dramatically promoted the expression of SIRT1 in a time-dependent manner in MDA453 and SKOV3 cells. SIRT1 was recently identified as a direct target of miR-199a-5p (34). Interestingly, the expression of miR-199a-5p was strikingly repressed in SKOV3 and MDA453 cells treated with epinephrine as determined by real-time RT-PCR analysis (Fig. 2c). The experimental study has demonstrated that β2-AR can activate an antiapoptotic signal through Gi-dependent coupling to phosphatidylinositol 3'-kinase/Akt pathway and the activated Akt is sufficient for inducing downregulation of miR-199a-5p in cardiac myocytes (35). We noticed that the phosphorylation of Akt was significantly enhanced by epinephrine stimulation, accompanied with reduction of miR-199a-5p in MDA453 and SKOV3 cells (Fig. 2b and 2c). Furthermore, the epinephrine-induced shedding of Her2 by ADAM10 was completely inhibited by ADAM10 specific siRNA (Fig. 2d), verifying that catecholamine modulates the cleavage of Her2 ECD by promoting ADAM10 expression through downregulation of miR-199a-5p and thus upregulation of SIRT1. It has been known that the stimulation of β2-AR with the agonists leads to the shedding of heparin-binding EGF-like growth factor by ADAM17, which is also the major sheddase for Her3 and Her4 (36, 37), and subsequent activation of EGFR in an autocrine/paracrine manner (4, 5). The findings in this study unambiguously demonstrate that ADAM10 activities induced by catecholamine stimulation mediate cleavage and activation of Her2 tyrosine kinase, indicating that transactivation of Her2 is mediated by β2-AR through a different mechanism.
γ-secretase activity induced by catecholamine-stimulation is responsible for the generation of p80Her2 ICD
Generation of CTFs from a transmembrane receptor involves cleavage within the transmembrane domain by γ-secretase-catalyzed proteolytic processing, whereas the activity of γ-secretase is proposed to be regulated by the ADAM-mediated ECD cleavage of transmembrane receptor (38). Soluble Her4 ICD is produced by the sequential activities of ADAM-17 and γ-secretase after binding of its ligand heregulin or activation of protein kinase C (PKC) by 12-O-tetradecanoylphorbol-13-acetate (39). However, proteolytic cleavage of Her2 by γ-secretase has not been reported. As mentioned previously, p80Her2, the product of Her2 cleavage triggered by catecholamine stimulation is somehow smaller than p95Her2. It was recently found that activation of β2-AR provoked γ-secretase activity in HEK293 cells (40). We speculated that the membrane-associated Her2 ICD may undergo the secondary cleavage after shedding of Her2 ECD, possibly by γ-secretase. The expression of γ-secretase components, presenilin 1 (PS1), PS2, nicastrin and PEN-2 was present in MCF-7, MCF-7/Her2 and SKOV3 cells (Fig. 3a). When we treated MCF-7 and SKOV3 cells with ISO, γ-secretase activities, determined by fluorogenic substrate assay, were prominently elevated in both cell lines (Fig. 3b and 3c). In comparison with the known γ-secretase substrates CD44, Notch and E-cadherin for homology of the amino acid sequences of the transmembrane domain Her2 dose not share high homology with these sequences. However, these transmembrane domains including that of Her2 frequently harbor several valine residues, some of which have been identified as the potential γ-secretase cleavage sites (Fig. 3d). To investigate the role of γ-secretase in Her2 intramembranous processing, we utilized the selective γ-secretase inhibitor L685,458 and a dipeptidic γ-secretase specific inhibitor N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester (DAPT). Pretreatment with L685,458 or DAPT significantly blocked the production of p80Her2 in a concentration-dependent manner (Fig. 3e). The polytopic membrane proteins PS1 and PS2 are suggested to be the catalytic components of an active γ-secretase complex. We employed the specific siRNAs targeting PS1 and PS2 to transfect SKOV3 cells. Knock-down of either PS1 or PS2 alone or both extraordinarily prevented the formation of p80Her2 (Fig. 3f and Fig. S1), demonstrating that γ-secretase activity induced by catecholamine-stimulation is responsible for the generation of p80Her2 ICD.
