Genetic Analysis of the Role of the Reprogramming Gene LIN-28 in Human Embryonic Stem Cells§



This article is corrected by:

  1. Errata: Corrigendum: Genetic Analysis of the Role of the Reprogramming Gene LIN-28 in Human Embryonic Stem Cells Volume 33, Issue 7, 2360, Article first published online: 19 June 2015

  • Author contributions: H.D.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing; N.B.: conception and design, data analysis and interpretation, manuscript writing.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLSExpress November 26, 2008.


LIN-28 is a gene recently shown to be involved in the conversion of somatic cells to induced pluripotent stem cells. We have previously shown that LIN-28 is highly expressed in human embryonic stem cells (HESCs); however, its role in these cells has not been investigated. We now show that, like OCT4, SOX2, and NANOG, LIN-28 is downregulated during differentiation of HESCs into embryoid bodies. In addition, we investigate the role of LIN-28 in HESCs by manipulation of its expression levels. LIN-28 overexpression impairs the ability of cells to grow at clonal densities, due to increased differentiation and decreased cell division. Analysis of cell differentiation under these conditions revealed that it is mostly towards the extraembryonic endoderm lineage. Moreover, we show that, during early mouse development, high levels of Lin-28 are also observed in the extraembryonic endoderm, and therefore it seems that, both in vitro and in vivo, high levels of LIN-28 may specify an extraembryonic endoderm fate. However, LIN-28 seems dispensable for self-renewal of HESCs; its downregulation neither impairs HESC proliferation nor leads to their differentiation. Thus, LIN-28 does not seem to be involved in the self-renewal of HESCs, but rather seems to be involved in their decision to switch from self-renewal to differentiation. STEM CELLS2009;27:352–362


Human embryonic stem cells (HESCs) are pluripotent cells derived from blastocyst-stage embryos [1, 2]. These cells have two unique properties. The first is their unlimited self-renewal that persists when kept in an environment preventing their differentiation. The second is their pluripotency, which enables them to differentiate and form cells of the three embryonic germ layers. Due to their theoretical ability to form all cell types of the body, these cells have been suggested as a potential cell source for regenerative medicine in multiple diseases. Much attention has been focused on understanding the regulation of their unique characteristics, arguing that understanding these processes might enable better regulation of their subsequent differentiation towards use in therapy. Studies such as these have analyzed the roles of genes like OCT4 [3], NANOG [4–6], and SOX2 [7] in the control of pluripotency and self-renewal. Recently, an exciting breakthrough enabled the derivation of induced pluripotent stem (iPS) cells from somatic cells, both in mice [8, 9] and more recently in humans [10–13]. These cells can be derived by expressing as few as four genes (and later it was shown that even three genes are sufficient) in somatic cells, followed by reprogramming to ESC-like cells. These iPS cells have been shown to be similar to ESCs, both in their markers and in their gene expression. Furthermore, they were shown to be pluripotent, since they differentiate into cells of the three embryonic germ layers, either by formation of chimeras in mice or teratomas (tumors composed from cells of the three embryonic germ layers) in both mice and humans. This breakthrough was possible as a result of work deciphering the unique proteins involved in governing the pluripotent state of ESCs, since the genes required for the derivation of iPS cells included genes formerly implicated in pluripotency, such as Oct4, Sox2 [8, 10, 13], and more recently also Nanog [11]. Another gene shown to facilitate the derivation of iPS cells, at least in humans, is LIN-28 [11]. LIN-28 has been shown to increase the efficiency of iPS formation when combined with OCT4, SOX2, and NANOG.

LIN-28 was first described as a heterochronic gene involved in the regulation of developmental timing control in Caenorhabditis elegans [14]. Heterochronic genes control the timing of development, and indeed, in C. elegans, mutation of lin-28 leads to precocious development, whereas the gain-of-function allele leads to retarded development. However, the molecular mechanism by which LIN-28 controls developmental timing is still unclear. LIN-28 has been suggested to be a post-transcriptional regulator due to its RNA binding domains: a cold shock domain (CSD) and a pair of retroviral type CCHC zinc fingers [14]. The CSD is known to be involved in translational control, and CSD proteins are known to be found bound to stored, nontranslating mRNA, predominantly during early development [15].

In this study, we examined the role of LIN-28 in the maintenance of self-renewal and pluripotency of HESCs. Although LIN-28 has been suggested to be enriched in undifferentiated ESCs, its role in these cells has not yet been investigated. Understanding its significance in the control of pluripotency seems imperative, especially since it has been shown to enhance the efficiency of iPS cell production. In this study, we aimed at dissecting out the role of LIN-28 in HESCs, by using loss and gain-of-function methodologies.


Cell Culture

HESCs from the H9 [1] and HES9 [16] cell lines were cultured on mitomycin-C treated mouse embryonic fibroblasts (MEFs) obtained from 13.5-day embryos at 85% KnockOut DMEM medium (Gibco-BRL, Gaithersburg, MD,, supplemented with 15% KnockOut SR (Gibco-BRL), 1 mM glutamine, 0.1 mM β-mercaptoethanol (Sigma-Aldrich, St. Louis,, 1× nonessential amino acid stock (Gibco-BRL), penicillin (50 units/ml), streptomycin (50 μg/ml), insulin-transferrin-selenium at a 1:200 dilution (Gibco-Invitrogen Corporation, Grand Island, NY,, and 4 ng/ml basic fibroblast growth factor (bFGF). To obtain a feeder-free culture, cells were plated on gelatin-coated (0.1%, Merck & Co., Whitehouse Station, NY, plates and grown in medium conditioned by MEFs. Differentiation in vitro into embryoid bodies (EBs) was performed by withdrawal of bFGF from the growth medium and allowing aggregation in Petri dishes, as previously described [17].

