miR-154* and miR-379 in the DLK1-DIO3 MicroRNA Mega-Cluster Regulate Epithelial to Mesenchymal Transition and Bone Metastasis of Prostate Cancer

Cancer cells often undergo epithelial to mesenchymal transition (EMT) before metastasizing to distant organs. Cancer cells activate embryonic programs and pathways that partially maintain stem cell identity, often referred as “stemness.” Recent studies suggest roles for miRNAs in both metastasis and cancer stem cell formation (1, 2). The miRNA members of the delta-like 1 homolog–deiodinase, iodothyronine 3 (DLK1-DIO3) clusters are activated in embryonic stem cells. During mouse embryogenesis, the miRNA members of the DLK1-DIO3 cluster have been shown to be elevated in the developing central nervous system, skeletal muscle, liver, and lung (3). Similarly, several transgenic mouse models of prostate, liver, and lung cancer also exhibit elevated levels of DLK1-DIO3 cluster miRNA members (4–8). The functions of many of these miRNAs within this imprinted region remain largely unexplored in cancer biology, specifically for EMT and cancer stemness.

Members of the DLK1-DIO3 cluster have been recently shown to be upregulated in the serum of patients with prostate cancer and lung cancer. Members of this cluster, including miR-379, miR-154*, and miR-409-3p, show increased levels in the circulating exosomes of patients with lung adenocarcinomas (9) and breast (10) and prostate (11) cancers. Our recent studies demonstrate the importance of miR-409-3p/5p in prostate cancer growth, EMT, and bone metastasis (12). In this study, we studied the role of miR-154* and miR-379, two key DLK1-DIO3 cluster miRNAs, in several EMT models of human prostate cancer. We demonstrated that these miRNAs modulate EMT and drive metastasis. Thus, these miRNAs are promising therapeutic targets for aggressive prostate cancer.

Three human prostate cancer bone metastatic progression models, ARCaP_E and ARCaP_M (13), LNCaP, and C4-2 (14), were used in our study. Prostate cancer cells and 293T cells were cultured in T-medium (GibcoBRL) supplemented with 5% heat-inactivated FBS (Bio-Whittaker) and 1% penicillin–streptomycin. All cells were tested for mycoplasma every three months and were negative. Small RNA preparations derived from embryonic stem cells and induced pluripotent stem cells (iPSC) were provided by Drs. Sareen and Svendsen of the Regenerative Medicine Institute at Cedars-Sinai Medical Center. Generation and characterization of these cells are described in Supplementary Materials and Methods.

Cells were trypsinized and total miRNA was extracted using a mirVana miRNA isolation kit (Ambion). miRNA expression analysis by quantitative real-time PCR (qRT-PCR) was performed separately for each miRNA using specific primer sets (Applied Biosystems) as previously described (15). RNU6B was used for normalization.

Total RNA was isolated from confluent monolayers of cells using the RNeasy Mini Kit (Qiagen). cDNA was made using SuperscriptIII reverse transcriptase (Life Technologies). mRNA primers were designed and synthesized at Integrated DNA Technologies. mRNA expression levels were determined by qRT-PCR assays and SYBR Green Dye (Applied Biosystems). Samples were analyzed using the ΔΔC_t method and normalized to 18S ribosomal RNA.

The dataset was published by the Memorial Sloan Kettering Cancer Center (MSKCC) team (16) and was obtained from cBioPortal (17). The expression levels of miR-154 and miR-379 were analyzed along with the survival data in the dataset. For the survival analysis of miR-379, the expression levels of miR-379 in patients with nonmetastatic disease were compared with the median expression level of normal individuals. The disease-free survival of patients with miR-379 expression levels higher than normal individuals (n = 29) was compared with those with lower miR-379 expression levels (n = 78). A Kaplan–Meier survival curve was done by log-rank test between high and low expression groups. The expression levels of miR-154* were not available in the dataset. For the analysis of miR-154, the expression levels of normal healthy individuals (n = 29) were compared with the expression levels of patients with primary (n = 99) and metastatic prostate cancer (n = 14). Two-tailed Student t tests were done between the normal group and two primary and metastatic groups for analysis of the differential expression of miR-154.

Stromal antigen 2 (STAG2) mutant luciferase activity: 3′untranslated region (3′UTR) STAG2 luciferase construct (Switchgear genomics) was used as the wild-type (WT) construct and it was further mutated as described below. miR-154* mimic and control miRNA were transiently transfected along with the WT or mutant (STAG2) construct into these 293T cells, and luciferase activity was determined 24 hours later using the Lightswitch luciferase assay system (Switchgear Genomics).

