To determine if a relationship exists between HER2 and cyclin E, we analyzed data from a cohort of 395 breast cancer patients that were originally studied to determine the relationship between total (full-length + LMW) cyclin E expression and survival (Keyomarsi et al., 2002). No patients in this cohort received trastuzumab therapy. Five-year DSS rates were significantly worse in patients with HER2-positive tumors (P < .0001) (Figure 1a). We stratified patients with HER2-positive tumors (n = 117) by total cyclin E levels and found those with high total cyclin E (n = 59) had a 5-year DSS of 14% compared to 89% for those with low total cyclin E (n = 58; P < .0001; Figure 1b). When stratified by LMW cyclin E levels, patients with high LMW cyclin E (n = 50) had a 5-year DSS of 10%; compared with 82% for patients with low LMW cyclin E (n = 67; P < .0001; Figure 1c). These data demonstrate that overexpression of both HER2 and cyclin E, particularly LMW cyclin E, contributes to an aggressive phenotype of breast cancer.

Activation of HER2 signaling promotes cellular proliferation by shortening the G1 phase of the cell cycle (Timms et al., 2002). Similarly, overexpression of cyclin E shortens the length of the G1 phase. Therefore, we hypothesized that HER2 overexpression may modulate cyclin E expression or activity. To this end, we transfected the HER2-overexpressing breast cancer cell line MCF-7-HER-18, with HER2 siRNA. MCF-7-HER-18 has exogenous HER2 overexpression established by stable transduction of the parental cell line, MCF7. MCF-7-HER-18 expresses both full-length and LMW cyclin E (Figure 2a). The HER2 siRNA was effective in knocking down HER2 expression (Figure 2b) and this was accompanied by a decrease in cyclin E expression (Figure 2c).

We repeated the transfections using a second HER2 siRNA and an additional cell line, SKBr3, a breast cancer cell line with endogenous HER2 overexpression that also expresses full-length and LMW cyclin E (Figure 2a). Immunofluorescence confocal microscopy revealed that transfection with HER2 siRNA decreased HER2 (Figure 2d) and cyclin E (Figure 2e) expression. Furthermore, western blot analysis confirmed a decrease in HER2 expression and revealed the decrease in cyclin E was primarily due to decreased expression of LMW isoforms (Figure 2f).

To determine if a feedback loop exists between HER2 and cyclin E, we investigated HER2 expression in MCF-7 cells engineered to overexpress full-length or LMW cyclin E. There were no changes in HER2 expression or its phosphorylated form, suggesting that cyclin E lies downstream of HER2 (Supplementary Figure 1).

To assess the effect of HER2 signaling on cyclin E activity, we altered HER2 levels using siRNA then examined cyclin E-associated kinase activity. HER2 downregulation decreased cyclin E–associated kinase activity (Figure 3a) which was more pronounced in SKBr3 cells (64% decrease in cyclin E-associated kinase activity) than in MCF-7-HER-18 cells which have higher baseline levels of LMW cyclin E (51% decrease). These data suggest that the extent of decrease in cyclin E-associated kinase activity is related to baseline levels of LMW cyclin E expression.

Since the functionality of cyclin E is crucial in cell cycle progression, we explored the effect of HER2 downregulation and subsequent decrease in LMW cyclin E expression on proliferation. MTT assays were used to assess cell viability as a marker of proliferation (Figure 3b). After transfection with HER2 siRNA, we noted a pronounced decrease in proliferation. Because decreased proliferation can be due to an increase in apoptosis or cell cycle arrest, we assessed both of these. As demonstrated in figure 3c, part of the decrease in proliferation was due to an increase in apoptosis as evidenced by an increased percentage of cells in the sub-G1 phase (Figure 3c). In three duplicate experiments, the average percent ± SEM in sub-G1 in MCF-7-HER-18 cells was 1.69%±0.58 in control cells versus 22.08%±3.95 in HER2 siRNA transfected cells (11.65%±0.33 in HER2 siRNA #2 transfected cells). In SKBr3 the average percent ± SEM in sub-G1 was 3.11%±0.17 in control cells versus 12.48%±1.17 in HER2 siRNA transfected cells ((12.45%±0.10 in HER2 siRNA #2 transfected cells). The increase in apoptosis after transfection with HER2 siRNA was confirmed using an annexin V assay (Supplementary Figure 2). Cell cycle analysis demonstrated a concomitant increase in the percentage of cells in G1 phase of the cell cycle in MCF-7-HER-18 cells using both siRNA sequences. In SKBr3 cells, a significant increase was seen in the average percent in G1 using HER2 siRNA #1 but not using siRNA #2 (Figure 3d). Taken together, these data suggest that the decrease in proliferation seen after HER2 downregulation is multifactorial with both an increase in apoptosis and alterations in cell cycle distribution contributing with the increase in apoptosis being the more consistent, pronounced effect.

