ROR1 is expressed in human breast cancer and associated with enhanced tumor-cell growth

Abstract

Receptor-tyrosine-kinase-like orphan receptor 1 (ROR1) is expressed during embryogenesis and by certain leukemias, but not by normal adult tissues. Here we show that the neoplastic cells of many human breast cancers express the ROR1 protein and high-level expression of ROR1 in breast adenocarcinoma was associated with aggressive disease. Silencing expression of ROR1 in human breast cancer cell lines found to express this protein impaired their growth in vitro and also in immune-deficient mice. We found that ROR1 could interact with casein kinase 1 epsilon (CK1ε) to activate phosphoinositide 3-kinase-mediated AKT phosphorylation and cAMP-response-element-binding protein (CREB), which was associated with enhanced tumor-cell growth. Wnt5a, a ligand of ROR1, could induce ROR1-dependent signaling and enhance cell growth. This study demonstrates that ROR1 is expressed in human breast cancers and has biological and clinical significance, indicating that it may be a potential target for breast cancer therapy.

Figure 1. ROR1 is expressed in human breast cancers and breast cancer cell lines.

(A) Representative images of fresh frozen tissue sections stained for immunohistochemistry with 4A5, a mAb specific for ROR1 (labeled anti-ROR1) or an IgG2b isotype-control mAb (labeled IgG). The bound antibody is red and the nuclear counterstain with hematoxylin is blue. Red arrow points to the tumor cells and blue arrow points to stromal cells. (B) Immunoblot analysis of lysates from normal human breast tissue or breast cancer tissue with rabbit antibodies specific for human ROR1 or β-Actin. CHO-ROR1 cells are CHO cells transfected to express human ROR1. (C) Fluorescence histograms of human breast cancer cell-lines (MCF-7, SKBR3, MDA-MB-431, MDA-MB-468, or MDA-MB-231) stained with 4A5 (open histograms) or control mAb (shaded histograms) (top panel). The expression of ROR1 mRNA in each tumor cell line relative to MCF-7 cells assessed via quantitative RT-PCR is indicated by the relative height of each bar below the corresponding histogram. The error bars represent the standard error about the mean of triplicate values. (D) Formalin-fixed, paraffin-embedded pellets of MCF-7 or MDA-MB-231 cells were stained with 4A5 or control IgG2b. Tissue-bound 4A5 is shown in red and the nuclear counterstain with hematoxylin is in blue (Scale bar in the top left picture represents 35 µm). (E) Representative images of breast normal tissues or cancer tissues stained with 4A5 A score of 0 indicates that none of the cells within the sample bound to the anti-ROR1 mAb; a score of 1 indicates low-level binding of the mAb to the tumor cells or low-to-moderate-level binding of the mAb on less than 50% of tumor cells; a score of 2 indicates moderate-level staining on more than 50% of tumor cells or high-level staining of the tumor cells. The scale bar in the top right picture represents 35 µM. Red arrows point to the tumor cells, blue arrow points to stromal cells and green arrow points to lymphocytes. (F) The proportion of breast tumor tissues or normal tissues found negative (Score 0) or having weak to moderate staining (Score 1) or strong staining (Score 2) for ROR1 are indicated by in each bar. The number of different cases examined for each tumor type is indicated in the parentheses. (G) The proportion of different histological types of breast tumor tissues found lacking staining (Score 0) or having weak to moderate staining (Score 1) or strong staining (Score 2) for ROR1 are indicated by in each bar. The number of different cases examined for each tumor type is indicated in the parentheses. (H) The proportion of poorly differentiated (grade 3) or well to moderately differentiated (grade 1–2) ductal breast adenocarcinoma tissues found lacking staining (Score 0) or having weak to moderate staining (Score 1) or strong staining (Score 2) for ROR1 are indicated by in each bar. The number of different cases examined for each tumor type is indicated in the parentheses. Statistical significance of the differences was analyzed using Kruskal-Wallis test (Grade 3 versus grade 1–2, P = 0.013).

Figure 2. Expression of ROR1 in human breast cancer is associated with aggressive disease.

(A) Each bar represents median-centered log₂ expression of ROR1 mRNA by breast cancer cell lines in Vene dataset. Cell lines with similar molecular characteristics are clustered together. (B) Each dot plot represents median-centered log₂ expression of ROR1 mRNA by tumor tissue from an individual patient. Patient samples with similar phenotypic or molecular characteristics are clustered together. The line indicates the median ROR1 expression level by the group. P indicates the statistical significance of the differences in the collective ROR1 expression between the two groups, as calculated using Student's t test. (C) Kaplan-Meier survival analysis for 295 breast cancer patients from Vandevijver dataset using defined cut-off values for ROR1 expression (patient sub-groups defined as samples that expressed ROR1 at levels that exceeded that of the mean level of ROR1 expressed by all samples (50% high: n = 148), highest 20% (20% high: n = 58) or highest 10% (10% high: n = 30) relative ROR1 expression. Statistical difference was determined by log-rank test. The each dataset used is indicated on the top of each graph.

