AKT inhibition in the central nervous system induces signaling defects resulting in psychiatric symptomatology
Cell & Bioscience volume 12, Article number: 56 (2022)
Changes in the expression and activity of the AKT oncogene play an important role in psychiatric disease. We present translational data assessing the role of AKT in psychiatric symptoms.
(1) We assessed the protein activity of an AKT3 mutant harboring a PH domain mutation (Q60H) detected in a patient with schizophrenia, the corresponding AKT1 mutant (Q61H), and wild-type AKT1 and AKT3 transduced in AKT-null mouse fibroblasts and modeled the Q61H mutation onto the crystal structure of the Akt1 PH domain. (2) We analyzed the results of earlier genome-wide association studies to determine the distribution of schizophrenia-associated single-nucleotide polymorphisms (SNPs) in the AKT3 gene. (3) We analyzed the psychiatric adverse events (AEs) of patients treated with M2698 (p70S6K/AKT1/AKT3 inhibitor) and with other PI3K/AKT/mTOR pathway inhibitors.
(1) Proteins encoded by AKT3 (AKT3Q60H) and AKT1 (AKT1Q61H) mutants had lower kinase activity than those encoded by wild-type AKT3 and AKT1, respectively. Molecular modeling of the AKT1-Q61H mutant suggested conformational changes that may reduce the binding of D3-phosphorylated phosphoinositides to the PH domain. (2) We identified multiple SNPs in the AKT3 gene that were strongly associated with schizophrenia (p < 0.5 × 10–8). (3) Psychiatric AEs, mostly insomnia, anxiety, and depression, were noted in 29% of patients treated with M2698. In randomized studies, their incidence was higher in PI3K/AKT/mTOR inhibitor arms compared with placebo arms. All psychiatric AEs were reversible.
Our data elucidate the incidence and mechanisms of psychiatric AEs in patients treated with PI3K/AKT/mTOR inhibitors and emphasize the need for careful monitoring.
The PI3K/AKT/mTOR (PAM) pathway is a highly conserved signaling cascade that plays an important role in several cellular processes. It mediates ribosome biogenesis, protein translation, autophagy, and actin dynamics. Activation of the PAM pathway by growth factors, hormones, and mitogen effectors regulates global and specific mRNA translation [1,2,3]. It regulates protein synthesis by directly modulating the activity of translation initiation factors and upregulating ribosome biogenesis [3, 4]. The pathway serves as a master regulator of epithelial cell metabolism, morphology, and function and integrates glycolysis with actin cytoskeletal dynamics . It regulates multiple steps in glucose uptake and metabolism  and cytoskeletal functions including cellular movement and attachment . The regulatory subunit p85α of PI3K plays a role in controlling actin dynamics associated with PDGF receptor–induced cytoskeletal changes and cell migration .
The discovery that inhibition of glycogen synthase kinase-3 (GSK3) stabilizes mood in patients with psychiatric illness had a major impact on neuroscience . Experiments addressing the regulation of GSK3 demonstrated that both known GSK3 isoforms, GSK3α and GSK3β, are inactivated by AKT via phosphorylation at Ser21 and Ser9, respectively , thereby confirming that inhibition of GSK3 depends on AKT activation [11, 12]. Earlier studies showed that the critical step in the catalytic activation of AKT is the binding of its PH (pleckstrin homology) domain to the D3-phosphorylated phosphoinositides PI(3,4)P2 and PI(3,4,5)P3 in the cell membrane . The abundance of these phosphoinositides, in turn, depends on the activity of phosphoinositide kinases and phosphatases, primarily the type I PI3 kinase, which phosphorylates PI(4,5)P2 and PI4P in the D3 position of the inositol ring to generate PI(3,4,5)P3 and PI(3,4)P2, respectively, and the phosphatase and tensin homolog (PTEN), which removes the D3 phosphate from PI(3,4,5)P3 and PI(3,4)P2 (Fig. 1). These findings place the PAM pathway at the center of a major axis that regulates brain function and whose deregulation may lead to the development of psychiatric illness .
A series of observations in animal models and humans supports the hypothesis that perturbations of the PAM pathway have a role in the pathogenesis of psychiatric illness. These include schizophrenia-linked genetic polymorphisms of ERBB4 and NRG1 (encoding the ERBB4 ligand, neuregulin-1), deregulation of PIK3CD (encoding PI3Kδ) , activation of the dopamine 2 receptor (D2R), which inhibits AKT by activating the AKT phosphatase PP2A [16, 17], and AKT phosphorylation in postmortem studies on the brains of patients with schizophrenia or major depression [18, 19].
Genetic studies have shown that an AKT1 haplotype associated with schizophrenia in humans resulted in low levels of AKT1 expression and greater sensitivity to sensorimotor gating disruption by amphetamine . Subsequent studies in mice showed that the ablation of AKT1 was associated with decreased proliferation of adult-born hippocampal progenitors and with defects in hippocampal long-term potentiation, along with defects in contextual fear conditioning, spatial learning, and other hippocampus-dependent functions . Ablation of AKT2 in mice was linked to anxiety and depression-like behaviors , while ablation of AKT3 resulted in small brains and schizophrenia-like symptomatology . Finally, similar psychiatric illness-like defects were induced by inactivating mutations of PI3K and overexpression of PTEN, both of which result in downregulation of the activity of AKT [23, 24].
