JNK/SAPK Signaling Is Essential for Efficient Reprogramming of Human Fibroblasts to Induced Pluripotent Stem Cells

Abstract Reprogramming of somatic cells to the phenotypic state termed “induced pluripotency” is thought to occur through three consecutive stages: initiation, maturation, and stabilisation. The initiation phase is stochastic but nevertheless very important as it sets the gene expression pattern that permits completion of reprogramming; hence a better understanding of this phase and how this is regulated may provide the molecular cues for improving the reprogramming process. c‐Jun N‐terminal kinase (JNK)/stress‐activated protein kinase (SAPKs) are stress activated MAPK kinases that play an essential role in several processes known to be important for successful completion of the initiation phase such as cellular proliferation, mesenchymal to epithelial transition (MET) and cell cycle regulation. In view of this, we postulated that manipulation of this pathway would have significant impacts on reprogramming of human fibroblasts to induced pluripotent stem cells. Accordingly, we found that key components of the JNK/SAPK signaling pathway increase expression as early as day 3 of the reprogramming process and continue to rise in reprogrammed cells throughout the initiation and maturation stages. Using both chemical inhibitors and RNA interference of MKK4, MKK7 and JNK1, we tested the role of JNK/SAPK signaling during the initiation stage of neonatal and adult fibroblast reprogramming. These resulted in complete abrogation of fully reprogrammed colonies and the emergence of partially reprogrammed colonies which disaggregated and were lost from culture during the maturation stage. Inhibition of JNK/SAPK signaling resulted in reduced cell proliferation, disruption of MET and loss of the pluripotent phenotype, which either singly or in combination prevented establishment of pluripotent colonies. Together these data provide new evidence for an indispensable role for JNK/SAPK signaling to overcome the well‐established molecular barriers in human somatic cell induced reprogramming. Stem Cells 2016;34:1198–1212


INTRODUCTION
Since 2007, it has been possible to reprogram human somatic cells back to an embryonic stem cell (ESC) like stage via introduction of four transcription factors, namely OCT3/4, SOX2, KLF4, and c-MYC (referred as OSKM; [1]). The reprogrammed cells termed human induced pluripotent stem cells (hiPSCs), akin to human embryonic stem cells (hESCs) are characterized by the ability to proliferate indefinitely and have the potential to give rise to all cell types found in the adult organism [2]. The aforementioned pluripotency factors have also been implicated in the initiation of tumorigenesis in various tissues [3][4][5] and are considered as potent oncogenes. The risk associated with the introduction of these oncogenes in normal human somatic cells is the likely activation of antioncogenic pathways through a process named oncogenic stress which can often result in cell cycle arrest as a means of protection from tumorigenesis [6]. Activation of cell cycle arrest, p53 activation and reduced cellular proliferation have been described as molecular barriers to efficient reprogramming [7][8][9]. In view of these findings, we hypothesised that transduction of dermal skin (Fibroblasts: Lonza Group Ltd, Basel, Switzerland; Lonza:http://www.lonza.com/) with OSKM would activate oncogenic stress and induce growth arrest thus providing an additional barrier to reprogramming. Hence inhibition of key sensors of this pathway could provide useful targets for removing this barrier and increasing its efficiency.
Stress activated MAP kinase signaling pathways are important mediators of cellular responses to intra-and extracellular signals such as growth factors, hormones, and environmental stresses. These pathways consist of triple kinase cascades comprising the MAP kinases which are phosphorylated and activated by MAP kinase kinases (MKKs) and then further phosphorylated and activated by MAP kinase kinase kinases (MKKKs) [10]. In mammals, three distinct MAP kinase pathways have been identified resulting in activation of ERK, c-Jun N-terminal kinase (JNK) and p38 [10]. One of the key sensors of oncogenic stress is the JNK signaling pathway which responds by phosphorylating and stabilising p53 via its downstream mediator MKK7 [11]. The JNK pathway, also named stress-activated protein kinase pathway (SAPK) is essential for providing a cellular response to extracellular changes such as ultraviolet and reactive oxygen species induced damage, mechanical stress and osmolarity changes. For the JNK/SAPK pathway 14 different MKKs have been described to date [12]. MKK7 exclusively activates JNKs and MKK4 is unique in its ability to phosphorylate and activate two MAP kinase groups: JNKs and p38. Both MKK4 and MKK7 are responsible for phosphorylation of JNK/SAPKs at Tyrosine (Tyr) and Threonine (Thr) residues located at the activation loop. In murine embryonic stem cells (mESCs), MKK4 is shown to phosphorylate JNK/SAPKs at Tyr 185 residue, while MKK7 phosphorylates the Thr 183 residue, and together they cause dual phosphorylation of JNK/SAPKs, thus leading to its optimal activation [13].