3′UTR mutant constructs: Mutated 3′UTR luciferase constructs were produced by site-directed mutagenesis. Briefly, primer pairs with two sequential base pair mutations in the miRNA seed sequence of the 3′UTR were generated. Following PCR amplification, parental methylated template DNA was digested for 1 hour with Dpn I. Two microliter of the reaction was then transformed into XL-10 Gold bacteria. Sixteen hours after transformation, colonies were picked for liquid culture. Plasmid DNA was isolated by the Zyppy Plasmid Miniprep Kit according to the manufacturer's directions (Zymo Research). Mutations were confirmed by sequencing before proceeding with luciferase assays.

ARCaP_E prostate cancer cell lines were transduced with lentivirus expressing control or miR-Zip-154* (miR-154*i) or miR-Zip-379 (miR-379i; System Biosciences) or cluster overexpression plasmid (custom made, miR-154*, miR-379, miR-409-3p/-5p) with GFP or control GFP plasmid. ARCaP_M prostate cancer cell lines were transduced with lentivirus expressing cluster inhibitor plasmid (custom made) with GFP. Lentiviral preparation and transduction of cell lines were performed per the manufacturer's instructions (System Biosciences). GFP-positive cells were FACS sorted and cultured in vitro before experiments were performed.

Cell viability assays were performed using trypan blue staining. Cancer cell invasion was assayed in companion 24-well plates (Becton Dickinson Labware) with 8 μm porosity polycarbonate filter membranes as described previously (18).

Whole cell lysates from cell lines were prepared using a modified RIPA lysis buffer (50 mmol/L Tris HCl, 1% NP-40, 0.5% Na-deoxycholate, 0.1% SDS, 150 mmol/L NaCl, 1 mmol/L EDTA, 10% glycerol) supplemented with 1:100 dilution of the protease inhibitor cocktail and the tyrosine phosphatase inhibitor (Sigma). Proteins were then separated on 4% to 20% or 10% acrylamide gels and transferred to nitrocellulose membrane. Membranes were probed with STAG2 (Cell Signaling Technology) antibody. β-Actin (Sigma) was used as the normalization control.

Mouse tibia was formalin-fixed and paraffin-embedded. The miRNA in situ hybridization (ISH) protocol was followed per the manufacturer's instructions (Exiqon). The scramble and miR-154* probes were 5′-biotin labeled. The probes were linked to streptavidin-conjugated quantum dot (QD). Tissue sections were deparaffinized, treated with proteinase-K, and dehydrated. ISH was performed for 1 hour at 55°C, followed by washes and streptavidin blocking and a reaction with streptavidin-conjugated QD at a specified wavelength. The QD staining procedure was followed as previously described (19). Single QD labeling was performed, and scramble or miR-154* probes were labeled with 625-nm QDs. Images were taken at 40×. Hematoxylin and eosin staining was performed on subsequent tissue sections.

A Gleason score tissue array was obtained from Vancouver Prostate Center. The use of tissue specimens was approved by the institutional review board of the Cedars-Sinai Medical Center (IRB# Pro21228). The tissues consisted of benign prostatic hyperplasia (BPH; N = 4), Gleason score 6 (N = 12), and Gleason 7 (N = 7). Each tissue had two sample cores. The tissue array was stained for H&E and graded by a pathologist to confirm the Gleason score. Single QD labeling was performed as previously mentioned (19). miR-154* was labeled with 625-nm QD, and signals were quantified as previously mentioned (19). The QD fluorescence intensity of each tissue section was determined and analyzed. Human prostate caner bone tissues were stained following the same procedure, except multiplexed ISH-QD was performed, where miR-154* (red) was stained first followed by miR-409-3p (green) or miR-409-5p (green) which was labeled with 565-nm QD.

All animal experiments were Institutional Animal Care and Use Committee approved and done in accordance with institutional guidelines. Luciferase-tagged ARCaP_M control and ARCaP_M-154*i cells were injected intracardially as previously described (20) in SCID/beige mice (Charles River Laboratories; N = 5). Mice were imaged using X-ray and bioluminescence via the IVIS Lumina Imaging system. Mice were given near infrared (NIR) dye (IR783) 48 hours before euthanasia. The tumor-specific NIR dye was used to detect metastatic tumors in the mice.

