Ikkb Enforces a LIN28B/TCF7L2 Positive Feedback Loop That

Published OnlineFirst March 5, 2015; DOI: 10.1158/0008-5472.CAN-14-2111 Cancer Tumor and Stem Cell Biology Research

IKKb Enforces a LIN28B/TCF7L2 Positive Feedback Loop That Promotes Cancer Cell Stemness and Metastasis Chong Chen1,2, Fengqi Cao1,2, Lipeng Bai1,2,Yan Liu1,2, Junling Xie1,2, Wei Wang1,2, Qin Si1,2, Jian Yang3, Antao Chang3, Dong Liu4, Dachuan Liu4, Tsung-Hsien Chuang5, Rong Xiang3, and Yunping Luo1,2

Abstract

Considerable evidence suggests that proinflammatory path- phenomena were driven through interactions with TCF7L2, ways drive self-renewal of cancer stem-like cells (CSC), but the which enhanced LIN28B expression by direct binding to intron underlying mechanisms remain mainly undefined. Here we 1 of the LIN28B gene, which in turn promoted TCF7L2 mRNA report that the let7 repressor LIN28B and its regulator IKBKB translation through a positive feedback loop. Notably, RNAi- (IKKb) sustain cancer cell stemness by interacting with the mediated silencing of LIN28B or pharmacologic inhibition of Wnt/TCF7L2 (TCF4) signaling pathway to promote cancer IKKb was sufficient to suppress primary and metastatic tumor progression. We found that LIN28B expression correlated with growth in vivo. Together, our results establish the LIN28B/ clinical progression and stemness marker expression in breast TCF7L2 interaction loop as a central mediator of cancer stem- cancer patients. Functional studies demonstrated that the ness driven by proinflammatory processes during progression stemness properties of LIN28B-expressing human breast and and metastasis, possibly offering a new therapeutic target for lung cancer cells were enhanced by IKKb,whereaslossof generalized interventions in advanced cancers. Cancer Res; 75(8); LIN28B abolished stemness properties in these settings. These 1–11. 2015 AACR.

Introduction Several studies have also revealed that the NF-kB pathway plays critical roles in sustaining the stemness of CSCs (11, 12) and Tumor-promoting inflammation has been recognized as one that targeting this pathway impairs self-renewal of CSCs (8); of the hallmarks of cancer (1). NF-kBactivationisoneofthe however, the molecular mechanisms by which NF-kB regulates pillars of inflammation that provides a mechanistic link CSCs are not fully understood. Wnt/b-catenin signaling con- between inflammation and cancer (2, 3). In fact, it is known trols virtually every aspect of embryonic development and that the IKKb-dependent NF-kB activation pathway provides a mediates homeostatic self-renewal in adult tissues (13, 14). critical molecular link between inflammation and colon cancer Recent studies have demonstrated the essential role of tran- growth in mouse models (4, 5) and that the specific inhibitor scriptional factor TCF7L2 in adult intestinal homeostatic self- for IKKb shows promising antineoplastic effects (6–8). Cancer renewal and oncogenic program (15–17). It is worth noticing stem cells (CSC) are highly tumorigenic, relatively resistant to that a recent study has revealed that interaction between NF-kB conventional chemotherapy and radiotherapy, and involved in and Wnt signaling promotes tumor-initiating cell traits during both tumor initiation and metastasis (9, 10). It has been intestinal tumorigenesis (18). proposed that the development of more effective cancer ther- Mammalian homologs of LIN28, such as LIN28A and apies may require specific targeting of the CSCs population. LIN28B, bind to terminal loops of the precursors of let-7 family miRNAs and block their processing into mature miRNAs (19, 1Department of Immunology, Institute of Basic Medical Science, Chi- 20). Thus, repression of let-7 is important in establishing the nese Academy of Medical Science and Peking Union Medical College, pluripotent state of embryonic stem cells. Several reports also 2 Beijing, China. Collaborative Innovation Center for Biotherapy, indicate that LIN28 can affect protein levels by regulating School of Basic Medical Science, Chinese Academy of Medical Science and Peking Union Medical College, Beijing, China. 3Department of mRNA stability and modulate cell proliferation by enhancing Immunology, Nankai University, Tianjin, China. 4Xuanwu Hospital, translation of various cell-cycle regulators in murine embryonic Capital Medical University, Beijing, China. 5Immunology Research stem cells (21, 22). Therefore, LIN28 may promote reprogram- Center, National Health Research Institutes, Zhunan, Miaoli, Taiwan. ming through both miRNA-dependent and independent path- Note: Supplementary data for this article are available at Cancer Research ways. In fact, it has been reported that LIN28A/LIN28B can Online (http://cancerres.aacrjournals.org/). promote transformation by repressing let-7 miRNAs, and that Corresponding Author: Yunping Luo, Institute of Basic Medical Science, Chinese activation of LIN28A/LIN28B occurs in many different human Academy of Medical Science and Peking Union Medical College, 5 Dongdan- tumors with a frequency of approximately 15% (23, 24). Recent santiao St., Beijing 10005, China. Phone: 86-01-0691-56475; Fax: 86-010-6525- studies have also suggested that LIN28 plays an important role 6546; E-mail: [email protected] in the maintenance of CSCs (25, 26) and Wnt signaling could doi: 10.1158/0008-5472.CAN-14-2111 directly activate LIN28 to augment breast cancer stem cell 2015 American Association for Cancer Research. expansion (27).

www.aacrjournals.org OF1

Chen et al.

