αCP1 Mediates Stabilization of hTERT mRNA by Autocrine Human Growth Hormone*

We herein demonstrate that autocrine human growth hormone production in human mammary carcinoma cells results in increased telomerase activity as a result of specific up-regulation of telomerase catalytic subunit (human telomerase reverse transcriptase (hTERT)) mRNA and protein. This increase in hTERT gene expression is not due to increased transcriptional activation of the hTERT promoter but is the result of increased stability of hTERT mRNA exerted by CU-rich cis-regulatory sequences present in the 3′-untranslated region of TERT mRNA. Autocrine human growth hormone up-regulates two poly(C)-binding proteins, αCP1 and αCP2, which bind to these cis-regulatory elements and stabilize hTERT mRNA. We have therefore demonstrated that post-transcriptional modulation of the level of hTERT mRNA is one mechanism for regulation of cellular telomerase activity.

We herein demonstrate that autocrine human growth hormone production in human mammary carcinoma cells results in increased telomerase activity as a result of specific up-regulation of telomerase catalytic subunit (human telomerase reverse transcriptase (hTERT)) mRNA and protein. This increase in hTERT gene expression is not due to increased transcriptional activation of the hTERT promoter but is the result of increased stability of hTERT mRNA exerted by CU-rich cis-regulatory sequences present in the 3-untranslated region of TERT mRNA. Autocrine human growth hormone up-regulates two poly(C)-binding proteins, ␣CP1 and ␣CP2, which bind to these cis-regulatory elements and stabilize hTERT mRNA. We have therefore demonstrated that post-transcriptional modulation of the level of hTERT mRNA is one mechanism for regulation of cellular telomerase activity.
Chromosomal ends consist of small DNA repeats that are not duplicated faithfully during replication resulting in loss of DNA during each replication. To faithfully replicate chromosomal ends, cells use a reverse transcriptase holoenzyme termed the telomerase complex (1). The template human telomerase RNA (hTR) 3 and the reverse transcriptase catalytic subunit (hTERT) are considered to be the major components of the human telomerase complex, because the expression of hTERT and hTR alone is able to restore the majority of telomerase activity (2). Telomerase activity is subject to stringent regulation and is regulated by multiple mechanisms such as hTERT mRNA transcription (3), splicing and maturation (4), protein phosphorylation and transport (5), and subcellular localization of each component of the telomerase complex (6). Telomerase activity is minimal in most somatic cells resulting in chromosome shortening after each proliferative cycle eventually resulting in crisis and cell death (7). However, telomerase activity is readily detected in 90% of all tumor cells (8), in all tissues during early development (9), and in tissues with continuous proliferation, such as the hematopoietic system or the epidermis (10,11). mRNA turnover is one important mechanism by which gene expression is regulated (13). Specific interactions between sequences within the mRNA (cis-acting regulatory elements) and cellular RNA-binding proteins (trans-acting factors) regulate ribonuclease action and subsequent mRNA decay rates. The majority of these cis-acting elements are present in the 3Ј-untranslated region (3Ј-UTR) of the mRNA. cis-Regulatory elements such as AU-rich elements (12), Iron-responsive element (14), and CU-rich elements (15) within the 3Ј-UTR have been demonstrated to possess a role in mRNA stability. ␣-Globin mRNA poly(C)-rich segment-binding proteins (␣CPs), also referred to as poly(C)-binding proteins, and hnRNP K bind to CU-rich cis-regulatory elements (16). Two of the five members of poly(C)-binding proteins, ␣CP1 and ␣CP2, are highly homologous (16). Both ␣CP1 and ␣CP2 bind to pyrimidinerich elements in the 3Ј-UTR and stabilize a variety of mRNAs including ␣-globin and erythropoietin (15,17).
We have previously demonstrated that autocrine production of hGH in immortalized human mammary epithelial cells concomitantly enhances proliferation and offers protection from apoptosis, forming the basis for abnormal mammary acinar morphogenesis, oncogenic transformation, and tumor formation in vivo (18). Furthermore, autocrine production of hGH, in mammary carcinoma cells with epithelial morphology, promotes mesenchymal cellular morphology, increased cell migration, and increased metalloprotease activity with subsequent acquisition of invasive behavior both in vitro and in vivo (19). Given the described oncogenic potential of autocrine hGH and that both the hGH receptor (20) and telomerase (21) are expressed highly in the human mammary stem cell population compared with differentiated mammary epithelial cells, we have examined the potential regulation of telomerase activity by autocrine hGH.
Herein, we demonstrate that autocrine hGH increases telomerase activity. Autocrine hGH increased hTERT mRNA stability by up-regulation of two poly(C)-binding proteins, ␣CP1 and ␣CP2, which bind to specific cis-regulatory sequences in the 3Ј-UTR of hTERT mRNA. Regulation of post-transcriptional stability of hTERT mRNA is therefore one mechanism for cellular modulation of telomerase activity.

EXPERIMENTAL PROCEDURES
Cell Lines and Cell Culture-The MCF-7 cell line was obtained from ATCC. The MCF-7 cells were stably transfected either with wild type hGH gene (pMT-hGH) (designated MCF-hGH) or with a translation-deficient hGH (designated MCF-MUT) (55). The MCF-7 cells were cultured at 37°C in 5% CO 2 in RPMI supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Serum is a stimulator of hTERT transcription because of the presence of mitogens such as epidermal growth factor, insulinlike growth factor, and estrogens (24,28,29). We have utilized serum deprivation to minimize hTERT transcriptional activation in the experiments reported herein unless indicated otherwise.
