Molecular Genetics and Metabolism 114 (2015) 467–473

Contents lists available at ScienceDirect

Molecular Genetics and Metabolism journal homepage: www.elsevier.com/locate/ymgme

Germline activating AKT3 mutation associated with megalencephaly, polymicrogyria, epilepsy and hypoglycemia Mark Nellist a, Rachel Schot a, Marianne Hoogeveen-Westerveld a, Rinze F. Neuteboom b, Elles J.T.M. van der Louw c, Maarten H. Lequin d, Karen Bindels-de Heus e, Barbara J. Sibbles e, René de Coo b, Alice Brooks a, Grazia M.S. Mancini a,⁎ a

Department of Clinical Genetics, Erasmus MC, Sophia Children's Hospital, Rotterdam, The Netherlands Department of Child Neurology, Erasmus MC, Sophia Children's Hospital, Rotterdam, The Netherlands c Department of Dietetics, Erasmus MC, Sophia Children's Hospital, Rotterdam, The Netherlands d Department of Radiology, Erasmus MC, Sophia Children's Hospital, Rotterdam, The Netherlands e Department of Pediatrics, Erasmus MC, Sophia Children's Hospital, Rotterdam, The Netherlands b

a r t i c l e

i n f o

Article history: Received 9 October 2014 Received in revised form 28 November 2014 Accepted 28 November 2014 Available online 5 December 2014 Keywords: AKT3 Hypoglycemia Ketogenic diet MPPH/MCAP

a b s t r a c t Activating germ-line and somatic mutations in AKT3 (OMIM 611223) are associated with megalencephalypolymicrogyria-polydactyly-hydrocephalus syndrome (MPPH; OMIM # 615937) and megalencephaly-capillary malformation (MCAP; OMIM # 602501). Here we report an individual with megalencephaly, polymicrogyria, refractory epilepsy, hypoglycemia and a germline AKT3 mutation. At birth, head circumference was 43 cm (5 standard deviations above the mean). No organomegaly was present, but there was generalized hypotonia, joint and skin laxity, developmental delay and failure to thrive. At 6 months of age the patient developed infantile spasms that were resistant to antiepileptic polytherapy. Recurrent hypoglycemia was noted during treatment with adrenocorticotropic hormone but stabilized upon introduction of continuous, enriched feeding. The infantile spasms responded to the introduction of a ketogenic diet, but the hypoglycemia recurred until the diet was adjusted for increased resting energy expenditure. A novel, de novo AKT3 missense variant (exon 5; c.548TNA, p.(V183D)) was identified and shown to activate AKT3 by in vitro functional testing. We hypothesize that the sustained hypoglycemia in this patient is caused by increased glucose utilization due to activation of AKT3 signaling. This might explain the efficacy of the ketogenic diet in this individual. © 2014 Elsevier Inc. All rights reserved.

1. Introduction Megalencephaly is a developmental disorder characterized by enlargement of the brain [1]. Megalencephaly can occur in association with several genetic syndromes and is often found as part of two overlapping disorders: megalencephaly-polymicrogyria-polydactylyhydrocephalus (MPPH) and megalencephaly-capillary malformation (MCAP) [2]. MPPH and MCAP are characterized by the occurrence of polymicrogyria (PMG), predominantly in the temporo-parietal and perisylvian areas of the cerebral cortex, but sometimes also diffuse through both hemispheres. Ventricular dilation and hydrocephalus can occur, requiring shunting to relieve the increased intracranial pressure. The cerebellum can be normal in size, but crowding of the posterior fossa is common and the Chiari malformation can occur. The corpus callosum is often thickened and enlarged. Capillary malformations, polydactyly and syndactyly are the variable features of MCAP/MPPH. ⁎ Corresponding author at: Department of Clinical Genetics, Erasmus Medical Center, Wytemaweg 80, 3015CN Rotterdam, The Netherlands. Fax: +31 10 7043200. E-mail address: [email protected] (G.M.S. Mancini).

http://dx.doi.org/10.1016/j.ymgme.2014.11.018 1096-7192/© 2014 Elsevier Inc. All rights reserved.