This study aims to obtain evidences that LIN28B and TCF7L2 ical and enzymatic dissociation of primary tumorspheres as are the key molecular players that mediate the function of NF-kB described previously (30). signaling in CSCs and to identify a novel molecular mechanism by which the LIN28B/TCF7L2 interaction loop, when activated by RNAi experiment IKKb, sustains the stemness of CSCs to promote cancer progres- RELA, RELB, IKKb, LIN28B, or TCF7L2 siRNA was transiently sion and metastasis. transfected using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions. LIN28B shRNA lentivirus Materials and Methods vectors were obtained from GenePharma and transfected into Cell lines and reagents HEK293T cells (ATCC) to generate retroviral particles. MDA-MB- The human breast cancer cell lines MDA-MB-231 and MCF-7, 231 or H1299 cells were transduced with lentivirus-containing m NSCLC cell line H1299, human monocytic cell line THP1, and media and selected with puromycin (2 g/mL). After a week, cell embryonic kidney cell line HEK293T cells were obtained from pools were obtained and expanded. ATCC (October 2011) and cultured according to guidelines. All the cell lines were recently authenticated by cellular morphology Western blotting and the short tandem repeat analysis at Microread Inc. (May Cells were lysed in protein extraction reagent (Pierce) and 2014) according to the guidelines from ATCC (28). IMD- loaded onto NuPAGE gels (Invitrogen). Blots were incubated 0354 was purchased from Sigma Aldrich and Taxol was obtained with the respective primary antibody diluted in TBST (contain- from Tocris. ing 0.1% Tween 20 and 5% BSA) overnight at 4 C. Then, blots were washed and incubated with appropriate secondary anti- Immunohistochemistry bodies (Santa Cruz Biotechnology) and detected using Super- Thetissuemicroarrayusedforanalysisofthecorrelationof Signal West Pico Chemiluminescent Substrate (Pierce). The fi LIN28B expression with stemness markers and clinical stages of band intensities in Western blotting were quanti ed with the breast cancer were purchased from Alenabio Inc. The tissue LAS4000 Image Analyzer. microarray used for analysis of the correlation of LIN28B expression with survival of patients with breast or lung cancer Luciferase assay was purchased from Shanghai Outdo Biotech. The pictures of The Cignal Finder Reporter Array (Qiagen) was used to identify immunohistochemistry were captured by microscope (Leica). the signaling pathways that were regulated by IMD-0354. To The expression level of LIN28B was based on the proportion of generate the luciferase reporter vectors, the LIN28B gene frag- positive cancer cells of 5 random fields in the tissues and the ments were amplified from human genomic DNA and cloned into cutoff defined as low (0%–30%), mediate (30%–60%), or the firefly luciferase plasmid pGL3-basic-IRES. For reporter assays, strong (60%–100%) expression of LIN28B. Tyramide Signal above distinct constructs with pRL-TK plasmid (reference) were Amplification (PerkinElmer, Inc.) was used for tissue immu- cotransfected into MDA-MB-231 or H1299 cells. To analyze the nofluorescence. Cy3, Cy5, and FITC-conjugated second anti- activity of Wnt/b-catenin signaling, Top-flash or Fop-flash (Milli- body (Sigma) was used to visualize LIN28B, ALDH1, and pore) with pRL-TK plasmid (reference) were cotransfected into OCT4, respectively, in immunofluorescence assay. The pictures MDA-MB-321 or H1299 cells. of tissue immunofluorescence were captured by confocal microscopy (Olympus). Animal study Female NOD/SCID mice, 6 to 8 weeks of age, were purchased Flow cytometry from Vital River Laboratories. Approximately 1 106 of MDA- The ALDEFLUOR assay (Stemcell Technologies) was per- MB-231 cells, mixed with Matrigel (1:1), were transplanted into formed according to the manufacturer's guidelines to identify 4th mammary gland fat pad of NOD/SCID mice. After 10 days, the cells with high ALDEFLUOR activity. Cells were incubated for NOD/SCID mice were treated with IMD-0354 or Taxol. Mice 40 minutes at 37C in the presence or absence of the ALDH were randomly divided into experimental groups with 5 mice in inhibitor diethylaminobenzaldehyde. CD44 and CD24 prima- each group. Primary and metastatic tumors in mice were detected ry antibodies were incubated with single cells in PBS/1% FBS by Bioluminescence (NightOWL LB983, Berthold Technologies) for 30 minutes at 4C. Cells were stained with 7-AAD to exclude or PET scan (MicroPET Focus 120, Siemens). Tissues of tumor and nonviable cells. lung were fixed in formalin and embedded in paraffin for histologic analysis. Tumorsphere assay Single-cell suspensions were seeded in 6-well or 24-well Statistical analysis ultra-low attachment plates (Corning) in sphere-culturing All results were presented as the SEM. The significance medium (Stemcell Technologies) for 7 to 14 days. For limited of correlation between LIN28B level and clinical stages in dilution assays, MDA-MB-231 cells were plated at 1, 10, 100, human breast cancer patients was analyzed by using the 1,000 cells/well or H1299 cells were plated at 10, 100, 1,000, Spearman rank correlation test. The significance of correlation 10,000 cells/well in 24 wells (6 wells/condition). The number between LIN28B and ALDH1, SOX2, or TCF7L2 was deter- of wells with spheres was recorded on day 14 after culturing. mined by using the c2 test. Patient survival curves were Tumorsphere formation was monitored using an inverted Leica plotted using Kaplan–Meier analysis, and the statistical para- microscope fitted with a camera. Data analysis performed using meters were calculated by log-rank (Mantel–Cox) test using the publicly available ELDA software as detailed in ref. 29. GraphPad Prism. Other experiments were analyzed with the Secondary tumorspheres were obtained following the mechan- Student t test. P < 0.05 was considered as the criterion for

OF2 Cancer Res; 75(8) April 15, 2015 Cancer Research

IKKb Activates LIN28B/TCF7L2 Loop to Maintain Cancer Stemness

statistical signiﬁcance (, P < 0.05; , P < 0.01; , P < (n ¼ 27). It was found that LIN28B expression was strongly 0.001). correlated with clinical stages in breast cancer patients (Fig. 1A; Supplementary Table S4). It was also found by IHC staining that LIN28B expression was associated with ALDH1 and SOX2, Results two stemness markers (Fig. 1B and Supplementary Fig. S1A; The expression of LIN28B is positively correlated with stemness Supplementary Tables S5 and S6). These results were extended markers and clinical progress of cancer patients by immunoﬂuorescence staining, which indicated that LIN28B To test the correlation between LIN28B expression and expression was correlated with OCT4 and ALDH1 expression clinical progress of breast cancer patients, immumohistochem- (Fig. 1C). Furthermore, statistical analysis revealed that higher ical (IHC) staining was used to detect the expression of the LIN28B level was associated with shorter overall survival of stemness gene LIN28B in 102 patients with different clinical patients with either breast cancer (n ¼ 159; Fig. 1D) or lung stages, including stage 1 (n ¼ 15), stage 2 (n ¼ 60), and stage 3 cancer (n ¼ 150; Supplementary Fig. S1B). Together with