DNA Microarray Analysis

Total RNA was extracted according to the manufacturer's protocol (Affymetrix, Santa Clara, CA, When extracting RNA from undifferentiated ESCs, the cells were grown for one passage on gelatin-coated plates with conditioned medium in order to avoid contamination by feeder cells. Hybridization to the U133 DNA microarray, washing, and scanning were performed according to the manufacturer's protocol, and expression patterns were compared between samples. Signals were normalized by dividing each probe by the average value of the DNA microarray to avoid differences among different arrays and experiments.

Plasmid Construction, Transfections, and Clone Establishment

A LIN-28 expression vector was constructed using a construct described previously [4]. This construct harbors a constitutive CAG promoter (a combination of chicken β-actin promoter and cytomegalovirus immediate-early enhancer) followed by a human NANOG transgene, an internal ribosomal entry segment, and a puromycin resistance gene. The NANOG open reading frame (ORF) was removed by digestion with BstXI (Fermentas, Burlington, ON, Canada, The vector was religated and used as an empty control vector to establish mock-transfected control clones. To create the LIN-28-overexpressing construct, the LIN-28 ORF was polymerase chain reaction (PCR) amplified from HESC cDNA using LIN-28-specific primers (supporting information Table 1A). The amplified ORF was inserted into a TOPO-TA cloning vector (Invitrogen, Carlsbad, CA, To insert LIN-28 ORF under the regulation of the CAG promoter, this vector was digested with EcoRI, end-filled, and ligated to the empty CAG expression vector, which was digested with XhoI and also end-filled. This vector was subsequently used for overexpression of LIN-28 (Fig. 2A). Wild-type ESCs were transfected with control plasmid or LIN-28 overexpression plasmid using the calcium phosphate method as described previously [18]. Stably transfected clones were established by puromycin selection (0.3 μg/ml, concentration determined following calibration; Sigma) (supporting information Fig. 1) following transfection.

Colony-Forming Assays

ESCs were trypsinized to a single-cell suspension and seeded at a density of 500 cells/cm2 in the presence or absence of conditioned medium (CM). After 8 days, the cells were fixed and assayed for alkaline phosphatase (AP) activity (86R kit, Sigma) according to the manufacturer's instructions. Each experiment was performed in triplicate, and at the end of the experiments the positively stained colonies were counted.

Immunostaining and Fluorescence-Activated Cell Sorting Analysis

For immunostaining, cells were washed once with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde. EBs were fixed with 4% paraformaldehyde and then incubated in 10% sucrose followed by incubation in 30% sucrose. Fixed EBs were embedded in OCT (Sakura Finetek, Torrance, CA,, frozen in liquid nitrogen, and cut into 6-μm sections. Embryonic day 7.5 (E7.5) mouse embryos were collected together with their deciduas, washed with PBS, and fixed in 4% paraformaldehyde for 1 hour 45 minutes. Following PBS washes, cells were precipitated in 30% sucrose. Fixed embryos were embedded in OCT, frozen in liquid nitrogen, and cut into 10-μm sections.

Blocking and permeabilization were performed with 2% bovine serum albumin (BSA), 10% low-fat milk, and 0.1% Triton-X in PBS. Staining with the first antibody was performed overnight. Secondary antibody staining was performed for 1 hour at room temperature, at a 1:200 dilution. For a list of antibodies see supporting information Table 1B. Nuclear staining was performed with Hoechst 33258 (Sigma). For TRA-1–60 expression, cells were trypsinized, washed with 3% BSA in PBS with 0.05% sodium azide, incubated with TRA-1–60 antibody for 1 hour on ice, incubated with secondary antibody for 1 hour on ice, and, after washing, fluorescence-activated cell sorting (FACS) analyzed. The cell cycle analysis was performed using propidium iodide (PI) staining. Cells were trypsinized and suspended in a buffer containing 50 μg/ml PI, 3 mM sodium citrate, 0.1% Triton-X, and 10 μg/ml RNase A. Staining and RNA digestion were performed for half an hour at 37°C, and the cells were taken for FACS analysis. Annexin-V staining for apoptosis detection was performed using an annexin-V kit (Bender MedSystems, Burlingame, CA, according to the manufacturer's instructions. The FACS analysis was performed using the FACSCalibur system (Becton, Dickinson and Company, Franklin Lakes, NJ, The analysis was performed on CellQuest software (Becton Dickinson). Forward and side-scatter plots were used to exclude dead cells and debris from the histogram analysis.