Recently, activating germline and somatic mutations in genes encoding components of the insulin signaling pathway, including the regulatory and catalytic subunits of phosphatidylinositol-3-kinase (PI3K), the vakt murine thymoma viral oncogene homolog (AKT) and the mechanistic target of rapamycin (mTOR) [3–5], were identified in individuals with MCAP/MPPH. In MPPH, mutations have also been described in CCDN2 that encodes a downstream target of insulin-AKT signaling [6]. The PI3K AKT–mTOR signaling cascade is critical for maintaining the balance between anabolic and catabolic metabolism [7] and regulates diverse cellular processes such as growth, proliferation, differentiation and survival [8]. AKT is activated by insulin and other growth and survival factors through the action of PI3K. The lipid products of PI3K bind the pleckstrin homology (PH) domain of AKT to recruit it to the membrane and allow 3-phosphoinositide-dependent protein kinase (PDK1) to phosphorylate a threonine residue in the activation segment of AKT. Together with phosphorylation of a serine residue in the Cterminal hydrophobic domain by mTOR complex 2 (TORC2), this activates AKT resulting in the phosphorylation of a wide range of downstream effectors, including glycogen synthase kinase 3β, TSC2 and the forkhead transcription factor FOXO3 [9] (Fig. 1).

468

M. Nellist et al. / Molecular Genetics and Metabolism 114 (2015) 467–473

Fig. 1. Schematic overview of signaling through AKT3. Receptors in the plasma membrane activate phosphoinositide 3-kinase (PI3K) that phosphorylates phosphatidylinositol (4,5) bisphosphate (PIP2) to produce phosphatidylinositol (3,4,5) trisphosphate (PIP3). PI3K consists of two subunits, the p85 regulatory subunit, encoded by PIK3R2, and the p110 catalytic subunit, encoded by PIK3CA. The N-terminal pleckstrin homology (PH) domain of AKT3 allows it to be recruited to the membrane by PIP3 where AKT3 is phosphorylated by phosphatidylinositol dependent kinase (PDK) at T305. Phosphorylation of AKT3 at T305 by PDK and at S472 by the mechanistic target of rapamycin (mTOR) complex 2 (TORC2) activates AKT3 kinase activity to promote growth, glucose metabolism, fatty acid synthesis, proliferation and survival. AKT3 phosphorylates many downstream targets. Phosphorylation of TSC2 inactivates the TSC1–TSC2 complex that restricts mTOR complex 1 (TORC1) signaling. Among other targets, TORC1 activates the p70 S6 kinase (S6K) that in turn phosphorylates ribosomal protein S6, leading to enhanced protein synthesis. Mutations in the genes PIK3R2, PIK3CA, MTOR, TSC1, TSC2, TBC1D7, CCND2 and AKT3, encoding components of this pathway, have been identified in individuals with (hemi)megalencephaly. The protein products of these genes are indicated in red in the diagram.

There are 3 closely-related isoforms of AKT. AKT1 is essential for growth [10] and somatic AKT1 mutations are associated with Proteus syndrome [11]. AKT2 is essential for glucose homeostasis [12] and AKT2 mutations have been described in individuals with diabetes [13, 14]. AKT3 is expressed in the brain and loss of AKT3 causes microcephaly [15,16] while duplication of AKT3 is associated with macrocephaly [17, 18] and focal cortical dysplasia [19]. Germline and somatic AKT3 mutations cause MCAP/MPPH and hemimegalencephaly, respectively [3–5]. To date, two de novo germline AKT3 mutations, c.1393CNT p.(R465W) and c.686ANG p.(N229S), have been reported [3,5]. In both cases, the proband had megalencephaly with variable connective tissue dysplasia but no polydactyly, syndactyly or vascular malformations. Here we describe an individual with megalencephaly, connective tissue laxity, profound and persistent hypoglycemia, antiepileptic polytherapy-resistant infantile spasms and a de novo germline AKT3 mutation. We employ in vitro assays to investigate the functional effects of AKT3 variants and demonstrate pathogenicity. In the patient, the seizures responded to the ketogenic diet, suggesting that switching to lipid-based metabolism might alleviate some of the symptoms associated with constitutive activation of insulin–PI3K AKT–mTOR signaling in other MPPH/MCAP patients with AKT3 mutations.