Figure 1. LIN28B expression correlated with clinical stages and stemness markers in breast cancer tissues. A, IHC analyses for LIN28B expression were classiﬁed as absent/weak, moderate, or strong in either human breast cancer tissues with different clinic stages or normal human breast tissue. Percentages of LIN28B expression at different levels in each stage are depicted on the right. The representative staining results are shown on the left. Scale bars, 200 mm; 20 mm (insets). B, IHC analyses were used for LIN28B and ALDH1 expression in 102 human breast cancer specimens. Analysis is shown for percentages of specimens with either low or high LIN28B expression relative to ALDH1 level. The representative cases are shown on the left. Scale bars, 200 mm; 20 mm (insets). C, immunoﬂuorescence staining for OCT4, ALDH1, and LIN28B expression in human breast cancer tissues. Scale bar, 10 mm. D, Kaplan–Meier analysis of LIN28B expression in survival of patients with breast cancer (n ¼ 159). , P < 0.05.

www.aacrjournals.org Cancer Res; 75(8) April 15, 2015 OF3

Chen et al.

clinical data, these results suggest that the expression of LIN28B 231 cells and H1299 cells were both significantly decreased after in tumor tissue correlates positively with their metastasis and LIN28B knockdown (Fig. 2E and Supplementary Fig. S5E). Fur- stemness markers, suggesting that LIN28B might be an impor- thermore, to address these effects in vivo, it is important to note tant player in driving cancer progress. that tumor grew distinctly slower in NOD/SCID mice injected with shLIN28B-MDA-MB-231 cells than with wild-type MDA- MB-231 cells in control mice (Fig. 2F). Similarly, tumors grew LIN28B is a key factor in maintaining stemness phenotypes of much slower in the shLIN28B-H1299 group compared with the CSCs and promotes tumor cell metastasis control group (Supplementary Fig. S5G), and the survival time of fl As reported previously, LIN28B was involved in in ammation tumor-bearing mice was significantly extended in shLIN28B leading to cell transformation (23). Here, we initially found groups (Supplementary Fig. S5H). A limiting dilution assay highly metastatic human breast cancer cell line MDA-MB-231 demonstrated that 1 of 5 of mice of the shLIN28B group formed – cells and human non small cell lung cancer cell line H1299 cells tumors compared with 5 of 5 of the control group when H1299 with high levels of LIN28B expression. On the basis of these cells were injected below one million (Supplementary Table S12). fi ndings, we established siRNA-LIN28B (siLIN28B) and two Moreover, pulmonary metastasis in groups of mice with reduced LIN28B knockdown to be stable pools in MDA-MB-231 cells and LIN28B was clearly suppressed as indicated by PET scan and H1299 cells (shLIN28B-1, shLIN28B-2; Supplementary Fig. S5A). hematoxylin and eosin (H&E) staining analysis (Fig. 2G and H As expected, low expression of LIN28B reduced the expression of and Supplementary Fig. S5F). Together with these data, it was SOX2 OCT4 NANOG stemness-related genes, including , , and ,in apparent that LIN28B played an important role in maintaining MDA-MB-231 and H1299 cells (Fig. 2A). Moreover, we also stemness phenotypes of cancer cells. found that low expression of LIN28B reduced the proportion of þ ALDH cells (Fig. 2B and Supplementary Fig. S5B) and restrained the tumorsphere forming ability of MDA-MB-231 cells and IKKb activates LIN28B and sustains the cancer stemness in a H1299 cells (Fig. 2C and D and Supplementary Fig. S5C and human breast cancer model S5D; Supplementary Tables S9 and S10). Inversely, the CSC During breast cancer progression, the NF-kB signaling pathway þ population CD44 CD24 cells in less malignant MCF-7 human was found to be markedly activated by immune inflammatory breast cancer cells was increased and its tumorsphere forming cells and other factors in the tumor microenvironment (TME) ability was enhanced after overexpression of LIN28B (Supple- such as tumor-associated macrophages (TAM; refs. 31–33). Here, mentary Fig. S4B, C). Moreover, the invasive ability of MDA-MB- we performed experiments to further explore the correlation

Figure 2. LIN28B maintains the stemness phenotype of CSCs and promotes tumor growth and metastasis in vivo. A, Western blot analysis for LIN28B, SOX2, OCT4, and NANOG in MDA-MB- 231 cells that were transfected with þ shLIN28B. B, percentage of ALDH cells of shLIN28B-MDA-MB-231 cells was detected by ﬂow cytometry. C, Primary mammosphere frequencies of shLIN28B-MDA-MB-231 cells were calculated and representative single sphere images were visualized by microscope (bottom). D, secondary mammosphere frequency of shLIN28B-MDA-MB-231 cells was calculated and representative sphere images were visualized by microscope (bottom). Scale bar, 200 mm. E, Matrigel invasion assay for shLIN28B- MDA-MB-231 cells were detected by microscope and OD value. Scale bar, 100 mm. F, the curve of tumor growth in NOD/SCID mice after injection of shLIN28B-MDA-MB-231 cells (n ¼ 5). G, the primary and metastasis tumor was monitored by PET scan. H, The pulmonary metastases in mice and lung metastatic foci were evaluated by H&E staining analyses (n ¼ 5). Scale bars, 200 mm; 50 mm (insets). Error bars, mean SD/SEM from three independent experiments. P < 0.01, P < 0.001.