Most of the data on JNK/SAPK signaling and activation of its targets has been obtained from work performed in somatic and cancer cells [14]; however in the last few years there have been a number of publications describing the role of this signaling pathway during embryonic development and in ESC function. Some of these have shown that MKK4 and MKK7 null mice die before E.12.5, highlighting the necessity of JNK/SAPK signaling pathway during embryonic development and suggesting that other pathways cannot substitute for MKK4 and MKK7 [15,16]. Furthermore, JNK1 and JNK2 have been shown to play a negative role in the reprogramming of murine fibroblasts by suppressing Klf4 activity [17]. In contrast, hESCs are characterized by high levels of JNK/SAPK activity which is important for maintenance of pluripotency; however a role for this signaling pathway during reprogramming of human somatic cells has not been described previously and forms the main focus of this manu-script. We report herein that transduction of human fibroblasts with OSKM (both as single and polycistronic Sendai based viruses) induces the activation of JNK/SAPK signaling during the initiation and maturation stage of reprogramming. Downregulation of JNK/SAPK with a specific chemical inhibitor or by RNA interference (RNAi) leads to the emergence of only partially reprogrammed colonies which disaggregate and are lost during the maturation stage of reprogramming. Our data suggest that JNK/SAPK signaling plays an important role in several key processes that are shown to be important during cellular reprogramming namely the induction of mesenchymal to epithelial transition (MET), activation of cellular proliferation as well the maintenance of the pluripotent phenotype. Hence lack of hiPSC colonies and loss of partially reprogrammed cells can be attributed to one or more of these three cellular processes which are tightly regulated by JNK/SAPK signaling.

JNK/SAPKs Kinases are Activated During the Course of Reprogramming
To understand the role of JNK/SAPKs during the generation of hiPSCs we assessed the expression of JNK1 and JNK2 and their upstream activators MKK4 and MKK7 in two different primary dermal skin fibroblasts (Neonatal/Neo1 and Adult/ Ad3), several hiPSCs clones derived therefrom (Fig. 1A, 1B, Supporting Information Fig. 1B) and hESCs (H9). Human ESCs are characterised by high levels of JNK/SAPK activity which has been shown to be important for maintenance of the pluripotent stem cell state [19]. In accordance with this, we found the highest levels of mRNA JNK1 expression in hESCs when compared to several hiPSCs clones derived from two adult fibroblast samples (Fig. 1A, 1B, Supporting Information Fig. 1B); however these differences were not maintained at the protein level across the iPSC clones examined (Fig. 1B). We also observed that neonatal fibroblasts had lower expression of all four kinases examined when compared to adult fibroblasts (Fig. 1A, 1B). These differences were in part maintained in the respective hiPSC lines with the adult derived hiPSC clones showing higher expression of JNK1 when compared to neonatal derived hiPSCs at both transcript and protein level (Fig. 1A, 1B, Supporting Information Fig. 1A, 1B).
Transduction of OSKM caused a significant increase in JNK1 expression in adult fibroblasts and a dual increase in JNK1 and JNK2 expression in neonatal fibroblasts as early as day 3 of reprogramming ( Together these data suggest an increased JNK/SAPK activity during the initiation and maturation stage of reprogramming. To determine how the four transcription factors (OSKM) individually contribute to JNK/SAPK activation during reprogramming, we performed transduction with each single factor in neonatal fibroblasts and substituted the rest of the factors with an equivalent number of control-GFP virus particles (Supporting Information Fig. 1C). Transduction with OCT4, KLF4 and c-MYC contributed mostly to an increase in JNK2 expression, while introduction of SOX2 increased both JNK1 and JNK2 expression with a preference for JNK1. Transduction with control viral particles alone did not lead to increased JNK1/ JNK2 expression or their phosphorylated form (data not shown), indicating that activation of JNK/SAPK pathway during reprogramming is not related to the transduction event, but specifically to introduction of OSKM in somatic cells.
