NFAT promotes carcinoma invasive migration through glypican-6

Abstract

Invasive migration of carcinoma cells is a prerequisite for the metastatic dissemination of solid tumours. Numerous mechanisms control the ability of cancer cells to acquire a motile and invasive phenotype, and subsequently degrade and invade the basement membrane. Several genes that are up-regulated in breast carcinoma are responsible for mediating the metastatic cascade. Recent studies have revealed that the NFAT (nuclear factor of activated T-cells) is a transcription factor that is highly expressed in aggressive breast cancer cells and tissues, and mediates invasion through transcriptional induction of pro-invasion and migration genes. In the present paper we demonstrate that NFAT promotes breast carcinoma invasion through induction of GPC (glypican) 6, a cell-surface glycoprotein. NFAT transcriptionally regulates GPC6 induction in breast cancer cells and binds to three regulatory elements in the GPC6 proximal promoter. Expression of GPC6 in response to NFAT signalling promotes invasive migration, whereas GPC6 silencing with shRNA (small-hairpin RNA) potently blocks this phenotype. The mechanism by which GPC6 promotes invasive migration involves inhibition of canonical β-catenin and Wnt signalling, and up-regulation of non-canonical Wnt5A signalling leading to the activation of JNK (c-Jun N-terminal kinase) and p38 MAPK (mitogen-activated protein kinase). Thus GPC6 is a novel NFAT target gene in breast cancer cells that promotes invasive migration through Wnt5A signalling.

Figure 1. Activation of NFAT1 enhances GPC6 expression and invasive migration

(A) Total RNA extracted from control or dox-treated SUM.N1#16 cells assayed for NFAT and GPC6 expression by RT-PCR. (B) Protein extracts from untreated cells or cells stimulated with dox (1 μg/ml) with or without PMA/ionomycin (ION) (100nM) immunoblotted with anti-HA, anti-GPC6 or anti-actin antibodies. (C) MDA-MB-231 cells transfected with empty vector (HA) or HA–GPC6 and assayed for Transwell migration and Matrigel invasion. GPC6 expression was confirmed by immunoblotting with anti-HA antibodies. (D) Migration and Matrigel invasion of HA–GPC6-transfected SUM-159PT cells. HA–GPC6 expression was revealed by immunoblotting with anti-HA antibody. Results are means±S.D. of triplicate measurements for at least two independent experiments. *P< 0.05 and #P< 0.01 between the control and transfected cells using an unpaired Student's t test.

Figure 2. Silencing GPC6 expression reduces invasive migration

SUM-159PT (A) and MDA-MB-231 (B) cells transfected with GPC6-specific shRNA in pLKO and assayed for migration and Matrigel invasion. GPC6 silencing by shRNA was confirmed by immunoblotting. Results are means±S.D. of triplicate measurements for at least two independent experiments. *P< 0.05 and #P< 0.01 between the control and transfected cells using an unpaired Student's t test. (C) SUM.N1#16 cells transfected with empty knockdown vector pLKO or GPC6-specific shRNA#1. Invasion of the transfected cells in Matrigel was determined in vitro with or without adding dox to induce exogenous NFAT expression. Results are means±S.D. of triplicate measurements for at least two independent experiments. *P< 0.05 between the cells transfected with pLKO or GPC6 shRNA in pLKO, with or without adding dox using an unpaired Student's t test. NFAT activation and GPC6 silencing were confirmed by immunoblotting.

Figure 3. NFAT regulates the transcription of GPC6

(A) Nucleotide sequence of the human GPC6 putative promoter region from −551 to +2. The transcriptional start site is indicated by an arrow and is assigned +1. The potential NFAT-binding sites are double underlined and putative AP-1 sites are single underlined. The numbers indicate the positions of the cis-regulatory elements relative to the transcriptional start site. (B) Schematic representation of the GPC6 deletion constructs. The filled boxes and stippled boxes represent putative NFAT- and AP-1-binding sites respectively. The results of luciferase assays are depicted on the right. (C) Schematic representation of alterations in the human GPC6 promoter region from −551 to +2. Deletions of the NFAT- and AP-1-putative binding sites are denoted by open boxes in the respective mutant constructs. Luciferase signals from the various constructs are shown on the right. Results are means±S.D. of triplicate measurements for at least two independent experiments. *P< 0.05 and #P< 0.01 between the Pr-551 and the promoter construct being examined using an unpaired Student's t test.