OF4 Cancer Res; 75(8) April 15, 2015 Cancer Research

IKKb Activates LIN28B/TCF7L2 Loop to Maintain Cancer Stemness

between inflammatory signals induced by TME and the degree of lung were much fewer in groups of mice subjected to IMD-0354 þ stemness in CSCs. ALDH cells, isolated from MDA-MB-231 and treatment (Fig. 3J). These results proved that IKKb in the NF-kB H1299 cancer cells, were used as CSC model. We initially found signaling pathway, which was activated by the TME such as TAMs, þ ALDH cells to be strongly tumorigenic, as indicated by tumors played a critical role in up regulating LIN28B and in sustaining þ growing more rapidly in NOD/SCID mice injected with ALDH cancer stemness properties in a MDA-MB-231 breast cancer cells (Supplementary Fig. S2D; Supplementary Table S8). In model. addition, mammosphere formation and expression of stemness-related genes, such as LIN28B, SOX2, OCT4, and NANOG, þ increased in ALDH cells (Supplementary Fig. S2A and S2B). We TCF7L2 is required for activation of LIN28B, which targets also used human macrophages THP1, which were treated with TCF7L2 directly via a positive feedback loop PMA and TH2 cytokines to induce M2 phenotype macrophages as As described above, activation of LIN28B is required for main- TAMs (34) that were cocultured with MDA-MB-231 and MCF-7 tenance of stemness and enhancement of cancer cell metastasis in cells (Supplementary Fig. S3A). It was found that this maneuver MDA-MB-231 breast cancer cells and H1299 lung cancer cells. To þ clearly increased the percentage of ALDH cells in MDA-MB-231 unveil the molecular mechanism behind the process, we per- þ cells and the percentage of CD44 CD24 cells in MCF-7 cells formed an experiment to test which signaling pathways were (Supplementary Fig. S3B). Similarly, tumorsphere formation of involved in regulation of IKKb or LIN28B in MDA-MB-231 cells. MDA-MB-231 cells was also increased in MDA-MB-231 cells that A pathway reporter array including 10 selected cancer-related were cocultured with treated THP1 cells (Supplementary Fig. signaling pathways was used to detect whether it could be regu- S3C). Results of Western blots indicated upregulated expressions lated by IMD-0354 on MDA-MB-231 cells. The result showed of p-IKKb, LIN28B in either MDA-MB-231 or MCF-7 cells that clearly that Wnt/b-catenin signaling pathway was significantly were cocultured with THP1-treated cells (Supplementary Fig. inhibited by IMD-0354 (Fig. 4A), suggesting a correlation S3D). Moreover, cell surface marker CD44 was upregulated in between Wnt/b-catenin and NF-kB signaling. Western blot anal- MCF-7 cells when cocultured with THP1 cells by flow cytometry yses further indicated that the expression of TCF7L2 (TCF4), a key þ (Supplementary Fig. S3E). These results suggest that ALDH cells transcription factor in the Wnt/b-catenin signaling pathway, was had CSCs' properties, and that the inflammatory TME could downregulated by either IKKb inhibitor IMD-0354 or siIKKb, but activate IKKb to enhance tumor stemness properties in cancer not by siRELA or siRELB, in MDA-MB-231 and H1299 cells (Fig. cells. 4B and Supplementary Fig. S6B and S6C). Furthermore, we NF-kBwasshowntobeimportantinTICsisolatedfrom investigated whether LIN28B could be regulated by TCF7L2. As þ HER2 breast cancer (35). On the basis of this fact, we inves- expected, LIN28B expression was reduced by siTCF7L2 (Fig. 4D tigated how NF-kB activation was involved in the maintenance and Supplementary Fig. S7A) and LIN28B expression was corre- of stemness in breast cancer cells. Initially, we found the lated with TCF7L2 in breast cancer tissues (Fig. 5F; Supplementary expressions of p-IKKb and LIN28B were much higher in Table S7). A further in depth investigation indicated the mech- þ ALDH of MDA-MB-231 and H1299 cells (Fig. 3A). We then anism underlying the important interaction between TCF7L2 and defined the factor in the NF-kB signaling pathway that played a LIN28B. In this regard, PROMO software analysis identified two leading role in this process. To this end, the main molecules in putative binding sites of TCF7L2E (one isoform of TCF7L2) in the NF-kB signaling pathway were knocked down by siRNA, intron 1 of the LIN28B gene, which were respectively named site 1 respectively, including siRELA, siRELB, or siIKKb. Only silence (AGCAAAG) and site 2 (CTGTACT; Fig. 4D). Chromatin immu- of IKKb could reduce the expression of the stemness gene noprecipitation (ChIP) found that only site 2 could be tested in LIN28B and SOX2 in MDA-MB-231 and H1299 cells (Fig. 3B the TCF7L2/DNA complex (Fig. 4E and Supplementary Fig. S7B). and Supplementary Fig. S6B). Furthermore, the above results We then constructed 4 plasmids containing both site 1/site 2, site were confirmed by an IKKb-selective inhibitor (IMD-0354) to 1, site 2 or mutant site 2 separately. A Luciferase reporter assay treat MDA-MB-231 or H1299 cells. As expected, it downregu- revealed that only site 2 had transcriptional activity that was in lated the expression of LIN28B and other stemness genes accord with the result of our ChIP assay (Fig. 4F and Supplemen- (SOX2, OCT4, NANOG) in a dose- dependent manner (Fig. tary Fig. S7C). Moreover, the transcriptional activity of the plas- 3F and Supplementary Fig. S6C). Moreover, treatment of MDA- mid containing both site 1 and site 2 was increased significantly MB-231 cells with 5 mmol/L IKKb inhibitor (IMD-0354) when b-catenin was overexpressed (Fig. 4G and Supplementary resulted in a decreased mRNA expression of LIN28B, SOX2 Fig. S7D). These results suggest that Lin28B was a novel target gene (Fig. 3G), and other important stemness properties, including of Wnt/b-catenin signaling pathway and TCF7L2 could promote þ the percentage of ALDH cells(Fig.3C),thefrequencyof LIN28B expression by directly binding to intron 1 of the LIN28B tumorsphere (Fig. 3D; Supplementary Table S11), and invasion gene. ability (Fig. 3E) among MDA-MB-231 cells. Similar results were Unexpectedly, we observed an interesting fact that expression obtained in H1299 cells treated with IMD-0354 (Supplemen- of TCF7L2, but not b-catenin, could be inhibited as well by tary Fig. S6C–S6F). shLIN28B (Fig. 5A). It indicated that downregulation of LIN28B Furthermore, an in vivo experiment was performed to verify our could reduce luciferase activity of TOP- flash (Fig. 5C and Sup- in vitro results (Fig. 3H). The tumor volume was decreased in the plementary Fig. S7E), as well as the expression of such TCF7L2 IMD-0354 group at the end of 7 weeks (Supplementary Fig. S8A). downstream genes as BCL2, CCND1, and MYC in MDA-MB-231 However, there was no difference in body weight of mice after cells (Fig. 5D). Furthermore, RNA immunoprecipitation (RIP) IMD-0354 treatment (Supplementary Fig. S8B). Pulmonary experiments showed that LIN28B could directly bind to the metastases in groups of mice treated with IMD-0354 were fewer mRNA of TCF7L2 in both of MDA-MB-231 and H1299 cells (Fig. than those in control groups (Fig. 3I). In fact, H&E staining of lung 5E). The down expression of LIN28B decreased the expression of specimens indicated clearly that the number of metastatic foci in TCF7L2 in MDA-MB-231 cells that was also confirmed by

www.aacrjournals.org Cancer Res; 75(8) April 15, 2015 OF5

Chen et al.

