Role of AKT signaling in DNA repair and clinical response to cancer therapy

Abstract

Effective cancer treatment has been limited by the emergence of resistant cancer cells. The results of many studies indicate that AKT activation plays an important role in the acquisition of resistance to anticancer therapy. AKT is a critical effector serine/threonine kinase in the receptor tyrosine kinase/phosphatase and tensin homolog/phospho-inositide 3-kinase pathway and controls a myriad of cellular functions. Activation of AKT not only supports tumor growth and progression but also contributes to tumor-cell evasion of the cytotoxic effects of cancer therapy through many avenues including the promotion of anti-apoptosis, proliferation, and migration and regulation of the cell cycle. Accumulating evidence has implicated AKT as a direct participant in the DNA damage response and repair induced by commonly used genotoxic agents. In this review, we discuss the molecular mechanisms by which genotoxic agents activate AKT and therefore contribute to resistance to cancer therapeutics, with particular emphasis on DNA repair.

Overview

AKT, also known as protein kinase B (PKB), belongs to the cAMP-dependent, cGMP-dependent, protein kinase C kinase family. AKT was originally identified in the transforming murine leukemia virus, AKT-8 provirus, in 1977 and is classified as an oncogene.¹ In humans, the AKT family has 3 evolutionarily conserved isoforms: AKT1 (PKBα) (including 3 splice variants), AKT2 (PKBβ), and AKT3 (PKBγ) (including 2 splice variants).² Although all 3 AKT isoforms have highly homologous sequences and structures, many findings have suggested that they possess redundant but unique functions (see³ for a review). AKT isoforms play key roles in a wide variety of cellular processes including anti-apoptosis, growth, proliferation, polarity, migration, DNA repair, glucose transport, metabolism, skeletal muscle and cardiomyocyte contractility, angiogenesis, and stem cell self-renewal.⁴ Irregular AKT activity is associated with cancer, cardiovascular disease, type 2 diabetes, muscle hypotrophy, and neurodegenerative disease (review in^5,6).

Growth factors and cytokines bind to the transmembrane receptor and stimulate the activity of lipid enzyme phosphatidyl-inositol 3-kinase (PI3K) family members, which phosphorylate phosphatidyl-inositol di-phosphate (PIP2) to generate PIP3 at the plasma membrane. PIP3 constitutes the binding sites for proteins that contain a pleckstrin homology (PH) domain, such as AKT and PDK1, recruiting them to the membrane. PDK1 phosphorylates AKT isoforms at a Thr residue in the catalytic domain (Thr308 in PKBα, Thr309 in PKBβ, and Thr305 in PKBγ), which results in most of the activity (see⁷ for a review^.). The mechanistic target of rapamycin complex 2 (mTORC2) phosphorylates the Ser residues at the C-terminal regulatory domain (Ser473 in PKBα, Ser474 in PKBβ, and Ser472 in PKBγ),⁸ which provides additional 10-fold increase in activation of AKT. The Ser residues can also be phosphorylated by other kinases such as integrin-linked kinase and DNA-dependent protein kinase (DNA-PK).⁹ Upon activation, AKT isoforms dissociate from the membrane and translocate to various subcellular compartments such as the mitochondria, Golgi, endoplasmic reticulum, and nucleus, where they phosphorylate a plethora of substrates or interact with other cell components. Phosphatase and tensin homolog (PTEN) and SH2-containing inositol phosphatase 2 dephosphorylate PIP3, preventing AKTs from plasma membrane translocation and activation (see¹⁰ for a review). Phosphorylation of AKT at the Thr residues is targeted by protein phosphatase 2A (PP2A),¹¹ whereas Ser residues are dephosphorylated by the pleckstrin homology domain leucine-rich repeat protein phosphatase (PHLPP) family, which includes 3 isoforms: the alternatively spliced PHLPP1α and PHLPP1β, and PHLPP2. These PHLPP isoforms target different AKT isoforms: PHLPP1 targets AKT2 and AKT3, and PHLPP2 dephosphorylates AKT1 and AKT3.¹²