Northern Blot Analysis-Total RNA was extracted using TRIzol reagent (Invitrogen) as described by the manufacturer. 30 g of total RNA were size-separated on a 1.2% denaturing agarose formaldehyde gel and blotted onto Hybond-N ϩ nylon membrane (Amersham Biosciences). The RT-PCR product of hTERT, ␤-actin, and luciferase were gel-purified and used as a template for generating the probes using a high prime DNA labeling kit (Roche Applied Science) according to the manufacturer's protocol. The blots were hybridized in ExpressHyb buffer (Clontech) at 60 -64°C using denatured hTERT or luciferase probe. The membranes were washed at high stringency (0.1ϫ SSC (0.015 M NaCl and 0.0015 M sodium citrate, pH 7.0) and 0.1% SDS) for 30 min at 65°C and exposed to x-ray film at Ϫ80°C for 1-7 days. The membranes were stripped in boiling 0.1% SDS and reblotted with the ␤-actin probe to measure the ␤-actin mRNA level, which acted as control.
mRNA Decay Measurements-To estimate the mRNA decay rates, transcription was inhibited by adding 5 g/ml actinomycin D in medium (56). RNA was extracted at the indicated times, and the endogenous hTERT and ␤-actin mRNA levels were analyzed by Northern blotting. The x-ray film was quantified by densitometer, and the ratio of hTERT to ␤-actin in each sample was calculated and used to determine the relative amount of specific mRNA remaining in each sample. Linear relationships were ensured by densitometric analysis of increasing amounts of RNA subject to Northern blot analysis and extending above the maximum densitometric reading used for data collection.
Western Blot Analysis and Confocal Laser Scanning Microscopy-The following primary antibodies were used. Rabbit anti-hTERT, rabbit anti-p-hnRNP K (Ser 302 ), goat anti-hnRNP K, goat anti-␣CP1, and goat anti-␣CP2 antibody were from Santa Cruz; rabbit anti-GH antibody was kindly provided by Dr. Parlow (National Institutes of Health NIDDK National Hormone and Peptide Program, Torrance, CA). Nuclear extracts or whole cell extracts were prepared, and Western blots were performed as previously described (38). For immunohistochemistry, the cells were transiently transfected with the indicated plasmids for 12 h and serum-starved for 48 h before fixing them for staining and scanning.
Transient Transfection and Luciferase Assay-The cells were cultured to 40 -60% confluence in six-well plates. Transient transfection was performed by use of Effectene (Qiagen) as described (38). Luciferase assays were performed as described previously (57). The results were normalized to the level of ␤-galactosidase activity and protein concentration in the samples.
Telomerase Assay-Telomerase activity was analyzed using TRAPeze telomerase detection kit (Intergen/Chemicon) according to the manufacturer's instructions. In brief, 100 g of cell extract protein from MCF-MUT or MCF-hGH in the presence or absence of 50 nM hGH was incubated with telomerase substrates, and the telomerase products were amplified by PCR. From 50 l of PCR product, 40 l were electrophoresed in a 12% polyacrylamide gel and visualized by ethidium bromide staining. 5 l of PCR product was quantified by ELISA according to the manufacturer's instructions.
RNA Electrophoretic Mobility Shift Assay (REMSA)-Cytoplasmic extracts were prepared as described (54). The cells were lysed in cytoplasmic extract buffer (10 mM HEPES, 3 M MgCl 2 , 40 mM KCl, 5% glycerol, 0.2% Nonidet P-40, 1 mM dithiothreitol), containing freshly added protease inhibitors mixture (Roche Applied Science). The lysates were cleared by centrifugation at 4°C for 20 min at 12,100 ϫ g, and the supernatant was utilized for electrophoretic mobility shift assay and UVXL assay. RNA probes were prepared from linearized vectors (XbaI-digested pCDNA-hTERT 3Ј-UTR: full length, 120 -554 and 360 -554 bp for probes 1, 3, and 4, respectively, and StyIdigested pCDNA-hTERT 3Ј-UTR-full length for probe 2) using T7 transcription kit (Fermentas). Binding reactions were performed with 20 g of cytoplasmic extract or 800 ng of recombinant protein and 10 5 cpm of in vitro transcribed 32 P-labeled RNA. Binding reactions were incubated at 22°C for 30 min, after which 2 units of RNase T1 (Roche Applied Science) was added for 10 min, followed by the addition of heparin (final concentration, 5 mg/ml) (Sigma) for 10 min. The complexes were separated by 7% native PAGE, visualized by exposing to x-ray film at Ϫ80°C for 1-3 days. In competition assays, the extracts were incubated with an unlabeled RNA transcripts (ϳ100-fold excess) or RNA homopolymers (10 ng/l-100 ng/l of final concentration) for 5 min prior to the addition of 32 P riboprobe. For supershift assay, 1-3-l antibodies (rabbit anti-␣CP1 and rabbit anti-␣CP2 used for supershift were a generous gift from Dr. Stephen A. Liebhaber, University of Pennsylvania) were added to the reaction mix after a step of RNase T1 treatment and incubated on ice for 1 h before adding heparin and run on 7% native PAGE.
