Cancer Hallmarks in Induced Pluripotent Cells: New Insights

MATERIALS AND METHODS

Cells and culture

Two hiPSC cultures IMR90-1, Foreskin -1 (WiCell; Madison, WI) and four hESC cultures H7, H14 (WiCell) and CM7, CM14, established at CMRC (Laurant et al., 2010), (currently pending approval for addition to the NIH Stem Cell Registry) were used for this study. The cells were grown in StemPro medium (Invitrogen; Carlsbard, CA) on a Matrigel substrate (BD Bioscience; San Jose, CA). The cultures were split mechanically using the StemPro EZ Passage tool (Invitrogen). For miR analysis, confluent cultures were lifted using trypsin and then washed in ice-cold PBS and pellet stored at −80 °C.

Western Blot, Real-time RT-PCR DNA methylation analyses

Thirty micrograms of total cell lysate from hiPSCs or hESCs were loaded per lane in pairs onto a 4–12% Tris-Bis SDS-PAGE (Invitrogen). After transblotting onto an Immobilon membrane (Millipore; Billerica, MA), the membrane was cut into thirds and each section probed for either Nodal (antibody Clone EP2058Y; Epitomics; Burlingame, CA), Lefty (antibody AF746; R&D Systems; Minneapolis, MN) or Cripto-1 (antibody 600-401-997; Rockland; Gilbertsville, PA). The membranes were then stripped and reprobed for actin (antibody MAB1501; Millipore) as a protein loading control. For Real-time RT-PCR, RNA was isolated using TRizol reagent (Invitrogen) and 1 μg reverse transcribed as previously described (Postovit et al., 2008). Real-time RT-PCR was performed as described (Postovit et al., 2008) using TaqMan (Applied Biosystems; Carlsbad, California) gene expression human primer/probe sets for Nodal (Hs00250630.s1), Lefty (Hs009996632.g1) and Cripto-1 (Hs02339499.g1) and gene levels normalized using HPRT-1 (433768F). Data were analyzed using Applied Biosystems’ Sequence Detection Software (V. 1.2.3) and error bars represent mean gene expression normalized to hESC values, +/−S.D. DNA from hiPSCs and hESCs was extracted by phenol-chloroform, bisulfite converted and sequenced for the Lefty B gene CpG island as previously reported (Costa et al., 2009). Six to ten positive clones were sequenced and percentages of DNA methylation were calculated.

miR analysis

Total mRNA isolation from the cell lines was performed with the PureZOL RNA isolation reagent (Bio-Rad; Hercules, CA), according to the manufacturer’s instructions. TaqMan Low-Density Arrays (TLDA Human MicroRNA Panel v1.0) were used to detect and quantify mature miRs in accordance with the manufacturer’s instructions (Applied Biosystems’ 7900HT Micro Fluidic Cards). The cards were processed in the ABI 7900 HT Fast Real Time PCR System (Applied Biosystems) and analyzed with Real-Time StatMiner (Integromics; Philadelphia, PA). The difference in miR expression between hiPSCs and hESCs was calculated by the comparative 2 ^{method with RNU44 and RNU48 as endogenous controls (}Livak and Schmittgen, 2001) (P<0.05 was considered as significant). Hierarchical clustering was performed by the Ward’s method using Pearson’s correlation for miR similarity measure. miRs with ΔCt<5 (RNU48 as endogenous controls) were considered to be at high level of expression.

To verify the accuracy of our TLDA data, we performed individual qRT-PCR experiments for representative miRs using TaqMan miR assays (Applied Biosystems) in triplicates, according to the manufacturer’s instructions (RNU48 as endogenous controls). miR expression levels were analyzed as above and the miRs were confirmed to be significantly up-regulated in the hiPSC compare to the hESC lines by the individual qRT-PCR experiments.