AKT is among the most activated oncoproteins in human cancer.¹³ The AKT pathway is particularly relevant to glioma, as nearly 90% of glioblastomas (GBMs) harbor activation of this pathway.¹⁴ AKT activity can be deregulated through multiple mechanisms (reviewed in¹⁵). The predominant mechanisms in glioma are oncogenic mutations or overexpression of growth factor receptors, mutational inactivation of PTEN, and mutational activation of the catalytic p110a subunit of PI3K.¹⁶ Epidermal growth factor receptor (EGFR) overexpression is found in approximately 60% of primary GBMs and 10% of secondary GBMs. PTEN mutations are found in about 25% of primary GBMs and 5% of secondary GBMs.¹⁷ Amplifications of the genes encoding AKT isoforms are less common but have been observed in a subset of human cancers.^18–20 Mutations of the genes that encode AKT isoforms are relatively rare; however, a transforming E17K PH domain mutation of AKT1 and AKT3 that increases the affinity for PIP3 has been identified.²¹

AKT has also been found to be activated in response to various anticancer therapies and is associated with poor prognosis and treatment resistance (reviewed in²²). In this review, we summarize the recent advances in understanding the roles of AKT in treatment resistance, with a particular emphasis on the DNA damage response and its effects on resistance to genotoxic cancer treatment.

Involvement of AKT in Resistance Mechanisms

Targeted Therapy

Increased AKT pathway signaling can lead to resistance to targeted therapies including mTOR-targeted agents, monoclonal antibodies against receptor tyrosine kinases, tyrosine kinase inhibitors, and hormone therapies. So far, targeted therapies have not been as effective in GBM as expected, including agents targeting EGFR, platelet-derived growth factor receptor, and mTOR.²³ First-generation mTOR-targeted agents in GBM have been shown to induce AKT activation through negative feedback loops including hyperactivation of mTORC2 and PI3K. Due to these limitations, more specific inhibitors and dual inhibitors, including dual mTOR/PI3K inhibitor and dual mTORC1/mTORC2 inhibitor, have been evaluated and have shown efficacy in GBM.²⁴

Multiple mechanisms are involved in acquired clinical resistance to EGFR inhibitors such as loss of the oncogenic mutant EGFRv3 residing on extrachromosomal elements,²⁵ loss of the activating EGFR-mutant gene, emergence of the EGFR T790M resistance mutation, amplification of MET, and hepatocyte growth factor-induced activation of MET.²⁶ Many of these mechanisms lead to reactivation of the AKT pathway. Inhibitors of the AKT pathway in patients with non-small cell lung cancer and acquired resistance to gefitinib or erlotinib showed promising results in a small prospective clinical study.²⁷ In HER2-positive breast cancer, activation of the AKT pathway was associated with significantly shorter overall survival and progression-free survival durations and was a significant independent risk factor for disease progression in patients undergoing trastuzumab and lapatinib therapy.²⁸ However, unlike HER2-positive resistant breast cancers, in which inhibition of PI3K/mTOR is sufficient to reverse resistance, the combined inhibition of PI3K/mTOR and MEK is required in EGFR-driven resistant lung cancers.²⁹ In hormone receptor-positive breast cancer, activation of the PI3K/AKT/mTOR pathway has been strongly implicated in acquired resistance to hormone therapy.³⁰

The broad participation of AKT in targeted therapy resistance provides an opportunity to target the AKT pathway to circumvent this resistance. Biological markers indicate that activation of the AKT pathway is useful for predicting response to targeted therapy or detecting the emergence of resistance before it becomes clinically apparent.

Multidrug Resistance

Beyond its role in mediating DNA repair, AKT has been reported to play a critical role in multidrug resistance (MDR) in diverse cancer cells. MDR is primarily based on the overexpression of drug efflux pumps in the cellular membrane. Drug efflux pumps, which belong to the ATP-binding cassette superfamily of proteins, include MDR1 protein, multidrug resistance-associated protein-1 (MRP1), lung cancer resistance protein, and breast cancer resistance protein.³¹ PI3K/AKT signaling plays a critical role in modulating the expression and function of MDR1, MRP1, and breast cancer resistance protein under a variety of conditions such as hypoxia and receptor tyrosine kinase activation.³² All of these findings confirm the importance of AKT in MDR and make AKT an attractive target for overcoming MDR.

