Oxidative stress can be induced by a vast range of agents, including xenobiotic, drugs, heavy metals, and ionizing radiation, and can lead to the generation of reactive oxygen species (ROS) and electrophiles. ROS often contribute to diseases such as cancer, cardiovascular complications, acute and chronic inflammation, and neurodegenerative condition [1
]. Therefore, cellular defense mechanisms are required to control constantly the levels of ROS and to prevent their accumulation. Many antioxidant and/or detoxification enzymes, such as NADPH: quinone oxidoreductase 1 (NQO1) [3
], glutathione S
-transferase (GST) [5
], and heme oxygenase-1 (HO-1) [6
] contribute to cellular defenses systems against ROS. The coordinated induction of the genes encoding the Phase II detoxification and antioxidant enzymes is governed by the core sequence 5'-G/A TGACNNNGC-3' located in their gene regulatory regions, and is termed the antioxidant responsive element (ARE) [7
]. Nuclear factor-erythroid 2-related factor 2 (Nrf2) is the key factor responsible for binding to AREs. This basic leucine zipper (bZIP) transcription factor belongs to the cap “n” collar (CNC) protein family (encompassing Nrf1, Nrf2, Nrf3, and p45 NF-E2) and forms heterodimer complexes with a small Maf protein (either MafG, MafK, or MafF), which provide high-affinity binding to AREs [5
]. The Nrf2-ARE axis is absolutely essential for the induction of genes encoding antioxidant and detoxification enzymes, and contributes to the regulation of cell growth, differentiation, and survival after oxidative stress.
The c-Jun dimerization protein 2 (JDP2) is a bona fide member of the AP-1 transcription factor family [9
] and is a bZIP repressor protein that is highly expressed in the brain and lung [11
]. JDP2 can homodimerize or form heterodimers with other AP-1 family members [9
]. The mechanism by which JDP2 represses AP-1 transcription involves competition for DNA binding, inactive heterodimer formation [9
], indirect recruitment of histone deacetylase 3 [13
], nucleosome assembly activity, inhibition of histone acetylation [14
], and potential competition with JNK phosphorylation [15
]. The knockout (KO) of Jdp2
gene affects adipocyte differentiation [16
], resistance to replicative senescence [17
], cell cycle arrest and regulation of cyclin A2 [18
]. Retroviral activation of alternative Jdp2
in T cell lymphomas of mice has been reported, providing the strong evidence for a gain-of-function of Jdp2 in cancer development in the hematopoietic system [18
]. Recent studies of tumor cells have demonstrated that JDP2 is a tumor suppressor [19
], suggesting that genomic alterations might be the underlying cause of cancer development. However, some studies have shown that JDP2 can potentiate cancer cell growth [21
]. It is not known whether these amplifications of JDP2 produce abundant amounts of normal JDP2 protein or the truncated JDP2 mRNAs, which are thought to be an oncogene [18
]. Other bZIP factors, such as JunD, PMF-1, and ATF4, bind to the ARE and can regulate ARE-driven transcription [23
]. The small Maf proteins can dimerize with CNC factors, such as Nrf2, and with other bZIP factors, including Fos, FosB, Bach 1, and Bach 2, via their leucine zipper domain [25
]. Because JDP2 is also a member of the bZIP family of transcription factors, we examined whether JDP2 binds to Maf-family and/or Nrf2 proteins, and whether it can regulate ARE-dependent genes encoding antioxidant and detoxification enzymes.
Somatic cells have been reprogrammed successfully into induced pluripotent stem cells (iPSCs) by ectopic overexpression of the transcription factors OCT4, SOX2, KLF4, and c-MYC [26
]. Other sets of transcription factors have also been reported to induce iPSCs from somatic cells [27
]. Similar approaches have been used for the reprogramming of cancer cells into induced pluripotent cancer cells (iPCCs) by different sets of transcription factors [29
]. Both types of pluripotent cells, iPSCs and iPCCs, share characteristic features with each other as well as with embryonic stem cells (ESCs) [33
]. During reprogramming of somatic or cancer cells, ROS are generated by metabolic stress, and increased ROS levels lead to DNA damage, cell senescence, and apoptosis. ROS may hinder the survival of reprogrammed cells, as suggested by observations of increased iPSCs generation during hypoxia [34
]. In addition, oxidative stresses repress the ability to generate or maintain iPSCs and human ESCs (hESCs) [36
], suggesting that ROS generation by reprogramming factors is unfavorable for generating iPSCs.