Figure 3. þ IKKb up regulates expression of LIN28B and sustains cancer stemness in cancer cells. A, Western blot analysis for LIN28B, p-IKKb, and IKKb in either ALDH or ALDH cells of MDA-MB-231 and H1299 cells. B, Western blot analysis for IKKb, RELA, RELB, and LIN28B in MDA-MB-231 cells that were transfected with siRELA, siRELB, or siIKKb, respectively. C, percentage of ALDHþ cells of MDA-MB-231 cells that were treated with IMD-0354 (2 mmol/L) was detected by ﬂow cytometry. D, mammosphere frequency of MDA-MB-231 cells that were treated with IMD-0354 (1 mmol/L) were calculated and representative images were visualized by microscope (below). Scale bar, 200 mm. E, Matrigel invasion assay for MDA-MB-231 cells that were treated with IMD-0354 (2 mmol/L) were detected by microscope and OD value. Scale bar, 100 mm. F, Western blots for p-IKKb,IKKb, p-RELA, RELA, LIN28B, SOX2, OCT4, and NANOG in MDA-MB-231 cells treated with IMD-0354 at doses of 1, 5, 10, and 50 mmol/L. G, qPCR for LIN28B, SOX2, OCT4, and NANOG in MDA-MB-231 cells that were treated with IMD-0354 (5 mmol/L). H, schedule of IMD-0354 treatment in vivo (n ¼ 6). I, bioluminescent imaging was used to detect the pulmonary metastasis at week 7 (n ¼ 6). J, lung metastatic foci were evaluated by H&E staining (n ¼ 6). Scale bars, 200 mm; 50 mm (insets). Error bars, mean SD from three independent experiments. , P < 0.05; , P < 0.01; , P < 0.001.

immunoﬂuorescence (Fig. 5B). We therefore proposed that TCF7L2 represents a novel mechanism underlying the regulation LIN28B protein binding with TCF7L2 mRNA occurred in the of LIN28B to Wnt/b-catenin signaling pathway via the protein/ cytoplasm, a ﬁnding consistent with our prior data indicating RNA interaction. that LIN28B protein appeared in both cytoplasm and nucleus in breast cancer tissues. Moreover, the expression of LIN28B and IMD-0354 blocks the p-IKKb/LIN28B/TCF7L2 signaling TCF7L2 was highly correlated in breast cancer tissues (Fig. 5F). It is pathway and suppresses stemness and lung metastasis of breast evident that TCF7L2, sustained by IKKb, promoted LIN28B tran- cancer cells. scription by binding site 2 (CTGTACT) in intron 1 of the LIN28B Recent evidences suggested that breast CSCs were relative- gene and LIN28B targets with mRNA of TCF7L2 directly as a ly resistant to chemotherapy and contributed to tumor recur- positive feedback loop. The important interaction of LIN28B and rence after therapy (36). Previous reports indicated Taxol, one of

OF6 Cancer Res; 75(8) April 15, 2015 Cancer Research

IKKb Activates LIN28B/TCF7L2 Loop to Maintain Cancer Stemness

Figure 4. TCF7L2 is required for IKKb-mediated activation of LIN28B. A, signaling pathway reporter array was used for seeking the signaling pathway associated with IKKb on MDA-MB-231 cells treated with IMD-0354 (5 mmol/L). B, Western blot analysis for TCF7L2 and b-catenin in MDA-MB-231 cells that were treated with IMD-0354 or siRNAs. C, software PROMO was used to identify two putative binding sites of TCF7L2E in intron 1 of the LIN28B gene. D, Western blots for TCF7L2 and LIN28B in MDA-MB-231 cells that were treated with siTCF7L2. E, ChIP was performed to verify TCF7L2 binding with the LIN28B gene in MDA-MB-231 cells. F, luciferase reporter system was used to identify the TCF7L2 binding site of the LIN28B gene in MDA-MB-231 cells that were transfected with plasmids containing both site 1/site 2, site 1, site 2, or mutated site 2 separately. G, reporter plasmid containing both site1/site2 and a b-catenin or control plasmid were cotransfected into MDA-MB-231 cells, and luciferase activity was detected. Error bars, mean SD from three independent experiments. , P < 0.001.

first-line clinical chemotherapy drug used in breast cancer treat- significant difference was found in the 4T1 model (Supplemen- ment, could upregulate CSC populations in vitro (37) or in tary Fig. S10B); however, tumor metastases in the lungs of these xenografts models (38, 39). To further investigate the therapeutic mice in both MDA-MB-231 and 4T1 models were not decreased in potential of IMD-0354, this study investigated the roles of the two the Taxol-only group (Fig. 6D and E and Supplementary Fig. þ drugs in CSCs and tumor progession. Results showed that Taxol S10C). The proportion of ALDH cancer stem cells in the IMD- þ could efficiently improve the proportion of ALDH cells and the 0354–only group and the combined treatment group were frequency of tumorsphere in MDA-MB-231 cells (Fig. 6A and reduced in both MDA-MB-231 and 4T1 models; however, it Supplementary Fig. S9A). On the basis of the inhibiting effect of increased in the Taxol-only group in both two cell models (Fig. IMD-0354 on CSCs phenotype and function in vitro, we designed 6F and Supplementary Fig. S10D). Furthermore, histologic anal- a combination therapy strategy. To test the effect of this ther- yses revealed that this combination therapy strategy could sup- apeutic strategy in vivo, MDA-MB-231 or 4T1 cells were trans- press the expression of p-IKKb, TCF7L2, LIN28B, ALDH1, SOX2, planted into the 4th mammary fat pads of NOD/SCID mice and NANOG in tumor tissues, which supported our hypotheses or BALB/C mice and tumor growth and lung metastasis were (Fig. 6G). Taken together, these results suggest that blockade of monitored. The mice were treated with DMSO (control group), IKKb activation with IMD-0354 may be able to inhibit TCF7L2/ IMD-0354 alone, Taxol alone, or both IMD-0354 and Taxol at LIN28B interaction, suppress cancer stemness properties, and the time when the tumor volume reached approximately reduce tumor metastasis. 100 mm3 (Fig. 6B and Supplementary Fig. S10A). Results showed that tumor volumes of the original site of the tumor in the combined treatment group were significantly smaller compared Discussion with those in the control group (Fig. 6C and Supplementary Fig. Our experiment showed that LIN28B was upregulated by S10B). Moreover, the degree of pulmonary metastasis in the IMD- inflammatory factor IKKb and together with TCF7L2 sustained 0354–only group and the combined treatment group significantly the stemness of cancer cells. Marotta and colleagues have dem- reduced (Fig. 6D and E and Supplementary Figs. S9B, S9CC, and onstrated previously that the inflammatory transcription factor þ S10C). In the Taxol-only group, tumor volumes in situ of the STAT3 was highly activated in the CD44 CD24 population (40). MDA-MB-231 model were obviously suppressed (Fig. 6C), but no Recent reports analogously indicated that STAT3 phosphorylation

www.aacrjournals.org Cancer Res; 75(8) April 15, 2015 OF7

Chen et al.