Values were expressed as mean ± SD. All experiments were done in triplicates at least two independent times. Statistical analysis was performed using the Student t test or ANOVA. Values of P < 0.05 were considered to be statistically significant.

We determined the levels of miR-154/154* and miR-379 in two bone metastatic models of prostate cancer, ARCaP, and LNCaP. Mesenchymal-type ARCaP_M cells upon intracardiac inoculation or orthotopic implantation have 100% bone metastatic capability compared with their isogenic epithelial-type counterpart ARCaP_E cells (13). Similarly, C4-2 prostate cancer cells have a high metastatic ability compared with their lineage-related LNCaP cells (21). Both miR-154* and miR-379 were elevated in ARCaP_M cells compared with ARCaP_E cells, and in C4-2 cells compared with LNCaP cells (Fig. 1A). We determined the expression of miR-154* in prostate cancer bone metastasis tissues by ISH-QD and observed higher expression in the metastatic prostate cancer cells (Fig. 1B). Previous studies demonstrated that several members of the DLK1-DIO3 miRNA cluster are elevated in human embryonic stem cells and iPSCs (22, 23). We measured the relative levels of these miRNAs in H9 embryonic stem cells and in iPSCs. Both H9 human embryonic stem cells and patient-derived iPS cells had elevated expression of miR-154* (Fig. 1C and D) but not miR-379 (23). We previously demonstrated that one of the predominant signaling pathways activated in prostate cancer bone metastasis is the β2-microglobulin/hemochromatosis E (HFE) pathway (20, 24). Inhibition of either of these proteins results in reversal of EMT (20). We observed that ARCaP_M HFE knockdown cells (ARCaP_M KD^HFE1cells) had significantly decreased expression of miR-154* and miR-379 similar to the epithelial ARCaP_E prostate cancer cells (Fig. 1E). In addition, other members, such as miR-409-3p/-5p of the DLK1-DIO3 cluster, were also decreased in ARCaP_M HFE knockdown cells. These data together demonstrate that miR-154* and miR-379 are elevated in prostate cancer EMT and bone metastatic models.

To test the hypothesis whether miR-154* and miR-379 are involved in prostate cancer EMT, we transiently depleted miR-154* and miR-379 in ARCaP_M and C4-2 cells using siRNA and determined the cell viability by the trypan blue exclusion test. miR-154* cells and miR-379 inhibited cells underwent cell death compared with control transfected cells in both of the prostate cancer cell line models studied (Fig. 2A). Next, we introduced an shRNA in the mesenchymal-type ARCaP_M prostate cancer cells to generate miR-154* knockdown (ARCaP_M-154*i) or miR-379 knockdown (ARCaP_M-379i) cells. In addition, we used a scrambled shRNA to generate a control shRNA vector (ARCaP_M-C). Reduced expression of miR-154* was detected in the ARCaP_M-154*i cells by qRT-PCR analysis compared with the shRNA control vector ARCaP_M-C. Reduced expression of miR-379 was also observed in the ARCaP_M-379i cells by qPCR analysis compared with the ARCaP_M-C cells (Fig. 2B). RNU6B was used for normalization. The miR-154* target gene, stromal antigen 2 (STAG2), was measured in the knockdown cells (TargetScan v6.2). STAG2 is a tumor suppressor protein (25). The mRNA and protein expression of STAG2 was increased in the ARCaP_M-154*i cells compared with the control ARCaP_M cells (Fig. 2C). We further tested if miR-154* directly binds to the 3′UTR of STAG2 using a luciferase reporter assay. Compared with control miRNA–treated cells, miR-154* mimic-treated cells had reduced basal luciferase activity (Fig. 2D). We mutated the 3′UTR of STAG2 at the miR-154* binding site and demonstrated restoration of luciferase activity in response to the miR-154* mimic (Fig. 2D). Interestingly, inhibition of miR-154* or miR-379 led to reversion of mesenchymal ARCaP_M cells to an epithelial phenotype (Fig. 2E) accompanied by increases in E-cadherin mRNA expression in ARCaP_M-154*i and ARCaP_M-379i cells but not in ARCaP_M-C cells (Fig. 2F). To determine if functional reversal of EMT (MET) occurred in addition to morphologic changes, we performed invasion assays on these cell lines. ARCaP_M-154*i and ARCaP_M-379i cells had significantly decreased invasive capacity compared with ARCaP_M-C cells (Fig. 2G). Taken together, these results demonstrate that miR-154* and miR-379 play an important role in the EMT and invasive capacity of prostate cancer cells.

