Dual Mechanisms of LYN Kinase Dysregulation Drive Aggressive Behavior in Breast Cancer Cells

Abstract

The SRC-family kinase LYN is highly expressed in triple-negative/basal-like breast cancer (TNBC) and in the cell of origin of these tumors, c-KIT-positive luminal progenitors. Here, we demonstrate LYN is a downstream effector of c-KIT in normal mammary cells and protective of apoptosis upon genotoxic stress. LYN activity is modulated by PIN1, a prolyl isomerase, and in BRCA1 mutant TNBC PIN1 upregulation activates LYN independently of c-KIT. Furthermore, the full-length LYN splice isoform (as opposed to the Δaa25-45 variant) drives migration and invasion of aggressive TNBC cells, while the ratio of splice variants is informative for breast cancer-specific survival across all breast cancers. Thus, dual mechanisms-uncoupling from upstream signals and splice isoform ratios-drive the activity of LYN in aggressive breast cancers.

Keywords: BRCA1; ESRP1; LYN kinase; PIN1; c-KIT; triple-negative/basal-like breast cancer.

Graphical abstract

Figure 1

LYN Is Positively Regulated by c-KIT in Normal Mammary Cells (A) Flow cytometry of primary mammary cells stained with CD45, CD24, and Sca-1 antibodies. CD45⁺ leukocytes (purple) were gated out (top plot), and CD45⁻ cells (bottom plot) were gated to define basal (CD24^+/low Sca-1⁻, red), luminal ER− (CD24^+/high Sca-1⁻, green), and luminal ER+ (CD24^+/high Sca-1⁺, blue) epithelial cell populations. (B) Expression pattern of c-Kit, Scf, and Lyn splicing transcripts in mouse mammary cell populations. Semiquantitative RT-PCR data are representative of two independent isolates (four mice for each). Amplicons of the expected size using primers spanning the alternative exon for each gene are indicated. Gapdh was used as a control. (C) qRT-PCR gene expression analysis of Scf in mouse mammary cell populations using probes for both total Scf (membrane bound and soluble) or soluble Scf (sSCF) only. Data are from two independent isolates (four mice for each), presented as relative expression levels with leukocytes as the comparators. (D) Schematic of LYN isoforms showing the 21-amino acid insertion (black residues) in the N-terminal domain of LYNA. (E) Representative western blot analysis and quantitation of c-KIT, JAK2, STAT3, AKT, and ERK1/2 phosphorylation levels in protein extracts from primary mouse mammary organoids cultured on Matrigel and stimulated with SCF for the indicated times. Tubulin was used as loading control. (F and G) Representative western blot analysis and quantitation of LYN autophosphorylation (Y397) (F) and immunoprecipitation (IP) LYN kinase assay of protein extracts (G) from primary mouse mammary organoids cultured on Matrigel and stimulated with SCF for 0, 15, 30, and 60 min. (H and I) Western blot of c-Kit expression and LYN autophosphorylation (Y397) in primary mouse mammary organoids after transduction with control (shScr) or c-Kit knockdown (shKit1 and shKit2) lentiviruses (H) or following treatment with c-KIT blocking (ACK2) or immunoglobulin G (IgG) isotype (IgG Ctr) antibodies (I). Unless otherwise stated, blots are representative of three independent experiments (mean and SD; two-tailed unpaired t tests) (in E and F, t tests are relative to time 0). ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001. See also Figure S1.