UV Cross-linking Assay-For UVXL assay, following the incubation with heparin, the samples were placed on ice and UV-irradiated for 10 min (58). After UVXL, the samples were incubated with RNaseT1 at 37°C for 15 min. The samples were boiled for 3 min in SDS sample buffer and electrophoresed on 10% SDS-polyacrylamide gels and exposed to x-ray film.
Construction of siRNA to ␣CP1-We generated two human ␣CP1 siRNA constructs in pRNAT-CMV3.1/Hygro vector from Genscript targeting the sequences 5Ј-TCGACAAGCTG-GAGGAAGATA-3Ј and 5Ј-CTACTCGATTCAAGGACA-ACA-3Ј. 5 ϫ 10 4 MCF-7 cells were seeded into six-well plates and were cultured as above. They were transiently transfected with 1 g of the scrambled or the siRNA constructs and grown for an additional 18 h. The efficiency of the constructs to knock down ␣CP1 was tested by semi-quantitative RT-PCR and by immunostaining with ␣CP1. The construct with the sequence 5Ј-TCGACAAGCTGGAGGAAGATA-3Ј was found to be effective and was used herein.

RESULTS
Autocrine hGH Up-regulates Telomerase Activity in a JAK2dependent Manner-To examine the potential contribution of autocrine hGH to neoplastic progression of the human mammary epithelial cell and to determine the mechanism by which it results in oncogenic transformation, we examined the effect of autocrine hGH on telomerase activity in human mammary carcinoma cells. MCF7 cells were stably transfected with the wild type hGH gene (MCF-hGH) or a translation-deficient hGH gene where the translation initiation is disabled (MCF-MUT) as described previously (22). Telomerase activity was determined using the telomere repeat amplification protocol using the commercial kit from Intergen/Chemicon. In this method, telomerase in the cell extract adds a number of telomeric repeats onto the 3Ј end of a biotinylated telomerase substrate oligonucleotide, which is extended and then amplified by PCR. This extension and amplification generates a ladder of products that is observed by agarose gel electrophoresis and ELISA. As observed by both agarose gel electrophoresis and ELISA ( Fig. 1, a and b), MCF-hGH cells exhibit increased telomerase activity compared with MCF-MUT cells. The extracts were treated with RNase A to ensure that the PCR product is generated by RNA-dependent telomerase. Exogenously added hGH (50 nM) was without effect on telomerase activity. We also observed autocrine hGH, but not exogenous hGH, regulation of telomerase activity in a nontransformed, but immortalized, human mammary cell line, MCF-10A (data not shown).
The cellular effects of hGH are primarily mediated by stimulation of JAK2 activity (23). The increased telomerase activity observed in MCF-hGH cells as a consequence of autocrine hGH production was completely inhibited by a specific inhibitor of JAK2 (AG490) (Fig. 1c). Thus, autocrine hGH enhancement of human mammary carcinoma cell telomerase activity is JAK2-dependent.
Both hTERT mRNA and Protein Levels Are Up-regulated by Autocrine hGH-Telomerase activity is directly correlated to the expression level of hTERT (8). We therefore examined whether autocrine hGH production in mammary carcinoma cells resulted in up-regulation of hTERT mRNA by semi-quantitative RT-PCR and Northern blot analysis (Fig. 2, a and b) and hTERT protein by Western blot analysis (Fig. 2c). Autocrine hGH production in MCF-hGH cells resulted in increased hTERT mRNA and protein compared with MCF-MUT cells (Fig. 2, b and c). To date, regulation of hTERT mRNA has been demonstrated to be predominantly exerted at the transcriptional level (3). We therefore examined whether the autocrine hGH-stimulated increase in hTERT mRNA was due to increased transcriptional activation of the hTERT promoter by use of a reporter construct containing 3.328 kb of the promoter hTERT mRNA 3-UTR Stabilization 5Ј to the hTERT start site (24). Surprisingly, autocrine hGH did not increase reporter activity from the hTERT promoter (Fig.  2d). In contrast both serum (10%) (Fig. 2e) and 17␤-estradiol (10 nM) stimulation (data not shown) of MCF-7 cells resulted in activation of the hTERT promoter. The autocrine hGH-stimu-lated increase in hTERT mRNA was therefore not due to transcriptional regulation of the hTERT gene.
