DNA-dependent protein kinase catalytic subunit (DNA-PKcs), encoded by PRKDC gene, is a member of the phosphatidylinositol 3-kinase-related kinase (PIKK) family, which also includes the two DNA damage repair proteins, ataxia telangiectasia mutated (ATM) and ATM and Rad3-related protein (ATR). The N-terminal region of DNA-PKcs is composed of HEAT (Huntington-elongation-A-subunit-TOR) repeats that likely serve protein-protein interaction and the C-terminal region containing the PI3 kinase domain, which is flanked N-terminally by the FAT (FRAP, ATM, TRRAP) domain and C terminally by the FATC (FAT C-terminal) domain (Davis et al., 2014).

DNA-PKcs has been extensively investigated for its role in repairing DNA double-strand breaks (DSB) in non-homologous end joining (NHEJ) pathway (Fig. 1). In cellular response to DNA DSB such as those produced by ionizing radiation, Ku heterodimer recognizes and localizes at the DNA damage sites, followed by the recruitment of DNA-PKcs (Yue et al., 2020). Once recruited, DNA-PKcs is activated through phosphorylation in a DNA-dependent manner (Yue et al., 2020). DNA-PK then forms a tight complex with Artemis and phosphorylates its C-terminal inhibitory region to stimulate the nuclease activity (Ma et al., 2002), which facilitates processing the incompatible mismatching overhangs of DNA DSB broken ends (Yue et al., 2020). Despite its clear role of DNA-PKcs in NHEJ, evidence suggests that DNA-PKcs is also involved in regulating homologous recombination (HR). For example Xie et al. (2020) have reported that increased DNA-PKcs activity can suppress HR repair in G1 phase of cells through increasing RBX1 protein expression, which prompts the neddylation and activity of cullin 1, a key component of the Skp1-Cullin1-F-box ubiquitin 3 ligase, consequently mediating ubiquitination and degradation of EXO1. This in turn limits the end resection of DNA DSB that is critical to initiate the HR (Xie et al., 2020). While NHEJ does not depend on the presence of sister chromatid in mediating DNA DSB repair thereby not being restricted to a particular phase of the cell cycle, experiments have demonstrated that NHEJ occurs much faster than HR (30 min for NHEJ compared to more than 7 h for HR) and that NHEJ accounts for approximately 75% of repair events (Zhang and Matlashewski, 2019).

DNA-PKcs is essential in the development of T and B lymphocytes by mediating V(D)J recombination (Fig. 1). V(D)J recombination is a somatic rearrangement of variable (V), diversity (D), and joining (J) segments of the genes coding immunoglobulins or T cell receptors thereby generating diversity of antigen receptors (Roth, 2014). Specifically, lymphocyte-specific endonucleases, namely recombination activating genes, RAG1 and RAG2 generate DNA DSB intermediates, in which hairpin coding ends should be joined via the end-ligation machinery of NHEJ whereby DNA-PKcs and Artemis open the hairpin structures (Menolfi and Zha, 2020). Immunoglobulin H chain genes during the B cell development further undergo class switch recombination (CSR), another genomic rearrangement to produce different Ig classes (e.g., IgG, IgE, or IgA) (Bosma et al., 2002). Although a number of studies have previously demonstrated defective CSR in B cells from mice lacking DNA-PKcs (Manis et al., 2002; Rolink et al., 1996). Bosma et al. (2002) have reported that DNA-PKcs activity is not required for CSR by using site-directed H and L chain transgenes. They reasoned that the previously reported studies are difficult to reconcile because B cells lacking DNA-PKcs are able to class switch to IgG1 normally but not IgG2a, IgG2b, or IgG3 while γ switch regions would share structural similarity hence common CSR mechanism for all isotypes (Bosma et al., 2002). In their study, they have demonstrated that DNA-PKcs-deficient severely combined immunodeficiency (SCID) mice B cells are not only able to class switch from IgM to IgG1, but also to IgG2, IgG2b, IgG3, IgA, and IgE (Bosma et al., 2002).

In innate immune cells, DNA-PKcs has been shown to regulate inflammatory defensive response (Fig. 1). For instance, bone marrow-derived dendritic cells deficient for DNA-PKcs exhibit delayed response to oligodeoxynucleotides (ODN) containing CpG motif (CpG-ODN) producing interleukin-6 (IL-6) and IL-16 (Ma et al., 2013). Macrophages with DNA-PKcs deficiency have also shown to exhibit defects in triggering inflammatory response. IL-10 (Yotsumoto et al., 2008) and IL-18 production (Morales et al., 2017) is reduced in thioglycollate-stimulated peritoneal macrophages and bone marrow-derived macrophages (BMDM) of SCID mice. Sun et al. (2020) have demonstrated that DNA-PK phosphorylates cGAS, suppressing its enzymatic activity and expression levels of IFNB1 and CXCL-10 in THP-1 monocytic cells, thereby making these cells more susceptible to viral infection. DNA-PKcs in BMDM has also been reported to activate Akt thereby nuclear translocation (Dragoi et al., 2005), which may be necessary to activate Akt-specific targets including inhibitor of nuclear factor-κB kinase (IKK) and NF-κB.