Figure 2

LYN Promotes Normal Mammary Cell Growth (A) Analysis of Lyn expression levels by qRT-PCR relative to shScr cells (top) and western blot (bottom) in primary mouse mammary organoids 4 days after transduction with control (shScr) or Lyn knockdown (shLyn1 and shLyn2) lentiviruses. (B) Growth of mammary organoids after transduction with control (shScr) or Lyn knockdown (shLyn1 and shLyn2) lentiviruses, assessed by cell number change (left bottom panel) or organoid size (right bottom panel) relative to shScr cells, day 4 after plating. Top panels: representative images show organoids at day 6 of culture. Scale bar, 75 μm. (C) Ki67 immunofluorescence staining (green) of control (shScr)- and shLyn-carrying mammary organoids 6 days after lentiviral transduction (DAPI nuclear counterstaining). Representative images and quantification of the percentage of Ki67-positive cells. Scale bar, 20 μm. (D) Colony-forming potential of unfractionated primary mammary epithelial cells (all epithelial) or basal, luminal ER−, and luminal ER+ subpopulations. 5,000 cells from each fraction were plated on Matrigel, and colony numbers were determined after 12–14 days. (E) Growth inhibition of unfractionated primary mammary epithelial cells (all epithelial) or the luminal ER− fraction transduced with control (shScr) or Lyn knockdown (shLyn1 and shLyn2) lentiviruses and seeded onto Matrigel. Cell growth was assessed after 12–14 days. Representative images, left (scale bar, 100 μm); quantitation, right. (F) Growth inhibition of MCF10A cells transduced with lentiviral vectors carrying control shRNA (shScr) or shRNA against LYN (shLyn1 and shLyn2). Transduced cells were grown in 3D on Matrigel, and relative cell numbers were assessed after 12 days of culture. Representative images of acinar structures (day 12) derived from shScr, shLyn1, or shLyn2 cells are shown (scale bar, 100 μm), together with CellTiterGlo quantitation and assessment of LYN knockdown by western blot (GAPDH loading control). (G) Confocal microscope analysis of Ki67 immunofluorescence-stained shScr, shLyn1, or shLyn2 knockdown MCF10A cells at day 4 of culture in 3D with quantitation. DAPI was used for counterstaining. Scale bar, 20 μm. Blots are representative of three independent experiments. Quantitation, mean and SD (n = 3; two-tailed unpaired t tests) except for gene expression analysis by quantitative real-time RT-PCR (mean ± 95% confidence intervals; significance of real-time RT-PCR data was determined from confidence intervals; n = 3 independent experiments for each of 3 technical replicates per sample) (Cumming et al., 2007). ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001. See also Figures S1–S3.

Figure 3

LYN Activity Is Required for Growth of Brca1 Tumor Cells (A) Primary cells isolated from three distinct BlgCre Brca1^fl/flp53^+/− mouse mammary tumors (1–3) were transduced with control (shScr) or Lyn knockdown (shLyn1 and shLyn2) lentiviruses, seeded at low density in adherent conditions (2D), and stained with crystal violet after 6 days. Viable cell density was determined by absorbance measurement following solubilization of the dye. Representative images of tumor cell colonies at day 6 of culture are shown. (B) shScr-, shLyn1-, or shLyn2-transduced BlgCre Brca1^fl/flp53^+/− tumor cells (1–3) seeded in Matrigel (3D) were assessed for growth after 6 days. Graphs show cell number assessed at day 5 of culture relative to shScr cells. (C) Ki67 immunofluorescence staining (green) of control (shScr)- and shLyn-transduced BlgCre Brca1^fl/flp53^+/− tumor cells in 3D culture 6 days after lentiviral transduction. Representative images and quantification of the percentage of Ki67-positive cells (n = 3). Scale bar, 20 μm. (D) Primary BlgCre Brca1^fl/flp53^+/− mouse mammary tumor cells were transduced with either lentiviral shScr and empty expression vectors (shScr), shLyn and empty expression vectors (shLyn), or shLyn and expression vectors carrying either an shLyn-resistant form (indicated by an asterisk) of wild-type LYNA (shLyn + LYNA^∗WT) or a kinase-dead LYNA mutant (shLyn + LYNA^∗KD). LYN protein levels determined by western blot 6 days after transduction. The graph shows cell number assessed at day 5 of culture relative to shScr cells. (E) HCC1937 cells were transduced with control (shScr) or Lyn knockdown (shLyn1 and shLyn2) lentiviruses and tested for LYN expression levels by western blot after 6 days. (F) shScr-, shLyn1-, or shLyn2-transduced HCC1937 cells were seeded at low density in adherent conditions. Viable cell density was determined after 7 days as in (A). Representative images show tumor cell colonies at day 7 of culture. (G) BRCA1 mutant PDX-derived cells (BCM 3887) were transduced with control (shScr) or LYN knockdown (shLyn1 and shLyn2) lentiviruses and tested for cell viability after 10–12 days of culture in 3D on Matrigel. (H) Primary mouse BlgCre Brca1^fl/flp53^+/− mammary tumor cells were transduced with pHIV-RFP-Tet repressor and pSEW-GFP-TO-H1 (carrying either shScr or shLyn) lentiviruses. Lyn levels were determined in cells transduced with either inducible shScr or shLyn and in either the presence or the absence of doxycycline (DOX) by qRT-PCR relative to shScr cells without DOX. (I) 250,000 inducible shScr- or shLyn-transduced cells were orthotopically injected into the fourth right mammary fat pad of nude mice. These were randomized to DOX treatment or normal diet, and tumor growth was monitored. Tumor volumes were calculated from caliper measurements of tumor width and length. Tumor growth curves (mean ± SEM) and representative images of endpoint tumors are shown. Blots are representative of three independent experiments. Unless otherwise stated, quantitation is shown as mean and SD (n = 3; for PDX cell experiments n = 3 cell isolations from 3 PDX implants in 3 mice; two-tailed unpaired t tests), except for gene expression analysis by quantitative real-time RT-PCR (mean ± 95% confidence intervals; significance of real-time RT-PCR data was determined from confidence intervals; n = 3 independent experiments for each of 3 technical replicates per sample) (Cumming et al., 2007). ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001. See also Figure S3.