Figure 5. LIN28B regulates Wnt/TCF7L2 signaling via direct binding to the mRNA of TCF7L2. A, Western blot analysis for LIN28B, b-catenin, and TCF7L2 in MDA-MB-231 cells that were transfected with shLIN28B. B, immunofluorescence staining for LIN28B and TCF7L2 in MDA-MB-231 cells that were treated with shLIN28B. Scale bar, 100 mm. C, TOP-flash/FOP- flash and Renilla pRL-TK plasmid were cotransfected into shLIN28B-MDA- MB-231 cells, and luciferase reporter activity was detected. D, Western blot analysis for LIN28B, BCL2, CCND1, and MYC in MDA-MB-231 cells that were treated with shLIN28B. E, RIP was performed and TCF7L2 mRNA was detected by RT-PCR. F, LIN28B expression was associated with TCF7L2 expression in 102 clinical breast cancer specimens. Shown are visualizations of two representative cases (left) and percentages of specimens showing low or high LIN28B expression relative to TCF7L2 level (right). Scale bars, 200 mm; 20 mm (insets). Error bars, mean SD from three independent experiments. , P < 0.01; , P < 0.001.

was positively correlated with ALDH1 expression in cancer cell TCF7L2, which was a Wnt/b-catenin effector. We identified for the lines and human breast cancer samples (41). However, this study first time that TCF7L2 directly bound to intron 1 of LIN28B and demonstrated a novel mechanism, indicating that activation of upregulated LIN28B expression. This finding suggests that IKKb,aninflammatory kinase in the NF-kB pathway, was strongly LIN28B a novel direct downstream target gene of the Wnt/b-cate- correlated with the expression of stemness genes, especially nin pathway. Interestingly, a recent research revealed that LEF1, LIN28B in human breast cancer cell lines or tissues. Previous another effector of Wnt/b-catenin pathway, could activate LIN28A studies also showed that LIN28B promoted tumor cell transfor- via direct binding to the LIN28A promoter (27). This report mation (23, 24) and had a functional role in the maintenance of provided compelling evidence that both LIN28A and LIN28B CSCs (25, 26). Consistent with these findings, our experiment were direct target genes of the Wnt/b-catenin pathway. Hence, our results showed that expression of LIN28B was highly positively results suggest that IKKb regulated LIN28B expression with correlated with tumor progress in human breast cancer patients, TCF7L2 assistance. However, we unexpectedly discovered that indicating that there were high levels of LIN28B in stage 3 patients LIN28B bound directly to mRNA of TCF7L2, thereby activating with lymphatic metastasis of breast cancer. LIN28B was shown to the Wnt/b-catenin signaling pathway. In this regard, it is well be a powerful predictor of poor clinical outcome. Besides, LIN28B known that LIN28B plays analogic biologic function due to their was also shown highly positively correlated with stemness mar- parallel structural domains (42). In fact, they bind to the terminal kers (ALDH1, SOX2, and OCT4) in breast cancer tissues. As far as loops of precursors of let-7 family miRNAs and block their mechanisms are concerned, we determined that selected blockade processing into mature miRNAs. Furthermore, another study of IKKb phosphorylation reduced the CSC population and reported that let-7–dependent effects of LIN28B may supersede LIN28B expression by using either siIKKb or IKKb inhibitor let-7–independent effects on intestinal tissue growth (43). There- IMD-0354. In this regard, Iliopoulos and colleagues found that fore, our study revealed a novel let-7–independent mechanism of LIN28B was activated by RELA to promote cell transformation LIN28B in regulating the function of CSC. (23). In contrast, results from our study indicated that RELA was The development of strategies to effectively target the CSC not the only player in regulating LIN28B expression; instead, IKKb population needs to be extended to improve outcome of did play a key role in our human breast and lung cancer model. cancer therapy. Here, we proposed such an approach, based Importantly, we noted that suppression of IKKb can inhibit the on blocking the IKKb kinase activation to eliminate CSCs. activity of Wnt/b-catenin signaling and reduced the expression of Meanwhile, we also designed a combination therapeutic

OF8 Cancer Res; 75(8) April 15, 2015 Cancer Research

IKKb Activates LIN28B/TCF7L2 Loop to Maintain Cancer Stemness

Figure 6. IMD-0354 blocks p-IKKb/TCF7L2/LIN28B signaling pathway and suppresses lung metastasis of MDA-MB-231 breast cancer cells in a NOD/SCID mouse model. A, mammosphere frequency of MDA-MB-231 cells that were treated with DMSO, IMD-0354 (2 mmol/L), Taxol (5 nmol/L), Taxol þ IMD-0354 were calculated and representative images were visualized by microscope (below). Scale bar, 200 mm. B, schedule of IMD-0354 or Taxol treatment in vivo (n ¼ 5). C, progression of tumor volumes during days 7–42 in mice treated with IMD-0354 or Taxol (n ¼ 5). D, bioluminescent imaging was used to detect the pulmonary metastasis at week 8 (n ¼ 5). E, lung metastatic foci were evaluated by H&E staining (n ¼ 5). Scale bars, 200 mm; 50 mm (insets). F, single-cell suspensions derived from MDA-MB-231 xenografts harvested on day 56 were analyzed for ALDH activity using the ALDEFLUOR assay. G, MDA-MB-231 xenografts from each group were collected, and immunohistochemistry staining was done to detect the expression of p-IKKb, TCF7L2, LIN28B, ALDH1, SOX2, and NANOG. Scale bar, 20 mm. Error bars, mean SD/SEM from three independent experiments. , P < 0.05; , P < 0.01; , P < 0.001. strategy of IKKb inhibitor IMD-0354 and the cytotoxic agent increased after Taxol treatment alone, which was analogous Taxol. We tested the effects of this treatment in NOD/SCID and with other studies (36, 37). Moreover, we found that expres- BALB/c mice and found that IMD-0354 treatment inhibited sion of TCF7L2, LIN28B, SOX2, and NANOG were all tumor metastasis to lung. In fact, the CSC population was decreased in tumor tissues following the combination therapy. signiﬁcantly reduced by either IMD-0354 treatment alone or in In fact, IMD-0354 did not result in any side effect in mice combined therapy with Taxol. However, the CSC population (Supplementary Fig. S8B). Our results suggested that blocking

www.aacrjournals.org Cancer Res; 75(8) April 15, 2015 OF9

Chen et al.