To further confirm the increase in p-JNK/SAPK expression at a cellular level, we used flow cytometric analysis as described previously [20]. This enabled us to follow three cellular subpopulations during the course of reprogramming; fully reprogrammed cells (TRA-1-601CD44-), partially reprog-rammed cells (TRA-1-601CD441) and fibroblasts (TRA-1-60-CD441; Fig. 1D, 1E). pSAPK was expressed in all three subpopulations (Fig. 1F, 1G). It is interesting to note that the partially reprogrammed cells showed the highest percentage of pSAPK 1 expressing cells; however this declined toward the end of the reprogramming period (Fig. 1G). Furthermore, the percentage of pSAPK expressing cells increased in the fully reprogrammed subpopulation from day 14 to day 18 (Fig. 1G) then decreased by day 28, corroborating the Western blotting analysis (Fig. 1B, 1C, Supporting Information Fig. 1A). We obtained the same profile of emergence of the three cellular subpopulations and the same trend of pSAPK activation (Supporting Information Fig. 2A, 2B) upon application of a polycistronic Sendai vector (Cytotune 2.0), demonstrating that activation of JNK/SAPK is independent of hiPSC transduction protocol. Together these data suggest an important role for the activity of JNK/SAPK in fully reprogrammed cells during the maturation stage of reprogramming.

Inhibition of JNK/SAPK Activation with Chemical Inhibitors Causes Disaggregation and Loss of hiPSC Colonies During Maturation Stage
To test the function of JNK/SAPKs in generation of hiPSCs we used a small molecule SP600125, which has been shown to significantly inhibit expression of all three JNK genes namely JNK1, JNK2, and JNK3 [21]. Our data show that 5 mM SP600125 (named SAPKi thereafter) is sufficient to downregulate the expression of JNK1 and JNK2 in hESCs ( Fig. 2A) corroborating previous data published in mouse ESCs [22]. To investigate the impact of JNK/SAPK inhibition in the reprogramming process, we used SAPKi for 24 hours at different time points as summarized in Figure 2B and Supporting Information MKK4, MKK7 and JNK1/SAPK1 are Indispensable for hiPSC Generation Application of chemical inhibitors can induce transient downregulation of JNK/SAPK which may mask some of their effects on cellular reprogramming. To investigate if permanent down-regulation of JNK/SAPK yields the same or more extensive impact we generated stable cell lines of neonatal and adult fibroblasts in which MKK4, MKK7 and JNK1/SAPK1 (JNK1) were downregulated by shRNA. Successful downregulation of each of the three genes was confirmed by qRT-RCR at the mRNA level (Fig. 3A, 3B) and protein level (Fig. 3C). We observed that downregulation of MKK4, MKK7 and JNK1 resulted in a more effective repression of downstream targets in neonatal fibroblasts when compared to their adult   (Fig. 3D, 3E). In part this may be explained by higher expression of these kinases in adult fibroblasts (Fig.  1A, 1B) and possibly by the existence of additional compensatory mechanisms.
Morphological observation of MKK4, MKK7 and JNK1 shRNA treated fibroblasts revealed differences between neonatal and adult cells (Supporting Information Fig. 4A, 4B). Neonatal fibroblasts continued to proliferate very slowly after passage 3, thus corroborating previously published literature reporting reduced proliferation and G2/M phase arrest in mouse embryonic fibroblasts [16]. We also observed that MKK4, MKK7 and JNK1 shRNA adult fibroblasts stopped growing after the first passage and demonstrated spindle shape and flattened morphology similar to senescent cells [23]. Despite this, we were unable to detect activation of senescence markers (b-gal, p15, p16, p19) in either neonatal or adult fibroblasts deficient in JNK signaling (data not shown).