While these data provide evidence implicating dysregulation of the PI3K/AKT/mTOR axis in the pathogenesis of psychiatric illness, it is unclear whether this is the result of developmental or signaling defects. Supporting the former possibility is the evidence that PI3K/AKT/mTOR inhibition and GSK3 activation are associated with decreased neuronal cell survival [25, 26] and that, while the ablation of AKT3 and overexpression of PTEN are associated with microcephaly (small brains) , high AKT3 activity and inactivating mutations of PTEN are associated with macrocephaly (large brains) and autism-like behaviors [23, 24]. The defects in hippocampal plasticity and function in AKT1−/− mice  also suggest the possibility of a developmental defect.
A strong association between single-nucleotide polymorphisms (SNPs) in the AKT3 gene and schizophrenia susceptibility was found by genome-wide association studies (GWAS) conducted by the Schizophrenia Working Group of the Psychiatric Genomics Consortium [28, 29]. The work presented in this report was initiated following the discovery of a missense mutation in the PH domain of the AKT3 gene in a patient with schizophrenia in an exome sequencing study organized by the Stanley Center for Psychiatric Research at the Broad Institute (https://schema.broadinstitute.org/gene/ENSG00000117020) . In that study, 97,322 healthy individuals and 24,248 patients with schizophrenia were analyzed for identification of rare genetic variants linked to schizophrenia . No enrichment of coding AKT3 variants in schizophrenia patients was identified. Heterogeneity for a missense mutation in the PH domain of AKT3 (AKT3Q60H) was found in one patient and in none of the healthy controls. While the loss-of-function variants of AKT3 did not confer risk for schizophrenia, we characterized AKT3-Q60H because of its location in the PH domain and potential functional consequences .
Additionally, our earlier studies had shown that AKT3 is a strong inducer of reactive oxygen species and that its expression activates the DNA damage response and reduces cellular proliferation. As a result, cells with high expression of active AKT3 and high AKT3 kinase activity are counter-selected .
Therefore, we hypothesized that the AKT3-Q60H mutant may be functionally defective and have a causative role in the pathophysiology of schizophrenia. To test this hypothesis, we employed site-directed mutagenesis to introduce this mutation in the Myc-tagged AKT3 gene, and we used retrovirus constructs of wild-type AKT3 and the AKT3 Q60H mutant to transduce a lung fibroblast cell line from AKT1fl/flAKT2−/−AKT3−/− mice we had established earlier [32, 33].
Finally, a phase I clinical trial with the brain-penetrant p70S6K/AKT1/AKT3 inhibitor M2698 (ClinicalTrials.gov: NCT01971515; EMD Serono) demonstrated that some of the M2698-treated patients developed psychiatric adverse events (AEs) and that all of these AEs were reversible . Importantly, similar AEs were noted in clinical trials with other PI3K/AKT/mTOR inhibitors . The reversibility of the symptoms elicited by these drugs suggests that low AKT activity in the central nervous system (CNS) may induce psychiatric illness because of signaling, rather than developmental, defects. In addition to providing information relevant to the pathophysiology of mental illness, the data in this report delineate a set of psychiatric symptoms that should be taken into consideration when designing and conducting clinical trials with inhibitors of AKT and other PAM pathway genes in human cancer.
The AKT3-Q60H mutant, which was detected in a patient with schizophrenia, exhibits reduced catalytic activity
We found that the phosphorylation of AKT3-Q60H at both the Thr305 and Ser472 sites was impaired (Fig. 2A), suggesting that the mutant protein exhibits reduced catalytic activity. In addition, the expression of the mutant protein was higher than that of the wild-type protein. In vitro kinase assays, using GSK3 as the kinase substrate, demonstrated that the kinase activity of AKT3-Q60H was impaired (Fig. 2B). In vitro kinase assays using AKT3 immunoprecipitated from cell lysates harvested 15 min later and GSK3 as the kinase substrate demonstrated that the activation of the AKT3-Q60H mutant by IGF1 was also impaired (Fig. 2C). Interestingly, the corresponding AKT1 mutant (Q61H) also exhibited reduced catalytic activity, as evidenced by the low level of phosphorylation of the protein at Thr308 and Ser473 (Fig. 2D).
