The findings that influenced the direction of HIPK2 research most profoundly for the next decade were published in 2002, when independent groups reported that HIPK2 is a DNA damage-responsive kinase that is partially co-recruited with the tumor suppressor p53 into PML bodies upon cytotoxic stress, most prominently by UV irradiation (D’Orazi et al. 2002; Hofmann et al. 2002). HIPK2 was shown to phosphorylate p53 on Serine 46, thereby enhancing the proapoptotic activity of the transcription factor p53 and programmed cell death. Moreover, it was shown in 2003 that HIPK2 is also able to promote p53-independent cell death by targeting the corepressor C-terminal binding protein (CtBP) for degradation, thus increasing the transcription of proapoptotic CtBP target genes. Since then, the important regulatory role of HIPK2 in determining cell fates upon genotoxic stress, but also in various developmental processes, has been elucidated in great detail.
The HIPK2 Protein
HIPK2 is a 130 kDa protein of 1191 amino acids in humans that is well conserved from flies to humans. It harbors an N-terminal kinase domain that was published to be very homologous to DYRK family kinases. It was shown that HIPK2 most probably autophosphorylates at multiple Serine/Threonine residues. Mutation of the conserved residue K221 within the ATP-binding site of the kinase domain results in an inactive kinase. HIPK2 preferentially targets Serine or Threonine residues that are followed by a Proline residue (SP/TP sites) in its substrates. Correlating with its observed localization in nuclear substructures, it contains several C-terminal NLS sequences and a speckle retention signal. HIPK2 was also shown to be covalently SUMOylated at several sites, the principal site being K25, and to have two adjacent SUMO interaction motifs (SIMs) within the C-terminus that influence its ability to be recruited to promyelocytic leukemia (PML) nuclear bodies (de la Vega et al. 2010 and references cited therein). Recently, it has been shown that HIPK2 is an unstable protein in unstressed cells which is kept low in cells via proteasomal degradation, and the region that is responsible for this degradation has also been mapped to the C-terminus (Calzado et al. 2009a, b; Sombroek and Hofmann 2009; Winter et al. 2008). Moreover, it has been speculated that the HIPK2 C-terminus may form an auto-inhibitory loop that might block HIPK2 activity in the absence of cofactors like Han11 (Ritterhoff et al. 2010) or Axin which was reported to bind to the HIPK2 C-terminus and to enhance the kinase activity of HIPK2 towards p53.
HIPK2 and the DNA Damage Response
The region within HIPK2 that is required for destabilization of HIPK2 by these proteins is located in its C-terminus, between amino acids 600 and 800 for Siah-1 and starting from amino acid 838 for MDM2. Upon DNA damage inflicted by various genotoxic stimuli like UV and ionizing irradiation or treatment with adriamycin or cisplatin, however, HIPK2 degradation is blocked, HIPK2 protein accumulates and phosphorylates its target genes. The process by which HIPK2 degradation is inhibited is best characterized for Siah-1, which is phosphorylated by the damage-activated checkpoint kinases ATM and/or ATR at Ser19, leading to the loosening of the HIPK2/Siah-1 interaction and therefore to HIPK2 stabilization. When this occurs in p53-proficient cells, HIPK2 phosphorylates p53 on Serine residue 46 (or 58 in mice) in a PML- and Sp100-dependent manner. Ser46 phosphorylation very likely causes enhanced acetylation of p53 on K382 by the HAT protein CREB-binding protein (CBP), another HIPK2 interactor, and these modifications lead to an altered promoter specificity of the p53 transcription factor and to the increased transcription of proapoptotic p53 target genes such as Puma and Noxa. Alternatively, HIPK2 can lead to derepression of proapoptotic genes by phosphorylating the transcriptional repressor CtBP, inducing its degradation via the proteasomal pathway. This results in upregulation of CtBP targets such as Bax, Noxa, and PTEN which then drive apoptosis in the damaged cells (for reviews, see Bitomsky and Hofmann 2009 and Puca et al. 2010). HIPK2-mediated apoptosis has also been linked to the presence of the transcriptional