Western Blot Analysis

Western blot analysis was performed according to standard protocols. For LIN-28 detection, a polyclonal rabbit antibody against human LIN-28 protein (a kind gift from Dr. Joel G. Belasco [19]), at a 1:1,500 dilution, was incubated overnight at 4°C. A secondary antibody (peroxidase-conjugated goat anti-rabbit IgG; Jackson Immunoresearch Laboratories, West Grove, PA, was incubated for 1 hour at a 1:20,000 dilution. For the loading control we used a mouse antibody against α-tubulin (Sigma). Quantification was performed using ImageJ software (National Institutes of Health, Bethesda, MD,

RNA Extraction and Reverse Transcription-PCR Analysis

RNA was extracted using TRI-reagent for total RNA isolation according to the manufacturer's instructions (Sigma). cDNA was synthesized using random hexamer primers. Amplification was performed using RedTaq ReadyMix PCR reaction mix (Sigma). PCR conditions included a first step of 3 minutes at 94°C, a second step of 25–30 cycles of 30 seconds at 94°C, a 45-second annealing step at 58°C–62°C, 1 minute at 72°C, and a final step of 7 minutes at 72°C. Glyceraldehyde-3-phosphate dehydrogenase was used as a housekeeping gene to evaluate and compare the quality of different cDNA samples. Primers are listed in supporting information Table 1A. The final products were examined by gel electrophoresis on 2% agarose ethidium bromide-stained gels. A real-time reverse transcription (RT)-PCR analysis was performed using TaqMan universal master mix (Roche Diagnostics, Basel, Switzerland, according to the manufacturer's instructions using the 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, For a description of the primers, see supporting information Table 1. TaqMan Human Stem Cell Pluripotency Array (Applied Biosystems) was used according to the manufacturer's instructions. RNA used for the array was extracted using the RNeasy Micro Kit (Qiagen, Hilden, Germany, according to the manufacturer's instructions.

Small Interfering RNA Knockdown Experiments

HESCs were plated at a density of 2 × 104 cells/cm2 on gelatin-coated plates in the presence of CM. Transfection of small interfering (si)RNA was carried out the following day using Oligofectamine reagent (Invitrogen) according to the manufacturer's instructions. Specifically, the Oligofectamine reagent was used at a final concentration of 1:450. siRNA oligos were used at a final concentration of 45 nM. siRNA for LIN-28 and OCT4 was obtained from Dharmacon Inc. (Lafayette, CO,; ON-TARGETplus SMARTpool). siRNA for green fluorescent protein (GFP) (Integrated DNA Technologies, Coralville, IA, was synthesized according to previously described sequences [20]. Transfection was performed over 24 hours in CM containing 5% SR and no antibiotics. The following day, the medium was changed to regular ESC CM. Cells were analyzed 72 hours post-transfection.


LIN-28 Is Expressed in Undifferentiated HESCs and Downregulated upon Their Differentiation

We previously identified LIN-28 as an ESC-specific gene [21]. Using microarray analysis, this work showed that LIN-28 is expressed at high levels in undifferentiated ESCs. We have now extended this initial observation. Since LIN-28 was recently shown to be involved in the reprogramming of human somatic cells to iPS cells [11], we compared its expression pattern with that of the other genes involved in the derivation of iPS cells: OCT4 [8, 10, 11], SOX2 [8, 10, 11], NANOG [11], KLF4 [8, 10], and MYC [8, 10]. We used transcriptome analysis [21] of undifferentiated ESCs and three differentiation stages—early EBs (2-day differentiation), mid-EBs (10-day differentiation), and late EBs (30-day differentiation)—and examined the expression pattern of these genes (Fig. 1A). Based on their expression pattern, these genes can be divided into two groups. The first includes OCT4, SOX2, NANOG, and LIN-28, genes expressed at high levels in HESCs, which are dramatically downregulated in mature EBs to low or absent levels. The second group consists of the two remaining genes, KLF4 and MYC, which are expressed at comparable levels in both undifferentiated and differentiated ESCs. The expression pattern observed using microarray data was further verified using real-time RT-PCR analysis (Fig. 1B). We compared the expression level of OCT4, NANOG, and LIN-28 in ESCs and late EBs. All three were abundantly expressed in undifferentiated ESCs. However, in late EBs, all three genes showed a marked decrease in expression during differentiation, with LIN-28 downregulated by about 5- to 10-fold.

Figure 1.

LIN-28 expression pattern during HESC differentiation. (A): The expression pattern of genes shown to be involved in the conversion of somatic cells to iPS cells was examined using microarray analysis, utilizing the U133A and U133B Affymetrix chips. The analysis was performed on RNA extracted from H9 undifferentiated HESCs and three stages of differentiation into EBs: early EBs (2 days), mid-EBs (10 days), and late EBs (30 days). The genes involved in iPS cell formation can be divided into two groups based on their expression pattern during ESC differentiation. First, genes expressed at high levels in undifferentiated ESCs and downregulated in the late stages of EB differentiation. This group includes OCT4, SOX2, and NANOG, as well as LIN-28 (see upper panel). The other group includes the remaining two genes, MYC and KLF4, the expression of which does not seem to significantly change during differentiation (see lower panel). (B): Using real-time reverse transcription-polymerase chain reaction (RT-PCR), we verified the results obtained from the microarray data using two ESC lines: H9 [1] and HES9 [16]. RNA from ESCs and late EBs was subjected to an RT reaction followed by real-time PCR. Expression of three genes was examined: OCT4, NANOG, and LIN-28. Real-time PCR confirmed the results shown by the microarray analysis, and showed that OCT4 and NANOG were downregulated dramatically during differentiation and LIN-28 also was downregulated by about 10-fold. The results shown are normalized to ubiquitin levels, and the expression level in undifferentiated ESCs was set to the value of one. Error bars represent the 95% confidence interval. (C): The expression pattern of LIN-28 during differentiation was verified using immunofluorescence on H9 and HES9 cells. Shown is staining of ESCs, early EBs, mid-EBs, and late EBs. The expression of LIN-28 (red staining) was examined in comparison to OCT4 (green staining), a known marker of undifferentiated cells, and to its cellular localization (using DAPI, blue staining). LIN-28 shows mostly cytoplasmic staining, whereas OCT4 is a nuclear protein. In ESCs LIN-28 (red staining) is predominantly expressed in cells with a typical ESC morphology, which are also positive for OCT4 staining (green staining). In early and mid-EBs LIN-28 is mostly localized to cells still expressing OCT4, and is absent from most OCT4 cells. In late EBs, most cells have lost expression of LIN-28 and OCT4. However, some cells still express both LIN-28 and OCT4. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; EB, embryoid body; ESC, embryonic stem cell; HESC, human embryonic stem cell; iPS, induced pluripotent stem.