Modeling of the binding of PI(3,4,5)P3 to the wild-type and mutant AKT PH domain
We found that the Q61 residue is located in the β5-strand of the PH domain, and its two major side-chain interactions are hydrophobic interactions with the (i,- > i + 2) residue M63 (in the β5 strand) and the R76 residue in the adjacent β6 strand (Fig. 3A). Importantly, the Q61 and R76 residues, as well as the E85 residue in the β7 strand, are conserved among all three AKT isoforms (Fig. 3B). In the crystallographic conformation, there were no favorable polar interactions between the Q61 and R76 side chain, but there was a very strong electrostatic salt bridge between R76 and E85 (Fig. 3C). However, two complementary modeling approaches, molecular dynamics simulations and analysis of a side-chain rotamer library, both show that the carbonyl group of the Q61 side chain may form favorable polar interactions with the guanidino side chain of R76 (Fig. 3D), while the Q61H histidine side chain rotamers are not able to form equivalent favorable interactions (see Additional file 1: Methods). In addition, partially charged protamers of the H61 imidazole side chain, expected over a pH range of 6.0 to 7.0, will result in repulsive charge-charge interactions with R76 (see Fig. 3E) that may significantly affect the structure and dynamics of the PH domain. Finally, observed structural changes to the conformational state of the Q61H binding pocket residues (K14 and R86) in molecular dynamics simulations support the hypothesis that Q61H may reduce the binding affinities for D3 phosphorylated phosphoinositides.
Common SNPs link AKT3 to the development of mental illness
Analyzing the publicly available dataset of the schizophrenia study [28, 29], we identified multiple SNPs with a genome-wide significant association with schizophrenia (p < 0.5 × 10–8). All schizophrenia-linked SNPs in the chromosomal region centered around AKT3 map within the AKT3 gene itself (Fig. 4). The peak risk association SNP in this region was rs61833239 (p = 5.22 × 10–12).
The use of PI3K/AKT/mTOR pathway inhibitors in patients with cancer gives rise to reversible psychiatric symptomatology
Table 1 illustrates the psychiatric adverse events noted in patients with advanced cancer treated on a phase I study of the brain-penetrant p70S6K/AKT1/AKT3 inhibitor M2698. Overall, 29% (29/101) of treated patients developed psychiatric AEs: M2698 monotherapy, 26% (16/62); M2698 plus tamoxifen, 35% (9/26); and M2698 plus trastuzumab, 31% (4/13). The most frequent adverse events were insomnia, anxiety, and depression. All psychiatric AEs were reversible and manageable with dose reduction. Some patients developed multiple psychiatric AEs. Five patients contributed to 40% of all AEs in the study. Of the 29 patients who developed psychiatric AEs during the study, 45% (13/29) had no prior psychiatric symptomatology. Details of psychiatric AEs in selected patients are described in Additional file 1: Results. Additional file 1: Table S2 describes patients who reported multiple (≥ 3) psychiatric AEs with M2698 treatment.
As of August 2020, more than 825 clinical trials with PAM pathway inhibitors, targeting single or multiple enzymes, have been conducted in patients with cancer. Table 2 summarizes the psychiatric AEs reported in selected randomized clinical trials, comparing treatment arms with or without PI3K/AKT/mTOR inhibitors. Although psychiatric AEs were not reported in clinical trials of alpelisib, capivasertib (AZD5363), or ipatasertib, they were noted in trials of buparlisib, taselisib, and MK-2206 in patients with breast and prostate cancer. Among these last three drugs, the highest incidence of mood disorders, including depression, mania, anxiety, and suicide attempts, was observed during treatment with buparlisib, which crosses the blood–brain barrier. In addition to the psychiatric AEs, whose incidence was higher in the PAM inhibitor arms than in the placebo arms, there was also an increased incidence of neurological symptoms such as dizziness and headache in the treatment groups across all studies. Headache was more frequent in taselisib- and MK-2206-treated patients.
The detection of a functionally hypoactive AKT3 mutation in a patient with schizophrenia, combined with the GWAS results linking AKT3 SNPs with schizophrenia, suggested a link between the AKT3 gene and schizophrenia. A third piece of evidence confirming this association was provided by the AE data from a phase I study of the brain-penetrant p70S6K/AKT1/AKT3 inhibitor M2698 (EMD Serono)  and by a review of psychiatric AEs reported in clinical trials of other PAM pathway inhibitors.
Our results suggest that decreased AKT activity in the CNS, caused by mutations (such as AKT3Q60H), by genetic polymorphisms in the AKT3 gene, or by treatment with brain-penetrant PAM pathway inhibitors, promotes the development of psychiatric symptoms in some individuals. The activity of the mutant AKT protein was reduced and its expression was higher than that of the wild-type protein in the cellular construct. In addition, psychiatric symptoms induced by pharmacological blockade of the pathway were reversible, suggesting that its emergence may depend on signaling, rather than developmental, defects. These data contribute to the understanding of the role of the AKT pathway in the pathophysiology of psychiatric illness. Furthermore, the association of psychiatric symptomatology with PAM pathway inhibitors suggests that efficient development of this important class of anticancer drugs requires standardized methods to assess, monitor, and mitigate the risk of these events in clinical trials.
Although the activity of AKT in the CNS was not monitored during the M2698 clinical trial, experiments in mouse models have shown that M2698 crosses the blood–brain barrier, resulting in CNS exposures sufficient to inhibit AKT1, AKT3, and p70S6K in vitro . Moreover, in the phase I study, M2698 was shown to strongly inhibit AKT1, AKT3, and p70S6K in the tumor tissue. Finally, M2698 was found to exhibit robust pharmacodynamic activity toward p70S6K and AKT in post-treatment brain tumor biopsies of orthotopic xenograft mouse models. These observations collectively suggest that M2698 inhibits AKT in the brain, although we understand that the tumor data may not be directly translatable to the normal brain as these tissues may differ in vascularity, which would affect the concentration of the drug, and in the genetic and epigenetic regulation of the target pathway, which may affect the response to the drug.