Figure 3. Expression of ROR1 promotes tumor-cell survival and growth.

(A–B) MDA-MB-231 cells transiently transduced with control shRNA (Ct-shRNA1 or 2) or ROR1-shRNA1 or 2, as indicated at the top of each lane, were examined for ROR1 expression by immunoblot analysis (A) or stained with Annexin V for apoptosis detection after 2 days (B). The height of each bar in the graph B indicates the mean proportion of cells that were Annexin-V positive. The error bars indicate the standard error of triplicate samples. P indicates the statistical significance as measured by Dunnett's multiple comparisons test. _* indicates P<0.05. (C) MDA-MB-231 cells were transduced with vectors encoding Ct-shRNA or ROR1-shRNA and then selected for stable expression of the shRNA. Equal numbers of cells indicated in the legend provided in the top left corner were cultured and monitored for growth over time. Each graph indicates the average number of viable cells assessed at 0, 1, 2, or 3 days using the WST-8 assay. The error bars indicate the S.E. of the mean for triplicate samples. _* indicates P<0.05 and _** indicates P<0.01(Student's t test). (D) Representative bioluminescent imaging of RAG^−/−/γ_c ^−/− mice engrafted with luciferase-labeled MDA-MB-231 cells transduced with Ct-shRNA1 (top row) or ROR1-shRNA1 (bottom row). The intensity of the fluorescence signal related to cell number is indicated by the scale on the bottom. The graph on the right provides the mean fluorescence activity of RAG^−/−/γ_c ^−/− mice engrafted with luciferase-labeled MDA-MB-231 cells transduced with Ct-shRNA1 (closed triangles) or ROR1-shRNA1 (open triangles) at various times after engraftment, as indicated by the legend in the upper left corner. The error bars indicate S.E.M. fluorescence activity of groups of mice (n = 5–7) at each time point. _** indicates P<0.01(Student's t test). (E) Tumors were extracted from mice engrafted with MDA-MB-231 cells transduced with Ct-shRNA1 or ROR1-shRNA1 and examined for ROR1 (left), TUNNEL (middle), or Ki-67 (right). The scale bar in the top right panel corresponds to 35 µm. The nuclear counterstain is in blue and antibody staining is in red. The height of each bar on the right graph indicates the proportion of Tunel-positive or Ki67-positive cells in tumors that formed from MDA-MB-231 cells transduced with either Ct-shRNA or ROR1-shRNA. Error bars indicate S.E.M (n = 7). * indicates P<0.05 (Student's t test).

Figure 4. ROR1 activates CREB to regulate tumor-cell growth.

(A) Relative expression of selected genes by MDA-MB-231 cells transduced with either Ct-shRNA (white bars) or ROR1-shRNA (black bars), as assessed via quantitative RT-PCR. The numbers on the y-axis represent fold difference in specific gene expression relative to GAPDH. Error bars indicate S.E.M. (n = 3 experiments). _* indicates P<0.05, _** indicates P<0.01 and _*** indicates P<0.001 (Student's t test). (B) Immunoblot analyses for proteins listed on the right margin, using lysates from MDA-MB-231 control cells versus cells silenced for ROR1. (C) Immunoblot analysis of CREB phosphorylation at ser-133 (p-CREB), total CREB (t-CREB), ROR1, or β-Actin for MDA-MB-231 cells with or without ROR1 silencing, as indicated on the top of each panel. The phospho-specific antibody recognizes phospho-serine-133 CREB. (D) Tumors extracted from mice engrafted with MDA-MB-231 control cells or ROR1-shRNA-transduced cells were examined for p-CREB by immunohistochemistry staining (n = 3) or Bcl-2, Cyclin D1 and β-Actin protein expression by immunoblot analysis. The scale bar in the left panels represents 35 µm. The antibody staining is in red and the nuclear counterstain is in blue. The intensity of p-CREB was quantified and data were represented as mean intensity ± S.E.M from 3 tumor tissues per group. (E) MDA-MB-231 cells infected with Ad-GFP or Ad-ACREB were examined for expression of proteins by immunoblot analyses. (F) MDA-MB-231tumor cells with or without ROR1 expression were transfected with Ad-GFP or Ad-ACREB and then monitored for cell growth. The numbers of viable cells (after 72 hours culture) are represented by the height of each bar in the graph. The error bars provide the S.E.M. of triplicate samples. P indicates the statistical significance as assessed by Student's t test. (G) Representative images of breast cancer tissues stained for phospho-CREB (p-CREB) (top panels) or ROR1 (bottom panels). Tissue-bound antibody is shown in red and the nuclear counterstain with hematoxylin is in blue. The scale bar in the top left picture represents 35 µM. Tissues were scored as 0 (none of the cells within the sample bound to the mAb); 1 (low-level binding of the mAb to the tumor cells or low-to-moderate-level binding of the mAb on less than 50% of tumor cells); 2 (moderate-level staining on more than 50% of tumor cells or high-level staining of the tumor cells). Red arrows point to the tumor cells, blue arrow points to stromal cells and green arrow points to lymphocytes.