2. Materials and methods 2.1. DNA analysis Written informed consent for diagnostic molecular screening and scientific research on patient material was obtained from the family before samples were taken. Genomic DNA was extracted from blood

leukocytes and cultured skin fibroblasts according to standard protocols. All coding exons of the PIK3CA, PIK3R2 and AKT3 genes were assessed by PCR followed by Sanger sequencing on an ABI3130 DNA Analyzer (Applied Biosystems, Foster City, U.S.A.). Primer sequences are available on request and were designed so that exon–intron boundaries could also be assessed. Paternity testing was carried out using the AmpFLSTR Identifiler PCR amplification kit (Life Technologies, Bleiswijk, The Netherlands). Identified variants were compared to variants listed in dbSNP, 1000 Genomes and the Exome Variant Server using ALAMUT version 2.3 software (Interactive Biosoftware, Rouen, France).

2.2. Constructs and antisera A wild-type AKT3 expression construct, encoding AKT3 fused to an N-terminal hemagglutinin (HA) epitope tag and myristoylation signal for membrane targeting was obtained from Addgene (Cambridge, U.S.A.). After removal of the sequence encoding the myristoylation signal sequence by site-directed mutagenesis (SDM), five different AKT3 variants, p.E17K, p.G171R, p.K177M, p.V183D and p.N229S, were derived by SDM. In each case, the complete open-reading frame was sequenced to ensure that no additional mutations were introduced during the mutagenesis procedure. The HA-tagged wild-type p70 S6 kinase expression construct (HA-S6K) was kindly provided by Dr. F. Zwartkruis (University of Utrecht); all other constructs have been described previously [20]. Antibodies were purchased from Cell Signaling Technology, except mouse monoclonals against the HA tag (Covance, Princeton, U.S.A.), GAPDH (Chemicon International, Billerica, U.S.A.) and TSC2

M. Nellist et al. / Molecular Genetics and Metabolism 114 (2015) 467–473

(Life Technologies). Anti-myc and anti-HA affinity beads were from Sigma-Aldrich (Poole, U.K.). 2.3. Functional analysis To AKT3 assess activation, HEK 293T cells were transfected with expression constructs using polyethyleneimine, as described previously [21]. Twenty-four hours after transfection the growth medium (DMEM + 10% fetal calf serum) was replaced with DMEM without serum and incubated for an additional 6 or 12 h. Cells were lysed (50 mM Tris– HCl (pH 7.5), 100 mM NaCl, 50 mM NaF, 1% Triton X-100 containing Complete™ protease inhibitor cocktail (Roche, Molecular Biochemicals, Woerden, The Netherlands)), and analyzed by immunoblotting. For the in vitro analysis of AKT3 kinase activity, affinity-purified TSC1–TSC2 complexes were prepared, as described previously [22] and HEK 293T cells were transfected with AKT3 expression constructs. Twenty-four hours after transfection the growth medium was replaced with DMEM without serum. The cells were lysed 24 h later and the expressed AKT3 variants immunoprecipitated with anti-HA affinity beads. After washing twice with N20 volumes of lysis buffer and twice with N 20 volumes of kinase buffer (25 mM Tris–HCl (pH 7.5), 5 mM β-glycerophosphate, 2 mM DTT, 0.1 mM Na3VO4, 10 mM MgCl2), the affinity beads were mixed with 200 μl purified TSC1–TSC2 complexes in kinase buffer containing 0.2 mM ATP and agitated at 37 °C. Aliquots (20 μl) of the reaction mixture supernatants were removed at the specified time-points and the reactions stopped by incubation for 5 min at 95 °C in Laemmli buffer prior to immunoblotting. 2.4. Molecular modeling predictions Structural figures were generated using PyMOL software, version 1.6 (Schrodinger, Mannheim, Germany). Amino acid prediction algorithms SIFT and PolyPhen-2 were implemented using ALAMUT. 3. Results 3.1. Case report The proband was born to non-consanguineous parents by cesarean section because of breech position after a full term pregnancy. Prenatal ultrasound data were not available. Head circumference (HC) at birth was 43 cm (5 standard deviations (SD) above the mean); weight (4.4 kg) and body length (57 cm) were both 2 SD above the mean. There was generalized hypotonia, joint and skin laxity, but no dysmorphic features. At 6 months, infantile spasms developed that were resistant to antiepileptic polytherapy (vigabatrin, clobazam, adrenocorticotropic hormone (ACTH) and levetiracetam). Glycemic controls during ACTH treatment revealed recurrent hypoglycemia (minimum 1.6 mmol/L). The insulin:glucose ratio was within the normal range, excluding hyperinsulinism. Growth hormone, insulin and cortisol responded normally to glucagon stimulation. Normal blood glucose levels were finally attained by continuous feeding, via a nasogastric tube, of a calorie-enriched diet containing extra complex carbohydrates (uncooked cornstarch). Partial central hypothyroidism was diagnosed due to low levels of free T4 and normal levels of thyroid stimulating hormone (TSH), and treated with thyroxine. No organomegaly was present at abdominal ultrasound. Lumbar puncture was not performed. There was no hyperammonemia and metabolic screening was normal. At 11 months head circumference had increased to 57 cm (7 SD above the mean), while body length remained within the normal range. Seizures did not respond to medication, there was generalized, severe hypotonia, developmental delay with lack of eye contact and abnormal swallowing. Magnetic resonance imaging (MRI) at 7, 18 and 21 months showed megalencephaly with enlarged ventricles, large corpus callosum and persistent cavum septum pellucidum. Furthermore, bilateral perisylvian PMG, a few scattered regions of