The psychiatric AEs observed with M2698 treatment are consistent with observations in clinical trials of other PAM pathway inhibitors . For instance, a phase Ib/II study of pictilisib (GDC-0941), a pan-PI3K inhibitor, combined with cisplatin in androgen receptor-negative, triple-negative metastatic breast cancer was closed early owing to gastrointestinal and neuropsychiatric toxicities, and the phase III studies of buparlisib, a pan-PI3K inhibitor, in hormone receptor-positive breast cancer demonstrated a high incidence of treatment-related psychiatric symptomatology, which halted further development of this molecule [37, 38]. An important factor that may be responsible for differences in the incidence and severity of symptoms induced by different inhibitors is the differential ability of these inhibitors to cross the blood–brain barrier. The host factors responsible for the heterogeneity in the penetrance and presentation of these AEs are unknown. To identify these factors, we need to discover biomarkers associated with predisposition to these events, which could include germline polymorphisms in AKT (primarily AKT3) and in other pathway-specific regulatory molecules. Such studies will benefit from the development of methodologies to monitor AKT activity in the CNS.
There are several challenges in interpreting psychiatric AEs in the context of oncology clinical trials and in comparing or combining data from different trials. First, it is difficult to attribute low-prevalence events to a study drug when there is a high background of these events in the patient population at baseline. Second, psychiatric AEs are not consistently reported across trials, and therefore standardization of reporting is needed for consistent monitoring of these events in order to mitigate drug toxicity and support the development of this class of therapeutics. Because behavioral AEs occur in patients who are already vulnerable to such symptoms owing to other organic and functional stressors, it will be beneficial to include experienced psycho-oncology specialists in the treatment team. Neuropsychological testing may also be useful to assess subclinical vulnerability before treatment, as well as the extent of the recovery from or the chronicity of deficits afterwards. Currently, our clinical data suggest that it is possible for patients with a history of psychiatric disorders to successfully participate in such clinical trials if they are properly monitored and managed, thereby expanding the benefit of these targeted therapies to a broader population of patients.
Our observations are in line with published data demonstrating an association between the PAM pathway and psychiatric illness . PAM pathway disruption of protein synthesis and actin dynamics can lead to abnormal neuronal morphology, learning/memory deficits, and psychiatric disease [19, 22, 40,41,42,43,44,45]. Proteins implicated in this pathway are downregulated in the dorsolateral prefrontal cortex in schizophrenia . In murine models, decreased AKT activity in the ventral tegmental area was associated with increased susceptibility to depressive behavior ; antipsychotic drugs were shown to activate GSK3β-AKT signaling [41,42,43]; and GSK3β inhibitors had anti-depressant effects in animal models . Furthermore, AKT isoform- and gender-specific effects were seen on levels of anxiety, spatial and contextual memory, and fear extinction . Selected AKT SNPs have been associated with psychiatric disease [20, 46,47,48,49]. For instance, the SNPs rs1130214 and rs3730358 have been associated with higher Akt1 level, bipolar disorder, and major depression [20, 50, 51]. In patients with depressive disorders, the polymorphism rs1130214 was associated with response to antidepressant therapy .
AKT3 may function as a key regulator of PI3K/AKT/mTOR signaling in the brain. Although there is no direct evidence that AKT3 protein levels are reduced in patients with schizophrenia, one study showed that the relative phosphorylation of GSK3β at Ser9 (downstream of all AKTs) was significantly lower in lymphocytes and in the frontal cortexes of individuals with schizophrenia compared with controls , providing indirect evidence that AKT3 could be reduced in their brains as well. AKT3 deletion associated with decreased GSK3α/β phosphorylation levels in various brain regions was restored with chronic administration of lithium, a mood stabilizer. Lithium treatment also rescued depressive and anxiety-like behaviors in AKT3-knockout mice . AKT3 may be the primary target for mTORC2 activity, as ablation of the mTOR2 complex proteins Rictor and Sin1 was shown in mice to decrease AKT Ser473 activation, causing microcephaly and alterations in neuronal morphology and function .
To date, only six PAM pathway inhibitors have been approved by the US Food and Drug Administration (FDA): the mTOR inhibitors everolimus and temsirolimus, the PI3K-δ inhibitors idelalisib and duvelisib, the PI3K-α inhibitor alpelisib, and the pan-PI3K inhibitor copanlisib (Additional file 1: Table S3). The main challenge for the development of these therapeutics is toxicity, which includes skin rash, hyperglycemia, and gastrointestinal and psychiatric symptoms . These adverse events require treatment interruption, with or without dose reduction, and in severe cases, treatment discontinuation.