LIN-28 expression was further investigated using immunostaining (Fig. 1C). In ESCs, LIN-28 (red staining) was predominantly expressed in cells with a typical ESC morphology, which are positive for OCT4 staining (green staining). Cryosections of three stages of EB differentiation were also examined. In early EBs and mid-EBs there are still high levels of expression of LIN-28 protein; however, it is mostly confined to cells expressing OCT4. Most cells that were negative for OCT4 expression (and therefore are differentiated cells) were also negative for LIN-28 expression. In late EBs, only a few cells still expressed OCT4. LIN-28 was also expressed at this stage in only a minor subset of cells. Therefore, like OCT4, LIN-28 is a marker of undifferentiated ESCs, dramatically downregulated upon their differentiation.

Establishment of LIN-28-Overexpressing Clones and Analysis of Their Growth

It has been shown that overexpression of NANOG enables propagation of ESCs in the absence of leukemia inhibitory factor for mouse ESCs [4, 5] or in the absence of CM for HESCs [6]. Therefore, we set out to examine whether overexpression of LIN-28 can create a similar phenotype. We obtained clones stably overexpressing LIN-28, as described in Materials and Methods (Fig. 2A), and verified expression of the LIN-28 transgene (Fig. 2B). In these clones, we also verified the upregulation of LIN-28 expression by real-time RT-PCR and Western blot analysis (Fig. 2C). The clones had characteristic ESC morphology, and when grown in the presence of CM were positive for TRA-1–60 expression, a marker of undifferentiated ESCs (Fig. 2D). These clones proliferated at comparable rates to parental cells and exhibited a similar percentage of apoptosis in culture, as assessed by annexin-V staining (data not shown). However, in the absence of CM, LIN-28-overexpressing cells showed somewhat lower self-renewal ability. In comparison with mock-transfected clones, LIN-28-overexpressing clones demonstrated fewer cells that underwent cell division, and the proportion of cells in the G1/G0 stage was about 10% higher than in mock-transfected clones (supporting information Fig. 1B). On examining their growth at clonal densities, these differences became even more pronounced. HESCs overexpressing LIN-28 and control cells were seeded at a clonal density (500 cells/cm2) in the presence and absence of CM, and after 8 days were assayed for the presence of undifferentiated colonies, as assessed by AP activity (a marker of undifferentiated cells). In the absence of CM, both parental cells and LIN-28-overexpressing clones hardly formed any undifferentiated colonies (0–1 colonies per 12 wells, data not shown). In the presence of CM, we observed that LIN-28-overexpressing clones created fewer undifferentiated colonies (Fig. 3A, AP activity shown as red staining). LIN-28-overexpressing clones formed around one third of the undifferentiated colonies that parental cells formed (Fig. 3B); similar results were obtained with mock-transfected clones (data not shown), and many of the colonies were smaller and only weakly stained for AP activity.

Figure 2.

Establishment of LIN-28-overexpressing clones. (A): Construct used for the establishment of the LIN-28-overexpressing cells. Human LIN-28 ORF was cloned and inserted into an expression construct under the control of the constitutive CAG promoter. To enable selection of stable clones, the LIN-28 ORF is followed by an IRES and a puromycin resistance gene, subsequently used for selection. (B): Puromycin-resistant clones were examined for expression of the LIN-28 transgene using LIN-28 transgene-specific primers. Shown are the PCR results of the LIN-28 transgene in clones resistant to puromycin and of the parental H9 cell line. PCR analysis for the GAPDH gene was used as a positive control. (C): Examination of LIN-28 expression at the level of RNA (I) and protein in LIN-28-overexpressing cells (II). α-Tubulin was used as loading control. (D): LIN-28-overexpressing clones are similar to parental cells both in morphology (left panel) and in expression of the undifferentiated marker TRA-1–60 (right panel). Shown are representative pictures. Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IRES, internal ribosomal entry segment; ORF, open reading frame; PCR, polymerase chain reaction; W.T, wild-type.

Figure 3.

LIN-28-overexpressing cells are inefficient in their ability to form colonies at low cell densities. About 500 cells/cm2 were seeded in the presence of conditioned medium, in 12-well plates coated with gelatin and fetal calf serum. After 8 days, the cells were fixed and stained for alkaline phosphates (AP) activity, a known marker of undifferentiated embryonic stem cells, to identify undifferentiated colonies. Each experiment was performed in triplicate. (A): Representative wells are shown. (B): Quantification of the results described in (A). Error bars represent the standard error of the mean.

