Human breast carcinoma cell lines MDA-MB-468 and MDA-MB-231 (both ERα^- and EGFR⁺ and derived from invasive breast carcinomas) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum under standard conditions as previously described 8. The antibodies used for immunoblotting and immunoprecipitation were Jab1 (1:3,000) (Sigma-Aldrich, Oakville, ON, Canada); p27 (1:200) (BD Biosciences, Mississauga, ON, Canada); Lamin A/C (1:1,000), pEGFR (phosphorylated EGFR) (1:1,000), extracellular signal-regulated kinase (ERK) (1:1,000), phosphorylated ERK (pERK) (1:1,000), AKT (1:2,000), and pAKT (1:1,000) (Cell Signaling Technology, Inc., Danvers, MA, USA); EGFR (1:1,000) (VWR International, Mississauga, ON, Canada); and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (1:4,000; Advanced ImmunoChemical Inc., Long Beach, CA, USA). The antibody to S100A7 was a rabbit polyclonal generated and described previously 621. Goat anti-mouse and goat anti-rabbit IgG secondary antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). All EGF treatments were for 4 hours and, with the exception of the EGF dose experiments, were 50 ng/mL (Millipore Corporation, Billerica, MA, USA). Treatments with ERK inhibitor PD98059 were at 20 μM for 4 hours (Cell Signaling Technology, Inc.).

Following treatment with selected reagent (EGF or PD98059), cells were fixed with 3.7% formaldehyde, permeabilized with 0.1% Triton X-100, and blocked with 0.2% bovine serum albumin. Cells then were stained for Jab1 (1:50) using the primary antibodies described above and Alexa Fluor 488-conjugated goat anti-rabbit IgG secondary antibody (1:100) (Invitrogen Corporation, Carlsbad, CA, USA). For double-immunostaining of Jab1 and pERK or p27, cells first were stained for Jab1 as described above and then were stained for pERK (1:100) or p27 (1:100) using the primary antibodies described above and Alexa Fluor 594-conjugated chicken anti-mouse IgG secondary antibody (1:100) (Invitrogen Corporation). Immunofluorescence images were captured using a Leica DM 6000B immunofluorescence microscope (Leica, Wetzlar, Germany), and image analysis was performed using OpenLab 4.0.4 software (Improvision Ltd., Coventry, UK).

Nuclear extracts were prepared by rinsing culture dishes with phosphate-buffered saline (PBS) and then resuspending cells in a nuclear extracting lysis buffer (10 mM HEPES [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid] [pH 7.9], 10 mM KCl, 0.1 mM EDTA [ethylene diamine tetraacetic acid], and 0.1 mM EGTA [ethylene glycol tetraacetic acid]) and incubating for 20 minutes on ice. Cells were further fractionated by adding 25 μL of Nonidet P-40 (10%), vortexing for 10 seconds, and centrifuging at 15,000 g for 10 minutes at 4°C. The pellet (nuclei) was then resuspended in 50 mM HEPES (pH 7.3), 150 mM NaCl, 2.5 mM EGTA, 10% glycerol, 0.1% Tween-20, 1 mM NaF, 1 mM DTT (dithiothreitol), 0.1 mM Na₃VO₃, and one tablet of EDTA-free protease inhibitor (Roche Diagnostics, Basel, Switzerland) per 10 mL, incubated 20 minutes on ice, and then boiled prior to loading.

Protein samples were separated by SDS-PAGE (4% to 12% acrylamide) and transferred to 0.2 μm nitrocellulose (Bio-Rad Laboratories, Inc., Hercules, CA, USA). After blocking in 5% skim milk powder in PBST (PBS + 0.05% Tween-20) for 30 minutes, blots were rinsed in PBST and then incubated with the primary antibody overnight in PBST at 4°C. Blots were washed in PBST for 10 minutes, three times, and then were incubated with the appropriate secondary antibody for 1 hour, followed by washing in PBST for 10 minutes, three times. Blots were developed by chemiluminescence (enhanced chemiluminescence; made in house) and were exposed to X-OMAT Kodak film (Eastman Kodak Company, Rochester, NY, USA). For all assays, at least three separate experiments were performed.