Given the complexity of the PAM pathway and clinical translation of our data, our study has several limitations. First, lung fibroblasts were utilized to study the impact of AKT mutations on kinase activity, whereas a more relevant model would be primary neurons. However, AKT activation via PIP3 binding to the PH domain occurs in all cells, including fibroblast and neuronal cells. A structural change in the PH domain will, therefore, affect a function that is common in all cell types. Consequently, the activity of this AKT3 mutant is expected to be similar in fibroblasts and neuronal cells. Second, we cannot determine whether patients treated with M2698 who experienced psychiatric AEs  also carried SNPs in AKT3, making them theoretically more susceptible to the study drug. SNP analyses should be planned prospectively in future clinical trials with PAM pathway inhibitors. Third, although our data suggest that the elicitation of reversible psychiatric symptoms by acute PI3K/AKT/mTOR inhibition may be due to signaling defects, this observation has to be interpreted with caution. Since AKT is a central node in a multitude of important cellular pathways, perturbation of brain development caused by a genetic mutation that impairs AKT kinase activity cannot be excluded.
In summary, we provide genetic data and data derived from clinical trials that further elucidate the role of AKT inhibition in psychiatric symptomatology. Our structure–function studies addressing the effects of an AKT PH domain mutation observed in a patient with schizophrenia deepen our understanding of AKT regulation by D3 phosphorylated phosphoinositides. Future research should focus on the identification of covariates that contribute to the development of psychiatric symptomatology in individuals treated with PAM pathway inhibitors and on the determination of the frequency of AKT activity downregulation among patients with psychiatric symptoms. To address these questions, methods to assess the risk and monitor the development of psychiatric symptomatology should be developed and standardized across clinical trials of PAM pathway inhibitors, and non-invasive technologies should be developed to monitor the activity of AKT in the CNS.
Cells and constructs
The methodology used was as previously published . The immortalized AKT1fl/flAKT2−/−AKT3−/− lung fibroblasts were transduced with pBabe-puro/Myc-Akt3, pBabe-puro/Myc-Akt3-Q60H, pBabe-puro/Myc-Akt1, or pBabe-puro/Myc-Akt1-Q61H constructs or with the pBabe-puro empty retroviral vector. The constructs used in the experiments in this report have been described previously . Cells transduced with the empty vector were used as a control. Subsequently, all cells were superinfected with a MigR1-Cre construct to ablate the floxed endogenous AKT1 gene . The triple-knockout cells that were not rescued with AKT3 did not proliferate, but they did remain alive for 7 to 10 days, so they could be used to determine response to external signals. Details about the experiments and sources of cells, constructs, etc. are provided in Additional file 1: Methods.
Site-directed mutagenesis was performed using polymerase chain reaction–based procedures, and pairs of overlapping oligonucleotide primers harboring the AKT3-Q60H or the AKT1-Q61H mutation were used to amplify pBabe-puro constructs of AKT3 and AKT1, respectively.
Cell lysis and protein extraction were performed using standard procedures. Western blots of lysates from cells maintained in complete serum-containing media were probed with antibodies to AKT1 and AKT3, the Myc-tag antibody, and antibodies that recognize the phosphorylated Thr305 and Ser472 sites of AKT3 (corresponding to the AKT1 phosphorylation sites Thr308 and Ser473). The antibodies used are listed in Additional file 1: Table S1.
In vitro kinase assay
To determine whether the AKT3-Q60H mutant responds to tyrosine kinase receptor-initiated signals, the triple-AKT-knockout lung fibroblasts rescued with wild-type or mutant AKT3 were serum-starved and stimulated with IGF1. Immortalized AKT1fl/flAKT2−/−AKT3−/− lung fibroblasts transduced with pBabe-puro/myc-Akt3 (wild-type or Q60H-mutant) or pBabe-puro/myc-Akt1 (wild-type or Q61H-mutant) and with MigR1-Cre were lysed with the Triton X-100 buffer. In vitro kinase assays on Myc-Akt3 immunoprecipitated from these lysates were performed as previously described . The phosphorylation substrate was recombinant GSK3α/β (Cell Signaling Technology, cat. no. 9237), and phosphorylation was detected with an antibody specific for GSK3β phosphorylated at Ser9. Western blots of the same lysates were probed with the anti-Akt3 or an anti-pan-Akt antibody (loading control). Quantification of the Western blot bands was performed using ImageJ software (for details see Fig. 2 legend).
For the kinase assays the ratio between pGSK3 and Myc-AKT3 or Myc-AKT1 has been calculated and the ratio volume for the WT forms was set as 1. For pAKTs at Ser or Thr, the ratio between pAKT and AKT3 or between pAKT and pan-AKT has been calculated. The calculated volume for the WT for was set as 1.
The structural effects of the Q61H mutation were explored with modeling, starting from the crystal structure of the Akt1 PH domain in complex with Ins(1,3,4,5)-tetrakisphosphate . Molecular modeling was performed using the program CHARMM and the generalized Born with molecular volume (GBMV) approach as previously described [57, 58].
Schizophrenia GWAS study
We analyzed the publicly available dataset of the large-cohort schizophrenia GWAS study to determine whether common polymorphisms link AKT3 to schizophrenia [28, 29]. These available data  were retrieved through the Ricopili portal (https://data.broadinstitute.org/mpg/ricopili/), and the p-values of SNP association with the AKT3 locus were plotted.