469

periventricular nodular heterotopia, and patchy areas of hyperintensity on T2 weighted MRI of the occipital, frontal white matter and basal ganglia were noted (Fig. 2). At the age of 13 months the patient was started on a classical ketogenic diet consisting of long chain fatty acids (88% energy intake), protein (10% energy intake) and carbohydrates (2% energy intake). Seizure activity reduced but hypoglycemia recurred and hyperketosis (N7 mmol/L), hypercholesterolemia (N 8 mmol/L) and hypertriglyceridemia (N 12 mmol/L; up to 21 mmol/L) were noted. Elevated (40% above normal) resting energy expenditure (REE) was measured. Consequently, caloric intake was corrected for increased REE and the glycemic index of the diet was lowered (0.6 g uncooked cornstarch/kg/day). The introduction of medium chain triglycerides into the diet corrected the hyperlipidemia. Blood glucose, ketone, cholesterol and triglyceride levels subsequently normalized and the patient gained weight. After 8 months on this diet, multifocal epileptic activity fell from 50–70% of the electroencephalographic (EEG) registration period to 5–15%. At 2.5 years of age, the individual had developmental delay and severe hypotonia, but showed adequate growth, was seizure free, made eye contact and reached for objects. To prevent hypoglycemia, a maximum interval of 3 h between feeds was maintained. We screened DNA isolated from blood leukocytes from the patient for mutations in AKT3, PIK3CA and PIK3R2. A novel heterozygous AKT3 variant (exon 5; c.548TNA, p.(V183D); reference sequence NM_181690) was identified (Fig. 2F). The variant was not detected in leukocyte DNA from either parent but was confirmed in DNA isolated from fibroblast cultures derived from a skin biopsy from the patient, indicating a probable germline origin. No changes in PIK3CA or PIK3R2 were identified; paternity was confirmed (data not shown).