Disclosure of Potential Conﬂicts of Interest No potential conﬂicts of interest were disclosed by the other authors.

Authors' Contributions Conception and design: C. Chen, L. Bai, Y. Liu, Dong Liu, R. Xiang, Y. Luo Development of methodology: C. Chen, W. Wang, A. Chang, Dong Liu Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Chen, F. Cao, L. Bai, Dong Liu, Dachuan Liu Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Chen, F. Cao, J. Yang, Dong Liu, R. Xiang Writing, review, and/or revision of the manuscript: C. Chen, Dong Liu, Y. Luo Figure 7. Administrative, technical, or material support (i.e., reporting or organizing The work model of interaction of LIN28B/TCF7L2 loop induced by IKKb data, constructing databases): C. Chen, L. Bai, Q. Si, J. Yang, Dong Liu, promotes cancer stemness and metastasis. The tumor inﬂammatory T.-H. Chuang environment activates the NF-kB signaling pathway in CSCs. Transcription Study supervision: C. Chen, L. Bai, J. Xie, W. Wang, Dong Liu factor TCF7L2, which is regulated by p-IKKb transfers into the nucleus Other (Material support): Y. Liu, Dong Liu from the cytoplasm, activates LIN28B expression by directly binding to intron 1 of LIN28B DNA. The upregulated LIN28B protein directly binds to the mRNA of TCF7L2 in the cytoplasm and promotes TCF7L2 mRNA translation. This Acknowledgments interaction constitutes a positive feedback loop to maintain the stemness of The authors thank Ralph R. Reisfeld (Department of Immunology, The CSC. Inhibition of inﬂammation (p-IKKb) could block this loop and impair the Scripps Research Institute) for critically reading this article. J. Yang was a Postdoc function of CSC thereby reduce the latent metastasis. Associate in New York University.

inflammatory activation may represent a novel and effective Grant Support strategy to target CSCs. This work was supported by grants from the Major State Basic Research Based on two cancer models (breast cancer and lung cancer), Development Program of China (973 program): grant no. 2013CB967202 (Y. this study propsoes a novel mechanism of the inflammatory gene Luo) and the National Natural Science Foundation of China (NSFC): grant no. IKKb regulating stemness and identifying the TCF7L2/LIN28B 91029734 and 81071711 (Y. Luo). The costs of publication of this article were defrayed in part by the payment of loop as the pivotal molecular bridge connecting inflammation page charges. This article must therefore be hereby marked advertisement in and stemness through their interaction with CSCs. This mecha- accordance with 18 U.S.C. Section 1734 solely to indicate this fact. nism suggests a novel therapeutic strategy with clinical potential, i.e., blocking activation of inflammation, that targets CSC and Received July 22, 2014; revised February 5, 2015; accepted February 24, 2015; suppresses tumor metastasis (Fig. 7). published OnlineFirst March 5, 2015.

References 1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell transgenic mice and tumor stem cell expansion. Cancer Res 2010; 2011;144:646–74. 70:10464–73. 2. Karin M. Nuclear factor-kappaB in cancer development and progression. 13. Clevers H, Nusse R. Wnt/beta-catenin signaling and disease. Cell 2012; Nature 2006;441:431–6. 149:1192–205. 3. Hayden MS, Ghosh S. Shared principles in NF-kappaB signaling. Cell 14. Kim W, Kim M, Jho EH. Wnt/beta-catenin signalling: from plasma mem- 2008;132:344–62. brane to nucleus. Biochem J 2013;450:9–21. 4. Greten FR, Eckmann L, Greten TF, Park JM, Li ZW, Egan LJ, et al. IKKbeta 15. Van Es JH, Haegebarth A, Kujala P, Itzkovitz S, Koo BK, Boj SF, et al. A links inﬂammation and tumorigenesis in a mouse model of colitis-asso- critical role for the Wnt effector TCF7L2 in adult intestinal homeostatic self- ciated cancer. Cell 2004;118:285–96. renewal. Mol Cell Biol 2012;32:1918–27. 5. Vlantis K, Wullaert A, Sasaki Y, Schmidt-Supprian M, Rajewsky K, Roskams 16. Korinek V, Barker N, Moerer P, vanDonselaar E, Huls G, Peters PJ, et al. T, et al. Constitutive IKK2 activation in intestinal epithelial cells induces Depletion of epithelial stem-cell compartments in the small intestine of intestinal tumors in mice. J Clin Invest 2011;121:2781–93. mice lacking Tcf-4. Nature 1998;19:379–83. 6. Tanaka A, Konno M, Muto S, Kambe N, Morii E, Nakahata T, et al. A novel 17. van de Wetering M, Sancho E, Verweij C, deLau W, Oving I, Hurlstone A, NF-kappaB inhibitor, IMD-0354, suppresses neoplastic proliferation of et al. The b-catenin/TCF-4 complex imposes a crypt progenitor phenotype human mast cells with constitutively activated c-kit receptors. Blood 2005; on colorectal cancer cells. Cell 2002;111:241–50. 105:2324–31. 18. Sarah Schwitalla, Fingerle AA, Cammareri P, Nebelsiek T, Goktuna€ SI, 7. Tanaka A, Muto S, Konno M, Itai A, Matsuda H. A new IkappaB kinase beta Ziegler PK, et al. Intestinal tumorigenesis initiated by dedifferentiation and inhibitor prevents human breast cancer progression through negative acquisition of stem-cell-like properties. Cell 2013;152:25–38. regulation of cell cycle transition. Cancer Res 2006;66:419–26. 19. Heo I, Joo C, Cho J, Ha M, Han J, Kim VN. Lin28 mediates the terminal 8. Zhou J, Zhang H, Gu P, Bai J, Margolick JB, Zhang Y. NF-kappaB pathway uridylation of let-7 precursor MicroRNA. Mol Cell 2008;32:276–84. inhibitors preferentially inhibit breast cancer stem-like cells. Breast Cancer 20. Viswanathan SR, Daley GQ, Gregory RI. Selective blockade of microRNA Res Treat 2008;111:419–27. processing by Lin28. Science 2008;320:97–100. 9. Visvader JE, Lindeman GJ. Cancer stem cells: current status and evolving 21. Qiu C, Ma Y, Wang J, Peng S, Huang Y. Lin28-mediated post-transcriptional complexities. Cell Stem Cell 2012;10:717–28. regulation of OCT4 expression in human embryonic stem cells. Nucleic 10. Visvader JE, Lindeman GJ. Cancer stem cells in solid tumours: accumu- Acids Res 2010;38:1240–8. lating evidence and unresolved questions. Nat Rev Cancer 2008;8:755–68. 22. Lei XX, Xu J, Ma W, Qiao C, Newman MA, Hammond SM, et al. Determi- 11. Kendellen MF, Bradford JW, Lawrence CL, Clark KS, Baldwin AS. Canonical nants of mRNA recognition and translation regulation by Lin28. Nucleic and non-canonical NF-kappaB signaling promotes breast cancer tumor- Acids Res 2012;40:3574–84. initiating cells. Oncogene 2014;33:1297–305. 23. Iliopoulos D, Hirsch HA, Struhl K. An epigenetic switch involving NF- 12. Liu M, Sakamaki T, Casimiro MC, Willmarth NE, Quong AA, Ju X, et al. kappaB, Lin28, Let-7 MicroRNA, and IL6 links inﬂammation to cell The canonical NF-kappaB pathway governs mammary tumorigenesis in transformation. Cell 2009;139:693–706.