repressor methyl-CpG-binding protein 2 (MeCP2) (Bracaglia et al. 2009), but the underlying mechanism has not been determined. Accordingly, it has been observed that HIPK2 can act as haploinsufficient tumor suppressor in a two-stage model of skin carcinogenesis in mice. Involvement of HIPK2 in tumor suppression, however, will be discussed in more detail in a separate chapter. Likewise, the involvement of HIPK2-mediated apoptosis in normal development will be discussed later. One interesting aspect of HIPK2 regulation upon genotoxic stress is the transient stabilization that occurs after sublethal damage but does not result in Ser46 phosphorylation. The transient nature of this stabilization is probably brought about by negative feedback loops involving the Siah-1 and MDM2 ubiquitin ligases, both of which are p53 target genes which accumulate upon DNA damage. However, the functions of the transiently increased HIPK2 levels after mild stress have not been elucidated so far. Another interesting aspect of HIPK2 regulation is the question whether HIPK2 is activated in addition to its accumulation, and if so, which modification and/or interactions might be required for that. One described mechanism activating HIPK2 after damage in addition to stabilization is caspase cleavage. It was shown that HIPK2 can be cleaved after D977 and D916 in a process involving caspases 3 and 6, and that the cleaved products are more active with respect to p53 phosphorylation than corresponding uncleavable HIPK2 constructs (reviewed in Sombroek et al. 2009). Nevertheless, the issue of HIPK2 activation by other means remains subject to intense investigation.
HIPK2 in Development
Apoptosis is an important process shaping many tissues and organs. For example, in brain development, neurons are produced in very large numbers and the excess is then eliminated by apoptosis. Indeed, it was shown that in sensory and sympathetic neurons for instance in trigeminal ganglia, HIPK2 overexpression induces massive apoptosis and thus seems to be important for the removal of excess neurons by programmed cell death, likely by regulating the pro-survival factor Brn3a. Somewhat paradoxically, another study suggested that HIPK2 knockout actually increases apoptosis in trigeminal neurons, which would suggest an as-yet-unrecognized pro-survival function of HIPK2 (reviewed in Rinaldo et al. 2007). The phenotypes of knockout mice generated independently in different laboratories (Isono et al. 2006; Trapasso et al. 2009; Hattangadi et al. 2010; Inoue et al. 2010) suggest that there is some degree of redundancy between HIPK2 and HIPK1. Analysis of double HIPK1/2 knockout mice (Isono et al. 2006) revealed disturbed hematopoiesis, vasculogenesis, and angiogenesis in one study, and embryonic death from embryonic day 9 onward and exencephaly, eye phenotypes, and fusion of dorsal root ganglia in another study. Moreover, the latter also revealed developmental defects like the formation of ectopic ribs, indicative of disturbed Hox gene regulation, and an influence of HIPKs on the Sonic hedgehog (Shh) pathway via Pax1 and -3. MEFs from double knockout were also reported to be more resistant to UV-induced apoptosis. HIPK2 single knockouts from independent groups showed a “clasping of hindlimbs” phenotype when suspended by the tail, albeit to greatly varying percentages. One study also reported shuffling gait and other motoric abnormalities resembling a Parkinson’s-like disease, and significantly enhanced apoptosis in neurons at E17.5 and P0. Another independent study (Trapasso et al. 2009) reported increased stem and progenitor cell compartments in mouse skin when HIPK2 was knocked out, suggesting that HIPK2 may restrict the proliferation of the stem and progenitor cells in the skin, maybe by interfering with the Wnt signaling pathway. Interestingly, the study reporting only 15% of “clasped hindlimbs” in their HIPK2 knockout observed a growth reduction of HIPK2-deficient animals in comparison to their littermates that was already observable at birth and persisted to adulthood, again indicating a role for HIPK2 in cell proliferation rather than cell death. However, the underlying mechanisms for a putative growth-stimulating role of HIPK2 have not been identified so far.