3.2. In vitro expression studies We were unable to detect AKT3 in human skin fibroblasts by immunoblotting (data not shown). Therefore, to investigate the effect of the V183D substitution on AKT3 activity we generated a series of AKT3 expression constructs encoding different AKT3 variants (Fig. 3A). Somatic AKT3 c.49CNT p.(E17K) mutations have been described in patients with hemimegalencephaly [4,23]; a germ-line c.686ANG p.(N229S) mutation was identified in an individual with MPPH [3]; a c.511GNC p.(G171R) change was detected in a glioblastoma [24]; and the c.530ANT p.K177M variant corresponds to the dominant negative p.K179M variant described for AKT1 [25]. The complete crystal structure of AKT3 has not yet been determined. Therefore, we mapped the variants onto the crystal structure of AKT1 [26] (Fig. 3B), and performed SIFT [27] and PolyPhen-2 [28] prediction analysis. The V183D substitution was predicted to be deleterious (SIFT) and probably damaging (PolyPhen-2). We compared the expression and activation of the AKT3 variants (Fig. 3) and estimated their in vitro kinase activity using purified TSC1–TSC2 complex as a substrate (Fig. 4). The signals for the variants were approximately equal, except for the p.K177M variant. This variant was detected at clearly reduced levels and had a negative effect on TORC1 signaling compared to wild-type AKT3, as shown by the reduction in S6-S235 phosphorylation in serum starvation conditions (Fig. 3B). This is consistent with previous observations with the AKT1 p.K179M paralog [25]. We estimated the activation state of the variants by comparing the signals for T305- and S472-phosphorylated AKT3 to the total AKT3 signals in full growth medium, after 6 h serum starvation and after 12 h serum starvation (Fig. 3C). Consistent with previous observations [29], PDK1-dependent (T305) phosphorylation of the p.E17K variant was increased compared to wild-type AKT3. T305 phosphorylation of the p.V183D and p.N229S variants was also increased, indicating that increased activation of these variants by PDK1 underlies their pathogenicity. We did not observe clear differences in S472 phosphorylation, suggesting that TORC2-dependent phosphorylation of the

470

M. Nellist et al. / Molecular Genetics and Metabolism 114 (2015) 467–473

Fig. 2. Clinical and genetic findings. (A) The affected individual at 10 months of age. Note the disproportionately large head. (B) Axial T2 cranial MRI image of the affected individual at 21 months. The cavum septum pellucidum is visible in the frontal areas, the lateral ventricles are mildly enlarged, periventricular frontal and occipital white matter is hyperintense, and the temporal perisylvian cortex shows abnormalities compatible with polymicrogyria. (C) Axial T2 cranial MRI of an age-matched control individual. (D) Sagittal T1 cranial MRI image of the affected individual at 21 months. The corpus callosum is enlarged and moderately thick, and the cerebellum is large in proportion to the size of the brain. (E) Sagittal T1 cranial MRI of an age-matched control individual. (F) Electropherograms showing the AKT3 c.548TNA p.(V183D) substitution detected in leukocyte DNA from the affected individual (top panel), compared to both parents (middle and bottom panels).

AKT3 hydrophobic domain was not affected by the amino acid substitutions tested. To compare the kinase activity of the variants, we expressed them in HEK 293T cells, as before. After 24 h serum starvation (Fig. 4A), the variants were immunoprecipitated (Fig. 4B) and incubated with purified TSC1–TSC2 complexes (Figs. 4C and D). AKT-dependent phosphorylation of TSC2 at multiple sites disrupts the association of the TSC1– TSC2 complex with RHEB at the lysosomal membrane to activate TORC1 [30] and is independent of TORC2-mediated AKT-S473 phosphorylation [9]. TSC2 was phosphorylated at the T1462 position by wild-type AKT3 and the p.E17K, p.G171R, p.V183D and p.N229S variants. In contrast, no TSC2-T1462 phosphorylation was detected after incubation with the p.K177M variant or S6K. We estimated the rates of TSC2-T1462 phosphorylation for wild-type AKT3 and the p.E17K, p.G171R, p.V183D and p.N229S variants (Fig. 4D). Compared to wildtype, the rate of TSC2-T1462 phosphorylation was increased in the presence of the p.E17K, p.V183D and p.N229S variants, but not in the presence of the p.G171R variant. p.E17K-, p.V183D- and p.N229Sdependent TSC2-T1462 phosphorylation was significantly increased compared to control values (ANOVA repeated measures; p b 0.05). Notably, T305 phosphorylation of these variants was still increased, relative to wild-type AKT3, even after 24 h serum-starvation. 4. Discussion We report megalencephaly, severe developmental delay, hypotonia, seizures and persistent hypoglycemia associated with a de novo germline AKT3 c.548TNA p.(V183D) mutation, leading to a diagnosis of MCAP [2,3]. In vitro functional studies demonstrated that the V183D substitution, as well as the previously reported E17K and N229S substitutions, activate AKT3. In agreement with a previous in vitro analysis of activating AKT mutations, we did not detect clear differences between wild-type AKT3 and the p.G171R variant [31]. Based on homology with AKT1 [26], the V183 residue maps to the PH-domain–kinase domain (KD) interface (Fig. 3B), supporting the hypothesis that amino acid changes that affect the PH-domain–KD interaction can activate AKT [31]. One possibility is that the V183D substitution disrupts hydrophobic contacts between the AKT3 KD and PH domains.