To determine if HER2 siRNA–mediated downregulation of cyclin E is transcriptionally regulated, RNA extracted from cells transfected with HER2 siRNA was subjected to qRT-PCR, after which comparative quantitation analysis was performed. In three duplicate experiments performed on MCF-7-HER-18 cells, the mean cyclin E/β-actin mRNA ratios were .006, .009 (P = .18), and .013 (P = .18) for controls, cells transfected with random sequence siRNA, and cells transfected with HER2 siRNA, respectively (Supplementary Table 1a). Experiments were repeated using SKBr3 cells, and mean cyclin E/β-actin ratios were .157, .288 (P = .46), and .128 (P = .24) (Supplementary Table 1b). These data suggest that HER2 does not regulate cyclin E transcription. This supports our finding that the effect of HER2 is greatest on LMW isoforms, which are the result of post-translational modification of the full-length protein.

Based on these results, we hypothesized that cyclin E lies downstream of HER2-mediated signaling and that decreased activity through these pathways effects cyclin E expression and activity. Using immunofluorescence confocal microscopy to assess cyclin E expression, we found a dose-dependent decrease in cyclin E expression in MCF-7-HER-18 and SKBr3 cells treated with different doses of trastuzumab. In particular, cyclin E expression in MCF-7-HER-18 and SKBr3 cells treated with 20 µg/mL trastuzumab decreased by 86.5% and 86.8%, respectively, compared to untreated cells (Figure 4a). Western blot confirmed the decrease was primarily in the LMW isoforms (Figure 4b). Similar to the effect seen after transfection with HER2 siRNA, decreased HER2-mediated signaling due to trastuzumab treatment resulted in decreased cyclin E-associated kinase activity (Figure 4c). Consistent with results from the kinase assays performed on HER2 siRNA-transfected cell lysates (Figure 3a), this effect was most pronounced in the SKBr3 cell line (60% decrease versus 26% decrease in MCF-7-HER-18 cells).

To better determine the effects of decreased HER2-mediated signaling on the G1 checkpoint, we assessed the effects of trastuzumab on cyclin D1, and the CDK inhibitors p21 and p27. Consistent with other reports (Lee et al., 2000) that cyclin D1 expression increases when cells are transfected with HER2, we found that trastuzumab treatment decreased cyclin D1 levels (Figure 4b). We did not find appreciable change in p21 or p27 expression (Figure 4b); however, confocal immunofluorescence microscopy revealed increased nuclear localization of both (Figure 4d).

Because one proposed mechanism of action of trastuzumab is decreased proliferation (Bacus et al., 1992; Brockhoff et al., 2007;), we assessed cell viability as a marker of proliferation using MTT assays in cells treated with trastuzumab and found that proliferation decreased in a dose-dependent manner. The number of viable MCF-7-HER-18 cells decreased by 56.8% after treatment with 20 µg/mL of trastuzumab, compared to untreated cells, and the number of viable SKBr3 cells decreased by 67% (Supplemental Figure 3a). Concomitant with the decreased proliferation was an increased percentage of cells in G1 phase (Supplementary Figure 3b). The mean percentage increase was greater in SKBr3 cells, which have endogenous HER2 overexpression. These data suggest that breast cancer cells with endogenous HER2 overexpression depend on mitogenic signaling through HER2 pathways to increase cellular proliferation. In contrast, MCF-7-HER-18 cells that were transfected to exogenously express HER2 may have additional oncogenic pathways affecting their rate of cellular proliferation.

Having demonstrated decreased proliferation after trastuzumab treatment, we investigated the effect of overexpression of cyclin E on this therapy. HER2-overexpressing SKBr3 cells were infected with adenoviruses overexpressing full-length cyclin E and the T1 LMW isoform. Cells were treated with trastuzumab and proliferation assessed. Expression of both full-length and LMW cyclin E resulted in inhibition of the anti-proliferative effect of trastuzumab treatment (Supplementary Figure 3c). In uninfected cells, growth was inhibited by trastuzumab by 22% compared to untreated cells (P=.03). In full-length cyclin E and cyclin E-T1 overexpressing cells, trastuzumab did not inhibit cell proliferation compared to untreated cells (8.4% increase in proliferation in full-length infected cells P=.23; 3.2% increase in T1 infected cells P=.65). These results demonstrate that trastuzumab-induced growth inhibition can be overcome by cyclin E overexpression.