Here we report that JDP2 indeed associates with the ARE and acts as a newly identified key cofactor of the Nrf2-MafK complex to regulate ARE-mediated gene expression and ROS production. In Jdp2
-Cre mice, Jdp2 was specifically expressed in granule cells in the cerebellum. We reprogrammed medulloblastoma cells [37
] to generate iPSC-like cells using the transduction of lentivirus-encoded JDP2 and OCT4 to characterize the reprogrammed medulloblastoma iPSC-like cells. We also discuss the role of JDP2 in cancer cell reprogramming and ROS homeostasis. Our results provide evidence that JDP2 plays a critical role in the cellular adaptive response to ROS and electrophiles generated by various cellular stimuli and in the generation of cancer progenitor cells with the phenotypes of stem cells and cancer cells.
In this study, we demonstrated the role of JDP2 in oxidative stress and ROS homeostasis, and the possible link between ROS homeostasis and nuclear reprogramming. During reprogramming, Yamanaka factors induced the chromatin changes to generate the stress response including ROS production [50
]. The increasing of ROS inhibited the survival of reprogramming cells which led to DNA damages, cellular senescence and apoptosis [7
]. Thus, the factor which regulates ROS homeostasis may play a critical role for efficient reprogramming [7
]. We showed that ROS production is higher in Jdp2
KO MEFs than in WT MEFs and that addition of an oxidative stress inducing reagent increased ROS production in Jdp2
KO MEFs compared with WT KO MEFs (Figure 1
). Conversely, ARE-luciferase activity increased more in WT MEFs than in Jdp2
KO MEFs (Figure 1
, Figure 2
A). These data indicate that JDP2 is a critical factor for inhibiting ROS production and DNA oxidation. In addition, it was reported that senescence impaired the reprogramming to iPSCs, and reprogramming triggered a stress response of senescence at the initial stage [53
]. In fact, senescence is the irreversible arrest during the G1 transition of the cell cycle elicited by replicative exhaustion or in response to stresses such as DNA damage, drugs, or oncogenes. This arrest is implemented primary through activation of p53 and the up-regulation of the cycling dependent kinase (CDK) inhibitors p16Ink4a
]. Introduction of Yamanaka factors triggers senescence by up-regulating p53, p16Ink4a
, and p21Cip1
at initial stage, and at later stage these expressions were repressed. This reprogramming-induced senescence (RIS) acts as an initial barrier limiting the efficiency of the reprogramming. The reprograming is slower and stochastic, suggesting the existence of barrier limiting its efficiency. In order to increase the efficiency of the reprogramming, the repression of RIS is definitely required for full reprogramming. Overexpression of JDP2 increases the expression of p16Ink4a
, thus generates the replicative senescence through the repression of H3K27me3 by polycomb complex [17
]. However, the reduced expression by shRNA against p16Ink4a
(Supplementary Figure S4
or p53, increased nuclear reprogramming to generate iPSCs at pre-iPC stage. Thus, JDP2 is assumed to be a possible gene to induce this RIS at the initial stage during the commitment to iPSCs.
Another merits of JDP2 as one of the candidates reprogamming factor is that JDP2 connects with WNT signaling. The recent report showed that oxygen is known to regulate stem cells through WNT signaling and HIF1α [55
]. Moreover, the GSK3β-TCF3 axis in WNT signaling controls self-renewal of stem cells [56
]. We previously demonstrated; (I) JDP2 regulates WNT signaling [11
], and (II) the expression of JDP2 is repressed under hypoxia conditions [17
] and JDP2 inhibits the ROS production. Thus, we hypothesize that JDP2 controls nuclear reprogramming with combination of OCT4. The transcription factor OCT4 interacts physically with various active and repressive chromatin complexes; raising the question whether OCT4 or another reprogramming factor is more important for reprogramming [57
]. It has been reported that high levels of OCT4 and low levels of SOX2 increase the efficiency of reprogramming [58
]. We tried the combination of OCT4 and JDP2 to induce reprogramming. In this study, we replaced SOX2, and KLF4 and c-MYC by JDP2 in the alternative set of transcription factors for reprogramming, in addition to the traditional four-factor reprogramming. We studied the characteristics of the reprogrammed medulloblastoma cells because JDP2 was expressed predominantly in the cerebellar granule cells (Supplementary Figure S2
Nanog is essential for the establishment of iPSCs and is expressed before many other pluripotency genes during the reprogramming process, suggesting that it may be required for their activation [60
]. Moreover, the interplay of Nanog and JDP2 was reported in P19 cells [61
]. Overexpression of NANOG was found in our iPSC-like cells (Figure 6
), reflecting the maintenance of the self-renewal and stemness-charactristics in iPSCs-like cells. The appearance of positive stem cell markers such as alkaline phosphatase, OCT4, SOX2, and NANOG in these cells indicates pluripotency. However, our teratoma study showed that only two germ layers formed after subcutaneous injection of either type of transfected DAOY cells into SCID mice, which are similar in DAOY cells. In conclusion, we cannot generate the iPSCs from DAOY meddullobastoma cells, instead we generated the iPCCs which have higher potency to generate the cancer-inducing ability than the original DAOY cells by JDP2 and OCT4. Moreover when we introduced four factors, we failed to generate the complete iPSCs from DAOY cells.