Our analysis suggests that functional assessment could be a useful tool to help establish the pathogenicity of AKT3 variants, especially as the ability of amino acid prediction algorithms such as SIFT and PolyPhen2 to identify activating or gain-of-function mutations is limited [32]. In mammals, three different genes encode distinct AKT isoforms. Mice lacking AKT1 die during embryonic development. In contrast, mice lacking AKT2 and AKT3 are viable, but exhibit impaired glucose homeostasis [33]. AKT3 is essential for mouse brain cell number and size [15], suggesting that AKT3 is the dominant isoform in the brain, and an important positive regulator of mTOR-dependent anabolic metabolism during brain development. In humans copy number variation at the AKT3 locus is associated with abnormal brain size. Individuals with AKT3 deletions show microcephaly [16], while duplications are associated with megalencephaly [17,18]. Furthermore, activating or gainof-function AKT3 mutations are associated with brain overgrowth [3–5]. Apart from thyroid hormone, which was supplemented, insulin and other hormone levels were normal. Therefore we think that it is unlikely that the sustained hypoglycemia in our patient is due to a general metabolic deregulation of carbohydrate metabolism. Instead, we hypothesize that increased glucose utilization in the cerebral compartment, as a result of brain overgrowth due to increased AKT3 activity caused the hypoglycemia in our patient. One possibility is that increased AKT3 activity resulted in increased translocation of GLUT4 from the intracellular endosomal compartment to the plasma membrane, thereby increasing glucose uptake and reducing extracellular levels of glucose [34]. Intracellular glucose is immediately used for anabolic synthesis, which is activated by AKT3 [7,8]. At birth, the head accounts for approximately 25% of the total body weight and the liver has hardly any reserves of glucose stored as glycogen. Therefore, it is possible that increased glucose utilization in the brain could destabilize glucose homeostasis. Although this mechanism might be common to megalencephalies of different causes, it does not explain why hypoglycemia is not observed more frequently in MCAP. Nonetheless, it might help explain the efficacy of the ketogenic diet in our patient. Ketones are an important energy source for the brain and the ketogenic diet is widely used for treating seizures. However, it has not yet been clearly established how the diet works. In addition to reductions in the levels of circulating carbohydrates, activation of ATP-sensitive potassium

M. Nellist et al. / Molecular Genetics and Metabolism 114 (2015) 467–473

471

Fig. 3. Comparison of the AKT3 variants. (A) Schematic overview of AKT3 showing the positions of the variants tested as part of this study, and the sites of PDK1- and TORC2-dependent phosphorylation. The ATP-binding domain (green), PH domain (lime) and kinase domain (KD) (orange) are indicated. (B) Mapping of the AKT3 variants onto the AKT1 crystal structure (Protein Data Bank 3O96). Variants associated with brain overgrowth are shown in red; other variant residues are indicated in light blue. In each case, the wild-type highlighted residue is represented as a stick model. For all other residues only the cartoon representation is shown. The ATP-binding domain is shown in green, the PH-domain in lime green and the KD in orange. (C–F) Transfected HEK 293T cells expressing either wild-type AKT3 (AKT3), the AKT3 variants (E17K, G171R, K177M, V183D or N229S), p70 S6 kinase (S6K), or no exogenous kinase (vector only; mock) were grown in full medium (DMEM + 10% fetal calf serum) or were serum starved for 6 or 12 h prior to immunoblotting (C). Expressed proteins were detected with an antibody against the N-terminal HA-tag; total AKT3 expression was detected with an AKT3-specific antibody. AKT3-T305 and -S472 phosphorylation was detected with phospho-specific antibodies. AKT3-dependent TORC1 activity was estimated using an antibody specific for S235-phosphorylated ribosomal protein S6 (S6-S235). After 12 h serum starvation, S6-S235 phosphorylation was reduced in the mock and AKT3-p.K177M lysates, compared to wild-type AKT3 , S6K and the AKT3 p.E17K, p.G171R, p.V183D and p.N229S variants. Signals for GAPDH are shown as a loading control. Note that phosphorylation of the hydrophobic motif (S473) of the endogenous AKT1 and AKT2 isoforms was detected in the lysates of S6K- and mocktransfected cells. Error bars represent the standard error of the mean. Signals for total (D), T305-phosphorylated (E) and S472-phosphorylated (F) AKT3 were quantified and compared to the wild-type signals (AKT3 = 1.0; indicated with dotted line) in 3 separate experiments. Note that the mean AKT3 p.K177M variant signal is weaker than the other AKT3 variants (D) and that T305 phosphorylation of the p.E17K, p.V183D and p.N229S variants is more than two-fold increased relative to wild-type AKT3 (E).