Catecholamine stimulation mediates nuclear translocation of Her2 ICD efficiently
Nuclear translocation of Her2 as a full-length molecule was investigated in certain cell lines (18). In contrast to other Her family members, the nuclear entry of Her2 was much less efficient. The regulatory mechanisms of nuclear translocation of full-length Her2 receptor under physiological conditions remain elusive. To illustrate the subcellular distribution of Her2, we traced the intracellular trafficking of GFP-tagged Her2 after treating MCF-7/Her2-GFP cells with ISO. In unstimulated cells ectopic overexpressed Her2-GFP was defined at the cytoplasmic membrane (Fig. 4a). No evidence for Her2-GFP in the nucleus was found. However, in the presence of ISO, nuclear Her2 was readily visualized (Fig. 4b). In addition, Her2-GFP proteins were also observed in the intracellular membrane systems. By immunofluorescence using the monoclonal antibodies against both N- and C-terminus of Her2 we observed that Her2 molecules were predominantly located at the cytoplasmic membrane and no substantial nuclear Her2 could be detected with either antibody in unstimulated SKBR3 breast cancer cells (Fig. 4b and 4c). The data are consistent with several previous observations that Her2 did not localize in the nucleus of several breast cancer cell lines spontaneously (16). However, after treatment with ISO for 60–90 min, Her2 was massively migrated into the nuclei (Fig. 4b). Notably, nuclear Her2 could be easily detected by the antibody against the C-terminus of Her2 (Fig. 4b) but not by the antibody against the N-terminus of Her2 (Fig. 4c). Consistent with these data, p80Her2 was detected in the nuclei of ISO-simulated SKOV3 cells by cellular fractionation and Western blot (Fig. 4d) but not in the unstimulated cells, suggesting that it was Her2 ICD liberated by γ-secretase cleavage that entered the nuclei. Moreover, the features for uniform distribution of nuclear Her2 was distinguishingly different from those in the previous observation that Her2 in the nuclear appeared as discrete punctate spots under conventional culture conditions. Altogether, these results indicate that catecholamine stimulation mediates nuclear translocation of Her2 ICD efficiently.
Her2 ICD physically binds to the promoter of COX2 gene and drives transactivation of COX2 gene
Sequential cleavage of transmembrane receptors can rapidly transform membrane-associated proteins into soluble effectors, which enter the nucleus and regulate the transcription of their target genes (31). To determine the functional significance of p80Her2 in the nucleus, we isolated the nuclear proteins from SKOV3 cells and performed oligonucleotide pull-down assay. A previous study identified the Her2-associated sequence (HAS), which was located at 1750 nucleotides upstream from the transcriptional initiation site in COX-2, a known target gene of Her2 (18). We utilized the sequence as an oligonucleotide probe. By oligonucleotide pull-down assay we could reproducibly detect the association of p80Her2, but not p185Her2, with the oligonucleotide probes containing the HAS sequence in ISO-simulated cells. Identical Western blot with the mutated HAS sequence or nonspecific oligonucleotides exhibited no specific signal (Fig. 5a). In addition, the binding of p80Her2 with the HAS sequence was strongly impaired or utterly abolished by the specific competitors in a dose-dependent manner. Furthermore, the expression of COX-2 at both mRNA and protein levels was remarkably upregulated by ISO stimulation (Fig. 5b and 5c). These data provide further evidence to confirm that under catecholamine stimulation Her2 ICD migrates into the nucleus, physically binds to the promoter of COX2 gene and drives transactivation of COX2 gene.
Catecholamine stimulation strongly promotes the invasive activities of cancer cells in vitro and spontaneous tumor lung metastasis in mice
In an effort to determine the effects of catecholamine stimulation on the biological behaviors of tumor cells and relevance of Her2 nuclear localization in tumor development and metastasis, we investigated whether catecholamine stimulation confers proliferation and invasion potential to Her2-overexpressing tumor cells in a human ovarian cancer xenograft model. The treatment with ISO daily did not accelerate tumor growth in mice, compared with the control group (Fig. S2). However, ISO stimulation significantly promoted the invasive capacity of SKOV3 cells in in vitro invasion assay using Matrigel invasion chamber and caused spontaneous lung metastasis in nude mice (Fig. 6a, 6b and Fig. S3). Metastatic colonization was observed by gross examination and microscopic inspection of tissue sections (Fig. 6b and Fig. S3). Surprisingly, the nuclear staining of Her2 was mainly observed in the metastatic tumor tissues by immunohistochemical labeling with the antibody against the C-terminus of Her2 (Fig. 6c), whereas nuclear Her2 was rarely seen in the primary tumors (Fig. S4). Moreover, using the antibody against the N-terminus of Her2 nuclear Her2 could not be detected, but only membrane-anchored Her2 was signaled (Fig. 6c).
To gain further insights into correlation of Her2 nuclear localization with β2-AR activation, we examined the expression of both Her2 and β2-AR in 55 Her2 positive human breast cancer tissues. In the tissues expressing relatively low level of β2-AR, Her2 molecules were predominantly distributed at the cytoplasmic membrane (Fig. 6d, case 1). Nevertheless, in the tissues expressing high level of β2-AR, nuclear Her2 was strongly positive (Fig. 6d, case 2). In 62.6% of the tissues with strong immunoreactivity for anti-β2-AR antibody (20/32) nuclear Her2 was positive. Only 13% of tissues expressing low level of β2-AR (3/23) displayed positive staining of nuclear Her2 (Table S1). The difference between two groups was highly significant (p < 0.0002). However, the kappa coefficient for β2-AR expression and Her2 nuclear localization was moderate (0.46), suggesting that other molecular mechanisms may be involved in the nuclear translocation of Her2. Simultaneous staining of β2-AR and nuclear Her2 was also observed in human ovarian cancer tissues (Fig. S5). The data demonstrate that nuclear localization of Her2 is intimately associated with the overexpression of β2-AR, activation of β2-AR-mediated signaling pathway and breast cancer metastasis.