Decrease in Clonality of LIN-28-Overexpressing Clones Results from Increased Differentiation and a Slower Cell Cycle

As noted above, the overexpression of LIN-28 hinders the ability of the cells to form colonies when seeded at low densities. Therefore, we decided to examine the processes that could lead to this phenotype. The first process is a shift in the balance between self-renewal and differentiation, by prevention of self-renewal or stimulation of differentiation. The second process is a change in the self-renewal or proliferation of the cells, since a slower proliferation rate can lead to formation of smaller colonies. The third process is increased apoptosis, which again can lead to smaller and fewer colonies. We examined this issue by performing a FACS analysis on cells grown at clonal density (Fig. 4). TRA-1–60 staining (a marker of undifferentiated cells) showed that cells overexpressing LIN-28 differentiated more readily at clonal densities, with around 30% differentiated cells in the culture, compared with only around 8% in mock-transfected clones (Fig. 4A). Cell cycle analysis also revealed a substantial change in the cells' profile, with a significantly higher proportion of cells being in the G1/G0 stage of the cell cycle upon LIN-28 overexpression (67% versus 27%) (Fig. 4B). The slower rate of the cell cycle could be a secondary effect of the higher differentiation rate observed in the culture, since it is known that once ESCs differentiate their proliferation is slower [22]. However, we cannot rule out that LIN-28-overexpressing cells, once seeded at a low density, have a different regulation of the cell cycle, independent of their differentiation state. This is also corroborated by the fact that when LIN-28-overexpressing cells are seeded at high densities, but without CM, although there are no substantial differences in the number of differentiated cells between them and control clones, LIN-28-overexpressing clones still show a slower cell cycle. Examination of apoptosis in the culture revealed a slight decrease in the percentage of apoptotic cells in LIN-28-overexpressing cells, as determined by annexin-V staining (Fig. 4C), but this could be the result of fewer cells undergoing cell division, and therefore fewer cells reaching a cell cycle checkpoint that might guide them to apoptosis. Altogether, these results suggest that when LIN-28-overexpressing cells are seeded at clonal densities, there is an increase in their differentiation and a decrease in cells undergoing cell division, compared with mock-transfected clones.

Figure 4.

LIN-28-overexpressing cells differentiate more readily and have a slower cell cycle at low cell densities. Cells overexpressing LIN-28 or mock-transfected clones were seeded at a density of about 500 cells/cm2, in plates coated with gelatin in the presence of fetal calf serum. After 8 days, a fluorescence-activated cell sorting analysis was performed on the cells to examine the following parameters: (A) their differentiation status by staining for TRA-1–60 expression, a marker of undifferentiated cells; (B) their cell cycle profile, using propidium iodide staining; and (C) apoptosis by annexin-V staining. A significant increase was observed in the percentage of differentiated cells in LIN-28-overexpressing clones compared with mock-transfected clones (A) together with a marked increase in cells in the G1 stage of the cell cycle (B). A slight decrease was observed in apoptosis of LIN-28-overexpressing cells compared with mock-transfected clones (C).

LIN-28-Overexpressing Cells Differentiate Mostly to Extraembryonic Endoderm When Seeded at Clonal Densities

To evaluate whether the increased differentiation mentioned above was directed to a specific lineage, RNA was extracted from cells grown at clonal densities and analyzed using the TaqMan Stem Cell Pluripotency Array (full data from the gene card analysis are in supporting information Table 2). This array is comprised of multiple genes assessing the differentiation status of the cells using both markers of undifferentiated cells and markers of multiple lineages of differentiation [23]. Comparison of the expression profile of LIN-28-overexpressing clones with that of mock-transfected clones revealed that both still expressed markers of undifferentiated cells, such as NANOG and SOX2, which is consistent with the FACS analysis, which showed that the majority of the cells in both populations were undifferentiated. However, LIN-28-overexpressing cells seemed to be more committed to differentiation, since REX1 and TERT were already downregulated. Further examination revealed that multiple differentiation markers were upregulated in LIN-28-overexpressing cells, consistent with the increase in TRA-1–60− cells observed in the FACS analysis. Examination of these markers revealed that almost all the markers of extraembryonic endoderm present on the card were upregulated in the LIN-28-overexpressing cells (Fig. 5A). This is in contrast to markers of other lineages known to form upon spontaneous monolayer differentiation of ESCs, such as neuronal markers, which did not significantly change upon overexpression of LIN-28 (Fig. 5B). In order to verify the gene card analysis, we performed a real-time RT-PCR analysis on a small subset of genes, including ESC markers, ectoderm markers, and endoderm markers. Since many endoderm markers are expressed by both the definitive and the primitive endoderm, we looked at three groups of markers: embryonic endoderm markers, panendoderm markers, and primitive endoderm markers, as defined by previous work [24, 25]. The panendoderm markers include markers expressed in both definitive and primitive endoderm, such as SOX17, FOXA2, GATA6, and GATA4. As definitive endoderm markers, we chose to examine the expression of PDX1, which is specifically expressed relatively late in differentiation [26] and marks definitive endoderm. In addition, the expression of markers upregulated earlier in differentiation was examined, including CXCR4, which is suggested to be expressed in definitive and not primitive endoderm, and N-cadherin, which is expressed in the primitive streak, definitive endoderm, and mesoderm, but not in the primitive endoderm. As markers expressed in primitive endoderm and absent from definitive endoderm, we examined the expression of SOX7, expressed in primitive endoderm, laminin B1, expressed in parietal endoderm, and thrombomodulin, expressed in parietal endoderm and trophoblast. As shown in Figure 5C, among the markers examined, the panendodermal markers and primitive endoderm markers were upregulated to a higher extent than the remaining groups examined. Therefore it seems that, under clonal conditions, LIN-28 predisposes the cells to increased differentiation, mostly towards extraembryonic endoderm. To asses if the increased differentiation of LIN-28-overexpressing cells towards extraembryonic endoderm is also relevant to development in vivo, we examined the expression pattern of Lin-28 during early mouse development and compared it with that of Oct4. This was performed using the GNF SymAtlas (Genomic Institute of the Novartis Research Foundation, San Diego, During mouse development, from the blastocyst-stage embryo to E10.5 of development, Oct4 and Lin-28 show different expression patterns. Oct4 is downregulated around E7.5 of mouse development, whereas Lin-28 is still expressed at this stage, and it is downregulated at E10.5, but to levels still comparable with those present in the blastocyst (Fig. 6A). The fact that Lin-28 is upregulated in later stages in which the pluripotent compartment has to differentiate implies that Lin-28 might be involved in the differentiation of the cells and their exit from the pluripotent state. To examine whether, in vivo, Lin-28 may be expressed at high levels in a specific cell type, we performed immunostaining of mouse embryos corresponding to the stage in which Oct4 is beginning to be downregulated and Lin-28 is still being expressed. In E7.5 mouse embryos Lin-28 protein is indeed still highly expressed, but its expression seems highest in the external layer of the embryo, which corresponds to the extraembryonic endoderm (Fig. 6B, 6C). These results point out that not only in vitro, but also in vivo, LIN-28 is involved in differentiation, and that high levels of LIN-28 predispose the cells to differentiation into extraembryonic endoderm.