RESULTS AND DISCUSSION

This study initially performed a comparative analysis of the major components of the embryonic Nodal signaling pathway in hESCs and hiPSCs. Western blot and Real-time RT-PCR results reveal slightly lower levels of Nodal (a TGF-β family member) and Cripto-1 (Nodal’s co-receptor) and a dramatic decrease in Lefty (Nodal’s inhibitor and TGF-β family member) in hiPSCs compared with hESCs (Fig. 1A). Based on the unanticipated noteworthy drop in hiPSC’s Lefty expression, we performed DNA sequence-based methylation analysis of Lefty B CpG island and found increased methylation (Fig. 1A), suggesting silencing of this critical regulator of Nodal. The implications associated with a significantly lower level of Lefty expression in hiPSCs vs. hESCs, together with our earlier findings of the re-emergence of aberrant Nodal signaling in metastatic tumor cells in the absence of Lefty (Postovit et al., 2008), prompted us to address a more fundamental question focused on the implications associated with the epigenetic reprogramming of hiPSCs, particularly related to the fidelity of these cells to maintain a stable ESC phenotype.

Figure 1

Differences in pluripotent markers and oncogenic-associated miRs in hiPSCs vs. hESCs

(A) Upper, Western blot and Real-time RT-PCR analyses of hiPSCs (IRM90-1) and hESCs (H9) for the expression of Nodal, Lefty and Cripto protein (relative values corrected against Actin for protein loading); Right, mRNA expression (normalized to hESC values). Lower, DNA methylation of the Lefty B CpG island. (B-1) Comparison of miR expression profiles between two hiPSC and four hESC lines. Unsupervised hierarchical clustering of 157 microRNAs (ΔCt, Pearson’s correlation, P<0.05): A-hiPSC (Foreskin-1), B-hiPSC (IMR90-1), C-hESC (CM7), D-hESC (H7), E-hESC (CM14), F-hESC (H14). (B-2) Supervised hierarchical clustering using 10-fold change between hiPSC and hESC lines as a cutoff (31 miRs-the same order of samples as in (B-1) (ΔCt, Pearson’s correlation, P<0.05): (B-3) Cancer related miRs highly expressed (ΔCt <5) in both hiPSC lines. (*based on literature search; **miR was also found to be differentially expressed between hiPSC and hESC lines [Chin et al., 2009]; ***verified by individual qRT-PCR experiments [P<0.05]).

We pursued a comparison of the expression profiles of 365 microRNAs (miRs) in two hiPSC (fibroblasts reprogrammed with Oct4, Sox2, Nanog and Lin28) and four hESC lines, recognizing that specific miRs are known to be associated with oncogenic pathways (Tong et al., 2009). Although the ability of hESCs and hiPSCs to form teratomas in immunocompromised mice is well documented (Yu et al, 2007; Thomson et al. 1998), particularly noteworthy are the observations in chimeric mice derived from iPSCs generated with exogenous c-myc, where malignant tumors developed in up to 20% of the mice (Okita et al., 2007) vs. mice derived from iPSCs reprogrammed without exogenous c-myc, where no tumors have been reported (Wernig et al, 2008). These disparate findings prompted further inquiry into the potential pathways employed by normal cells resulting in pluripotency vs. oncogenic transformation.

An unsupervised hierarchical clustering analysis of 157 miRs that were expressed in at least one of the six cell lines tested (Fig.1B-1) revealed 72 miRs expressed at statistically different levels in hiPSCs vs. hESCs (P<0.05), 31 exhibiting greater than 10-fold difference (Fig. 1B-2; Table 1). Further statistical analysis of the 31 miRs indicated that 15 were expressed at high levels (ΔCt<5), 10 of which have been shown to be cancer related (Fig. 1B-3). Specifically, differential expression of these 10 miRs have been shown to regulate critical checkpoints in Hodgkin’s lymphoma, multiple myeloma, and breast, pancreatic and prostatic carcinoma (Tong et al. 2009; Pichiorri et al., 2008; Griether et al., 2010; Mertens-Talcott et al., 2007; Gibcus et al., 2009; Yan et al., 2008). The miR differences found in this study between hiPSCs and hESCs further support the recent findings of Doi and colleagues (Doi et al., 2009), who indicated that the target loci involved in epigenetic reprogramming to pluripotency parallels aberrant oncogenic transformation programming, and advances the observations of Feng and coworkers reporting early senescence of hiPSCs derivatives (Feng et al., 2010). Our investigation also revealed that both hiPSCs fibroblasts -- isolated from either fetal origin (IMR90) or newborn foreskin hiPSCs resulted in a similar miR expression profile between them as did hESCs miR expression among cell lines of different ethnic origin. Collectively, these data demonstrate cancer hallmarks expressed by hiPSCs, which will require further elucidation for their impact on clinical applications, especially with respect to the fate of precancerous stem cells.

Acknowledgments