In Drosophila, knockout of the only existing HIPK gene results in defective eye formation and frequent death at the pupal or embryonal stages. Drosophila HIPK phosphorylates Groucho at multiple residues, resulting in upregulation of the Notch pathway (Lee et al. 2009). This connection has not (yet) been described for vertebrates. In addition, it could be shown that HIPK is required for collective programmed cell death of epithelial cells in the developing wing, similar to previously determined apoptotic factors like Dronc or Dark (Link et al. 2007).
It has also been reported that HIPK2 can interact with the corepressor c-ski and with SMAD1 and thereby counteract bone morphogenetic protein (BMP) signaling, which is implicated in embryo polarity, heart and CNS development as well as bone formation.
HIPK2 and HIPK1 depletion also impacts very strongly on fetal liver hematopoiesis (Hattangadi et al. 2010), preventing upregulation of many erythropoietic and heme-biosynthesis-associated genes, but the exact target(s) of the HIPK2 in this system has not been identified to date.
HIPK2 and Cancer
Given the observations that HIPK2 can induce apoptosis in both p53-proficient and p53-deficient cells upon genotoxic stress, it is highly likely that HIPK2 can act as a tumor suppressor in many tissues (Krieghoff-Henning and Hofmann 2008; Sombroek and Hofmann 2009 and references cited therein). This question has been addressed directly in mice in a skin carcinogenesis model, where deletion of only one HIPK2 allele already renders the mice more sensitive to skin cancer formation, with an even stronger phenotype in full knockouts. Thus, HIPK2 acts as haploinsufficient tumor suppressor in mouse skin. Moreover, there are a number of human tumors, such as thyroid and breast carcinoma, that show a high incidence of HIPK2 downregulation on the mRNA level. In one case each of myelodysplastic syndrome and AML, point mutations in the HIPK2 sequence were found: R868W and N958I, and these point mutants were found to localize aberrantly, affecting their ability to transactivate p53 target genes. Moreover, there is evidence for a tumor-promoting effect of the HIPK2 interactor HMGA1, which is highly overexpressed in many breast tumors and seems to recruit HIPK2 to the cytoplasm, thereby preventing its association with PML-NBs and hence p53 Ser46 phosphorylation. Interestingly, HIPK2 protein is degraded in a possibly HIF1- and Siah-2-dependent manner under hypoxic conditions, which occur very frequently in larger tumors. HIPK2 also co-localizes with the tumor suppressor PML in PML nuclear bodies, and while PML may actually stabilize HIPK2, HIPK2 was shown to phosphorylate PML IV at several residues, increasing PML SUMOylation and its ability to induce apoptosis. Finally, injection of colorectal cancer RKO cells into mice causes tumors that grow much faster if the cells are HIPK2-depleted. Moreover, depletion of HIPK2 in RKO cells leads to integrin subunit ß4 upregulation (Bon et al. 2009), which is strongly associated with increased migration and metastatic potential. In breast cancer, HIPK2 nuclear positivity was reported to be inversely correlated with ß4 expression. These observations all suggest that in many tumors HIPK2 is indeed inactivated, strengthening the hypothesis that HIPK2 is a tumor suppressor protein. Reports on pilocytic astrocytoma, however, indicate that HIPK2 may also be upregulated by amplification in a subset of human tumors, although the driving factor of cancer formation in these samples may be BRAF which is located in the same region (Cin et al. 2011). In cervical cancer, HIPK2 nuclear positivity was positively correlated with tumor stage and tumor dedifferentiation, suggesting that HIPK2 may also have a growth-promoting role. However, the molecular basis of this potential role has not been determined so far.