OF10 Cancer Res; 75(8) April 15, 2015 Cancer Research

IKKb Activates LIN28B/TCF7L2 Loop to Maintain Cancer Stemness

24. Viswanathan SR, Powers JT, Einhorn W, Hoshida Y, Ng TL, Toffanin S, et al. cell proliferation and apoptosis. Proc Natl Acad Sci U S A Lin28 promotes transformation and is associated with advanced human 2004;27:10137–42. malignancies. Nat Genet 2009;41:843–8. 34. Sica A, Schioppa T, Mantovani A, Allavena P. Tumour-associated macro- 25. Hayashi S, Tanaka J, Okada S, Isobe T, Yamamoto G, Yasuhara R, et al. phages are a distinct M2 polarised population promoting tumour pro- Lin28a is a putative factor in regulating cancer stem cell-like properties in gression: potential targets of anti-cancer therapy. Eur J Cancer side population cells of oral squamous cell carcinoma. Exp Cell Res 2006;42:717–27. 2013;319:1220–8. 35. Cao Y, Luo JL, Karin M. IkappaB kinase alpha kinase activity is required for 26.ChengSW,TsaiHW,LinYJ,ChengPN,ChangYC,YenCJ,etal. selfrenewal of ErbB2/Her2-transformed mammary tumor-initiating cells. LIN28B is an oncofetal circulating cancer stem cell-like marker asso- Proc Natl Acad Sci U S A 2007;40:15852–7. ciated with recurrence of hepatocellular carcinoma. PLoS ONE 2013;8: 36. Gupta PB, Onder TT, Jiang G, Tao K, Kuperwasser C, Weinberg RA, et al. e80053. Identification of selective inhibitors of cancer stem cells by high-through- 27. Cai WY, Wei TZ, Luo QC, Wu QW, Liu QF, Yang M, et al. The Wnt–b-catenin put screening. Cell 2009;138:645–59. pathway represses let-7 microRNA expression through transactivation of 37. Tam WL, Lu H, Buikhuisen J, Soh BS, Lim E, Reinhardt F, et al. Protein Lin28 to augment breast cancer stem cell expansion. J Cell Sci 2013; kinase C alpha is a central signaling node and therapeutic target for breast 126:2877–89. cancer stem cells. Cancer Cell 2013;24:347–64. 28. American Type Culture Collection. Cell line verification test recom- 38. Ginestier C, Liu S, Diebel ME, Korkaya H, Luo M, Brown M, et al. CXCR1 mendations; ATCC recommends cell line verification tests and guide- blockade selectively targets human breast cancer stem cells in vitro and in lines for publishing. ATCC Technical Bulletin No. 8. Manassas, VA: xenografts. J Clin Invest 2010;120:485–97. ATCC; 2007. 39. Bhola NE, Balko JM, Dugger TC, Kuba MG, Sanchez V, Sanders M, et al. 29. Hu Y, Smyth GK. ELDA: extreme limiting dilution analysis for comparing TGF-b inhibition enhances chemotherapy action against triple-negative depleted and enriched populations in stem cell and other assays. J Immu- breast cancer. J Clin Invest 2013;123:1348–58. nol Methods 2009;347:70–8. 40. Marotta LL, Almendro V, Marusyk A, Shipitsin M, Schemme J, Walker SR, þ 30. Al-Hajj M, Wicha M, Benito-Hernandez A, Morrison S, Clarke M. Prospec- et al. The JAK2/STAT3 signaling pathway is required for growth of CD44 tive identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S CD24 stem cell–like breast cancer cells in human tumors. J Clin Invest A 2003;100:3983–3988. 2011;121:2723–35. 31. Cogswell PC, Guttridge DC, Funkhouser WK, Baldwin AS Jr. Selective 41. Lin L, Hutzen B, Lee HF, Peng Z, Wang W, Zhao C, et al. Evaluation of STAT3 þ þ þ activation of NF-kappa B subunits in human breast cancer: potential roles signaling in ALDH and ALDH /CD44 /CD24 subpopulations of breast for NF-kappa B2/p52 and for Bcl-3. Oncogene 2000;9:1123–31. cancer cells. PLoS ONE 2013;8:e82821. 32. Sovak MA, Bellas RE, Kim DW, Zanieski GJ, Rogers AE, Traish AM, et al. 42. Viswanathan SR, Daley GQ. Lin28: A microRNA regulator with a macro Aberrant nuclear factor-kappaB/Rel expression and the pathogenesis of role. Cell 2010;140:445–9. breast cancer. J Clin Invest 1997;12:2952–60. 43. Madison BB, Liu Q, Zhong X, Hahn CM, Lin N, Emmett MJ, et al. LIN28B 33. Biswas DK, Shi Q, Baily S, Strickland I, Ghosh S, Pardee AB, et al. NF- promotes growth and tumorigenesis of the intestinal epithelium via Let-7. kappa B activation in human breast cancer specimens and its role in Genes Dev 2013;27:2233–45.

www.aacrjournals.org Cancer Res; 75(8) April 15, 2015 OF11

IKKβ Enforces a LIN28B/TCF7L2 Positive Feedback Loop That Promotes Cancer Cell Stemness and Metastasis

Chong Chen, Fengqi Cao, Lipeng Bai, et al.

Cancer Res Published OnlineFirst March 5, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-14-2111

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2015/03/05/0008-5472.CAN-14-2111.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2015/04/02/0008-5472.CAN-14-2111. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.