LIFR is a breast cancer metastasis suppressor upstream of the Hippo-YAP pathway and a prognostic marker

Abstract

There is a pressing need to identify prognostic markers of metastatic disease and targets for treatment. Combining high-throughput RNA sequencing, functional characterization, mechanistic studies and clinical validation, we identify leukemia inhibitory factor receptor (LIFR) as a breast cancer metastasis suppressor downstream of the microRNA miR-9 and upstream of Hippo signaling. Restoring LIFR expression in highly malignant tumor cells suppresses metastasis by triggering a Hippo kinase cascade that leads to phosphorylation, cytoplasmic retention and functional inactivation of the transcriptional coactivator YES-associated protein (YAP). Conversely, loss of LIFR in nonmetastatic breast cancer cells induces migration, invasion and metastatic colonization through activation of YAP. LIFR is downregulated in human breast carcinomas and inversely correlates with metastasis. Notably, in approximately 1,000 nonmetastatic breast tumors, LIFR expression status correlated with metastasis-free, recurrence-free and overall survival outcomes in the patients. These findings identify LIFR as a metastasis suppressor that functions through the Hippo-YAP pathway and has significant prognostic power.

Figure 1

LIFR is a target of miR-9 and mediates its effect on migration, invasion and metastasis. (a) Immunoblotting of LIFR and β-actin in SUM159 cells infected with the miR-9–expressing vector or empty vector. LE, long exposure; SE, short exposure. (b) Luciferase activity of a reporter fused to a wild-type or mutant LIFR 3′ UTR in SUM159 cells with ectopic expression of miR-9. Mock, mock-infected control cells. (c) Immunoblotting of E-cadherin, LIFR and heat shock protein 90 (HSP90) in human breast cancer cell lines. Metastatic cell lines are defined as cell lines that are capable of launching metastases when growing as primary tumors in mice. (d) Quantitative PCR (qPCR) of miR-9 in the same cell lines used in c. (e) Immunoblotting of LIFR and β-actin in SUM159 cells infected with LIFR shRNA (clones A8 and F3, alone or in combination) and in miR-9–overexpressing SUM159 cells (SUM159_miR-9) infected with human LIFR. Scr, the pGIPZ-GFP lentiviral vector with a scrambled sequence that does not target any mRNA. (f) Transwell migration and Matrigel invasion assays of SUM159 cells infected with LIFR shRNA (shLIFR; clones A8 + F3) and of SUM159_miR-9 cells infected with LIFR. (g) Number of GFP-positive foci in the lungs of mice with orthotopic injection of cells described in f at week 12 after implantation. (n = 10 mice per group). Statistical significance was determined by unpaired, two-tailed Student's t test. Data are means ± s.e.m.

Figure 2

Restoring LIFR expression in highly malignant breast cancer cells suppresses metastasis. (a) H&E staining of primary breast tumors isolated from mice with orthotopic injection of LIFR-transduced 4T1 cells (4T1_LIFR) or mock-infected 4T1 cells (4T1_mock) at day 25 after implantation. Scale bar, 100 μm. The weight of the primary tumor is indicated in parentheses. (b,c) Bright-field imaging (b) and H&E staining (c) of lungs isolated from mice with orthotopic injection of LIFR-transduced or mock-infected 4T1 cells at days 25 and 31 after implantation. Scale bars: b, 2,000 μm; c, 500 μm. Arrows and circles in b indicate visible metastatic nodules. Insets in c are high-magnification (×200) images of specific areas in the corresponding low-magnification (×25) images. (d) Number of metastatic nodules in the lungs of mice with orthotopic injection of LIFR-transduced or mock-infected 4T1 cells at days 25 and 31 after implantation. Data are means ± s.e.m. (n = 7, 9, 10 and 8 mice, respectively, in the four groups shown from left to right). (e,f) Human-specific vimentin immunohistochemical staining (e) and the number of vimentin-positive foci (f) in the lungs of mice with orthotopic injection of LIFR-transduced MDA-MB-231 cells (MDA-MB-231_LIFR) or mock-infected MDA-MB-231 cells (MDA-MB-231_mock) at week 10 after implantation. Scale bar, 600 μm. Insets in e are high-magnification (×600) images of vimentin-positive foci in the corresponding low-magnification (×40) images. Data in f are means ± s.e.m. (n = 9 and 8 mice in the control and LIFR groups, respectively). (g,h) Bright-field imaging (g, left and middle), H&E staining (g, right) and the number of metastatic nodules (h) in the lungs of mice with tail vein injection of LIFR-transduced or mock-infected 4T1 cells at day 21 after implantation. Scale bar, bright-field images (left and middle), 2,000 μm; H&E staining (right), 500 μm. White nodules in g are macroscopic metastases. Data in h are means ± s.e.m. (n = 7 mice per group). Statistical significance was determined by unpaired, two-tailed Student's t test.