Figure 5. ROR1 interacts with CK1ε to activate PI3K/AKT, leading to activation of CREB.

(A) MDA-MB-231tumor cells with or without ROR1 expression were examined for phosphorylated AKT at ser-473 (p-AKT) or total AKT (t-AKT). (B) MDA-MB-231 cells were cultured with (+) or without (−) LY294002 and examined for p-AKT, t-AKT, p-CREB, or t-CREB by immunobot analysis after 16 hours. (C) MDA-MB-231 cells (Ct-shRNA) or cells silenced for ROR1 (ROR1-shRNA) were cultured with different concentrations of LY294002 and monitored for cell growth after 48 hours using the WST-8 assay. The graphs depict the mean numbers of viable cells, ± S.E.M of triplicate samples, which are representative of three independent experiments. *P<0.05 and ***P<0.001). (D–E) MCF7 control cells (Ct-Vector) or cells made to express ROR1 (pCDNA3-ROR1) were examined for protein expression (D) or (E) monitored for growth over 4 days using the WST-8 assay. The graph depicts the mean numbers of viable cells ± S.E.M. over time in each of these cell lines, as indicated in the legend at the top of the figure, which are representative of three independent experiments. *P<0.05, **P<0.01 and ***P<0.001 by Student's t test. (F) Protein lysates from MDA-MB-231 cells or cells silenced for ROR1 were immunoprecipitated (IP) with ROR1 antibody. The bound immunoprecipitated products and whole cell lysate (WCL) were probed by the antibodies indicated in the Probe column. (G–H) MDA-MB-231 cells were treated without (−) or with CK1ε siRNA (CK1ε-siRNA) or control siRNA (Ct-siRNA) for 72 hours and examined for protein expression (G) or cell growth (H). The height of each bar in the graph F provides the mean number of viable cells ± S.E.M., which are representative of more than three independent experiments. P indicates the statistical significance.

Figure 6. Wnt5a is involved in ROR1-mediated signaling and cell growth.

(A) MDA-MB-231 cells (Ct-shRNA) or cells transduced for ROR1-shRNA (ROR1-shRNA) were transfected with Cre-luciferase reporter construct and/or increasing amounts of Wnt5a expression vector. The bars provide the mean relative luciferase activity of the various cell populations. *P<0.05 by Dunnett's multiple comparisons test. (B–D) MDA-MB-231 cells or cells silenced for ROR1 were stimulated with/without 250 ng/ml exogenous recombinant Wnt5a protein (rWnt5a) and examined for p-AKT, AKT, p-CREB, CREB at 15 minutes (B), BCL-2, ROR1 and β-actin expression at 48 hours (C) and cell growth at 48 hours (D). The graph depicts the mean numbers of viable cells ± S.E.M., which is representative of more than three independent experiments. (E) MDA-MB-231 cells were pretreated with/without 10 µM LY294002 for 2 hours and then stimulated with/without 250 ng/ml rWnt5a for 15 minutes and examined for protein expression. Relative phospho-protein levels were normalized to total-protein. (F–G) MDA-MB-231 cells were treated with/without 10 µM LY294002, or infected with either Ad-GFP or Ad-ACREB, or cultured in the media with/without 250 ng/ml rWnt5a for 48 hours as indicated on the top, and examined for proteins (F) or cell growth using WST-8 assay (G). The height of each bar in the graph provides the mean number of viable cells ± S.E.M. in each cell population. P indicates the statistical significance as determined using Student's t test. (H) Model of ROR1 promoting tumor-cell growth. Wnt5a expressed by tumor or tumor microenviroment (e.g. stromal cell) interacts with ROR1 to activate PI3K/AKT through CK1ε, leading to activation of CREB to enhance expression of genes that enhance resistance to tumor cell apoptosis and/or promote tumor cell growth.