channels and the inhibition of excitatory glutamatergic synaptic transmission, it has been proposed that the diet inhibits mTOR activity [35,36]. 5. Conclusions The AKT3 c.548TNA p.(V183D) substitution results in activation of AKT3 and PI3K–AKT–mTOR signaling, and causes megalencephaly, PMG and refractory epilepsy. We advise careful control of blood

glucose in individuals with activating AKT3 mutations and MCAP/ MPPH to prevent possible brain damage due to hypoglycemia, as well as early consideration of the ketogenic diet for seizure treatment. Metabolic parameters should be closely monitored to detect hypercholesterolemia and hypertriglyceridemia. Additional studies are required to demonstrate whether the ketogenic diet will be safe and effective for seizure control in other forms of MCAP/MPPH and in megalencephaly in general.

472

M. Nellist et al. / Molecular Genetics and Metabolism 114 (2015) 467–473

Fig. 4. Comparison of the in vitro kinase activity of the AKT3 variants. Transfected HEK 293T cells expressing wild-type AKT3 (AKT3), the AKT3 variants (E17K, G171R, K177M, V183D or N229S), p70 S6 kinase (S6K), or with vector only (mock) were serum-starved for 24 h prior to lysis and immunoprecipitation with anti-HA affinity beads. The washed beads were mixed with purified TSC1–TSC2 complexes and agitated at 37 °C for 1, 5, 10, 15, 20 and 25 min followed by immunoblotting. (A) Expressed proteins were detected with an antibody against the HA-tag; total AKT3 expression was detected with an AKT3 -specific antibody; AKT3 T305 and S472 phosphorylation were estimated with T305 and S472 phospho-specific antibodies. (B) Immunoblot analysis of the anti-HA immunoprecipitates. Immunoprecipitated HA-tagged proteins were detected with an antibody against the HA-tag. Signals for total and T305and S472-phosphorylated AKT3 were estimated as in (A). Total and T1462-phosphorylated TSC2 were also estimated in the supernatant fraction remaining after the kinase assay. (C) In vitro kinase activity of the AKT3 variants. The anti-HA immunoprecipitates were mixed with purified TSC1–TSC2 complexes and agitated at 37 °C for 1, 5, 10, 15, 20 and 25 min followed by immunoblotting for total (TSC2) and T1462-phosphorylated (T1462) TSC2. (D) In vitro kinase activity of the AKT3 variants was determined as shown in (C) in 3 separate experiments. The graph shows the ratios of the T1462 and TSC2 signals in the presence of the different AKT3 variants, relative to wild-type AKT3 (AKT3 = 1.0 at 25 min). Note that TSC2-T1462 phosphorylation occurs more rapidly in the presence of the AKT3 p.E17K, p.V183D and p.N229S variants (red) compared to wild-type AKT3 and the p.G171R variant (blue). No TSC2-T1462 phosphorylation was detected in the presence of immunoprecipitates of the AKT3 p.K177M variant or S6K (green). Error bars indicate the standard error of the mean of three independent experiments.

Conflict of interest statement The authors declare that they have no conflict of interest.

