MCF-7 cells were maintained in medium consisting of equal parts phenol red-containing DMEM and Ham's F12, 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA). Two MCF-7 stable clones expressing PI3K p110αCAAX were generated in the laboratory of Dr. James McCubrey at East Carolina University  and maintained in 89% RPMI (Invitrogen, Carlsbad, CA), 10% FBS, and 1% penicillin/streptomycin. In experiments requiring treatment with 1 μM trans-retinoic acid and 0.1 μM hydrocortisone (tRAH; Sigma, St. Louis, MO), both experimental and control groups of MCF-7 and MCF-7/PI3K p110αCAAX cells were transferred to medium composed of equal parts DMEM and Ham's F12, 5% charcoal-stripped FBS and 1% penicillin/streptomycin.
PI3K p110αCAAX was cloned into the BamHI site in the pcDNA3 vector. Flag-tagged human NIS open reading frame was cloned into the pcDNA3 vector as described in Zhang et al. . Fugene6 reagent (Roche, Alameda, CA) was used for transient transfection according to the manufacturer's protocol. Co-transfection of a GFP expression plasmid was used to assess transfection efficiency.
Cell surface biotinylation
Cell surface NIS expression was determined as described in Vadysirisack et al. . Cells were surface biotinylated and lysed, followed by avidin pull-down using 5 mg of total cell lysate. Pulled-down biotinylated surface protein derived from MCF-7/PI3K p110αCAAX cells and MCF-7 cells treated with tRAH for 48 hours  was used for Western Blot as described below.
Western blot analysis
Western blot was performed as described in Jhiang et al.  using 100 μg of total cell lysates or membrane-enriched fraction (after removal of nucleus and cytosol) for total NIS, β-actin or Na+K+ATPase detection, or the pulled-down biotinylated surface protein, as described above, for surface NIS and Na+K+ATPase detection. Akt and pAkt Ser473 levels were assessed using 100 μg of whole cell lysate. Experiments utilized anti-hNIS antibody (#331, 1:1000), anti-actin antibody (Santa Cruz Biotechnology, Santa Cruz, CA; 1:200), anti-Akt and anti-pAkt Ser473 antibody (Cell Signaling Technology, Danvers, MA; 1:250) or anti-Na+K+ATPase antibody (Santa Cruz Biotechnology, Santa Cruz, CA; 1:1000). Secondary antibodies were anti-rabbit IgG conjugated to horseradish peroxidase for NIS, actin, Akt and pAkt Western blot (GE Healthcare Bio-Sciences Corp, Piscataway, NJ; 1:4000) or anti-mouse IgG conjugated to horseradish peroxidase for Na+/K+ATPase Western blot (Cell Signaling Technology, Danvers, MA; 1:4000). NIS antibody specificity was verified by PNGase F deglycosylation of NIS protein to a sole 50 kDa band, the molecular weight of underglycosylated NIS. Densitometry was performed by scanning with the Scion Image program (Scion Corp., Frederick, Md.).
NIS and Flag immunofluorescence were performed essentially as described in Marsee et al. . Briefly parental MCF-7 cells and MCF-7/PI3K p110αCAAX stable clones were seeded in 4-well chamber slides. Forty-eight hours after Flag-hNIS or empty vector transfection or tRAH treatment (when applicable), cells were fixed using 1% paraformaldehyde in PBS and labeled with anti-NIS antibody (#331; 1:1000) or anti-Flag M2 monoclonal antibody (Sigma, St. Louis, MO; 1:750), followed by incubation with CyTM3-conjugated affinipure F(ab')2 fragment donkey anti-rabbit IgG (1:1400) for NIS immunofluorescence or anti-mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA; 1:500) for Flag immunofluorescence. DAPI nuclear stain (Invitrogen, Carlsbad, CA) was applied before mounting and coverslipping. Secondary antibody only negative controls were included in the experiments. Images were obtained at 63× magnification using a Leica TCS SP2 AOBS Confocal Laser Scanning Microscope in The Ohio State University College of Veterinary Medicine.
Radioactive iodide uptake assay
RAIU was performed as described in La Perle et al.  using MCF-7 or MCF-7/PI3K p110αCAAX cells after forty-eight hours of tRAH treatment, Flag-hNIS or activated PI3K p110αCAAX transient transfection (when applicable). In experiments involving transfection of MCF-7 cells and subsequent treatment with tRAH, the transfection time was set at 6 hours, followed by medium change and addition of tRAH for 42 hours. After incubation with 125I, washing and cell lysis, the supernatant was counted in a γ-radiation counter. Values were recorded in counts per minute (cpm) per microgram of DNA. Assays were performed in triplicate.
Immunohistochemical staining technique was performed as described in Knostman et al  with the following modifications. A paraffin-embedded formalin-fixed human breast cancer tissue microarray consisted of 2 mm punches from 50 patient samples sectioned to 4 μm thickness and affixed to glass slides. Slides were incubated with rabbit polyclonal anti-human NIS primary antibody (#331; 1:500) followed by HRP-conjugated goat anti-rabbit IgG secondary antibody (Bio-Rad, Hercules, CA; 1:250) and chromogen DAB detection (Dako, Carpinteria, CA). Immunostaining for pAkt was performed in The Ohio State University Department of Pathology using mouse monoclonal anti-pAktSer473 antibody (Cell Signaling Technology, Danvers, MA; 1:100). Thirty-three of the 36 tissues in the array had suitable integrity for interpretation of NIS positivity and pAkt expression and were included in the statistical analyses.
Statistical analysis consisted of Student's t-test (in vitro studies) or Fisher's Exact Test (in vivo correlational studies) and was performed using GraphPad software. Only breast cancer tissue punches available for NIS, pAkt and PTEN immunohistochemical staining were used in statistical analyses. Additional comparison of NIS expression and estrogen receptor, progesterone receptor and Her-2/neu expression was performed, but no statistical relationship was evident.