Adverse psychiatric events in clinical trials of PI3K/AKT/mTOR pathway inhibitors, including M2698
We performed a comprehensive literature review of the neuropsychiatric effects of other PAM pathway inhibitors and surveyed the AE data from randomized controlled trials of these drugs published in the U.S. National Library of Medicine clinical trials database (www.clinicaltrials.gov). We focused on psychiatric AEs and categorized them based on clinical presentation. Serious and non-serious AEs were reported as the proportion of patients affected. The AEs were reviewed by a psychiatrist (AV) with expertise in cancer clinical trials. The events were grouped into four general categories (impaired reality testing; delirium; mood disorders; other). We also systematically reviewed the database of patients with advanced cancer treated on a phase I study of the brain-penetrant p70S6K/AKT1/AKT3 inhibitor M2698 as monotherapy (n = 62) and in combination with tamoxifen (n = 26) or trastuzumab (n = 13) for all treatment-emergent AEs of the system organ class psychiatric disorders .
Availability of data and materials
All the primary data are available.
Central nervous system
Dopamine 2 receptor
Generalized Born with molecular volume
Glycogen synthase kinase-3
Genome-wide association studies
Single nucleotide polymorphisms
Gingras AC, Raught B, Sonenberg N. Regulation of translation initiation by FRAP/mTOR. Genes Dev. 2001;15(7):807–26.
Raught B, Peiretti F, Gingras AC, et al. Phosphorylation of eucaryotic translation initiation factor 4B Ser422 is modulated by S6 kinases. EMBO J. 2004;23(8):1761–9.
Ruggero D, Pandolfi PP. Does the ribosome translate cancer?. Nat Rev Cancer. 2003;3(3):179–92.
Ruggero D, Sonenberg N. The Akt of translational control. Oncogene. 2005;24(50):7426–34.
Hu H, Juvekar A, Lyssiotis CA, et al. Phosphoinositide 3-Kinase Regulates Glycolysis through Mobilization of Aldolase from the Actin Cytoskeleton. Cell. 2016;164(3):433–46.
Rathmell JC, Fox CJ, Plas DR, et al. Akt-directed glucose metabolism can prevent Bax conformation change and promote growth factor-independent survival. Mol Cell Biol. 2003;23(20):7315–28.
Cantley LC. The phosphoinositide 3-kinase pathway. Science. 2002;296(5573):1655–7.
Jimenez C, Portela RA, Mellado M, et al. Role of the PI3K regulatory subunit in the control of actin organization and cell migration. J Cell Biol. 2000;151(2):249–62.
Jope RS, Roh MS. Glycogen synthase kinase-3 (GSK3) in psychiatric diseases and therapeutic interventions. Curr Drug Targets. 2006;7(11):1421–34.
Cross DA, Alessi DR, Cohen P, et al. Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature. 1995;378(6559):785–9.
Paraskevopoulou MD, Tsichlis PN. A perspective on AKT 25-plus years after its discovery. Sci Signal. 2017;10:486.
Manning BD, Toker A. AKT/PKB Signaling: Navigating the Network. Cell. 2017;169(3):381–405.
Franke TF, Yang SI, Chan TO, et al. The protein kinase encoded by the Akt proto-oncogene is a target of the PDGF-activated phosphatidylinositol 3-kinase. Cell. 1995;81(5):727–36.
Beaulieu JM. A role for Akt and glycogen synthase kinase-3 as integrators of dopamine and serotonin neurotransmission in mental health. J Psychiatry Neurosci. 2012;37(1):7–16.
Law AJ, Wang Y, Sei Y, et al. Neuregulin 1-ErbB4-PI3K signaling in schizophrenia and phosphoinositide 3-kinase-p110delta inhibition as a potential therapeutic strategy. Proc Natl Acad Sci U S A. 2012;109(30):12165–70.
Goldman-Rakic PS, Castner SA, Svensson TH, et al. Targeting the dopamine D1 receptor in schizophrenia: insights for cognitive dysfunction. Psychopharmacology. 2004;174(1):3–16.
Sotnikova TD, Beaulieu JM, Barak LS, et al. Dopamine-independent locomotor actions of amphetamines in a novel acute mouse model of Parkinson disease. PLoS Biol. 2005;3(8): e271.
Balu DT, Carlson GC, Talbot K, et al. Akt1 deficiency in schizophrenia and impairment of hippocampal plasticity and function. Hippocampus. 2012;22(2):230–40.
Karege F, Perroud N, Burkhardt S, et al. Alteration in kinase activity but not in protein levels of protein kinase B and glycogen synthase kinase-3beta in ventral prefrontal cortex of depressed suicide victims. Biol Psychiatry. 2007;61(2):240–5.
Emamian ES, Hall D, Birnbaum MJ, et al. Convergent evidence for impaired AKT1-GSK3beta signaling in schizophrenia. Nat Genet. 2004;36(2):131–7.
Leibrock C, Ackermann TF, Hierlmeier M, et al. Akt2 deficiency is associated with anxiety and depressive behavior in mice. Cell Physiol Biochem. 2013;32(3):766–77.