Figure 4. NFAT directly binds to the GPC6 proximal promoter

Oligonucleotides containing NFAT- and AP-1- binding sites in the IL (interleukin)-2 (A) and GPC6 (B) promoter with a NFAT site at −354 and an AP-1 site at −337, incubated with nuclear extracts from SUM.N1#16 cells. DNA–protein complexes were resolved and detected by chemiluminescence. Nuclear extracts (NX) used were from untreated (UT) or dox–treated cells. Molar excesses (100-fold) of unlabelled double-stranded oligonucleotides as indicated at the top of the lanes were used as competitors. – signifies no added competitor; NFAT, NFAT-AP1 sites in the IL-2 promoter; G507, GPC6 probe with an NFAT site at −507 and an AP-1 site at −498; G354, GPC6 probe with an NFAT site at −354 and an AP-1 site at −337; G141, GPC6 probe with an NFAT site at −141; κE2, immunoglobulin κ chain enhancer E box; Sp1, consensus sequence of the transcription factor SP1. The bracket at the left of the blot indicates the protein–DNA complexes formed by NFAT. The arrow at the right-hand side denotes non-specific binding (NS). (C) Biotinylated oligonucleotide probes indicated at the bottom incubated with nuclear extracts from dox-treated SUM.N1#16 cells in the presence of non-immune IgG or antibodies specific against NFAT1 or HA. – signifies no IgG or antibody; IgG, 2 μg of non-immune IgG; α-NFAT, 2 μg of NFAT1 polyclonal antibody; α-HA, 2 μg of monoclonal HA antibody. The brackets at the left-hand side of the blots indicate the NFAT–DNA complexes, whereas the arrowheads on the right-hand side denote the supershifted DNA–protein complexes.

Figure 5. GPC6 and NFAT1 inhibit canonical Wnt/β-catenin signalling

(A) SUM 159-PT cells transiently transfected with TOPflash and S37A mutant β-catenin, HA, GPC6 or NFAT1 plasmids, along with pCS2-(n)-βgal to control for transfection efficiency. Luciferase activities were determined 24 h after transfection. (B) Mutant S37A β-catenin transiently transfected with increasing amounts of GPC6. Luciferase activities from TOPflash were determined 24 h after transfection. (C) Similar to (B), except plasmids encoding NFAT1 were used. Results are means±S.D. of triplicate measurements for at least two independent experiments. #P< 0.01 in luciferase activities between the S37A and HA, GPC6 and NFAT1 plasmids, between HA control and GPC6 plasmids, and between HA control and NFAT1 plasmids as indicated by the brackets (unpaired Student's t test).

Figure 6. Activation of NFAT induces Wnt5A signalling through GPC6

(A) Lysates prepared from the indicated cell lines with or without overnight incubation with PMA and ionomycin, immunoblotted with anti-GPC6, anti-Wnt5A/B or anti-actin antibodies. (B) MDA-MB-231 stable pools harbouring control tet-pLKO or NFAT1 shRNA were induced with dox (100 nM) for 72 h and lysates immunoblotted with anti-NFAT1, anti-GPC6, anti-Wnt5A/B or anti-actin antibodies. (C) MDA-MB-231 or SUM-159PT cells transiently transfected with empty vector (HA) or HA–GPC6. Lysates were immunoblotted with anti-GPC6, anti-Wnt5A/B or anti-actin antibodies. (D) SUM 159-PT cells transiently transfected with empty vector control or HA–GPC6, together with control siRNA or Wnt5A/B siRNA, then assayed for migration. HA–GPC6 expression and Wnt5A/B silencing were confirmed by immunoblotting. (E) SUM-159-PT cells transiently transfected with either scrambled siRNA or Wnt5A/B siRNA were left unstimulated or stimulated with PMA and ionomycin (100 nM) and assayed for migration in Transwell assays. A portion of the lysates were immunoblotted with anti-Wnt5A/B or anti-actin antibodies. #P< 0.05 (unpaired Student's t test). All experiments are representative of at least three independent experiments.

Figure 7. GPC6 regulates JNK and p38 MAPK signalling

(A–C) SUM-159-PT cells infected with GPC6 shRNA #1 or control empty vector pLKO, serum-starved and then stimulated with EGF for 20 min, and analysed by immunoblotting with anti-GPC6 (A), anti-phospho-JNK (p-JNK) and total JNK (B), anti-phospho-p38 MAPK (pMAPK p38) and total p38 MAPK (C) antibodies. (D) SUM-159PT cells transfected with vector (HA) or HA–GPC6, serum starved and then stimulated with EGF for 20 min, and lysates analysed by immunoblotting for phospho-Akt (pSer⁴⁷³) (pAkt) or total Akt. Anti-actin antibody served as control in all cases. Results are representative of at least three independent experiments.