We performed transduction of neonatal and adult control and MKK4, MKK7, JNK1 shRNA fibroblasts with OSKM and as in the case of SAPKi we observed greatly reduced numbers of emerging hiPSCs (TRA1-601CD44-) and partially reprogrammed cells (TRA-1-601CD441; Fig. 4A, 4B). Quantitative RT-PCR analysis at 7 days post transduction indicated that downregulation of MKK4, MKK7 and JNK1 shRNA was maintained (Supporting Information Fig. 5). On a morphological level, partially reprogrammed cells were detected from neonatal fibroblasts transduced with JNK1, MKK4 and MKK7 shRNA as early as day 14 (Fig. 4B); however these started to disaggregate by day 18, corroborating the data observed with SAPKi. Transduced adult fibroblasts exhibited a few minor morphological changes during the first 10 days of reprogramming; however they did not form either partially reprogrammed or fully reprogrammed colonies (data not shown). The same profile of colony disaggregation and a significant reduction in the percentage of true emerging hiPSC (TRA1-601CD44-) was observed when JNK1, MKK4 and MKK7 shRNA neonatal fibroblasts were reprogrammed with a Sendai polycistronic vector (Cytotune 2.0; Supporting Information Fig. 6). To understand whether the absence of a true hiPSCs from JNK1 shRNA transduced cells was associated with cellular differentiation we performed qRT-PCR analysis of the expression of three germ layer differentiation markers on cells collected from remaining colonies at day 18. This analysis indicated the upregulation of al differentiation markers examines apart from FGF5, suggesting that the onset of differentiation can be an additional reason for the absence of hiPSC from JNK1 shRNA cells (Fig. 4C). In accordance with the above results, no alkaline phosphatase positive hiPSC colonies were obtained from either neonatal or adult fibroblasts, indicating that MMK4, MKK7 and JNK1 are indispensable for successful reprogramming.

Impact of JNK/SAPK Downregulation on Pluripotent Stem Cells
To understand the deleterious effect of JNK/SAPK downregulation during the reprogramming process, we inhibited JNK/ SAPK pathway in hESC via administration of SAPKi. Human ESCs have been characterized by high JNK activity [24] and inhibition of this pathway with chemical inhibitors has been shown to result in cellular differentiation, decreased OCT4 and NANOG expression [19]. Consistent with published results, we observed loss of typical pluripotent stem cell morphology (Supporting Information Fig. 7A) downregulation of OCT4 expression (Supporting Information Fig. 7B) and an increase in KLF4 expression, corroborating previously published data on mESCs [17]. Also, in agreement with previous reports [17] we found increased expression of differentiation markers characteristic for neuroectoderm (SOX1, NESTIN), trophoectoderm (CDX2) and primitive endoderm (AFP; Supporting Information Fig. 7C). No change in cell survival was observed as a result of SAPKi application (data not shown); however an increase in G1 phase at the expense of S phase was observed upon cell cycle analysis by flow cytometry (Supporting Information Fig. 7D). Together these data suggest that JNK activity is required for the maintenance of pluripotency; hence its downregulation during the reprogramming process can provide one of the underlying factors for the emergence of colonies with partially reprogrammed phenotype and lack of fully reprogrammed hiPSCs.

STEM CELLS
A recent report has suggested that heat shock treatment of hESC results in activation of SAPK/JNK signaling and reduced expression of OCT4 due to increased binding of heat shock factor 1 to the OCT4 promoter [25]. Although this finding may appear contradictory to our data presented above, we would like to point out that the difference may be due to physiological stimuli. While hESC data presented in this work were collected from normal culture conditions (378C and normoxic oxygen levels), the Buyn et al. findings relate to the "heat shock" conditions which represent stress conditions and are likely to affect many other signaling pathways in hESC in addition to the JNK/SAPK pathway.