Howell KR, Floyd K, Law AJ. PKBgamma/AKT3 loss-of-function causes learning and memory deficits and deregulation of AKT/mTORC2 signaling: Relevance for schizophrenia. PLoS ONE. 2017;12(5): e0175993.
Riviere JB, Mirzaa GM, O’Roak BJ, et al. De novo germline and postzygotic mutations in AKT3, PIK3R2 and PIK3CA cause a spectrum of related megalencephaly syndromes. Nat Genet. 2012;44(8):934–40.
Hobert JA, Embacher R, Mester JL, et al. Biochemical screening and PTEN mutation analysis in individuals with autism spectrum disorders and macrocephaly. Eur J Hum Genet. 2014;22(2):273–6.
Fricker M, Tolkovsky AM, Borutaite V, et al. Neuronal Cell Death. Physiol Rev. 2018;98(2):813–80.
Mishra R, Barthwal MK, Sondarva G, et al. Glycogen synthase kinase-3beta induces neuronal cell death via direct phosphorylation of mixed lineage kinase 3. J Biol Chem. 2007;282(42):30393–405.
Boland E, Clayton-Smith J, Woo VG, et al. Mapping of deletion and translocation breakpoints in 1q44 implicates the serine/threonine kinase AKT3 in postnatal microcephaly and agenesis of the corpus callosum. Am J Hum Genet. 2007;81(2):292–303.
Schizophrenia Working Group of the Psychiatric Genomics. Biological insights from 108 schizophrenia-associated genetic loci. Nature. 2014;511(7510):421–7.
Consortium TSW, Ripke S, Walters JT, et al. Mapping genomic loci prioritises genes and implicates synaptic biology in schizophrenia. medRxiv. 2020:2020.09.12.20192922.
Singh T, Poterba T, Curtis D, et al. Exome sequencing identifies rare coding variants in 10 genes which confer substantial risk for schizophrenia. medRxiv. 2020:2020.09.18.20192815.
Polytarchou C, Hatziapostolou M, Yau TO, et al. Akt3 induces oxidative stress and DNA damage by activating the NADPH oxidase via phosphorylation of p47(phox). Proc Natl Acad Sci U S A. 2020;117(46):28806–15.
Iliopoulos D, Polytarchou C, Hatziapostolou M, et al. MicroRNAs differentially regulated by Akt isoforms control EMT and stem cell renewal in cancer cells. Sci Signal. 2009;2(92):62.
Sanidas I, Polytarchou C, Hatziapostolou M, et al. Phosphoproteomics screen reveals akt isoform-specific signals linking RNA processing to lung cancer. Mol Cell. 2014;53(4):577–90.
Tsimberidou AM, Shaw JV, Juric D, et al. Phase 1 study of M2698, a p70S6K/AKT dual inhibitor, in patients with advanced cancer. J Hematol Oncol. 2021;14(1):127.
Li X, Dai D, Chen B, et al. Efficacy of PI3K/AKT/mTOR pathway inhibitors for the treatment of advanced solid cancers: A literature-based meta-analysis of 46 randomised control trials. PLoS ONE. 2018;13(2): e0192464.
Machl A, Wilker EW, Tian H, et al. M2698 is a potent dual-inhibitor of p70S6K and Akt that affects tumor growth in mouse models of cancer and crosses the blood-brain barrier. Am J Cancer Res. 2016;6(4):806–18.
Baselga J, Im SA, Iwata H, et al. Buparlisib plus fulvestrant versus placebo plus fulvestrant in postmenopausal, hormone receptor-positive, HER2-negative, advanced breast cancer (BELLE-2): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol. 2017;18(7):904–16.
Di Leo A, Johnston S, Lee KS, et al. Buparlisib plus fulvestrant in postmenopausal women with hormone-receptor-positive, HER2-negative, advanced breast cancer progressing on or after mTOR inhibition (BELLE-3): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol. 2018;19(1):87–100.
Shalimova A, Babasieva V, Chubarev VN, et al. Therapy response prediction in major depressive disorder: current and novel genomic markers influencing pharmacokinetics and pharmacodynamics. Pharmacogenomics. 2021;22(8):485–503.
Chadha R, Meador-Woodruff JH. Downregulated AKT-mTOR signaling pathway proteins in dorsolateral prefrontal cortex in Schizophrenia. Neuropsychopharmacology. 2020;45(6):1059–67.
Kitagishi Y, Kobayashi M, Kikuta K, et al. Roles of PI3K/AKT/GSK3/mTOR Pathway in Cell Signaling of Mental Illnesses. Depress Res Treat. 2012;2012: 752563.
Roh MS, Seo MS, Kim Y, et al. Haloperidol and clozapine differentially regulate signals upstream of glycogen synthase kinase 3 in the rat frontal cortex. Exp Mol Med. 2007;39(3):353–60.
Pan B, Huang XF, Deng C. Aripiprazole and Haloperidol Activate GSK3beta-Dependent Signalling Pathway Differentially in Various Brain Regions of Rats. Int J Mol Sci. 2016;17(4):459.