Genotoxic Agents

AKT signaling is integral to cell survival, particularly in the transformed setting. In addition to the classic model of AKT activation via growth factor signaling or loss of tumor suppressor phosphatases, AKT may also be activated in response to genotoxic insults induced by traditional chemotherapies. In glioma, highly activated AKT is associated with suppression of temozolomide-induced senescence, apoptosis, and cell cycle arrest, thus contributing to temozolomide resistance.³³ Furthermore, temozolomide treatment can activate AKT by inducing genetic alterations and drive the evolution of temozolomide-resistant glioma cells to a higher malignant potential. Johnson et al. sequenced the exomes of initially low-grade gliomas and recurrent tumors resected from the same patients and found that the recurrent tumors were hypermutated and harbored driver mutations in the retinoblastoma and Akt-mTOR pathways after temozolomide treatment.³⁴ Constitutively active AKT has been found in cisplatin-selected chemoresistant lung cancer, glioma, and ovarian cancer cell lines.³⁵ Taxols, which lead to destabilization of mitotic microtubules, have been reported to activate AKT in ovarian cancer cells and lead to constitutive activation in cells with acquired taxol resistance.³⁶ AKT activity can be induced by anthracyclines, doxorubicin, and daunorubicin in gastric cancer, leukemia, and breast cancer.³⁷ In a series of studies of human breast cancer and pancreatic adenocarcinoma with constitutively active high AKT activity, treatment with gemcitabine, a nucleoside analog, failed to induce cell death.³⁸

Activation of AKT is also associated with radioresistance in many cancers including those of the brain, head and neck, colon, bladder, and prostate.³⁹ Ionizing radiation (IR) induces AKT activation in many cancer cell lines. Inhibition of the AKT pathway has been shown to impair DNA repair after IR, resulting in radiosensitization in a variety of cancers.⁴⁰

Activation of AKT in DNA Damage

Most commonly used anticancer agents, including the classic genotoxic anticancer drugs and IR, attack the cellular DNA, thereby inducing DNA lesions that prevent DNA replication and transcription. Apoptosis induced by DNA lesions results almost exclusively from double-strand breaks (DSBs) and DNA replication blocking; therefore, cells have evolved DNA damage checkpoints to detect and repair DSBs and replication blocks. The DNA damage checkpoint network consists of DNA damage sensors, signal transducers, and effector pathways. The central sensors are the phosphoinositide 3-kinase-related kinases ataxia telangiectasia mutated (ATM), ataxia telangiectasia and Rad3-related (ATR), and DNA-PK. Upon activation by DSBs and DNA replication blocks, these sensors activate specific substrates that mediate cell cycle arrest, replication fork stability, DNA repair, and cell death.⁴¹

ATM, ATR, and DNA-PK are involved in AKT activation in the DNA damage response pathway (Figure 1).^42,43 AKT and DNA-PKcs are tightly coregulated in both checkpoint response and DNA repair. AKT1 knockout mice show impaired DNA damage-dependent induction of p21 and increased tissue apoptosis, similar to the DNA-PK deficiency phenotype.⁴⁴ Treating cells with doxorubicin, a DNA damage agent, increased the Ser-473 phosphorylation of AKT in DNA-PK+/+ mouse embryonic fibroblasts (MEFs) but not DNA-PK−/− MEFs.⁴⁵ Park et al. further demonstrated that DNA-PK can directly phosphorylate AKT at Ser-473 in vitro and that the hydrophobic motif of AKT plays a key role in defining the substrate specificity of DNA-PK.⁴⁶ This pathway is specific for DNA damage because DNA-PK is dispensable for AKT Ser473 phosphorylation upon insulin or growth factor stimulation.⁴⁴ Nevertheless, Viniegra et al. also provided clear evidence that ATM is required for activation of AKT in response to insulin or radiation.⁴⁷ Fraser et al. demonstrated that DNA DSBs promote AKT phosphorylation that is dependent on MRE11-ATM-RNF168 signaling and independent of DNA-PKcs, PI3K, and ATR.⁴⁸ The discrepancy between these observations is most likely due to different cell types. In the same situation, IR-induced phosphorylation of AKT was DNA-PKcs-dependent in both the MO59K and MO59J cell lines but not in the HCT116 cell line.⁴⁹ In a study using human fibroblasts and glioma cells, Golding et al. found that ATM is important for controlling AKT phosphorylation. When they tested ATM−/− mouse embryonic fibroblasts for their ability to modulate S473 AKT phosphorylation in response to radiation, they found that ATM no longer controlled AKT phosphorylation. The author suggested that there are mechanistic differences in how human and mouse cells control DNA damage-induced and subsequent AKT signaling.⁵⁰ ATR, another member of the phosphoinositide 3-kinase-related kinase family, may act as an upstream activator of AKT in response to DNA damage induced by O6-guanine methylating agents,⁵¹ but there is no evidence that ATM or ATR can directly phosphorylate AKT in response to DNA damage. The manner by which AKT is phosphorylated is not clear in those situations.