These data suggested potential utility in targeting cyclin E and HER2. We therefore investigated the effects of combining trastuzumab with roscovitine, an olomucine-related purine that preferentially inihibits CDK1 and CDK2. High-throughput clonogenic assays were used to compare cytotoxic effects of trastuzumab alone, roscovitine alone, or the combination in SKBr3 and BT474 (breast cancer cell line with endogenous HER2 overexpression) cells. When given individually, both agents showed a dose-dependent reduction in cell viability (Figure 5a). With the combination, there was significantly decreased cell viability compared with either agent alone. The combination index showed synergistic cytotoxicity between the two agents (Figure 5b).

To evaluate effects of HER2 on cyclin E expression in vivo, we created a HER2-overexpressing breast cancer xenograft model by injecting MCF-7-HER-18 breast cancer cells into the mammary fat pads of nude mice. After tumors reached 100 mm³, mice received intraperitoneal injections of trastuzumab or PBS twice weekly for 3 weeks.

Immunohistochemical analyses of tumors showed that trastuzumab-treated mice had lower levels of phosphorylated HER2 expression, confirming a treatment effect (percentage of pHER2 positive membranes: control group 41.7 ± 3.3% vs. treated group 22.4 ± 1.7%, p<0.01). In addition, there was a decrease in cyclin E expression (percentage of cyclin E positive nuclei: control group 57.1 ± 12.3% vs. treated group 17.5 ± 4.1%, p<0.01) (Figure 6a). Western blot for cyclin E on lysates from these tumors demonstrated that the decrease in cyclin E was primarily due to a decrease in the LMW forms (Figure 6b). These data provide in vivo confirmation of the effects of decreased HER2-mediated signaling on cyclin E expression.

MCF-7, SKBr3, and BT474 breast cancer cells were obtained from American Type Culture Collection (Manassas, VA). MCF-7-HER-18 was a gift from Dr. Mien-Chie Hung (M. D. Anderson Cancer Center). Cells were passaged in culture less than 6 weeks in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum, 2 mM L-glutamine, 100 U/mL penicillin, and 100 µg/mg streptomycin (Gibco; Invitrogen Corp, Grand Island, NY) in 5% CO₂ at 37°C. Media for MCF-7-HER-18 included 0.5 mg/ml G418.

FLAG-tagged constructs for full-length, N-terminal truncated and LMW cyclin E isoforms were generated as previously described (Harwell et al., 2000). These included cyclin EL (EL1), T1 (EL2/EL3), and T2 (EL5/6) isoforms stably overexpressed in MCF-7 cells. Validated HER2/neu siRNA (ID#540 and ID#42836) and Silencer negative control #1 siRNA were purchased from Ambion (Austin, TX).

Transfections with HER2 siRNA were performed with 5 × 10⁴ – 1 × 10⁵ cells/well in 6-well plates at a concentration of 100 nM. Briefly, 12.5 µL of 50 µM HER2 siRNA was added to 500 µL of opti-MEM media (Gibco; Invitrogen, Carlsbad, CA) and 7.5 µL of X-treme GENE transfection reagent (Roche Applied Science, Basel, Germany) was added to 500 µL of opti-MEM. Mixtures were combined and 200 µL added to wells. After 4 hours, media was added to ensure a final siRNA concentration of 100 nM.

Cyclin E adenoviruses were constructed using the AdEasy XL adenoviral vector system (Strategene; La Jolla, CA). Tumor cells were infected at a multiplicity of infection selected to ensure > 70% transduction efficiency.

Trastuzumab (Herceptin, Genentech, San Francisco, CA) reconstituted in normal saline (21 mg/mL) was diluted in media to concentrations of 10 and 20 µg/mL. Twenty-four hours after plating cells, media were changed to low-serum and 24 hours later, cells were treated with trastuzumab for 24 to 72 hours prior to harvesting.

Cells were plated at 4 × 10⁵ cells/well overnight before treatment. Cells were washed with PBS, permeabilized with .2% Triton X-100 for 20 minutes at 4 °C, blocked with 1% normal goat serum for 1 hour, and incubated with primary antibody overnight at 4 °C {(monoclonal HER2 antibody (Cell Signaling, Danvers, MA) and polyclonal cyclin E, p21, and p27 antibodies (Santa Cruz Biotechnology Inc, Santa Cruz, CA)}. Cells were incubated with secondary antibody: FITC–conjugated goat anti-mouse IgG for monoclonal antibodies, and rhodamine-conjugated goat anti-mouse IgG for polyclonal antibodies. Cells were washed, and TO-PRO-3-iodide added for nuclear staining. Cells were visualized using the confocal immunofluorescence microscope (Olympus FV 500 confocal microscope, Melville, NY) with a 40× oil immersion lens. The multi-line argon laser was used to stimulate green and red fluorescence, after which images were obtained by merging green and red channels.