hGH Specifically Up-regulates hTERT mRNA-Human telomerase is a complex structure comprised of a core enzyme and a number of telomerase-associated proteins. Based on reconstitution experiments in Saccharomyces cerevisiae two essential subunits required for telomerase function are hTR, which serves as the template for telomeric DNA synthesis and the hTERT (25). However, fully functional telomerase complex comprises other components such as telomerase-associated protein 1 (TP1), Hsp90, p23, DKC1, STAU, RPL22, NOLA1, and SSB (26). Such additional components interact with telomerase and assist in proper ribonucleoprotein assembly and formation of an active telomerase enzyme complex (27). We therefore examined the possibility that autocrine hGH may regulate other components of the telomerase enzyme complex in addition to hTERT. The mRNA level of the different telomerase subunits from MCF-MUT and MCF-hGH cells were examined by semi-quantitative RT-PCR. Of the 11 subunits and associated proteins examined, only hTERT mRNA exhibited signifi-FIGURE 1. Autocrine hGH increases telomerase activity in a JAK2-dependent manner. MCF-MUT (control, where the ATG start site in the hGH gene for translational initiation is disabled) or MCF-hGH (wild type hGH gene) cells were serum-starved for 24 h and were cultured with or without 50 nM hGH and with or without 20 M AG490 for an additional 24 h. The protein extracts were analyzed by telomere repeat amplification protocol assay for telomerase activity using electrophoresis (a) or ELISA (b and c). Telomerase activity is increased upon expression of hGH (a). MCF-MUT and MCF-hGH cells were grown in serum-free medium supplemented with 20 M AG490 for 48 h and analyzed for telomerase activity (c). FIGURE 2. Autocrine hGH specifically increases hTERT mRNA and protein levels in human mammary carcinoma cells. mRNA expression of telomerase subunits and telomerase-associated proteins in MCF-MUT and MCF-hGH cells serum-deprived for 24 h and cultured with or without 50 nM hGH for an additional 24 h as determined by semi-quantitative RT-PCR (a). Most telomerase subunits and telomerase-associated proteins remain unchanged (hTR, TP1, Hsp90␣, Hsp90␤, STAU, DKC1, NOLA1, and P23) or decreased (RPL22 and SSB), whereas hTERT mRNA was increased by autocrine hGH. hTERT expression was also determined by Northern blot (b) and Western blot (c) analysis of MCF-MUT and MCF-hGH cells. hTERT promoter activity (3.328 kb) in MCF-MUT and MCF-hGH cells (d). e, 10% serum stimulation of MCF-7 cells resulted in activation of hTERT promoter, whereas hGH did not increase the reporter activity. The asterisk denotes significance.
Autocrine hGH Stabilizes hTERT mRNA-The lack of transcriptional regulation of hTERT by autocrine hGH is indicative that the autocrine hGH-stimulated increase in hTERT mRNA will be achieved by post-transcriptional regulation such as stabilization of hTERT mRNA. To examine this possibility we first utilized serum deprivation to minimize hTERT transcriptional activation. Serum is a stimulator of hTERT transcription because of the presence of mitogens such as epidermal growth factor, insulin-like growth factor, and estrogens (24,28,29). We quantified hTERT mRNA levels by Northern blot analysis in both MCF-MUT and MCF-hGH cells in different time points after serum deprivation. The level of hTERT mRNA in MCF-MUT cells decreased more than 2-fold after 6 h of serum deprivation. In comparison, the level of hTERT mRNA in MCF-hGH cells exhibited only a slight decrease, even after 12 h of serum deprivation (Fig. 3a). To verify this potential stabilization of hTERT mRNA by autocrine hGH, we inhibited mRNA transcription by use of actinomycin D and determined the halflife of hTERT mRNA in MCF-MUT and MCF-hGH cells. As presented in (Fig. 3b), the calculated half-life of hTERT mRNA in MCF-hGH cells was 10 h compared with a 6-h half-life for hTERT mRNA in MCF-MUT cells. Thus, autocrine hGH stabilizes hTERT mRNA and results in an increase in hTERT mRNA expression.
Autocrine hGH-stimulated hTERT mRNA Stability Is Due to cis-Acting Sequences in the 3Ј-UTR Region-3Ј-UTRs are most frequently linked with regulation of mRNA stability (30). Given that autocrine hGH promoted stabilization of hTERT mRNA, we next localized potential 3Ј-UTR cis-acting sequences responsible for autocrine hGHdependent hTERT mRNA stabilization. To this end, we generated a series of reporter plasmids in which the full-length (1-554 bp) and truncated (120 -554 and 360 -554 bp) versions of the 3Ј-UTR of hTERT mRNA were fused to the 3Ј end of the luciferase open reading frame and cloned into the pCDNA3 ϩ expression vector (Fig. 4a). MCF-MUT and MCF-hGH cells were transiently transfected with the indicated constructs, and the halflife of luciferase mRNA was estimated by Northern blot analysis after actinomycin D chase experiments. Luciferase mRNA lacking the surrogate 3Ј-UTR of hTERT exhibited a similar half-life of ϳ6 h in both MCF-MUT and MCF-hGH cells (Fig. 4b). In contrast, luciferase mRNA combined with the full-length hTERT 3Ј-UTR degraded with a half-life of approximately 5 h in MCF-MUT cells in contrast to MCF-hGH cells, where the half-life was prolonged for Ͼ10 h (Fig. 4c). The last 434 bp (nucleotide positions 120 -554) of hTERT 3Ј-UTR also conferred stability to the luciferase transcript in MCF-hGH cells (half-life Ͼ10 h) as compared with that of MCF-MUT cells (half-life ϳ5 h) (Fig. 4d), similar to that observed with the fulllength 3Ј-UTR. In contrast, the last 194 bp (position 360 -554) of hTERT 3Ј-UTR failed to provide the differential stability to luciferase mRNA between MCF-MUT and MCF-hGH cells with a half-life of 5 h in both cell lines (Fig. 4e). To further verify the differential stability of luciferase mRNA with the surrogate 3Ј-UTR of hTERT observed between MCF-MUT and MCF-hGH cells, we measured the luciferase activity of the various constructs in both cell lines under serum deprivation. Both MCF-MUT and MCF-hGH cells were transfected with the indicated constructs for 6 h followed by serum deprivation for 48 h Ϯ 20 M AG490, and the luciferase activity was determined. Consistent with that observed by Northern blot analysis after actinomycin D chase, luciferase activity was higher by ϳ2-fold in MCF-hGH cells compared with MCF-MUT cells when transfected with the luciferase vectors containing either the full-length or the 120 -554-bp fragment of hTERT 3Ј-UTR. Concordantly, there was no difference in luciferase activity between MCF-hGH and MCF-MUT cells when transfected with the control luciferase vector or the luciferase vector containing the 360 -554-bp fragment of hTERT 3Ј-UTR. Further, FIGURE 3. hTERT mRNA is stabilized upon hGH expression. a, MCF-MUT and MCF-hGH cells were serumdeprived as indicated. Total RNA was isolated, and the levels of hTERT and ␤-actin mRNAs were determined by Northern blot analysis. The ratio of hTERT to ␤-actin in each sample was calculated by densitometric analysis and the relative amount of specific mRNA remaining in each sample was determined. The graph shows the means Ϯ S.D. of three independent experiments. b, MCF-MUT and MCF-hGH cells were grown in serum-free medium with or without 5 g/ml actinomycin D, and the levels of hTERT and ␤-actin mRNAs were determined by Northern blot analysis. The ratio of hTERT to ␤-actin in each sample was calculated by densitometric analysis, and the relative amount of specific mRNA remaining in each sample was determined. The graph shows the means Ϯ S.D. of three independent experiments.