Figure 4

LYN Activity Is Regulated by BRCA1 via the Prolyl Isomerase PIN1 (A) Protein extracts from TNBC cells with either wild-type BRCA1 or impaired BRCA1 expression were analyzed for phospho-LYN (p-LYN) (Y397), total LYN, and GAPDH levels by western blot. Scatterplot shows quantification of p-LYN levels normalized to total LYN levels. (B) Primary mouse mammary organoids were transduced with control (shScr) or Brca1 knockdown (shBrca1) lentiviruses. Knockdown was assessed by qRT-PCR relative to comparator shScr cells (left). shScr and shBrca1 cells were assessed for levels of phospho-c-KIT (Y719), phospho-LYN (Y397), LYN, and GAPDH by western blot after 4 days (middle). (C) Western blot analysis and quantitation of LYN autophosphorylation levels in primary mouse mammary organoids transduced with control (Ctr) or HA-tagged BRCA1 (HA BRCA1) expression lentiviruses. (D) Examples of PIN1 immunohistochemistry scores in breast cancer TMAs: (i) 0, (ii) 1, (iii) 2, (iv) 3, and (v) 4. DAB staining of PIN1, and blue counterstaining of nuclei. Scale bar in main panels, 500 μm; scale bar in inset, 50 μm. (E) Quantitation of PIN1 scoring in BRCA1 mutant and sporadic TNBC TMAs. Blots in (B) and (C) are representative of three independent experiments. Quantitation is shown as mean and SD (n = 3; two-tailed unpaired t tests) except for gene expression analysis by quantitative real-time RT-PCR (mean ± 95% confidence intervals; significance of real-time RT-PCR data was determined from confidence intervals; n = 3 independent experiments for each of 3 technical replicates per sample) (Cumming et al., 2007). ^∗p < 0.05; ^∗∗p < 0.01. See also Figure S4.

Figure 5

LYN Is Activated in BRCA1 Null Cells by the Prolyl Isomerase PIN1 (A–C) Primary cells from BlgCre Brca1^fl/flp53^+/− mouse mammary tumors (A), human HCC1937 BRCA1-deficient breast cancer cells (B), and BRCA1 mutant PDX-derived cells (C) were transduced with control (shScr) or Pin1 knockdown (shPin1#1 and shPin1#2) lentiviruses and lysed after 72 hr. Protein extracts were assessed for levels of PIN1, LYN, phospho-LYN (Y397), and c-KIT (Y719) (PDX samples were not probed for phospho-KIT). Representative western blots and quantitation of phospho-LYN (Y397) levels are shown. GAPDH was used as loading control. shScr-, shPin1#1-, and shPin1#2-transduced BlgCre Brca1^fl/flp53^+/− tumor cells and HCC1937 cells were also seeded at low density in adherent conditions and stained with crystal violet after 6 days. Cell number was determined by absorbance measurement following solubilization of the dye. PDX-derived transduced cells were cultured for 10–12 days in 3D on Matrigel and then assayed for cell viability. (D) Protein extracts from primary BlgCre Brca1^fl/flp53^+/− mouse tumor cells transduced with vectors carrying wild-type LYNA were subjected to immunoprecipitation by anti-PIN1 or control (IgG) antibodies. Total extracts (input) and immunoprecipitates (IPs) were probed for PIN1 and LYN by western blot. (E) Schematic of LYN showing the position of PIN1 consensus recognition sequences and the proline > isoleucine mutants generated. (F) Representative western blot analysis of LYN phosphorylation levels at the negative regulatory phosphorylation site (Y508) in primary BlgCre Brca1^fl/flp53^+/− transduced with vectors carrying wild-type LYNA or LYNA proline mutants (LYN P229I, LYN P197I, or LYN P197I P229I). (G) Western blot analysis of LYN autophosphorylation and PIN1 levels in human HCC1937 cells transduced with either control (Ctr) lentivirus or virus-carrying HA-tagged wild-type or mutant BRCA1 (C61G, A1708E, or L1407P). Blots are representative of three independent experiments. Quantitation is shown as mean and SD (n = 3; two-tailed unpaired t tests). ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001. See also Figures S5 and S6.