Fig. 1.

Activation and function of AKT in DNA damage response. After DNA damage caused by genotoxic agents, DNA damage sensors (including ATM, ATR, DNA-PK, and PARP) are involved in AKT activation because of different types of DNA damage and different cell types. Exactly how AKT is activated by ATM or ATR is not clear. DNA-PK can phosphorylate AKT at Ser-473. PARP causes NAD+ depletion and AMP accumulation. NAD+ depletion causes SIRT1 inactivation and releases AKT from SIRT1 inhibition. AMP accumulation causes AMPK activation and AKT activation. Activated AKT phosphorylates substrates and participates in the cellular response to DNA damage, including cell cycle arrest, anti-apoptosis, metabolism adjustment, drug efflux, and DNA repair.

Mounting evidence indicates that poly ADP ribose polymerase (PARP) is an important sensor in DNA damage checkpoints. PARP1 is a highly conserved DNA-binding protein and is the most abundant member of the PARP family. In addition to its important role in the base excision repair (BER) and single-strand break repair pathways, PARP1 appears to be involved in the nonhomologous end-joining (NHEJ) and homologous recombination (HR) pathways.⁵² PARP family proteins are also involved in regulating AKT activation due to DNA damage response. Ethier et al.⁵³ observed a significant increase in Ser473 phosphorylation of AKT at 1–3 hours after exposure to the alkylating agent MNNG, followed by a shutdown at 4 hours. Inhibition of PARP-1 activation by AG14361 led to a much more durable and significant activation of AKT phosphorylation. After inducing certain types of DNA damage, PARP1 is rapidly recruited to the altered DNA, and its catalytic activity increases 10- to 500-fold, resulting in the synthesis of protein-conjugated long-branched PAR chains 15–30 seconds after the damage has occurred.⁵⁴ PARP1 hyperactivation and PAR synthesis drive the nearly complete depletion of nicotinamide adenine dinucleotide (NAD+) and ATP pools. Increased adenosine monophosphate (AMP) levels, caused by ATP depletion, lead to activation of AMP-activated protein kinase, which can phosphorylate Rictor of the mTORC2 and activate mTORC2. NAD+ levels regulate SIRT1, an NAD-dependent longevity-promoting deacetylase. SIRT1 negatively regulates AKT phosphorylation in some cellular models by deacetylating PTEN.⁵⁵ Conversely, SIRT1 may also activate AKT and PDK1 via deacetylation at lysine residues in their PH domains, enhancing the binding of AKT and PDK1 to PIP3 and promoting their activation.⁵⁶ Thus, the influence of acetylation on AKT activation needs further investigation. It has also been reported that PARP1 inhibitors attenuate AKT-FOXO3A signaling through the activation of PHLPP1.⁵⁷

Activation of Nuclear AKT

Traditionally, AKT is phosphorylated in the cytomembrane; this is followed by translocation to various subcellular compartments, where it exerts its effects. However, AKT phosphorylation induced by DNA damage most likely occurs in the nucleus. The mechanism of AKT nuclear translocation is unknown, and the nuclear localization sequence motif of AKT has not yet been identified. However, the proto-oncogene T-cell leukemia-1 protein family may be involved in AKT nuclear localization.⁵⁸ It has been documented that kinase-dead AKT mutants (T308A and S473A) can migrate to the nucleus in HEK293 cells, suggesting that phosphorylation of AKT is not required for its nuclear localization.⁵⁹

Boehme et al. found that nuclear AKT was phosphorylated at Ser 473 much earlier than cytoplasmic AKT after IR. The overall increase of nuclear AKT phosphorylation was also significantly higher than that of cytoplasmic AKT. Although cytoplasmic AKT was phosphorylated at Thr 308 in response to IR, phosphorylation of Thr 308 was consistently weaker than phosphorylation of Ser 473. Moreover, phosphorylation of Thr 308 in the nucleus did not increase after IR. Such observations suggest that phosphorylation at Ser 473 of AKT after IR occurs within the nuclear compartment, as opposed to the entire population of activated AKT translocating to the nucleus.⁶⁰ Fraser et al.⁴⁸ also reported that pAKT-S473, but not pAKT-T308 or total AKT, accumulated in the vicinity of IR-induced DSB sites and colocalized with γH2AX and ATM-pSer1981.