hTERT mRNA 3-UTR Stabilization
the JAK2 inhibitor AG490 prevented the hTERT 3Ј-UTR mediated up-regulation of luciferase activity observed in MCF-hGH cells with the full-length and the 120 -554-bp fragment of the hTERT 3Ј-UTR (Fig. 4f). We conclude that the cis-acting regulatory elements contained within nucleotide positions 120 -360 bp of the 3Ј-UTR of hTERT are responsible for the autocrine hGHdependent stabilization of hTERT mRNA.

cis-Acting Regulatory Elements from the hTERT 3Ј-UTR Region Are Bound by Multiple RNA-binding
Proteins-We next proceeded to identify the protein(s) binding to the potential cis-acting regulatory region in the 3Ј-UTR region of hTERT mRNA. We performed REMSA to determine the formation of RNA-protein complexes by use of different regions of the hTERT 3Ј-UTR as probes (Fig. 5a). Binding reactions were incubated at 22°C for 30 min, after which 2 units of RNase T1 (Roche Applied Science) was added for 10 min, followed by the addition of heparin (final concentration, 5 mg/ml) (Sigma) for 10 min for all of the REMSA experiments. When REMSA was performed with cytoplasmic extract from MCF-hGH cells using the fulllength hTERT 3Ј-UTR, we observed the formation of five complexes (Fig. 5b). The uppermost complex was observed with all hTERT 3Ј-UTR probes, despite no overlap, and was considered to be nonspecific. The other four bands (designated as RNA-protein binding complexes (RPC) 1-4) were considered specific because they were observed only with the full-length and 120 -554-bp hTERT 3Ј-UTR probes and were absent when the 3Ј 120-bp (360 -554 bp fragment) of hTERT 3Ј-UTR was used as probe. Also, when the 5Ј 201-bp (1-201-bp fragment) of hTERT 3Ј-UTR was used as probe, no shift bands were seen. Thus, the region between nucleotides 201-360 of the hTERT 3Ј-UTR is a potential cis-regulatory region and is bound by multiple RNA binding proteins. We therefore next performed REMSA with cytoplasmic extracts from MCF-MUT and MCF-hGH cells using the full-length hTERT 3Ј-UTR as probe to determine whether autocrine hGH would alter the binding of proteins to the hTERT 3Ј-UTR. All four RPCs were present at higher intensity in MCF-hGH cell extract compared with that of the MCF-MUT cell extract, whereas the nonspecific band exhibited no difference between these two extracts (Fig. 5c,  3). A 100 molar excess of unlabeled probe incorporating the nucleotide region between 120 and 554 bp of the hTERT 3Ј-UTR completely abrogated the RNA-protein complex formation on the full-length hTERT 3Ј-UTR (Fig. 5c). Thus, autocrine hGH promotes the formation of a complex between cytoplasmic proteins and the 3Ј-UTR of hTERT.
Having determined that the hTERT 3Ј-UTR contained cisregulatory elements, we examined for potential regulatory sequences by searching the UTR data base (31). We observed multiple CU-rich regions within the 201-554-bp fragment of the 3Ј-UTR (Fig. 6a). These sequences were entirely homologous with the sequences demonstrated to bind poly(C)-binding proteins involved in the regulation of mRNA stability (15). We therefore first determined whether the four RPCs we observed indeed contain poly(C)-binding proteins. We competed the formation of RPCs with an excess of RNA homopolymers. As observed in Fig. 5c (lanes 5 and 6) both low (10 ng/l) and high (100 ng/l) concentration of poly(C) completely abrogated the formation of RPCs1-4. The poly(U) homopolymer partially abrogated the formation of RPC1 but did not affect formation of RPCs2-4. Poly(A) homopolymer had no effect on the four RPCs indicative that proteins binding to the hTERT 3Ј-UTR are bona fide poly(C)-binding proteins.