Duda P, Hajka D, Wojcicka O, et al. GSK3beta: A Master Player in Depressive Disorder Pathogenesis and Treatment Responsiveness. Cells. 2020;9:3.
Krishnan V, Han MH, Mazei-Robison M, et al. AKT signaling within the ventral tegmental area regulates cellular and behavioral responses to stressful stimuli. Biol Psychiatry. 2008;64(8):691–700.
Wong H, Levenga J, LaPlante L, et al. Isoform-specific roles for AKT in affective behavior, spatial memory, and extinction related to psychiatric disorders. Elife. 2020;9:897.
Ikeda M, Iwata N, Suzuki T, et al. Association of AKT1 with schizophrenia confirmed in a Japanese population. Biol Psychiatry. 2004;56(9):698–700.
Schwab SG, Hoefgen B, Hanses C, et al. Further evidence for association of variants in the AKT1 gene with schizophrenia in a sample of European sib-pair families. Biol Psychiatry. 2005;58(6):446–50.
Xu MQ, Xing QH, Zheng YL, et al. Association of AKT1 gene polymorphisms with risk of schizophrenia and with response to antipsychotics in the Chinese population. J Clin Psychiatry. 2007;68(9):1358–67.
Magno LA, Miranda DM, Neves FS, et al. Association between AKT1 but not AKTIP genetic variants and increased risk for suicidal behavior in bipolar patients. Genes Brain Behav. 2010;9(4):411–8.
Pereira PA, Bicalho MA, de Moraes EN, et al. Genetic variant of AKT1 and AKTIP associated with late-onset depression in a Brazilian population. Int J Geriatr Psychiatry. 2014;29(4):399–405.
Losenkov IS, Vyalova NM, Simutkin GG, et al. An association of AKT1 gene polymorphism with antidepressant treatment response. World J Biol Psychiatry. 2016;17(3):239–42.
Bergeron Y, Bureau G, Laurier-Laurin ME, et al. Genetic Deletion of Akt3 Induces an Endophenotype Reminiscent of Psychiatric Manifestations in Mice. Front Mol Neurosci. 2017;10:102.
Chia S, Gandhi S, Joy AA, et al. Novel agents and associated toxicities of inhibitors of the pi3k/Akt/mtor pathway for the treatment of breast cancer. Curr Oncol. 2015;22(1):33–48.
Ezell SA, Polytarchou C, Hatziapostolou M, et al. The protein kinase Akt1 regulates the interferon response through phosphorylation of the transcriptional repressor EMSY. Proc Natl Acad Sci U S A. 2012;109(10):E613–21.
Thomas CC, Deak M, Alessi DR, et al. High-resolution structure of the pleckstrin homology domain of protein kinase b/akt bound to phosphatidylinositol (3,4,5)-trisphosphate. Curr Biol. 2002;12(14):1256–62.
Lee MS, Feig M, Salsbury FR Jr, et al. New analytic approximation to the standard molecular volume definition and its application to generalized Born calculations. J Comput Chem. 2003;24(11):1348–56.
Feig M, Onufriev A, Lee MS, et al. Performance comparison of generalized born and Poisson methods in the calculation of electrostatic solvation energies for protein structures. J Comput Chem. 2004;25(2):265–84.
Dr. Philip N. Tsichlis wishes to thank Dr. Edward Scolnick (Stanley Center for Psychiatric Research) for initially bringing the Q60H AKT3 mutation to his attention and for facilitating the collaboration of Dr. Tsichlis’ laboratory with Dr. Jeffrey R. Cottrell and Dr. Jen Q. Pan.
This work was supported by the following grants from the National Institutes of Health/National Cancer Institute: R01CAN186729 to Philip N. Tsichlis (Ohio State University), P30 CA016672 (The University of Texas MD Anderson Cancer Center), and P30 CA016058 (Ohio State University). This work was also supported in part by donor funds from Mr. and Mrs. Steven McKenzie, Mr. and Mrs. Zane W. Arrott, and Jamie’s Hope for Dr. Tsimberidou’s Personalized Medicine Program.
Ethics approval and consent to participate
Consent for publication
The authors have the following financial relationships to disclose:
Apostolia M. Tsimberidou: Clinical Trial Research Funding (received through the institution): OBI Pharma, IMMATICS, Parker Institute for Cancer Immunotherapy, Agenus, Tempus, Tvardi, Boston Biomedical, Karus Therapeutics; Consulting or Advisory Role: Vincerx, Diaccurate.
Alan Valentine: Common stock in Merck and Co.
Jamie Shaw is an employee of EMD Serono. Inc. (a business of Merck KGaA, Darmstadt, Germany).
Ursula Hering is an employee of Merck KGaA, Darmstadt, Germany.
Drs. Skliris, Vo, Chan, Armen, Cottrell, and Pen declare no relevant conflict of interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Tsimberidou, AM., Skliris, A., Valentine, A. et al. AKT inhibition in the central nervous system induces signaling defects resulting in psychiatric symptomatology. Cell Biosci 12, 56 (2022). https://doi.org/10.1186/s13578-022-00793-8