The mechanism of nuclear AKT phosphorylation is not clear. The nucleus contains the machinery necessary for phosphorylating AKT, including PI3K, PIP3, and PDK1.⁶¹ mTORC2 and Ser 473 p-AKT have also have been found to colocalize to the nucleus in papillary thyroid carcinomas. Therefore, it was hypothesized that mTORC2 phosphorylates AKT at Ser 473 within the nucleus of thyroid neoplastic cells.⁶² In contrast, Fraser et al.⁴⁸ demonstrated that DSBs activates a signaling cascade that directly promotes a PI3K-independent pathway of AKT phosphorylation. Moreover, mTORC2 is not involved in the activation of AKT induced by cisplatin.⁶³ PML specifically recruits Thr 308 p-AKT phosphatase PP2A and p-AKT into PML nuclear bodies, indicating that phosphatases also play an important role in AKT nuclear activity.⁶⁴ Elucidating the specific regulation of nuclear AKT activation will provide insight into its role in DNA repair.

AKT's Role in DNA Adduct Repair

DNA damage is an umbrella term used to describe different types of DNA modification. There are 8 major DNA repair pathways that eukaryotes use to address different types of DNA damage: direct lesion reversal, BER, nucleotide excision repair (NER), mismatch repair (MMR), Fanconi anemia (FA) pathway, HR, NHEJ, and translesion DNA synthesis.⁶⁵ Different DNA repair pathways can overlap in function, and the loss of elements of one pathway may be compensated for by the increased activity of other elements or pathways (Figure 2). Traditional DNA damage-inducing anticancer drugs take advantage of defective DNA repair pathways in cancer cells. In some cases, the inhibition of additional pathways may have a stronger effect on cancer cells, an observation called synthetic lethality.

Fig. 2.

DNA repair pathway choice and the effect of AKT (green lines: successful pathway; red lines: failing pathway). Critical DNA lesions are induced by methylating agents, IR, cisplatin, and TOPs inhibitors. Eukaryotes have developed different DNA repair pathways to deal with the different types of DNA damage. The thick grey lines represent the pathway that successfully repairs DNA adducts, and the thin grey lines represent the process for forming DSBs, which is necessary to trigger apoptosis or cell senescence. AKT participates in this process on many levels.

Therapeutic methylating agents methylate DNA at 13 positions, forming 12 types of base adducts. Among these, the main products are N7-methylguanine (N7MeG; about 62% of adducts formed), N3-methyladenine (N3MeA), and N3-methylguanine (N3MeG; 8%–18% of total alkyl adducts). O6-methyl-guanine (O6MeG) and O4-methyl-thymine (O4MeT) comprise 7% and 1%–2% of adducts, respectively, and are repaired by O⁶-alkylguanine DNA alkyltransferase (MGMT) (for reviews, see ^66,67). The amount of MGMT in cells and the rate of resynthesis of MGMT are therefore directly proportional to the resistance of cells to methylating agents that induce O6MeG. Apoptosis is not directly induced by O6MeG but requires MMR and DNA replication. In the absence of MMR, cells acquire GC to AT site mutations after 2 rounds of DNA replication.⁶⁸ AKT can repress MMR by affecting the stability and nuclear localization of hPMS2, a component of heterodimer MutL, which binds to the mismatch recognition complexes that facilitate repair. Immunoprecipitation and protein stability assays using cycloheximide have revealed that phosphorylated AKT1 S473 can bind to hPMS2 directly and induce its degradation. Blocking AKT activity has also resulted in accumulation of hPMS2 protein in the nucleus.⁶⁹ Therefore, cells with hyperactivated AKT escape from apoptosis and senescence caused by the futile MMR, resulting in resistance to treatment and accumulation of mutations.