Figure 6

LYNA Drives Migration and Invasion in Breast Cancer Cells (A) Western blot of LYN protein levels and growth of control (shScr) and LYNA knockdown (shLynA) MDA-MB-231 cells relative to shScr cells at day 0. (B) Migration and invasion of shScr- and shLynA-MDA-MB-231 cells, assessed by transwell assay. Representative images show endpoint assays, and quantification of results compares the percentages of cells per field to shScr cells. (C) Total LYN knockdown and LYNA or LYNB reconstitution using shLyn-resistant forms (LYNA/B^∗) in MDA-MB-231 cells. Western blot shows LYN knockdown cells forced to express either LYNA or LYNB and growth relative to shScr cells at day 0. (D) Migration and invasion of control MDA-MB-231 (shScr) cells, LYN-depleted (shLyn) cells, cells expressing LYNA only (shLyn + LYNA^∗), or cells expressing LYNB only (shLyn + LYNB^∗). Representative images show endpoint assays, and quantification of results compares the percentages of cells per field to shScr cells. (E) siControl (siCtr) and siESRP1 MCF7 cells were analyzed for LYNA::LYNB transcript ratio and LYN protein levels by semiquantitative RT-PCR (top, with quantitation) and western blot (bottom), respectively. (F) MDA-MB-231 cells were transduced with control (Ctr) or FLAG-tagged ESRP1 (FLAG-ESRP1) lentiviruses. ESRP1 overexpression was assessed by qRT-PCR (fold expression over comparator Ctr cells; upper left panel). Ctr and FLAG-ESRP1 cells were analyzed for LYNA:LYNB transcript ratio by semiquantitative RT-PCR (bottom panels), and LYN and ESRP1 protein levels were analyzed by western blot (upper right panel). Blots are representative of three independent experiments. Quantitation is shown as mean and SD (n = 3; two-tailed unpaired t tests), except for gene expression analysis by quantitative real-time RT-PCR (mean ± 95% confidence intervals; significance of real-time RT-PCR data was determined from confidence intervals; n = 3 independent experiments for each of 3 technical replicates per sample) (Cumming et al., 2007). ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001. See also Figures S7 and S8.

Figure 7

The LYNA::LYNB Isoform Ratio Is Prognostic in Breast Cancer (A) Representative semiquantitative RT-PCR expression analysis of ESR1 and LYN levels in breast cells from reduction mammoplasty tissue separated into stromal cells, basal cells, luminal progenitors, and mature luminal populations (Iriondo et al., 2015). Plot is representative of outcomes of four independent sorts. (B) Quantitation of ESR1 expression levels, confirming ESR1 is most highly expressed in the mature luminal population (mean + SD relative to mature luminal cells; n = 3 independent cell preparations; unpaired two-tailed t test; ^∗∗p < 0.01). (C) Quantitation of the relative LYNA::B ratio (n = 4 independent cell preparations; paired two-tailed t test; ^∗p < 0.05) in breast cell populations. (D) Expression of LYNA and LYNB isoforms in The Cancer Genome Atlas (TCGA) breast cancer data split by immunohistochemistry (IHC)-defined TNBC (n = 112) and non-TNBC (n = 470) tumors. Tumors are also color-coded based on their PAM50 molecular subtype. (E) For each sample shown in (D), the LYNA::B ratio was established based on log2 expression of LYNA over LYNB. The distribution of LYNA::B ratios among samples derived from normal breast tissue (data also from TCGA), TNBC and non-TNBC were comparable when tested by Wilcoxon rank-sum test. (F) Density distributions of LYNA::B log2 RSEM expression ratios in the TCGA breast cancer dataset. A LYNA::B ratio cutoff of 7.3 is indicated. (G) Breast cancer-specific survival of patients based on LYNA::B isoform expression ratio from TCGA data in patient groups dichotomized at the 7.3 ratio boundary. The IHC phenotype of each tumor (where known) is indicated by the color of each data point. A high LYNA::B ratio selects patients with a shorter survival time independent of the breast cancer subtype (Wilcoxon rank-sum test). See also Figures S9 and S10.