Jab1 expression was inhibited by transfecting cells with a pool of four different Jab1-specific short interfering RNA (siRNA) duplexes (Dharmacon, Inc., Chicago, IL, USA). Scrambled siRNA was used as a non-targeting control (Dharmacon, Inc.). siRNA transfection was carried out using DharmaFECT 1 transfection reagent/vehicle (Dharmacon, Inc.) according to manufacturer recommendations. siRNA was transfected at a concentration of 100 nM, after which cells were cultured for 48 hours prior to lysis and protein harvest. Densitometry of Western blots was conducted using Adobe Photoshop (Adobe Systems Incorporated, San Jose, CA, USA). Densitometry results for p27 were normalized to GAPDH within each treatment. Statistical analysis of p27 densitometry was performed with JMP software (version 7.0) (SAS Institute Inc., Cary, NC, USA) using t tests.

After the institutional research ethics board gave ethical approval, a tissue microarray (TMA) was obtained from the Manitoba Breast Tumor Bank (Winnipeg, MB, Canada) to investigate the relationship between Jab1 and EGFR and S100A7 in breast tumors in vivo. The TMA was constructed from duplicate 0.6-mm tissue cores that were removed from the central portion of a representative paraffin block from each tumor and arrayed within one of seven paraffin blocks, using a tissue arrayer (Beecher Instruments, Inc., Sun Prairie, WI, USA). The TMA included interpretable cores from 424 invasive breast carcinomas. Case selection was designed to mirror the distribution of major prognostic clinical-pathological features (size, grade, and lymph node and ER status) that the entire tumor bank collection accrued over the period 1992 to 2002 and was also based on the following criteria: (a) a minimum patient follow-up of 60 months and tumors that had (b) an invasive component of greater than 20% of the tissue section and (c) less than or equal to 10% of the normal epithelial content. ER^- status was defined by ligand-binding assay (LBA) criteria of less than 3 fmol/mg protein. The criteria for interpretation of the variables were as follows: (a) PR⁺ status was defined as greater than or equal to 15 fmol/mg protein by LBA, (b) tumor grading was according to the Nottingham system, and (c) tumor size was classified as either small (≤ 20 mm) or large (>20 mm). Patients received a range of treatments, including local radiotherapy (n = 163) and systemic hormonal and/or chemotherapy (n = 375). Patient outcome was defined as the time from initial surgery to the date of death attributable to breast cancer only.

Immunohistochemistry (IHC) staining for Jab1, EGFR, and S100A7 was performed using an automated tissue immunostainer (Ventana Medical Systems, Inc., Tucson, AZ, USA) and using bulk reagents supplied by the manufacturer (IHC protocol previously described 621). Primary antibody incubation for Jab1 and S100A7 was 32 minutes. Tumor cell staining was scored for each protein in semi-serial sections by a single observer (PHW) but in independent sessions for each protein to ensure blinded independent scoring. For Jab1 and S100A7, only nuclear expression was scored as cytoplasmic signals were generally weak and difficult to quantify. IHC staining was scored using a semi-quantitative IHC score (IHC score = [percentage of positive neoplastic epithelial cells] × [staining intensity ranked from 0 to 3]) that ranged from 0 to 300. In univariate analysis, cut-points for Jab1 and S100A7 were those used in previous studies 629 to distinguish low from high expression (Jab1 IHC scores of greater than 50, corresponding to nuclear expression in greater than 50% of tumor cells, and S100A7 IHC scores of greater than 0) or EGFR IHC scores of greater than 100, corresponding to 2+ or 3+ intensity as used for the clinical assessment of Her2 30. Statistical analysis was performed with JMP software (version 7.0) (SAS Institute Inc.) and GraphPad Prism (version 4.2) (GraphPad Software, Inc., San Diego, CA, USA) using Spearman correlation, chi-square, Mann-Whitney t test, or log-rank test as appropriate.

Jab1 has been shown previously to exist in both the nucleus and cytoplasm of different cell types. However, it has been shown that interactions between Jab1 and many of its downstream targets are associated with translocation of Jab1 to the nucleus. These include interaction with AP-1 31, NF-κB 32, and p27 32. To determine whether Jab1 translocation is affected by EGFR signaling, we first used immunofluorescence microscopy to look for changes in cellular localization of Jab1 following treatment with EGF. We observed that EGF treatment was followed by increased translocation of Jab1 to the nucleus in both MDA-MB-231 and MDA-MB-468 breast cancer cell lines (Figure 1b). This effect is particularly evident in the merged images. Quantitative analysis of Jab1 nuclear expression confirmed that Jab1 levels were approximately two-fold higher following EGF treatment compared with untreated cells (Figure 1c). This difference was statistically significant in both cell lines tested (P < 0.01). These results were confirmed by immunoblots of extracts prepared from EGF-treated MDA-MB-231 cells, which showed increased nuclear Jab1 and a corresponding decrease in cytoplasmic Jab1 following EGF treatment (Figure 1d), as well as EGF dose-response experiments that showed increased nuclear Jab1 with increased EGF treatment concentration (Figure 1e). Similar observations were made in MDA-MB-468 cells (data not shown).