Figure 5.

High levels of LIN-28 direct differentiation to the extraembryonic endoderm fate. (A): The TaqMan Human Stem Cell Pluripotency Array was used to evaluate differentiation of LIN-28-overexpressing cells grown at clonal densities. Shown are expression levels of extraembryonic endoderm markers. Plotted are the changes in the cycle threshold of each gene (ddCt) in the LIN-28-overexpressing clones compared with mock-transfected clones. (B): Heat map depicting the results obtained by the TaqMan Human Stem Cell Pluripotency Array showing the dCT of extraembryonic endoderm markers (upper panel) and neuronal markers (lower panel). Shown are the results of control and LIN-28-overexpressing clones. (C): In order to verify the results obtained by the TaqMan Human Stem Cell Pluripotency Array, real-time reverse transcription-polymerase chain reaction was performed with markers of the undifferentiated state (embryonic stem [ES] cell) and ectoderm, embryonic endoderm, panendoderm, and primitive endoderm. Shown is the fold induction compared with control clone, which was normalized to the value of one. Abbreviations: dCT, delta cycle threshold; ddCt, delta delta cycle threshold; THBD, thrombomodulin.

Figure 6.

High levels of LIN-28 are correlated with differentiation. (A): Expression pattern of Lin-28 and Oct4 during early mouse development. The data were adopted from the GNF SymAtlas by the Genomic Institute of the Novartis Research Foundation. During the early stages of mouse development, Lin-28 continues to be highly expressed at least until day 10.5 of embryonic development (E10.5), at levels comparable with its expression during the blastocyst stage. This is in contrast to the expression pattern of Oct4, which is starting to be downregulated after day 7.5 of development. (B): Immunostaining for Lin-28 expression of early mouse embryos. Analysis of Lin-28 protein expression pattern in early mouse embryos (E7.5) revealed that Lin-28 is most highly expressed in the extraembryonic endoderm layer. Shown are hematoxylin and eosin stainings of serial sections (I, IV), a sketch of the different layers in the embryo (extraembryonic endoderm shown in red color, II, V), and immunostaining (Lin-28 is shown as red staining and 4′,6-diamidino-2-phenylindole is shown as blue staining, III, VI). (C): Control staining for LIN-28 antibody. Left panel, staining using the LIN-28 antibody, a polyclonal anti-human lin-28 rabbit antiserum; right panel, control staining using rabbit serum.

Knockdown of LIN-28 in HESCs Does Not Impede Their Self-Renewal

To further investigate the role of LIN-28 in HESCs, we utilized siRNA directed against LIN-28 to downregulate its expression. Initial siRNA experiments were performed using a previously established enhanced GFP (EGFP)-expressing clone and siRNA directed against EGFP, showing that the siRNA efficiency using our protocol leads to downregulation of EGFP in 50%–70% of the cells (data not shown). We first determined that the siRNA we used indeed downregulates LIN-28. Western analysis showed a marked decrease in LIN-28 level (Fig. 7A, I), and immunostaining showed that most cells were negative for LIN-28, with a few cells still expressing it, most likely cells not transfected (Fig. 7A, II; LIN-28 shown as red staining). Interestingly, although the LIN-28 level was markedly reduced, no change was observed in the level of OCT4 in the culture (Fig. 7A, I). When examining the morphology of the cells, no change was observed upon downregulation of LIN-28, compared with control cells, and the vast majority of the cells retained the characteristic morphology of HESC colonies (Fig. 7B). This is in striking contrast to a control experiment utilizing siRNA directed against OCT4, whose downregulation is known to lead to differentiation of ESCs [3, 27]. siRNA for OCT4 led to the appearance of multiple cells reminiscent of differentiated cells. Immunostaining for OCT4 (Fig. 7C, upper panel) or TRA-1–60 (Fig. 7C, lower panel) again showed that downregulation of LIN-28 did not lead to differentiation of the cells, since they continued to express markers of pluripotent cells. FACS analysis further emphasized that, indeed, LIN-28 downregulation did not affect any of the characteristics of undifferentiated HESCs (Fig. 7D). TRA-1–60 expression persisted in the culture, at a similar intensity to control cells and in a similar proportion of the cells (around 90% positive cells in both cases), the cell cycle profile remained unchanged, and no increase in apoptosis, as determined by annexin-V staining, was observed in the culture (around 20% positive cells in both cases). Therefore, we conclude that LIN-28 is not essential for the self-renewal of HESCs. In the absence of LIN-28, cells remain undifferentiated, they continue to proliferate at comparable rates, and there is no increase in cell death. Similar results were obtained with the HES9 cell line (data not shown).