Cisplatin is a major class of bifunctional drugs. Cisplatin can bind to N7 of adjacent guanine or adenine or nonadjacent guanine.⁷⁰ Adjacent intrastrand crosslinks comprise 70%–80% of lesions, intrastrand and interstrand crosslinks (ICLs) of 2 nonadjacent guanines comprise 8%–10% of lesions, and monoadducts comprise 2%-3% of lesions.⁷¹ The intrastrand crosslinks are repaired by NER, whereas the ICLs are repaired by the FA pathway.⁷² Eukaryotic NER includes 2 major branches: transcription-coupled NER (TC-NER) and global genome NER (GG-NER).⁷³ Different proteins are involved in the initial recognition and binding steps of the 2 types of pathways.

P300, a histone acetyltransferase, acetylates histones to facilitate chromatin remodeling and thus recruit transcription factor or repair complexes to DNA damage sites. AKT phosphorylates P300 at Ser 1834,⁷⁴ which is essential to NER. AKT-mediated P300 phosphorylation enhances its histone acetyltransferase activity, which permits initial damage recognition factors to access damaged DNA and subsequently promotes P300 degradation to allow the sequential recruitment of downstream repair proteins for the successful execution of NER.⁷⁵ Although AKT promotes TC-NER, there is evidence that activated AKT can reduce the capacity of GG-NER by negatively regulating the expression of XPC, a specific GG-NER protein. AKT activation promotes p130 phosphorylation and increases the nuclear abundance of the E2F4-p130 repressor complex, which can repress XPC transcription (Figure 3).⁷⁶ The opposing roles of AKT in TC-NER versus GG-NER may be due to the functional role of XPC in the NER pathway, which is not yet fully understood. However, a possible link is evident in the genetic disorders Cockayne syndrome (CS) and xeroderma pigmentosum (XP). CS is linked to mutations that are specific to TC-NER, and XP is a disease caused by mutations in XP genes that renders cells highly sensitive to UV light.⁷⁷ The fact that there is no increased incidence of cancer in CS patients as a result of TC-NER deficiency suggests that, the apoptotic response after UV exposure is enhanced without functional TC-NER.⁷⁸ XP patients with mutated XPC are deficient only in GG-NER and are less sensitive to UV-induced cell death, which leads to increased mutation rates and cancer incidence.⁷⁹ Thus, activated AKT can suppress GG-NER by downregulating XPC expression, leading to the accumulation of mutations and the promotion of cancer cell transformation.

Fig. 3.

AKT's role in the NER pathway. AKT phosphorylates P300 and increases its HAT activity, which is important for histone acetylation and initial damage recognition factors to access damaged DNA. AKT reverse-regulates XPC expression through p130 and p38. Moreover, AKT is essential for the basal expression of ERCC1, which cuts 5′ to the damaged DNA with XPE.

AKT is also involved in the repair of ICLs. When DNA replication forks are stalled by ICLs, the FA core complex binds to the ICLs and acts as a platform to recruit multiple nucleases that coordinate the nucleolytic incisions flanking the ICL. Incisions near ICLs leave crosslinked nucleotides tethered to the complementary strand and enable a translesion DNA polymerase to pass the lesions. The DNA lesion can be removed by NER and the DSB of the other chromosome, which is created by ICL incisions that are rejoined by HR.⁸⁰ The mTORC1-S6K1 pathway controls the transcription of FANCD2, an essential protein for the FA pathway, by suppressing phosphorylation and nuclear translocation of nuclear factor-κB. AKT can increase the expression of FANCD2 by activating mTORC1, leading to a proficient FA pathway to repair ICLs and thus promoting chemoresistance.⁸¹

AKT's Roles in Double-strand Break Repair

The tumoricidal activity of radiation largely depends on the type of DNA lesions it induces. One gray of IR induces approximately 2000 base modifications, 1000 DNA single-strand breaks, and 35 DSBs per cell.⁸² Along with the ionization tracts, reactive oxygen species (ROS) are generated in the nucleus. The DNA bases are particularly susceptible to ROS. Among the 20 different base lesions induced by ROS, 5,6-dihydroxy-5,6-dihydrothymine, and 8-oxo-7,8-dihydroguanine (8-oxoG) are the most well known.⁸³ Thymine glycol can block DNA replication and may account for replication-dependent apoptosis activation if it is not repaired.⁸⁴ 8-oxoG mispairs with adenine during DNA replication, giving rise to GC→TA transversion mutations. The apoptosis caused by IR may come from a mix of DSBs that induce apoptosis, both directly and from the base modifications, and BER repair intermediates that block DNA replication and induce DSBs indirectly.