The effects of EGF are known to be mediated through the EGFR and by mitogen-activated protein (MAP) kinases 33. We therefore examined whether the effect of EGF on Jab1 translocation is dependent on selective activation of the MAP kinases: p38, c-jun N-terminal kinase (JNK), and ERK. Experiments in our breast cancer cell lines showed that EGF treatment significantly increased phosphorylation of ERK as measured by immunofluorescence (Figure 2a). Minimal effects of EGF treatment were observed on phosphorylation of p38 and JNK (data not shown). We next looked at the localization of Jab1 and phosphorylated ERK (pERK). Double-immunostaining for these proteins showed that, following EGF treatment, there was an increase in both Jab1 and pERK and that these proteins were colocalized in the nucleus (Figure 2a, bottom row). ERK inhibitor, PD98059, was used in conjunction with EGF stimulation and was shown to effectively block increased nuclear Jab1 expression in MDA-MB-231 cells by both immunofluorescence (Figure 2b) and immunoblotting (Figure 2c). Similar observations were made in MDA-MB-468 cells (data not shown). These results indicate that EGF-induced Jab1 translocation can be mediated through the ERK signaling pathway.

To investigate whether EGFR signaling has a functional effect on Jab1 activity, we performed immunoblotting and double-immunostaining for the Jab1 downstream target, p27. In both MDA-MB-231 and MDA-MB-468 cell lines, Western blot assay showed that EGF treatment and phosphorylation of EGFR resulted in a significant decrease in p27 expression (Figure 3a). Additional observed changes following EGF treatment included increased pAKT (Figure 3a). The inverse correlation between nuclear Jab1 and p27 expression was also observed in double-immunostaining for these proteins (Figure 3b). To confirm that Jab1 was necessary for the effect of EGF on p27, we performed Jab1 knockdown using an siRNA approach in MDA-MB-231 cells in conjunction with EGF treatment. In addition to re-confirming that cells treated with EGF have reduced p27 (P < 0.05), we found that Jab1 knockdown restored p27 to EGF-untreated levels compared with cells treated with EGF and control siRNA (P < 0.0001) (Figure 4a,b). In cells treated with Jab1 siRNA, EGF had no effect on p27 levels (P = 0.68). Taken together, these results indicate not only that EGFR signaling affects Jab1 translocation but that it may regulate targets downstream of Jab1 and that the effect of EGF on p27 levels is mediated by Jab1.

To further explore the relationship between EGFR and Jab1 expression in vivo, we examined the expression of these genes in a series of 424 invasive breast tumors using TMAs. The characteristics of the cohort are outlined in Table 1. The relationship between nuclear expression of Jab1 and the level of EGFR was assessed, together with the level of S100A7, because of the previously established strong relationship between S100A7 expression and Jab1. In analysis of the entire tumor cohort, high levels of Jab1, EGFR, and S100A7 were seen in 154/424 (36%), 42/424 (10%), 144/424 (34%) cases, respectively (Figure 5). Jab1 was not associated with prognostic factors or biomarkers, including grade, axillary nodal status, tumor size, ER, PR, EGFR, or S100A7, or with overall patient survival when examined in the entire cohort. In subgroup analysis of the ERα⁺ subgroup, no significant associations were observed. However, in subgroup analysis of the ERα^- subgroup (n = 154), Jab1 levels were associated with axillary node-positive status (P = 0.019, t test) and higher levels of Jab1 nuclear expression were associated with both EGFR (P = 0.019, t test) and S100A7 (P = 0.036, t test) (Table 2). Notably, higher Jab1 levels were more strongly associated with combined EGFR⁺/S100A7⁺ versus EGFR^-/S100A7^- status within this subgroup (P = 0.004, t test).

Outcome analysis of the ERα^- subgroup showed no significant association between survival and Jab1⁺, EGFR⁺, or S100A7⁺ status when each marker was analyzed independently. Comparison of the subset of ERα^- tumors that were positive for all three markers, EGFR/S100A7/Jab1, showed that this phenotype was associated with worse outcome compared with EGFR/S100A7/Jab1^- tumors (n = 10 versus 13; P = 0.07).