DNA-PKcs was originally discovered and characterized as part of Sp1 transcriptional complexes (Jackson et al., 1990) and as a regulatory component of transcriptionally poised RNA polymerase II (Dvir et al., 1992). Previous studies have suggested that DNA-PKcs is recruited to active transcription sites (Ju et al., 2006) and modulate the function of various transcription factors including AIRE (autoimmune regulator), p53, Ets (erythroblast transformation-specific)-related gene as well as nuclear receptors such as glucocorticoid, progesterone, estrogen, and androgen receptors in cancer cells (Goodwin and Knudsen, 2014). In breast cancer cells, DNA-PKcs phosphorylates serine118 residue of the estrogen receptor-α (ERα), which promotes cancer cell proliferation, to prevent ubiquitinylation and degradation of the ERα (Medunjanin et al., 2010). Under hypoxic condition, DNA-PKcs has been shown to form a hetero-trimer with hypoxia-inducible factor-1α (HIF-1α) and tripartite motif-containing 28 (TRIM28) by phosphorylating serine 824 of TRIM28 thereby binding to hypoxia-responsive element of HIF target genes (Yang et al., 2022).

DNA-PKcs has also been shown to regulate transcription activation via phosphorylation of transcription factors. For instance, early growth response 1 (EGR1) (Waldrip et al., 2021) and nuclear factor of activated T-cells (NFAT) (Kim Wiese et al., 2017) have been shown to be phosphorylated hence being translocated to the nucleus influencing interleukin-2 production in activated human Jurkat T cells. Interestingly, upstream stimulatory factor (USF) 1/2 heterodimer has been shown to be phosphorylated by DNA-PKcs resulting in fatty acid synthase (FASN) promoter activation (Wong et al., 2009). Consistent with this, Dylgjeri et al. (2022) have also found FASN as one of the novel DNA-PK partners, indicating an essential role of DNA-PKcs in regulating the cellular metabolism.

Despite the role of DNA-PKcs in the nucleus, recent evidence suggests that approximately 20% of the total DNA-PKcs protein levels have been detected in the cytoplasm of castration-resistant prostate cancer cells (Dylgjeri et al., 2022). Furthermore, this study has demonstrated that DNA-PKcs promotes Warburg effect through phosphorylation of various enzymes involved in the glycolytic pathway including glyceraldehyde 3 phosphate dehydrogenase (GAPDH), phosphoglycerate kinase 1 (PGK1), enolase 1, and pyruvate kinase M2 (PKM2), hence suggesting that inhibition of DNA-PK in conjunction with aerobic glycolysis would increase antitumor efficacy (Dylgjeri et al., 2022) (Fig. 1). Given that HIF-1 is a well-known master regulator of glycolysis, cancer cells may utilize DNA-PKcs to increase glycolysis in the situation where HIF-1 activation is limited (e.g., in well-oxygenated conditions).

In normal tissues such as muscle, DNA-PKcs has been shown to be responsible for lowering mitochondrial biogenesis through phosphorylation of heat shock protein 90, which then dampens the activation of AMP-activated protein kinase (AMPK) (Park et al., 2017). Interestingly Park et al. (2017) have shown that phosphorylated DNA-PKcs levels increase in aged mice and monkeys and that inhibition of DNA-PKcs by genetic or pharmacological approaches can improve AMPK activation, mitochondrial functions, and physical fitness. Although this is a very elegantly performed study, it is not yet clear whether these effects are derived from (1) cytoplasm versus nuclear role of DNA-PKcs; (2) any tissue-specific role of DNA-PKcs; and (3) other DNA damage-sensing kinases such as ATM and ATR. Nonetheless, the fact that DNA-PK inhibitor, NU7441 can significantly improve the physical ability on exercise and protects from diet-induced obesity through the whole body glucose metabolism (Park et al., 2017) suggests novel roles of DNA-PK in metabolic homeostasis.

The development of DNA-PKcs inhibitors has been challenging due to the close similarity to PI3K and PI3K-related family of proteins including ATM and ATR. NU7441 is a specific DNA-PKcs inhibitor with minimal activity against PI3K and mTOR (Hardcastle et al., 2005) and has been shown to potentiate the antitumor activities of DNA damage-inducing chemotherapy or radiotherapy in many different types of cancers (Mohiuddin and Kang, 2019). Despite the favorable antitumor activities and pharmacokinetics, the poor water solubility of NU7441 has prevented its clinical use (Zhao et al., 2006). M3814 (peposertib), an orally administered DNA-PKcs inhibitor, has also shown to potentiate the effect of ionizing radiation (Zenke et al., 2020) or topoisomerase II inhibitors in multiple cancer models (Wise et al., 2019), and is in phase I/IIa clinical trials (NCT02516813). Recently developed AZD7648 has also been reported the selectivity towards DNA-PKcs over other kinases including PI3K isoforms, mTOR, ATM, and ATR (Fok et al., 2019). Interestingly, a recent study has demonstrated that AZD7648 when combined with radiation results in activation of type I interferon signaling in cancer cells, which in turn activates CD8α-positive T cells for antitumor immunity (Nakamura et al., 2021). Currently, AZD7648 is in phase I/IIa clinical trials recruiting patients (NCT03907969).