Most unrepaired or incorrectly repaired DNA modifications and lesions lead to DSBs before apoptosis or cell senescence is triggered. The DSBs can be repaired through 2 distinct mechanisms: the error-prone NHEJ pathway and the error-free HR pathway. In NHEJ, DSB ends are held in close proximity by a double-stranded DNA end-binding protein complex, the Ku70/80 DNA-binding complex. After minimal processing, mainly by the Artemis:DNA-PKcs complex, the double-stranded DNA ends are directly ligated by a complex of XLF, XRCC4, and DNA ligase IV (for details, see reviews^84,85). In contrast, the multifunctional Mre11-Rad50-Nbs1 (MRN) protein complex recruits ATM checkpoint kinase and CtIP to DSBs. HR repair initiates the 5′-to-3′ resection of DNA ends by Exo1, which is recruited and activated by a complex of BRCA1, MRN, and CtIP. The resection generates 3′ single-strand DNA overhangs that are rapidly coated with the replication protein A (RPA) complex. With the assistance of BRCA2, RPA is replaced by Rad51 recombinase for homology search within adjacent sister chromatids.⁸⁶ Limited DNA end resection by MRN and CtIP can result in an alternative NHEJ pathway, which may involve ligation by the XRCC1/DNA ligase III complex.⁸⁷ The cellular choice between the HR and NHEJ repair pathways is not well understood, although both the 53BP1 and BRCA1 proteins may play a key role in this choice.⁸⁸

Fraser et al.⁴⁸ showed that AKT was directly involved in NHEJ-mediated DSB repair using a cell-free reporter assay (Figure 4). Antibodies against AKT effectively attenuated end-joining of the linearized plasmid that is mediated by nuclear extracts of HCT116 colon cancer cells. Co-immunoprecipitation experiments showed a complex formation of activated AKT and DNA-PKcs in the DSB sites, indicating that AKT plays an important regulatory role in activating DNA-PKcs and NHEJ repair. Inhibitors or siRNA of AKT markedly reduced the DNA-PKcs phosphorylation at T2609 and S2056 that had been induced by IR, which is required for activation of DNA-PKcs.⁸⁹

Fig. 4.

AKT suppresses the HR pathway and activates the NHEJ pathway via many mechanisms. AKT can phosphorylate Brca1 and induce its cytoplasmic translocation. Brca1 inhibits the interaction of 53BP1 and methylated histone, which is important for NHEJ pathway choice. AKT also inhibits many other essential components of the HR pathway including rad51 and CtIP. AKT can directly phosphorylate and activate DNA-PK. Activated AKT forms a complex with DNA-PK and stimulates the autophosphorylation of DNA-PK.

AKT can regulate DNA-PKcs-dependent DNA-DSB repair in many ways. First, complex formation of AKT and DNA-PKcs stimulates binding of DNA-PKcs with Ku dimers in DNA duplex ends. Second, AKT promotes the kinase activity of DNA-PKcs, which is necessary for efficient DNA-DSB repair. Third, autophosphorylation of DNA-PKcs, stimulated by AKT, facilitates the release of DNA-PKcs from the damage necessary for ligation and termination of DSB repair.⁹⁰

Highly activated AKT can also suppress ATR signaling and HR. AKT has been reported to phosphorylate TopBP1 at Ser1159, a consensus-conserved sequence. Ser1159 phosphorylation prevents the association of ATR with TopBP1 after DNA damage, thus inhibiting ATR activation and G2/M cell cycle arrest.⁹¹ AKT1 can inhibit HR by inducing the cytoplasmic translocation of Brca1 and Rad51. Biopsies of breast tumors show a strong correlation between cytoplasmic Brca1 and Rad51 localization and pAKT1-S473 Brca1 and Rad51 foci formation. IR-induced HR is more impaired in cells with high AKT activity than in cells with low AKT activity.⁹² Although AKT can phosphorylate Thr509 of Brca1 in nuclear localization sequences, cytoplasmic retention of Brca1 does not require Thr509 phosphorylation. Thus, Thr509 phosphorylation may prevent Brca1 from interacting with other proteins that localize to DNA damage foci.⁴³ Brca1 plays an important role in facilitating HR pathway choice by antagonizing 53BP1 chromatin interactions, which are essential for the NHEJ pathway.⁸⁸