Impact of JNK/SAPK Downregulation on Cellular Proliferation and Cell Cycle Regulation in Somatic Cells
To better understand the changes induced in neonatal and adult fibroblasts by downregulation of the JNK/SAPK pathway, we administered SAPKi to culture media. Similarly to our observations obtained with JNK1 shRNA (Supporting Information Fig. 4), we observed that SAPKi treated fibroblasts became more elongated, flattened and in some cases spindle shaped (Supporting Information Fig. 7E). Growth curve analysis also indicated a slower rate of proliferation in SAPKi treated fibroblasts (Supporting Information Fig. 7F) supporting previously published data [26]. This was further confirmed by cell cycle analysis which demonstrated that downregulation of JNK/SAPKs resulted in cell accumulation in G1 phase at the  Fig. 7G). Western blot analysis also indicated the downregulation of the downstream target, c-Jun known to be important for cellular proliferation (Supporting Information Fig. 7H).
To understand the molecular changes underpinning JNK/ SAPK inhibition, we performed quantitative RT-PCR analysis of JNK target genes involved in cellular proliferation, namely c-Jun, E2F2 and ATF2. Since the effect of SAPKi may be transient, we performed this analysis on neonatal and adult fibroblasts treated with MKK4, MKK7 and JNK1 shRNA described in Figure 3A and 3B. We found that c-JUN, ATF2 and E2F2 were significantly downregulated in all cases (Fig. 5A). Of interest are also the findings that key cyclin dependent kinases (CDK6, CDK2 and CDK1) were downregulated in at least one of the shRNA treated samples. Furthermore CCNB1, a key cyclin involved in both G1/S and G2/M regulation was downregulated in MKK4, MKK7 and JNK1 shRNA treated samples, while Cyclin D1 (CCND1) was downregulated in MKK4 and JNK1 shRNA cells (Fig. 5B). Flow cytometry cell cycle analysis revealed that in contrast to fibroblasts treated with control shRNA, which demonstrated a characteristic somatic cell cycle distribution at day 0, MKK4, MKK7 and JNK1shRNA treated fibroblasts showed an increased cell accumulation at G1 phase at the expense of S and G2/M phase prior to reprogramming (day 0) and at day 7 of reprogramming (Fig. 5C). Together these data suggest an impaired cell cycle regulation and proliferation as a result of JNK/SAPK downregulation.

Importance of JNK/SAPK Signaling for the MET
MET is an important process during somatic cell reprogramming and is orchestrated by the suppression of pro-epithelial to mesenchymal transition (EMT) signals and activation of an epithelial signature in reprogrammed cells [27]. During this process, two of the key reprogramming factors SOX2 and OCT4 suppress the EMT mediator SNAIL, while KLF4 induces epithelial gene expression including E-cadherin [27]. Ultrastructural visualization of this process has shown that until day 6 of reprogramming, human fibroblasts retain their mesenchymal characteristics. The MET occurs between days 6 and 12 and this is followed by maturation of the epithelial phenotype from day 12 to day 18 [28]. Although traditionally JNKs have been thought to be activated in response to apoptotic, proliferation and stress signals, more recently it has been shown that they can also act as cell junction regulators [29] and can be involved in cell migration which is tightly linked to dynamic formation as well as dissolution of cell-cell junctions [29,30]. Given the importance of MET in hiPSC formation we asked whether inhibition of JNK/SAPK signaling leads to loss of E-cadherin expression-dependent adhesion, followed by loss of cell-to-cell communication resulting in disintegration of partially reprogrammed colonies and their loss from culture. Immunocytochemical analysis of OSKM treated control fibroblasts indicated formation of a bright network of E-cadherin and b-catenin plasma membrane staining in TRA-1-60 positive cells (Fig. 6A, 6B), indicating formation of adherent junctions between the cells. In contrast, E-cadherin and b-catenin staining was observed in very few of the cells present in the partially reprogrammed TRA-1-60 positive colonies obtained from reprogramming of JNK1 shRNA treated fibroblasts (Fig. 6A,  6B). Furthermore, the bright E-cadherin and b-catenin networks which ensure the cell-cell adhesion and communication were missing in reprogrammed JNK1 shRNA treated fibroblasts. Assessment of N-cadherin expression, a marker of fibroblasts that fail to undergo MET and further convert into hiPSC [31] indicated a higher number of expressing cells in JNK deficient cells (Fig. 6C), further supporting the disruption of MET during the reprogramming of JNK deficient samples. Quantitative RT-PCR analysis indicated downregulation of epithelial marker (E-cadherin) and upregulation of mesenchymal markers (SNAIL, N-cadherin, TWIST, VIMENTIN, ZEB1 and SLUG) under both Cytotune 1 and 2 Sendai reprogramming (Fig. 6D, Supporting Information Fig. 8), further supporting the disruption of MET during the reprogramming of JNK deficient samples.