Figure 3

LIFR activates Hippo signaling and leads to phosphorylation and functional inactivation of YAP in breast cancer cells. (a) Immunoblotting of phosphorylated YAP (pYAP), YAP and β-actin in mock-infected and LIFR-transduced 4T1 and MDA-MB-231 cells in the presence or absence of LIF stimulation. (b) Immunoblotting of pYAP, YAP and cyclophilin B (CypB) in SUM159 cells infected with LIFR shRNA (shLIFR) or the pGIPZ-GFP lentiviral vector with a scrambled sequence (Scr) that does not target any mRNA. (c) Immunoblotting of YAP, histone H3 (nuclear marker) and HSP90 (cytoplasmic marker) in cytoplasmic and nuclear fractions of 4T1 cells infected with the LIFR-expressing vector or empty vector (mock). The asterisk indicates a nonspecific band that is enriched in the nuclear fraction. (d) Immunofluorescent staining of YAP (green) in mock-infected and LIFR-transduced 4T1 cells. The images on the right are the overlay of YAP and nuclear DAPI (blue) staining of the same field. Scale bar, 20 μm. (e) Immunofluorescent staining of YAP (red) in SUM159 cells infected with the LIFR shRNA (SUM159_shLIFR) or the pGIPZ-GFP lentiviral vector with a scrambled sequence (SUM159_scr). The images on the right are the overlay of YAP and nuclear DAPI (blue) staining of the same field. Scale bar, 20 μm. (f) Percentage of cells with exclusively nuclear (N) YAP and cells with both nuclear and cytoplasmic (N + C) YAP. No cells showed YAP that was localized exclusively in the cytoplasm. Data are means ± s.e.m. (n = 5 fields per group). (g,h) qPCR (g) and immunoblot (h) of CTGF in 4T1 cells infected with the LIFR-expressing vector or empty vector. Data in g are means ± s.e.m. Statistical significance was determined by unpaired, two-tailed Student's t test. (i) Immunoblotting of YAP upstream kinases (phosphorylated MST1 and MST2 (pMST1/2), MST1, phosphorylated LATS1 (pLATS1) and LATS1) in 4T1 cells infected with the LIFR-expressing vector or empty vector. (j) Immunoblotting of total Scribble and plasma membrane-localized Scribble in 4T1 cells infected with the LIFR-expressing vector or empty vector and in SUM159 cells infected with the miR-9–expressing vector or empty vector. Na/K ATPase and HSP90 are markers of the plasma membrane and cytoplasm, respectively. LE, long exposure; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Figure 4

Inhibition of YAP and CTGF mediates the metastasis-suppressing effect of LIFR. (a) Immunoblotting of LIFR, YAP and HSP90 in SUM159 cells transduced with LIFR shRNA (shLIFR) alone or in combination with YAP shRNA (shYAP). (b,c) Bright-field imaging (b) and the number of metastatic nodules (c) in the lungs of mice with tail vein injection of SUM159 cells transduced with LIFR shRNA alone or in combination with YAP shRNA at day 30 after implantation. Scale bar, 2,000 μm. Data in c are means ± s.e.m. (n = 6 mice per group). (d) Immunoblotting of YAP, CTGF and HSP90 in LIFR-expressing 4T1 cells transduced with wild-type (WT) YAP, the S112A YAP mutant or CTGF. (e,f) Primary mammary tumor weight (e) and the number of metastatic nodules in the lungs (f) of mice with orthotopic injection of mock-infected 4T1 cells and LIFR-expressing 4T1 cells transduced with wild-type YAP, the S112A YAP mutant or CTGF at day 22 after implantation. Metastatic nodules were counted under a stereomicroscope and subsequently verified by H&E staining. Data are means ± s.e.m. (n = 9, 10, 10, 9 and 10 mice, respectively, in the five groups shown from left to right). (g) H&E staining of lungs isolated from the mice described in f. Scale bar, 500 μm. Insets are high-magnification images (×200) of specific areas in the corresponding low-magnification (×25) images. Statistical significance was determined by unpaired, two-tailed Student's t test.

Figure 5

LIFR is downregulated in human breast cancer and correlates with clinical outcomes. (a) Immunohistochemical staining of LIFR in representative normal breast, DCIS and invasive breast carcinoma specimens on the NCI Progression TMAs. Brown staining indicates LIFR immunoreactivity. Scale bar, 50 μm. LN, lymph node metastasis. (b) Immunohistochemical staining of LIFR in two representative breast tumor specimens (left, LIFR positive; right, LIFR negative) on the NCI Prognostic TMAs. Brown staining indicates LIFR immunoreactivity. Scale bar, 50 μm. The survival times of both patients are listed. (c–e) Kaplan-Meier graphs representing the probability of cumulative metastasis-free (free of distant metastasis) survival (c), recurrence-free (recurrence indicates tumor relapse at the primary site, the metastatic site or both) survival (d) and overall survival (e) in patients with breast cancer stratified according to LIFR expression status in their primary tumors. Survival time data are presented as means ± s.e.m. (mean survival time is estimated as the area under the survival curve). The log-rank test P value reflects the significance of the correlation between LIFR positivity and longer survival outcome. (f) Model of two metastasis suppressor pathways that are negatively regulated by miR-9 in breast cancer cells. Green indicates oncogenic and/or pro-metastatic factors; pink indicates tumor-suppressing and/or metastasis-suppressing factors. MST1/2, mammalian Hippo homologs 1 and 2; LEF/TCF, lymphoid enhancer-binding factor/T cell-specific factor; LATS1/2, large tumor suppressor homologs 1 and 2; TEAD, TEA domain. “p” in the circles indicates phosphorylation.