Figure 7.

LIN-28 is dispensable for HESCs self-renewal. (A): Knockdown of LIN-28 in H9 cell line using siRNA was confirmed using Western blot analysis (I) and immunostaining using a LIN-28 antibody (II). Both confirmed a dramatic downregulation of LIN-28 protein compared with control cells. Western blot analysis showed no change in the level of OCT4 protein upon downregulation of LIN-28. (B): LIN-28 downregulation does not lead to a change in the morphology of HESCs. Whereas OCT4 downregulation led to the appearance of cells with differentiated morphology, LIN-28 knockdown did not affect the morphology of the cells compared with control cells. (C): LIN-28 knockdown does not lead to downregulation of pluripotency-associated markers. Whereas siRNA directed against OCT4 led to downregulation of both OCT4 (upper panel) and TRA-1–60 (lower panel), LIN-28 downregulation did not lead to a decrease in either the intensity or the proportion of cells positive for either OCT4 or TRA-1–60, two markers of undifferentiated HESCs, compared with control cells. (D): A fluorescence-activated cell sorting analysis was performed on control cells (upper panel) and cells transfected with siRNA against LIN-28 (lower panel) to examine the following parameters: their differentiation status (using TRA-1–60 staining), their cell cycle profile, and their apoptosis (by annexin-V staining). No change was observed in the HESCs upon downregulation of LIN-28. Abbreviations: DAPI, 4′,6-diamidino- 2-phenylindole; HESC, human embryonic stem cell; si, small interfering.


ESC-Enriched Genes Are Involved in the Reprogramming of Somatic Cells

We show that LIN-28 is highly enriched in undifferentiated HESCs compared with their differentiated derivatives, at both the mRNA and the protein level. In addition, the reprogramming genes can be divided into two groups based on their expression pattern. The first, containing OCT4, SOX2, NANOG, and LIN-28, are genes expressed at high levels in undifferentiated cells, and downregulated upon differentiation. The second group, comprised of MYC and KLF4, are genes whose expression does not seem to drastically change during differentiation. MYC is an oncogene and KLF4 was suggested to function as both an oncogene and a tumor suppressor. Whereas OCT4 and SOX2 have been suggested to be required for the reprogramming of somatic cells to ESC-like cells, exogenous MYC and KLF4 may be required primarily for an initial immortalization of the somatic cells, prior to their reprogramming to an embryonic state [28]. Exogenous MYC is dispensable for the derivation of iPS cells, and chimeric mice derived from iPS cells lacking exogenous c-Myc do not develop tumors [29]. Exogenous KLF4 is also dispensable for reprogramming [13]; however, it has not been examined whether KLF4 omission reduces the frequency of tumors formed by iPS cells. An intriguing question is whether the use of different genes for the derivation of iPS cells leads to different iPS cells. Since NANOG and LIN-28 expression seems unique to pluripotent cells, iPS cells generated using NANOG and LIN-28 may be more similar to ESCs than to those generated with MYC and KLF4. Such a comparison might shed light on the role of these genes in reprogramming and whether different iPS cells are established by their utilization.

The Sensitivity of HESCs to High Levels of LIN-28 and Other ESC-Specific Genes

At high cell densities and in the presence of feeder support, no differences were observed between control and LIN-28-overexpressing cells. However, at low cell densities, LIN-28-overexpressing cells had lower proliferation and higher differentiation rates, with differentiation mostly toward the extraembryonic endoderm. At low cell densities, there may be a lack of autocrine support, which HESCs depend on [30]. Therefore, LIN-28 overexpression may lead to a severe response to the absence of autocrine support, and this might imply that LIN-28 responds to exogenous signals and fine-tunes the cells decision to self-renew or differentiate. In addition, examination of the expression pattern of LIN-28 during early mouse embryogenesis also revealed that Lin-28 is expressed and even upregulated in stages in which the pluripotent compartment in the embryo initiates differentiation and loses its pluripotency. Furthermore, since high levels of Lin-28 were observed in the extraembryonic endoderm layer of mouse embryos, it seems that high levels of LIN-28 are indeed correlated with differentiation into this cell lineage, and are not the result of expressing nonphysiological levels of LIN-28 in vitro.

Increased differentiation upon LIN-28 overexpression initially seems puzzling. However, overexpression of additional genes enriched in undifferentiated ESCs and involved in reprogramming leads to differentiation. OCT4 upregulation leads to differentiation to primitive endoderm and mesoderm [3]. Overexpression of NANOG in HESCs leads to a transition from cells resembling the inner cell mass (ICM) to cells resembling primitive ectoderm [6]. Sox2 upregulation causes differentiation into a variety of cell types, excluding endoderm [31]. These results demonstrate that pathways governing self-renewal are sensitive to the levels of the proteins regulating them, and a slight increase in their level might tilt the balance between self-renewal and differentiation. This may also have implications for iPS cell derivation. Since only a narrow range of gene levels confers pluripotency, silencing of transgenes introduced during iPS formation may be crucial for the pluripotent state. Otherwise, overexpression may lead to subsequent differentiation. Therefore, during derivation of iPS cells, when colonies form and grow at clonal densities, the precise levels of LIN-28 may be critical.