Poly(C)-binding Proteins ␣CP1 and ␣CP2 Bind the hTERT 3Ј-UTR cis-Acting Regulatory Region-There exist two identified groups of mammalian poly(C)-binding proteins, the hnRNPs K (66 kDa) and the ␣CP proteins (ϳ39kDa) (16). To identify which of these poly(C)-binding proteins bind to the 3Ј-UTR of hTERT mRNA and maintain its stability, we performed an UVXL experiment to assess the molecular masses of proteins binding to hTERT 3Ј-UTR using the full-length (1-554 bp) 3Ј-UTR as a probe. Multiple, distinct RNA-protein complexes with approximate molecular masses of 110, 80, and 45 kDa were observed in cytosolic extracts of MCF-hGH cells. The predominant RNA-protein complex had a molecular mass of 45 kDa, which is closer to the molecular mass of ␣CP proteins, which is 39 kDa (Fig. 6b, lane 2). This indicated that ␣CPs may be the predominant RNA-binding proteins interacting with the hTERT 3Ј-UTR. The possibility that ␣CP proteins may bind to the hTERT-3Ј-UTR was determined by repetition of the UVXL experiment with recombinant GST-␣CP1 and GST-␣CP2 proteins. Although both GST-␣CP1 and GST-␣CP2 bound to the hTERT 3Ј-UTR (Fig. 6b, lanes 3 and 4), the binding observed with GST-␣CP2 was minimal in comparison. We further confirmed the binding of the ␣CP1 protein to the 3Ј-UTR of hTERT by REMSA. GST and GST-␣CP1 proteins were analyzed with full-length hTERT 3Ј-UTR as the probe for RNA-protein complex formation. No RNA-protein complex was formed with GST, but GST-␣CP1 (Fig. 6c) formed complexes with hTERT 3Ј-UTR. The formation of this complex was completely abrogated by excess unlabeled poly(C) and repressed partially by poly(U) but not by poly(A) homopolymers (Fig. 6c). Similar data were observed with GST-␣CP2 (data not shown). Furthermore, ␣CP1 and ␣CP2 antibodies each separately formed supershifted complexes from the hTERT 3Ј-UTR full-length (1-554-bp 3Ј-UTR) probe. As observed in Fig. 6d, both ␣CP1 and ␣CP2 antibodies produced a prominent supershift; ␣CP1 antibody significantly reduced RPC2 and 3, whereas ␣CP2 antibody significantly reduced RPC4. An antibody to hnRNP K was without effect on the protein complexes binding to the hTERT 3Ј-UTR. Thus, we conclude that ␣CP1 and ␣CP2 are hTERT 3Ј-UTR-binding proteins.  lanes 7 and 8), and poly(A) (lanes 9 and 10). *, nonspecific band.

hTERT mRNA 3-UTR Stabilization
Poly(C)-binding Proteins ␣CP1 and ␣CP2 Stabilize hTERT mRNA-␣CPs have been demonstrated to regulate mRNA stability by binding to the 3Ј-UTR of various mRNAs, such as ␣-globin (15), collagen ␣1 (32), and erythropoietin (17). We therefore examined the effect of ␣CP1 and ␣CP2 on hTERT mRNA stability. We forced the expression ␣CP1 and ␣CP2 proteins in MCF-7 cells by transient transfection and assessed the half-life of hTERT mRNA by Northern blot analysis. Forced expression of ␣CP1 or ␣CP2 extended the half-life of hTERT mRNA (ϳ9 h) compared with that of vector transfected control (ϳ5 h) after an actinomycin D chase. Thus, both ␣CP1 and ␣CP2 separately stabilized hTERT mRNA (Fig. 7, a and b). We have further verified these results by reporter assay using the luciferase-hTERT 3Ј-UTR chimeric RNA. MCF-7 cells were co-transfected with either control vector or ␣CP1 or ␣CP2 expression vectors together with the luciferase reporter containing the surrogate hTERT 3Ј-UTR. Forced expression of either ␣CP1 or ␣CP2 increased luciferase activity 3-fold when compared with the vector control. In contrast, no change in luciferase activity was observed between control or ␣CP1or ␣CP2-expressing cells co-transfected with luciferase cDNA without hTERT 3Ј-UTR (Fig.  7c). Therefore, ␣CP1 and ␣CP2 regulate hTERT mRNA stability through the hTERT 3Ј-UTR.
Autocrine hGH Regulates the Poly(C)-binding Proteins ␣CP1 and ␣CP2-We next examined the regulatory interactions between autocrine hGH and the poly(C)-binding proteins ␣CP1, ␣CP2 as well as hnRNP K. The mRNA and protein levels of ␣CP1, ␣CP2, and hnRNP K from MCF-MUT and MCF-hGH cells were analyzed by Northern and Western blotting (Fig. 8, a and b). The levels of ␣CP1 predominantly but also minimally ␣CP2 mRNA and protein were increased in MCF-hGH cells expressing autocrine hGH compared with MCF-MUT cells (Fig. 8, a and b). However, autocrine hGH exerted no effect on either the mRNA or the protein level of hnRNP K (Fig. 8, a and b). Because Ser 302 phosphorylation of hnRNP K has been demonstrated to decrease its RNA binding activity (33), we also examined Ser 302 phosphorylation of hnRNP K in MCF-MUT and MCF-hGH cells by use of a Ser 302 phosphorylation-specific hnRNP K antibody. Autocrine hGH did not alter the serine phosphorylation of hnRNP K (Fig. 8b). We also verified the protein expression of ␣CP1 and ␣CP2 by confocal laser scanning microscopy after transient transfection of MCF-7 cells with an hGH expression plasmid. In cells with expression of hGH, we observed an increased expression of both ␣CP1 and ␣CP2 (Fig. 8c). The increased ␣CP1 expression observed in MCF-hGH cells as a consequence of autocrine hGH production was abrogated by a specific inhibitor of JAK2 (AG490) (Fig. 8d). Thus, autocrine hGH regulates the expression of ␣CP1 and ␣CP2.