AKT has a well-established role in cell cycle control and is involved in DNA repair-induced cell cycle regulation. Xu et al.⁹³ showed that AKT suppresses DNA damage processing and checkpoint activation in late G2 phase. In HCT116 cells, DNA damage-induced RPA, CtIP, and Rad51 foci formation are markedly suppressed by constitutively activated AKT and fail to resect DSBs to generate single-stranded DNA, which leads to failed initiation of the subsequent events required for HR or activation of ATR. Conversely, inhibition of AKT by selective chemical inhibitors or AKT siRNA restores the DNA damage-induced recruitment of RPA, CtIP, Rad51, and Chk1 activation.

p21-mediated inhibition of CDK activity interrupts CDK-mediated BRCA2 S3291 phosphorylation and thus facilitates the interaction between RAD51 and the carboxy-terminus of BRCA2 that promotes HR DNA repair.⁹⁴ AKT regulates CDK activity via the CDK inhibitors p21 and p27. AKT directly phosphorylates both T145 and S146 near the carboxyl terminus of p21/Cip1. Thr 145 phosphorylation results in cytoplasmic localization of p21Waf1/Cip1 with PCNA and release of CDK2 from the inactive PCNA/p21Cip1/Cdk/cyclin quaternary complex that promotes the cell cycle,⁹⁵ whereas Ser146 site phosphorylation enhances stability of p21 and further increases assembly of the cyclin D-CDK4 G1/S transition complex.⁹⁶ AKT can directly phosphorylate p27Kip1 on T157 and thereby cause relocation of p27Kip to the cytoplasm, subsequently relieving the nuclear substrates (CDK2/cyclin E and CDK2/cyclin A) from p27Kip and prompting cell cycle progression.⁹⁷

Cell cycle arrests induced by DNA damage are coordinated primarily by 2 distinct kinase signaling cascades: the ATM-Chk2 and ATR-Chk1 pathways. Chk1 is directly phosphorylated by AKT at Ser280, a modification that results in cytoplasmic sequestration.⁹⁸ Phosphorylation by AKT also inhibits the interaction between Chk1 and Claspin, which results in attenuated Chk1 activity and checkpoint proficiency and subsequently disrupts the interaction of Chk1 and Rad51 necessary for Rad51 foci formation and the HR pathway.⁹⁹ AKT can also reduce the recruitment of Chk2 to sites of DNA damage and inhibit Chk2 activation.³³

AKT promotes chromosome instability and cell-cycle progression by suppressing HR and activating the NHEJ pathway because NHEJ is faster and more error prone than HR. Thus, cancer cells with high levels of activated AKT can avoid apoptosis or senescence caused by HR pathway failure and accumulate the mutations necessary for malignant progression.

Conclusions

Aberrant AKT activation may account for the important limitation of successful cancer therapy. Activation of the PI3K/AKT pathway is also a consequence of the administration of many types of genotoxic agents and accounts for the development of acquired resistance and hypermutation caused by genotoxic agents. Besides the traditional role of AKT in cell proliferation, cell cycle regulation, and anti-apoptosis, accumulating evidence indicates that AKT is involved in DNA repair; however, the role of AKT in DNA repair is under ongoing investigation. The mechanism of DNA damage-induced nuclear AKT activation is unclear because its activation differs from the classic activation pathway that occurs at the plasma membrane. Identifying the mechanisms involved in nuclear entrance and activation of AKT by the DNA damage response will provide new targets for AKT inhibitor development. Moreover, many investigations have demonstrated that AKT isoforms have different functions in specific situations.¹⁰⁰ Strnach et al. found that PEO23 and SKOV3 require AKT1 for cisplatin resistance in platinum-resistant ovarian carcinoma cell lines: PEA2 requires AKT2, whereas PEO4 requires AKT3.⁶³ The precise roles of the 3 AKT isoforms need further clarification. Future studies may identify the exact mechanisms of AKT involved in different DNA repair pathways and provide a new combination therapy aimed at circumventing or decreasing treatment resistance in cancer patients with different phenotypes.