Cancer cells gain their metastatic potential by undergoing EMT. Previous studies from our laboratory using the ARCaP model demonstrate the close association between EMT and prostate cancer bone metastasis (26). Because miR-154* levels are elevated in metastatic cancer cells, we determined whether inhibition of miR-154* would lead to decreased metastasis in vivo. Consistent with our hypothesis, we found that inhibition of miR-154* resulted in increased cell death, mesenchymal to epithelial transition (MET), and decreased invasion in vitro. To determine if miR-154* plays a role in cancer metastasis in vivo, we inoculated luciferase-tagged ARCaP_M-C control cells or ARCaP_M-154*i (miR-154* inhibited) cells via the intracardiac route into SCID/Beige mice (N = 5/group) to mimic in vivo metastasis. Tumor growth was monitored by luciferase imaging. Mice that received ARCaP_M-154*i cells had a significantly decreased incidence of metastasis (0/5) compared with mice that received ARCaP_M-C control cells (4/5) at 15 weeks postinoculation. X-ray and luciferase imaging of representative mice from both groups are shown in Fig. 3A. The tumors were detected by IR783 (NIR dye) in all mice (Fig. 3A, right panels). The control mice developed metastatic tumors at 1 to 5 sites in the body, whereas ARCaP_M-154*i–injected mice did not develop any tumors. Bone tumors sites included the tibia, femur, humerus, and mandible, and had mixed osteoblastic and osteolytic lesions (27). Mice inoculated with ARCaP_M-C control cells had decreased survival compared with those inoculated with ARCaP_M-154*i cells as shown in the Kaplan–Meier survival curve (Fig. 3B). Taken together, these studies demonstrate that miR-154* is essential for the development of bone metastasis of human prostate cancer cells and that knockdown of miR-154* reduces bone metastasis and increases the survival of mice.

We next determined the expression of miR-154* in human prostate cancer tissues using ISH and QD analysis. miR-154* probes were biotin-labeled (Exiqon) and further labeled to a streptavidin-conjugated QD at a specified wavelength. The tissues were separated into three groups, BPH (N = 4), Gleason 6 (N = 12), and Gleason 7 (N = 7). Each tissue sample has two sections. Gleason 7 tumor tissues had significantly higher staining of miR-154* compared with BPH (Fig. 3C). A representative image of Gleason 7 demonstrates higher staining in the tumor tissues (Fig. 3C). We also determined the expression of miR-154* in other cancers and found increased expression in other urological cancers (Supplementary Fig. S1). Using the publicly available database (MSKCC), we demonstrated that elevated expression of miR-379 is associated with progression-free survival and miR-154 (the opposite strand of miR-154*) is downregulated in both primary and metastatic prostate cancer tissues compared with normal individuals (Fig. 3D and E; refs. 28–30). These results are consistent with previous studies demonstrating that miR-154 is decreased in prostate cancer cell lines and clinical samples (28–30). These databases did not have information on miR-154*. Nevertheless, our observations in bone metastatic prostate cancer cell line models are clinically relevant and our experimental observations correlated well with tumor progression. These results collectively demonstrate that miR-154* and miR-379 are highly expressed in prostate cancer and miR-379 expression correlates with progression-free survival in patients.