Cancer cells exhibit several characteristics that also appear in ESCs or iPSCs [33
]. Genomic instability, which leads to numerical and structural aberrations, is observed frequently in cancer and cultured pluripotent stem cells. These findings imply that iPSCs or iPSC-derived cells are potentially carcinogenic. Several human gastrointestinal cancer cell lines have been reprogrammed into iPCCs sharing the characteristic features with ESCs by overexpressing the established Yamanaka factors OCT4, KLF4, SOX2, and c-MYC [31
]. Reprogramming of cancer cells in vitro
may reflect a more complex mechanism than the reprogramming of normal cells because of aneuploidy, deregulation of signaling pathways, and the high proliferation rate of cancer cells. Moreover, iPCCs carry the burden of their cancer cell descent, requiring more work for the maintenance of iPCCs because cells differentiated from iPCCs can remain immortal and overgrow in cell culture [33
Our data showed a more aggressive phenotype after 28 days in the DAOY iPCCs inoculated into SCID mice compared with the untransfected DAOY cells they were derived from. The questions arising from this observation are: to what extent does reprogramming alter cancer cell tumorigenicity to obtain the stemness characteristics; how the resistance capacity of ROS and the antioxidation ability of defined factors contribute to the generation of full reprogramming or the stability of genome; and which transcription factors to regulate the ROS homeostasis contribute to the differences in cell characteristics from those in iPSCs reprogrammed from somatic cells; and does reprogramming and ROS regulation in stem cells offer new understanding of cancer cell tumorigenicity or possibilities for cancer treatment?
Our data demonstarted that 2F such as JDP2 and OCT4 generated the cancer stem-like cells with stemness-characteristics. How does JDP2 induce the cancer phenotype during reprogramming? One possible explanation is the mutation of p53
gene. JDP2 regulates the expression of p53
]. We reported that Jdp2
KO MEFs showed the reduced expression of p53 and p21Cip1
]. However, others reported the contradicted results [21
]. Moreover, Aronheim et al.
] demonstrated that the initial expression of JDP2 is critical for tumor suppressor function, after which it potentiates hepatocellular carcinoma with higher mortality and increased number and size of tumors in mice. However, we do not know whether the reprogramming could shut down the expression of this mutant p53 in DAOY-iPSCs or maintain its mutation of p53
gene, even though the epigenomes are resetted. During the reprogramming, the new mutation of p53 might also be arisen in the genome of DAOY iPSCs. Assuming p53 remains to be mutated in iPSCs, it could be easily produce the cancer when it meets the second hit. We need further study to clarify the mechanism how the DAOY-iPSC-like cells generate tumors.
Miyoshi et al
. reported a lower tumorigenic potential and increased the sensitivity to 5-fluorodeoxyuridine compared with the untransfected cancer cell line after subcutaneous transplantation of post-iPCCs into the dorsal flank of immunodeficient mice [32
]. Similarly, Carette et al
. found that reprogrammed iPCCs from chronic myeloid leukemia cells developed resistance to treatment [31
]. Decreased susceptibility might explain why some chemotherapeutics or targeted therapies cannot cure certain types of cancers fully, assuming that putative cancer stem cells or defined populations of cells within the tumor exhibit a less differentiated state. Thus, the further study is required for complete understanding of the pluripotency, self-renewal, gene expression, epigenetic changes, mutation of tumor suppressor genes, and immune surveillance among iPSCs, ESCs and CSCs.
Hopefully, this approach will help to identify new tumorigenesis-related genes or epigenetic changes that can be explored for the development of new anticancer drugs, which targets on the control of ROS generation. Because iPCCs can differentiate into cancer cells or may acquire more carcinogenic characteristics after reprogramming, development of suitable model for studying tumorigenesis in vitro is necessary. Reprogramming of cancer cells and the regulation of ROS are an exciting new approach for basic and therapeutic research and might offer new possibilities in future.