HIPK2, Aging and Metabolism
HIPK2 may not be regulated solely by DNA damage, but also by the metabolic state of the cell. Vice versa, HIPK2 activity also appears to contribute to metabolic regulation in the presence or absence of genotoxic stress. Interestingly, it is becoming more and more obvious that metabolism, genotoxic stress, and cellular aging in particular are closely interlinked. Caloric restriction is probably extending the organismal lifespan, as shown in C. elegans and mice, most likely by limiting the amount of DNA-damaging reactive oxygen species (ROS).
As mentioned previously, HIPK2 plays an important regulatory role in pancreas development and function. One important HIPK2 target in this context is the Insulin Promoter Factor/Pancreatic Duodenal Homeobox 1 (IPF-1/PDX-1)-dependent upregulation of the insulin gene, glut2, glucokinase, etc., in the developing pancreatic epithelium and in ß-cells (Boucher et al. 2009). HIPK2 seems to regulate the nucleocytoplasmic distribution of IPF-1 (An et al. 2010). HIPK2 can also regulate the proglucagon, somatostatin, and insulin promoters via Pax6, which also impacts on pancreas development and function. HIPK2 may also play a role in ROS metabolism, by virtue of its ability to suppress transcription of Nox1, encoding the catalytic subunit of a NADPH oxidase which can generate ROS (Puca et al. 2010).
Conversely, it was shown that HIPK2 is upregulated in aging neurons, which can be counteracted by caloric restriction, arguing that the presence of ROS induces HIPK2, and placing HIPK2 in the aging-associated insulin-like growth factor (IGF1)-R signaling cascade (Li et al. 2009).
A very general way in which HIPK2 regulates metabolic and other processes within the cell is its ability to recruit and/or interact with the HAT enzyme p300 and its paralog CBP, which are important co-transcription factors for a variety of transcription factors, including the CREB protein, which itself was shown to be a HIPK2 target (Sakamoto et al. 2010), and which has also been linked to lifespan extension in C. elegans.
Although discovered relatively recently, HIPK2 has already been shown to regulate several important processes within eukaryotic cells. The best studied function of HIPK2 is to induce programmed cell death in response to genotoxic stress, either by p53 phosphorylation at Serine residue 46 or by degrading the antiapoptotic transcriptional repressor CtBP. Furthermore, HIPK2 has been shown to act as transcriptional regulator in various transcription complexes, most prominently those formed by homeobox transcription factors. Especially with respect to the developmental functions of HIPK2, a certain degree of redundancy with the related kinase HIPK1 seems to occur, so that developmental defects are in some cases only clearly detectable when both proteins are depleted. Single and double knockout animals suggest roles for HIPK2 in brain development, hematopoiesis, pancreas development, and body patterning, in part by regulating apoptosis, but also by modulating the activity of several important signaling pathways such as canonical Wnt, TGFß, or BMP. In part, the detailed consequences of the described interactions and phosphomodifications of the HIPK2 targets identified so far still remain to be elucidated on the organismal level.
Moreover, the prominent role of HIPK2 in the DNA damage response and first hints obtained by analyzing human tumors as well as a two-step carcinogenesis model in mice suggest that HIPK2 may act as a tumor suppressor in many tissues, and indicate that cancer patients might benefit from specific activation of HIPK2 in tumors.
- Isono K, Nemoto K, Li Y, Takada Y, Suzuki R, Katsuki M, et al. Overlapping roles for homeodomain-interacting protein kinases hipk1 and hipk2 in the mediation of cell growth in response to morphogenetic and genotoxic signals. Mol Cell Biol. 2006;26(7):2758–71.PubMedPubMedCentralCrossRefGoogle Scholar
- Sakamoto K, Huang BW, Iwasaki K, Hailemariam K, Ninomiya-Tsuji J, Tsuji Y. Regulation of genotoxic stress response by homeodomain-interacting protein kinase 2 through phosphorylation of cyclic AMP response element-binding protein at serine 271. Mol Biol Cell. 2010;21(16):2966–74.PubMedPubMedCentralCrossRefGoogle Scholar