Previously published studies by our lab demonstrate that miR-409-3p/5p, two key miRNAs in the DLK1-DIO3 cluster, are elevated in prostate cancer cells (12). Because all four members (miR-409-3p/5p, miR-154*, and miR-379) of the DLK1-DIO3 cluster are elevated in prostate cancer bone metastatic cells, we stably transduced ARCaP_E cells, a marginally metastatic epithelial cell line, using a lentivirus that carries a GFP control vector (ARCaP_E-C) or a lentiviral overexpression vector that carries GFP and a combination of all four miRNAs—miR-409-3p, -5p, miR-154*, and miR-379 (ARCaP_E-cluster overexpression). The miRNA expression of the members of the DLK1-DIO3 cluster was determined by qRT-PCR assay (Fig. 4A). In the ARCaP_E-cluster overexpression cells, miR-409-5p and miR-154* were highly expressed (100-fold) followed by moderate expression of miR-379 and miR-409-3p (5- to 10-fold) when compared with ARCaP_E-C prostate cancer cells (Fig. 4A). Next, we measured the expression levels of the target genes of these miRNAs by qRT-PCR. Several of the target genes of these four miRNAs include tumor suppressors that are shared by these miRNAs (Prediction software: Targetscan v6.2 June 2012 and Pictar). Overexpression of these miRNAs resulted in a decrease in mRNA levels of several of the target genes that are commonly shared by these miRNAs. STAG2 (target gene of miR-154* and miR-409-5p) and Ras suppressor protein 1 (RSU1; target gene of miR-409-3p and miR-409-5p) were significantly decreased in the ARCaP_E-cluster overexpression cells compared with control (Fig. 4B). miR-409-5p targets, including STAG2, retinoblastoma-like 2 (RBL2), nitrogen permease regulator-like 2 (NPRL2), and RSU1, were decreased in the cluster overexpression cells (Fig. 4B). Targets of miR-409-3p, including RSU1, polyhomeotic homolog 3 (PHC3), and tumor suppressor candidate 1 (TUSC1), were decreased in cluster overexpression cells (Fig. 4B). These results demonstrate that overexpression of miRNA members of the DLK1-DIO3 miRNA mega-cluster results in decreased expression of several tumor suppressor genes. We also observed EMT changes in ARCaP_E-cluster overexpression prostate cancer cells. The cells appeared spindle shaped and had decreased expression of E-cadherin and increased expression of vimentin compared with ARCaP_E-C cells (Fig. 4C). To evaluate the combinatorial effects of knockdown of all four miRNAs (miR-409-3p, miR-409-5p, miR-379, and miR-154*) in metastatic ARCaP_M prostate cancer cells, we introduced shRNA that targets all four miRNAs and found that successful shRNA-mediated knockdown of miR-154* and miR-409-3p (the knockdown of miR-409-5p and miR-379 was not effective in this experiment) led to reversal of EMT. ARCaP_M cells underwent morphologic, biochemical, and functional EMT changes upon knockdown of miR-154* and miR-409-3p (Supplementary Fig. S2A–S2D).

Analyses of the target pathways regulated by these miRNAs include oncogenic pathways such as E2F signaling, the Ras pathway, hypoxia inducible factor signaling, and the WNT and TGFβ pathways. These pathways also activate EMT and the cancer stem cell phenotype (Fig. 5). miR-154* targets STAG2 and SMAD7. STAG2 is known to induce aneuploidy (25, 31). SMAD7 is an inhibitor of the TGFβ pathway (32). miR-379 is predicted to inhibit forkhead box F2 (FOXF2), which has been shown to inhibit the WNT pathway in colon cancer development (33). We previously demonstrated that miR-409-3p and miR-409-5p inhibit RSU1, which is a known inhibitor of the oncogenic Ras pathway (12, 34, 35). Other targets of miR-409-3p include von Hippel–Lindau (VHL) tumor suppressor, E3 ubiquitin protein ligase, and PHC3. VHL is a tumor suppressor and degrades HIF1α (36). Inhibition of VHL results in the stabilization of HIF1α, which is known to induce resistance to radiation and chemotherapeutic agents (37). PHC3 is a tumor suppressor protein and is lost in osteosarcoma (38). PHC3 and ephrin receptors have been predicted to exhibit protein–protein interactions (www.biograph.be/). miR-409-5p has been shown to target STAG2, RSU1, NPRL2, and RBL2 (12). NPRL2 activates AKT pathway (39, 40) and RBL2 activates the E2F pathway (41). Thus, these miRNAs activate oncogenic proteins by targeting tumor suppressors. Because these miRNAs are elevated in prostate cancer bone metastatic models, we determined the levels of these miRNAs in human prostate cancer bone metastatic samples using multiplexed ISH-QD labeling. We observed increased miR-154*, miR-409-3p, and miR-409-5p staining in human metastatic tumor tissues in the bone (Fig. 6). Taken together, these results demonstrate that miR-154*, miR-409-3p/-5p, and miR-379 induce EMT in prostate cancer cells, show upregulated expression in human prostate cancer bone metastasis tissues, and correlate with progression-free survival in patients with prostate cancer.