␣CP1 Mediates Autocrine hGH-stimulated hTERT mRNA Stability-To determine whether ␣CP1 mediated the autocrine hGH-stimulated increase in hTERT mRNA stability, we generated two siRNA constructs to ␣CP1. Transient transfection of MCF-7 cells with either the scrambled or the ␣CP1 siRNAs demonstrated that the siRNA construct 5Ј-TCGACAAGCTG-GAGGAAGATA-3Ј abrogated ␣CP1 mRNA expression in MCF-7 cells by more than 70% (data not shown). Using this construct we subsequently determined whether ␣CP1 mediates autocrine hGH-stimulated hTERT mRNA stabilization by hTERT mRNA 3-UTR Stabilization depletion of the autocrine hGH-stimulated increase in ␣CP1 using semi-quantitative RT-PCR, immunohistochemistry, as well as luciferase assay. MCF-MUT and MCF-hGH cells were transfected with either the scrambled or ␣CP1 siRNA construct, and the mRNA levels were analyzed by semi-quantitative RT-PCR, and the protein expression was analyzed by confocal laser scanning microscopic analysis. For confocal microscopy the transfected cells were identified by the positive staining for GFP, which is present in the siRNA vector. Use of the ␣CP1 siRNA abrogated the autocrine hGH-stimulated increase in ␣CP1 mRNA and protein (Fig. 9, a and b). MCF-7 cells were subsequently co-transfected with either hGH expression plasmid or control vector, with or without ␣CP1 siRNA, together with the luciferase reporter containing the surrogate hTERT 3Ј-UTR. Forced expression of hGH increased luciferase activity 3.5-fold when compared with the vector control. Depletion of endogenous ␣CP1 with the ␣CP1 siRNA prevented this autocrine hGH-stimulated increase in luciferase reporter activity (Fig. 9c). Therefore, ␣CP1 mediates the hGH-stimulated increase in mRNA stability conferred by the 3Ј-UTR of hTERT. We conclude that autocrine hGH up-regulation of ␣CP1 protein stabilizes hTERT mRNA via its 3Ј-UTR.

DISCUSSION
Herein we have demonstrated another mechanism by which autocrine hGH in human mammary carcinoma cells regulates FIGURE 7. Poly(C)-binding proteins ␣CP1 and ␣CP2 stabilize the hTERT mRNA. a and b, MCF-7 cells were transfected transiently with the vector or with ␣CP1 and ␣CP2 expression plasmids, and after 6 h the culture was divided into four. Each subculture was treated with 5 g/ml actinomycin D for the indicated times. Total RNA was isolated, and the levels of hTERT and ␤-actin mRNAs were determined by Northern blotting. The ratio of hTERT to ␤-actin in each sample was calculated by densitometry, and the relative amount of specific mRNA remaining in each sample was determined. The graphs show the means Ϯ S.D. of three independent experiments. c, MCF-7 cells were co-transfected with either control, ␣CP1, or ␣CP2 expression plasmids and with the wild type luciferase or luciferase conjugated to hTERT 3Ј-UTR reporters. Luciferase activity was determined as described under "Experimental Procedures." FIGURE 8. Autocrine hGH up-regulates ␣CPs and stabilizes hTERT mRNA via its 3-UTR. a and b, MCF-MUT and MCF-hGH cells were serum deprived for 48 h, and the hnRNP K, ␣CP1, and ␣CP2 mRNA and protein levels were assayed by Northern (a) and Western blot (b) analysis. Serine phosphorylation of hnRNP K protein in MCF-MUT and MCF-hGH cells was also determined. c, MCF-7 cells were transiently transfected with hGH and were stained for hGH (green) and ␣CP1 (red) or hGH (green) and ␣CP2 (red) expression. There is an up-regulation of ␣CP1 and ␣CP2 in the cytoplasm when there is overexpression of hGH. d, the level of ␣CP1 mRNA expression in MCF-hGH and MCF-MUT cells determined by semi-quantitative RT-PCR after treatment with JAK2 inhibitor AG490. There is a decrease in the level of ␣CP1 mRNA after treatment with JAK2 inhibitor AG490. ␤-Actin served as loading control. telomerase activity. It is shown that specific growth hormonereleasing hormone (GHRH) antagonists decrease telomerase activity in a glioblastoma cell line (34). It is of interest to note that the GHRH antagonist-mediated decrease in telomerase activity is also associated with down-regulation of hTERT mRNA and without change in the level of hTR or TP1 mRNA as is observed with autocrine hGH in this study. GHRH antagonists have previously been demonstrated to inhibit autocrine GH production in MXT mouse mammary xenografts (35). Autocrine GHRH in the mammary epithelial cell may therefore stimulate telomerase activity by regulation of autocrine hGH production. Although it is apparent that hTERT is hormonally regulated, the hGH-mediated increase in hTERT gene expression is the result of increased stability of hTERT mRNA exerted by CU-rich cis-regulatory sequences present in the 3Ј-UTR of TERT mRNA. We note here that autocrine and not exogenously added hGH regulates telomerase activity. Specific prooncogenic actions of autocrine hGh have previously been described, and mechanisms relating to the specific effect of autocrine hGH have been delineated (18,19).