DISCUSSION
In this manuscript, we investigated the expression of key components of JNK/SAPK signaling and found that this pathway is activated as early as day 3 of reprogramming with both partially reprogrammed cells (TRA-1-601CD441) and fully reprogrammed cell populations (TRA-1-601CD44-) showing an increase in the number of pSAPK expressing cells until day 18, regardless of the reprogramming protocol used. In contrary to our first expectations, we found that inhibition of JNK/SAPK signaling pathway either by chemical inhibitors or RNAi mediated downregulation of MKK4, MKK7 or JNK1 resulted in the complete abrogation of hiPSC colony formation and emergence of partially reprogrammed cells which were lost during the maturation stage of reprogramming. Together these data highlight the importance of this signaling pathway in human somatic cell induced reprogramming which to the best of our knowledge has not been reported previously. In contrast to our findings, it has been reported that Jnk1-/-and Jnk2-/murine fibroblasts exhibit a greater potency for reprogramming due to increased Klf4 activity which is phosphorylated by Jnk1 and Jnk2 through inhibitory phosphorylation at Thr residues 224 and 225 [17]. These differences in the role of JNK signaling during the reprogramming of mouse versus human somatic cells can reflect different modes of actions for this pathway in each species; however this needs to be investigated further. We were also surprised to see the lack of functional redundancy between various components of key signaling molecules involved in JNK signaling as downregulation of MKK4, MKK7 and JNK1 in fibroblasts led in all cases to complete absence of hiPSC colonies. Notwithstanding this, we would like to point out that all single component downregulation by RNAi led to further changes in the expression of additional components of this pathway (e.g., MKK4 shRNA resulted in downregulation of MKK7 and JNK1) and/or parallel pathways such as p38, thus outlining the complexity of regulation of this pathway which has been reported previously and must be taken into account when investigating the function of separate signaling components [32][33][34]. Furthermore, we observed some differences in the expression of key components of JNK/SAPK signaling between neonatal and adult fibroblasts as well as their response to JNK/SAPK inhibition which indicates differences in the mode of operation of this pathway in these two cell types. Nevertheless, the activation window of JNK/SAPK signaling during the reprogramming process and the outcome of JNK/SAPK downregulation in the reprogramming of both cell types were very similar, suggesting that at least the JNK/SAPK functions related to reprogramming are likely to be conserved in both cell types. Human cord blood and hair follicle epithelial cells have emerged as new sources of somatic cells for reprogramming because of their accessibility. Recent reports suggest that JNK/SAPK signaling is required in cord blood cells to prevent the onset of senescence [35], while regulating apoptosis in hair follicle epithelial cells [36]. While our experiments have only been performed in dermal skin neonatal and adult fibroblasts thus   preventing to make sound conclusions about the applicability of our findings in other cell types, the well-known role of JNK/SAPK signaling in maintaining cell survival and proliferation may suggest a critical role for this pathway during the reprogramming process; however this has to be further investigated experimentally.