To understand the biology of the miRNAs of the DLK1-DIO3 mega-cluster in EMT and cancer bone metastasis and to identify novel biomarkers and/or therapeutic targets, we probed for specific miRNAs of this mega-cluster in unique EMT models of human prostate cancer developed in our laboratory. Specifically, the miRNAs miR-154* and miR-379, located within the DLK1-DIO3 cluster, were highly upregulated in two prostate cancer cell lines with a mesenchymal phenotype and with bone metastatic potential (Fig. 1). The miRNA members of the DLK1-DIO3 cluster have been shown to be important for totipotency during embryogenesis and induced pluripotent stem cell formation. We uncovered a surprising role for miR-154* and miR-379, which is expressed by embryonic stem cells and pluripotent stem cells, promoting prostate cancer development and metastasis. Specifically, we show that (i) miR-154* and miR-379 are elevated in human prostate cancer tumor tissues and miR-379 expression correlates with progression-free survival of patients with prostate cancer, (ii) miR-154* and miR-379 can promote EMT of prostate cancer cells in vitro, and (iii) inhibition of miR-154* results in decreased bone metastatic tumor growth and increased survival. Thus, miR-154* and miR-379 appear to be promising new biomarkers for cancer detection and attractive new therapeutic targets for prostate cancer treatment.

The primary functions of miR-154* are to repress tumor suppressors and modulate the expression of EMT and stemness to mediate downstream convergent signal axes. Inhibition of miR-154* resulted in increased E-cadherin and decreased invasion in vitro. miR-154*-depleted prostate cancer cells had decreased bone metastasis and increased survival. miR-154* mediates its effects by downregulating its target gene, STAG2, which is a tumor suppressor. STAG2 plays a critical role in the cohesion complex, and a decrease in STAG2 has been correlated with aneuploidy and cancer (25, 31). Other miR-154* target genes may be involved such as SMAD7, which inhibits the TGFβ pathways (32) involved in EMT (42). Interestingly, the sense strand of miR-154* is undetectable or decreased in prostate cancer models and in patients with prostate cancer (28, 29, 30). miR-154* is upregulated in both patients with lung squamous cell carcinoma and lung adenocarcinoma (9). miR-379, another member of the DLK1-DIO3 cluster, located upstream of the miR-154 gene cluster, is moderately elevated in prostate cancer cells (Fig. 1). Inhibition of miR-379 also resulted in reversal of EMT (MET), increased E-cadherin, and decreased invasion. miR-379 expression in clinical specimens correlates with disease-free survival of patients with prostate cancer. Consistent with our observations, previous reports demonstrate that the miRNAs of the DLK1-DIO3 cluster are elevated in the serum of patients with cancer compared with healthy patients (9, 11). Accordingly, it was shown that miR-379 is also elevated in patients with metastatic prostate cancer compared with patients with localized disease (43). Interestingly, other members of the DLK1-DIO3 mega-cluster share similar mRNA target genes. STAG2 is targeted by both miR-154* and miR-409-5p, whereas RSU1 is targeted by both miR-409-5p and miR-409-3p. Thus, the members of the cluster work synergistically and are elevated in the mesenchymal-type prostate cancer cells to promote EMT and metastasis (Fig. 5). We conclude that the miRNA members of the DLK1-DIO3 cluster promote EMT, stemness, and bone metastasis in prostate cancer. These findings have potential clinical implications in both the biomarker field and the therapeutic arena.

No potential conflicts of interest were disclosed.

Conception and design: M. Gururajan, S. Josson, E.M. Posadas, L.W.K. Chung

Development of methodology: M. Gururajan, S. Josson, H.E. Zhau, C. Liu, L.W.K. Chung

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Gururajan, S. Josson, G.C.-Y. Chu, C.L. Haga, J. Lichterman, P. Duan, E.M. Posadas, L.W.K. Chung

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Gururajan, S. Josson, C.-L. Lu, Y.-T. Lu, L.W.K. Chung

Writing, review, and/or revision of the manuscript: M. Gururajan, S. Josson, Y.-T. Lu, H.E. Zhau, J. Lichterman, E.M. Posadas, L.W.K. Chung

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Gururajan, S. Josson, C.-L. Lu, Y.-T. Lu, E.M. Posadas, L.W.K. Chung

Study supervision: M. Gururajan, S. Josson, E.M. Posadas, L.W.K. Chung

The authors thank Drs. Svendsen and Sareen for sharing the iPSC and embryonic stem cell cultures and Gary Mawyer for editing the article.

This study was supported by grants P01-CA98912, DAMD-17-03-02-0033, and RO1-CA122602 (to L.W.K. Chung).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.