Cell lysates were prepared for western blot as previously described (Rao et al., 1998). Primary antibodies were HER2 (Cell Signaling; Danvers, MA), cyclin E and cyclin D1 (Santa Cruz Biotechnology Inc., Santa Cruz, CA), p21 (Oncogene Research Products, Boston, MA), and p27 (Transduction Laboratories, Lexington, KY). Lysates were run on SDS-page gels then transferred to polyvinylidene fluoride membranes. Blots were washed and incubated with secondary antibodies then developed and signals detected using chemiluminescence detection reagents (Amersham Biosciences, Buckinghamshire, England).

Kinase assays using cyclin E polyclonal antibody were performed as previously described (Porter et al., 2001). For quantitation of relative kinase activity, bands corresponding to histone H1 were analyzed on a phosphoimager Typhoon 9400 machine (Amersham Biosciences, Sunnyvale, CA).

Cell viability was determined by MTT assays (Sigma, St. Louis, MO). Cells were transfected with HER2 siRNA or treated with trastuzumab, then fixed with dimethyl sulfoxide and stained with MTT solution. Absorbance was read with a spectophotometer (EL808 Ultramicroplate reader; Bio-Tek instruments, Inc., Winooski, VT) at 570 nm. Values were normalized and plotted as percentage change relative to control cells (mean ± SEM). The MTT assay was modified for use as a high-throughput clonogenic assay (HTCA) to determine cell viability (Lambert et al., 2008). Briefly, cells were treated with media, trastuzumab (10 or 20 µg/ml), roscovitine (5 to 15 µmol/L) (provided by Dr. Laurent Meijer, Centre national de la Recherche Scientifique, Roscoff, France), or a combination. Media was replaced every 48 hours and MTT solution added ten days after plating. Resultant crystals were solubilized and absorbance read at 590 nm. Drug interactions were assessed using CalcuSyn software (Biosoft, Inc.; Ferguson, MO).

Total cellular RNA was extracted isolated using Qiagen RNeasy kits (Qiagen Inc., Valencia, CA). Briefly, cells were lysed then homogenized in guanidine-thiocyanate-containing buffer to inactivate RNases. Ethanol was added to provide appropriate binding conditions and samples were applied to an RNeasy spin column. RNA was eluted and quantified. cDNA was synthesized from 1 µg of total RNA using the Roche Transcriptor First Strand cDNA Synthesis kit (Roche Applied Science, Indianapolis, IN). All reverse transcriptase reactions were carried out with anchored oligo(dT)18 primers to target transcription of polyadenylated mRNA and generate full-length cDNAs.

RT-PCR reactions were performed on a Rotor-Gene 2000 Real-Time cycler (Corbett Research, Sydney, Australia). The primer sequences for cyclin E (forward primer 5'-TTCTTGAGCAACACCCTCTTCTGCAGCC-3', reverse primer 5'-TCGCCATATACCGGTCAAAGAAATCTTGTGCC-3') yielded a 138-bp product. The primer sequences for β-actin, an endogenous control, (forward primer 5'-TCACCCACACTGTGCCCATCTACGA-3', reverse primer 5'-TGAGGTAGTCAGTCAGGTCCCG-3') yielded a 155-bp product (obtained from Integrated DNA Technologies, Inc., Coralville, IA). PCR products were detected using SYBR Green Jumpstart Taq Ready Mix (Sigma, St. Louis, MO). Reactions were performed in triplicate and data analyzed using Rotor-Gene Analysis software, version 5.0 (Corbett Research).

Nude mice obtained from Charles River Laboratories (Wilmington, MA) were injected with 0.5-mg estrogen pellets. 5 × 10⁶ MCF-7-HER-18 cells were injected into the mammary fat pad. When tumors reached 100 mm³, mice were divided into groups receiving intraperitoneal treatments twice weekly for 3 weeks: group 1 110 µg trastuzumab; group 2 PBS. Animals were cared for and euthanized according to institutional guidelines.

Tumors were divided and processed for western blot or immunohistochemistry (IHC). For IHC, rabbit polyclonal antibodies to cyclin E (Santa Cruz Biotechnology) and p-HER2/ErbB2 (Cell Signaling) were diluted 1:500 and 1:300 respectively in 1% goat serum.

Clinical data from 395 breast cancer patients were previously reported (Keyomarsi et al., 2002). Full-length and LMW cyclin E were evaluated by western blot analysis of tumor tissue lysates. Protein levels were measured by densitometry and LMW and total (full-length plus LMW) cyclin E scored as low (≤ the level of protein found in normal breast epithelium) or high (>normal-cell controls).

DSS was calculated from date of surgery to date of death or last follow-up. Patients dying from causes other than breast cancer were censored at time of death. DSS survival curves were computed by the Kaplan-Meier method. Univariate analyses of DSS survival according to levels of HER2, total and LMW cyclin E were performed with a two-sided log-rank test. Continuous data obtained from experiments analyzing cell cycle profiles and qRT-PCR reactions was compared using a Student t-test.