LIN-28 has been suggested to be an RNA-binding protein [14], and it has been suggested to function as a translational enhancer [32]. Translational control is an area that has not been researched in depth in ESCs. However, recent research examined this issue and showed that protein translation in ESCs changes upon their differentiation, with differentiation inducing a switch leading to an increase in translational efficiency [33]. It seems that ESCs are prone to a rapid elevation in mRNA translation and protein production in response to differentiation signals. This seems to be in agreement with our results suggesting a role for LIN-28 in ESC differentiation. LIN-28 has been shown to increase translational efficiency during myoblast differentiation, and it has been shown to shift from monosomal fractions to polysomal ones upon initiation of differentiation [32], suggesting that it may function in a similar manner during ESC differentiation.

How can overexpression of LIN-28 lead to increased differentiation at clonal densities? It has been suggested that a critical factor in the function of CSD proteins is their density on the mRNA [15], which can determine the transition between active and nonactive translation of the mRNA. Such transition can also be regulated, for instance, by phosphorylation of the CSD protein [34], and this provides the possibility of signaling pathways to regulate the process. In mouse myoblasts, Lin-28 is required for differentiation [32]. Overexpression of Lin-28 in a myoblast cell line had a stimulating effect on terminal differentiation, whereas its downregulation decreased the differentiation efficiency. Although examined in a system in which LIN-28 shows an opposite expression pattern (low expression in proliferating undifferentiated muscle cells and upregulation during terminal differentiation), these data agree with our results, which indicate that, in HESCs as well, LIN-28 may be involved not in the regulation of self-renewal, since it is dispensable for this, but in the timing of differentiation. KLF4 was also shown to act as a switch from proliferation to differentiation in skin [35] and colon [36], and therefore this might define a new group of genes involved in iPS cell derivation—genes involved in the transition between different differentiation stages of cells. Lin-28 has been shown to reside in both polysomal fractions, in which mRNA is translated, and in P-bodies, storage sites for nontranslating mRNA [37]. Therefore, the transition between different differentiation stages may be achieved by switching from enhancing translation of mRNA by loading it with polysomes to preventing translation by directing mRNA to P-bodies.

LIN-28 and Other ESC-Specific Genes Are Not Essential to ESC Self-Renewal

siRNA experiments showed that LIN-28 is not essential for HESC self-renewal. Knockdown of LIN-28 did not change the morphology of the cells, the expression of pluripotency markers, or the proliferation or apoptosis rates. This implies that LIN-28 is not essential for the maintenance of self-renewal. Nanog and Sox2 are also not essential for ESC maintenance. Nanog was considered essential for ESC self-renewal [5]; however, although ESCs lacking Nanog are prone to differentiation, they can self-renew indefinitely in its absence and remain pluripotent [38]. Still, it seems indispensable for the establishment of the pluripotent compartment, whether for the formation of germ cells or construction of the ICM. Sox2 was also considered essential for ESC self-renewal [39]; however, its essential function is maintaining the requisite levels of Oct4 expression [7], and forced expression of Oct4 can rescue the pluripotency of Sox2-null ESCs.

Since LIN-28 has been shown to raise the efficiency of iPS cell formation, it is important to analyze its function in HESCs. This work is an effort to dissect its role in the network of proteins involved in governing the pluripotency of ESCs. It was recently shown that LIN-28 is involved in the selective blockade of pri-let-7 microRNA processing in ESCs [40]. let-7 has been suggested to have a tumor suppressor role in multiple cancers by repression of oncogenes such as Hmga2 [41] and Ras [42]. Therefore, it has been suggested that, in a manner similar to what was previously suggested to be the role of MYC and possibly KLF4 in reprogramming, LIN-28 might also improve the efficiency of reprogramming through promoting an oncogenic hyperproliferative state through downregulation of the tumor suppressor let-7 and upregulation of HMGA2 and RAS [40]. However, our results demonstrate that LIN-28 does not simply shift the cell to a hyperproliferative state. Such an explanation cannot account for the observed increased differentiation and decreased cell division upon upregulation of LIN-28.

In addition, we show that, although LIN-28 is dispensable for the maintenance of self-renewal of HESCs, it seems to be involved in the decision to exit this state. This raises the question of its potential role in iPS cell formation. Since we show that Lin-28 is involved in the efficiency with which undifferentiated ESCs leave the pluripotent state, it might be possible that it also has the opposite effect, that is, directing cells back into the pluripotent state. The question of the mechanism that accomplishes this remains unresolved. Presumably, it can enhance the translation of genes important for reprogramming, or decrease the translation of somatic genes expressed in the somatic cell during reprogramming. Future work examining mRNAs and microRNAs targeted by LIN-28, both in ESCs and in different stages of iPS cell formation, will aid in understanding LIN-28 action during reprogramming of iPS cells and in ESC maintenance and differentiation.


We thank Dr. Joel G. Belasco for providing us with the LIN-28 antibody, Oded Kopper and Dr. Inbar Friedrich Ben-Nun for help with embryoid body sections, Tamar Golan-Lev for help with mouse embryo sections and with graphic design of the figures, and Dr. Marjorie Pick for help with FACS analysis. We also thank Drs. Danny Kitsberg and Rachel Eiges for critically reviewing the manuscript. This research was partially supported by funds from “Bereshit Consortium” The Israeli Ministry of Trade and Industry (grant number 37675), by the European Community (ESTOOLS, grant number 018739), and by the Legacy fund.


The authors indicate no potential conflicts of interest.