The functions of JNK/SAPK signaling are complex and encompass a wide range of cellular processes which yield different outcomes in specific cell types [37]. Given the increased expression of key components of JNK/SAPK signaling during the initiation and maturation phases of reprogramming, we sought to identify its involvement in several processes shown to be important during this time window namely induction of MET [38], silencing of tumor suppressor and cell cycle inhibitor genes [8], induction of apoptosis [39] and induction of cell proliferation [40,41]. Consistent with previous reports [42], we found that inhibition of JNK signaling did not have any impact on the apoptosis of fibroblasts undergoing reprogramming (data not shown); hence the link between apoptosis induction and JNK signaling can be excluded. JNK signaling has been shown to be required for the expression of c-Jun and JunD, two essential components of the AP1 transcription factor which is essential for cell proliferation and prevention of senescence often associated with impaired JNK signaling [43]. Despite repeated investigations, we did not observe expression of senescence associated bgalactosidase in fibroblasts with deficient JNK signaling achieved through chemical inhibition or RNAi. Neither did we see an increase in the expression of cell cycle inhibitors including p15, p16, p19 which have been linked to onset of senescence in various cell types [44,45]. These findings could reflect differences in JNK outcomes in vivo versus in vitro experiments (tissue culture) which have been reported previously, or the time lag that is needed from cell cycle arrest to detectable expression of these markers. Despite lack of senescence marker expression, we observed a significant decrease in cellular proliferation, increased percentage of cells in G1 phase of the cell cycle at the expense of S phase and reduced ). It is important to note that such expression changes were specific for each component of the JNK signaling pathway downregulated; however in all cases simultaneous downregulation of at least one CDK and one partnering Cyclin, an event that has been shown to be important for efficient reprogramming [40] was observed. These findings were further corroborated by decreased expression of cellular proliferation markers c-Jun, E2F2 and ATF2 in all three shRNA treated samples consistent with previous reports indicating an important role for MKK4, MKK7 and JNK1 for cell cycle progression [16,46]. High cellular proliferation rate akin to ESC state is an important early event for cellular reprogramming [40]. This leads us to suggest that suppressed cellular proliferation and cell accumulation at the G1 phase (both known as reprogramming barriers; [47]) at the onset of reprogramming of JNK1, MKK4 and MKK7 deficient neonatal and adult fibroblasts is a key reason for the complete lack of hiPSC observed during reprogramming of these samples.
The efficiency of reprogramming can be promoted by the onset of epithelial expression markers concomitantly with the repression of mesenchymal markers in cells undergoing reprogramming during the 6th until the 12th day of reprogramming, known as initiation phase [48,49]. This is precisely the window during which the JNK signaling is activated; hence we asked the question of whether MET transition is affected in JNK deficient cells. Our data suggest that while control cells showed well organized TRA-1-60 1 colonies characterized by plasma membrane staining of E-cadherin and b-catenin indicating formation of adherent junctions typical of epithelial cells, JNK deficient cells displayed a reduced percentage of cells expressing these markers at both cellular and mRNA levels, which suggests that fewer cells are able to undergo MET transition, thus leading to inefficient reprogramming. Furthermore, the cel-lular networks that guarantee cell-cell communication while present and well organized in control cells were lacking in JNK/SAPK deficient fibroblasts. It has been suggested that expression of E-cadherin and the presence of intact adherent junctions are essential for maintenance of pluripotency in hESC by ensuring access to critical autocrine signals and by promoting cell-cell exchange of signals through the gap junctions [49,50]. In this context, it is interesting to note that all colonies that emerged during reprogramming of JNK/SAPK deficient fibroblasts were partially reprogrammed and completely disaggregated and subsequently lost during the maturation stage of reprogramming thus leading us to suggest that deficient JNK/SAPK signaling followed by impaired cell to cell contact is likely to be one of the causative factor that counts for the partially reprogrammed phenotype and for the loss of colonies.
Our inhibition experiments were performed during the initiation stage of reprogramming. In all cases despite the time of JNK/SAPK downregulation we obtained no hiPSC colonies. However, we did not perform JNK/SAPK downregulation studies at the stabilization stage where hiPSC colonies are present and continue to grow and develop. There were two main reasons behind this decision: (1) JNK/SAPK activation occurs during the initiation and maturation stage and shows a drop during the stabilization stage and (2) our data presented in this manuscript together with published reports indicate loss of pluripotency in hESCs upon JNK downregulation [19], which would suggest that downregulation of JNK/SAPK signaling after the emergence of hiPSC colonies would result in their differentiation and thus provide a further push toward loss of true bona fide hiPSC colonies.

CONCLUSION
Together our data indicate an increase in JNK/SAPK activity during the initiation and maturation phases of reprogramming regardless of the reprogramming protocol and an indispensable role for the generation of hiPSC colonies. Furthermore, we have shown that inhibition of JNK/SAPK signaling results in reduced cell proliferation, disruption of MET and loss of pluripotent phenotype which either singly or in combination prevent establishment of pluripotent colonies as shown in the summary (Fig. 7).