Telomerase has been reported to possess cellular functions in addition to its role in maintaining the length of the telomere (36). Evidence indicates that abrogation of telomerase activity inhibits cell proliferation via pathways that are not dependent on telomere erosion (37), and it is possible that autocrine hGH regulation of hTERT mRNA stability may participate in these processes. Indeed, autocrine hGH has been demonstrated to promote both proliferation (22) and survival (38) of human mammary carcinoma cells. In this regard it is interesting that uterine endometrium displays altered telomerase activity in a menstrual cycle-dependent manner despite its somatic origin (41). hGH has also been detected in endometrial epithelial cells (39). Thus, it may be possible that autocrine hGH participates in the regulation of telomerase activity required for temporal specific or cyclical proliferative activity. Concordant with this hypothesis, autocrine GH is expressed in the morula and blastocyst before implantation (40), and we have also detected maximal autocrine GH expression in mouse mammary epithelial cells at puberty (42). Furthermore, the expression of the GH receptor is enriched in adult stem cell populations derived from mammary, hematopoietic, and nervous systems (43). Increased expression of autocrine hGH in mammary neoplasia (44) and consequent up-regulation of telomerase activity may therefore simply be a recapitulation of the embryonic or stem cell phenotypic effect of autocrine hGH.
The 3Ј-UTR of hTERT was replete with CU-rich elements that bind with poly(C)-binding proteins (15). Two distinct subsets of poly(C)-binding proteins are found in mammalian cells, hnRNPs K/J and the ␣CP proteins (␣-complex proteins) (45). The precise mechanism by which these proteins bind mRNA and result in its stabilization is not clear, although they are proposed to protect the endoribonuclease-sensitive site from cleavage by the sequence-specific endoribonuclease (46). Alternatively they are proposed to favor the binding of poly(A)-binding proteins to the poly(A) tail of mRNA (47). All poly(C)-binding proteins are characterized by the presence of three KH (hnRNP K homology) domains that bind to CU-rich RNA sequences (16). Use of REMSA herein demonstrated that all four RPCs formed on hTERT 3Ј-UTR are abolished in the presence of unlabeled poly(C) homopolymers suggesting that all contain poly(C)-binding proteins. Poly(C)-binding proteins may exist as homodimers or oligomerize with other proteins (48). RPCs 1 and 3 contained ␣CP1, whereas RPC4 contained ␣CP2. Based on the molecular masses, there must be other proteins present in RPC1. The identity of other proteins in RPC1 and their functional importance for hTERT mRNA stability remain to be determined.
Regulation of mRNA stability by hGH has hereto not been reported. However, the role of mRNA stability in control of gene expression by other hormones has been demonstrated. For example, estrogen increases the stability of vitellogen mRNA by 30-fold (49), and the hGH related hormone, prolactin, stabilizes casein mRNA by 20-fold (50). In contrast, luteinizing hormone decreases the stability of hCGR mRNA (51). Autocrine hGH-stimulated stabilization of hTERT mRNA reported herein is indicative that mRNA stability will be one mechanism by which hGH regulates gene expression. It is therefore likely that other mRNAs utilizing ␣CP1 and ␣CP2 for stabilization will also be modulated by autocrine hGH.
Creation of a tumor cell from a normal cell requires both immortalization and oncogenic transformation (52). Forced expression of hTERT in primary human mammary epithelial FIGURE 9. ␣CP1 mediates autocrine hGH-stimulated hTERT mRNA stability. The level of ␣CP1 mRNA and protein expression in MCF-hGH and MCF-MUT cells determined by semi-quantitative RT-PCR (a) or by confocal laser scanning microscopy (b), after transient transfection of scrambled or ␣CP1 siRNA. For RT-PCR, ␤-actin served as loading control. For confocal microscopy, the transfected cells were identified by the positive staining to GFP, which is present in the siRNA vector (Vect). c, MCF-7 cells were transiently transfected with the indicated constructs in serum-free medium, and the luciferase activity was determined as described under "Experimental Procedures. " The asterisk denotes significance (p Ͻ 0.01).
hTERT mRNA 3-UTR Stabilization cells (HMECs) has been demonstrated to result in their immortalization (53), although the addition of both insulin and epidermal growth factor were required to maintain proliferation. Immortalization is not sufficient to create an oncogenically transformed cell (54). Together with the fact that the oncogenic transformation in immortalized mammary epithelial cells can be achieved by simple forced expression of the hGH gene (18), the autocrine hGH regulation of telomerase activity described herein is therefore an indication that autocrine hGH may constitute a complete oncogene for the HMEC. However, it remains to be determined whether the autocrine hGH-stimulated increase in telomerase activity will be sufficient for immortalization and allow for subsequent oncogenic transformation.
In conclusion, we have demonstrated that autocrine hGH regulates telomerase activity in human mammary carcinoma cells by increasing hTERT mRNA stability, an effect mediated by binding of poly(C)-binding proteins ␣CP1 and ␣CP2 to specific cis-regulatory sequences in the 3Ј-UTR of hTERT mRNA.