While this is all exciting, one should bear in mind that DNA-PKcs inhibitors would also inhibit DNA-PKcs in normal cells. When combined with radiation or DNA damage-inducing chemotherapy, this would lead to more serious consequence: all those normal cells will also be very sensitive to DNA-PKcs inhibition to the same extent as tumors (Brown, 2019). Brown (2019) has stated that this may explain some of the enhanced normal tissues reactions including dysphagia, prolonged mucosal inflammation/stomatitis, and radiation skin injury in phase I/IIa clinical trial of M3814 combined with radiotherapy (Van Triest et al., 2018). In this regard, selective activation of DNA-PKcs would be an attractive strategy. SN38023 has been reported as a potential bioreductive prodrug, which can be activated by one electron reductase under hypoxic conditions releasing a DNA-PKcs inhibitor, IC87361 (Wong et al., 2019) (Fig. 2). Although this compound demonstrated efficient radiosensitization under hypoxic conditions in vitro, in vivo antitumor efficacy when combined with radiation was comparable to the radiation alone (Wong et al., 2019), suggesting prodrug design needs to be improved by incorporating more efficient reductive linker and more potent DNA-PK inhibitor.

While we have long been appreciating the role of DNA-PKcs in DNA DSB repair mechanism, recent evidence highlighting a significant amount of DNA-PKcs being present in the cytoplasm (Dylgjeri et al., 2022) suggests the novel role of DNA-PKcs in cells. By performing DNA-PK protein interactome, Dylgjeri et al. (2022) have reported novel interacting partners of DNA-PK, including those involved in ribonucleic acid metabolism (DOT1L [histone methyltransferase], RPLs [L ribosomal proteins], HNRPs [heterogeneous nuclear ribonucleoprotein Ks], eEFs [eukaryotic elongation factors]), energy metabolism (PGK1, PKM2, FASN, VDAC-2 [voltage-dependent anion channel-2]), heat shock proteins (HSPs), zinc finger proteins (ZNFs), and inflammation (MIF [macrophage migration inhibitory factor]). This suggests that inhibition of DNA-PK is likely to inhibit multi-substrates at multiple locations (nucleus and cytoplasm), which may yield a complex outcome.

The availability of SCID mice has been widely used to study the functions of DNA-PKcs in various types of cells. However, cells derived from SCID mice may not reflect the results of DNA-PKcs deficiency (knockout or kinase mutation). SCID mice have the homozygous nonsense mutation at amino acid position 4046 resulting in a truncated protein lacking the extreme C-terminal region leaving the PI3K lipid kinases remain intact (Beamish et al., 2000). Later-developed DNA-PKcs (Prkdc) knockout mice have demonstrated a similar phenotype (the signal join-proficient, coding join-deficient phenotype) to SCID mice (Taccioli et al., 1998). While Dragoi et al. (2005) have observed similar results of defective Akt activation between BMDM from SCID and those from DNA-PKcs knockout mice, Menolfi and Zha (2020) have observed drastically different results—SCID and DNA-PKcs knockout mice are viable and fertile, DNA-PKcs kinase-dead mice have been shown to be embryonically lethal, which can only be rescued by KU deletion.

Genetic alterations of PRKDC such as point mutations and copy number amplifications have been shown to be common in a variety of cancer types including stomach adenocarcinoma, hepatocellular carcinoma, lung adenocarcinoma, and colorectal adenocarcinoma (Yang et al., 2020). A recent study has revealed that PRKDC mutations are strongly associated with the increased immune cell (CD8+ T cells and NK cells) infiltration, tumor mutational burden, and better survival in patients after immune checkpoint inhibitors (Chen et al., 2020). This is very interesting because it is not currently known how DNA-PK can increase the efficacy of anticancer immunotherapy: either through canonical functions regulating DNA DSB repair NHEJ pathway or through non-canonical functions such as regulation of cellular metabolism and/or inflammation. It will be also interesting to examine how and why PRKDC mutations differentially regulate the tumor microenvironment. While there are a number of DNA-PK inhibitors currently being evaluated to increase the efficacy of radiotherapy and/or chemotherapy, we should still carefully understand the role of DNA PKcs at various subcellular localization, multiple substrate specificity, and many different types of cells in order to maximize the clinical benefit in cancer patients.

This work was supported by the New Faculty Startup Fund from Seoul National University, the Research Institute for Veterinary Science, the Ministry of Science and Information, Communication, and Technology, Korea (RS-2022-00144495 and NRF-2020M2D9A2092374) and BK21 Four Future Veterinary Medicine Leading Education and Research Center, the Ministry of Education, Korea.

Y.R.L., G.S.K., T.O., H.J.P., and G.O.A. wrote the manuscript. H.J.J. drew the original figures.

The authors have no potential conflicts of interest to disclose.