Acknowledgments We thank the family for their cooperation and Dr. Renske Oegema (Erasmus MC, Rotterdam), and Dr. W.B. Dobyns and Dr. G. Mirzaa (University of Washington, Seattle) for useful discussions. Dr. F. Zwartkruis (University of Utrecht) is thanked for providing the HA-S6K expression construct.

References [1] A.J. Barkovich, R. Guerrini, R.I. Kuzniecky, G.D. Jackson, W.B. Dobyns, A developmental and genetic classification for malformations of cortical development: update 2012, Brain 135 (2012) 1348–1369. [2] G.M. Mirzaa, J.-B. Riviere, W.B. Dobyns, Megalencephaly syndromes and activating mutations in the PI3K-AKT pathway: MPPH and MCAP, Am. J. Med. Genet. C: Semin. Med. Genet. 163C (2013) 122–130. [3] J.-B. Riviere, G.M. Mirzaa, B.J. O'Roak, et al., De novo germline and postzygotic mutations in AKT3, PIK3R2 and PIK3CA cause a spectrum of related megalencephaly syndromes, Nat. Genet. 44 (2012) 934–940. [4] J.H. Lee, M. Huynh, J.L. Silhavy, et al., De novo somatic mutations in components of the PI3K–AKT3–mTOR pathway cause hemimegalencephaly, Nat. Genet. 44 (2012) 941–945.

[5] K. Nakamura, M. Kato, J. Tohyama, et al., AKT3 and PIK3R2 mutations in two patients with megalencephaly-related syndromes: MCAP and MPPH, Clin. Genet. 85 (2014) 396–398. [6] G.M. Mirzaa, D.A. Parry, A.E. Fry, et al., De novo CCND2 mutations leading to stabilization of cyclin D2 cause megalencephaly-polymicrogyria-polydactylyhydrocephalus syndrome, Nat. Genet. 46 (2014) 510–515. [7] C.C. Dibble, B.D. Manning, Signal integration by mTORC1 coordinates nutrient input with biosynthetic output, Nat. Cell Biol. 15 (2013) 555–564. [8] M. Laplante, D.M. Sabatini, mTOR signaling in growth control and disease, Cell 149 (2012) 274–293. [9] D.A. Guertin, D.M. Stevens, C.C. Thoreen, et al., Ablation in mice of the mTORC components raptor, rictor, or mLST8 reveals that mTORC2 is required for signaling to AKT-FOXO and PKC, but not S6K1, Dev. Cell 11 (2006) 859–871. [10] W.S. Chen, P.Z. Xu, K. Gottlob, et al., Growth retardation and increased apoptosis in mice with homozygous disruption of the AKT1 gene, Genes Dev. 15 (2001) 2203–2208. [11] M.J. Lindhurst, J.C. Sapp, J.K. Teer, et al., A mosaic activating mutation in AKT1 associated with the Proteus syndrome, N. Engl. J. Med. 365 (2011) 611–619. [12] H. Cho, J. Mu, J.K. Kim, et al., Insulin resistance and a diabetes mellitus-like syndrome in mice lacking the protein kinase AKT2 (PKB beta), Science 292 (2001) 1728–1731. [13] S. George, J.J. Rochford, C. Wolfrum, et al., A family with severe insulin resistance and diabetes due to a mutation in AKT2, Science 304 (2004) 1325–1328. [14] K. Hussain, B. Challis, N. Rocha, et al., An activating mutation of AKT2 and human hypoglycemia, Science 334 (2011) 474. [15] R.M. Easton, H. Cho, K. Roovers, et al., Role for AKT3 /protein kinase B in attainment of normal brain size, Mol. Cell. Biol. 25 (2005) 1869–1878. [16] E. Boland, J. Clayton-Smith, V.G. Woo, 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. 81 (2007) 292–303. [17] D. Wang, S. Zeesman, M.A. Tarnopolsky, M.J.M. Nowaczyk, Duplication of AKT3 as a cause of macrocephaly in duplication 1q43q44, Am. J. Med. Genet. A 161A (2013) 2016–2019.

M. Nellist et al. / Molecular Genetics and Metabolism 114 (2015) 467–473 [18] B.K. Chung, P. Eydoux, C.D. An Karnebeek, W.T. Gibson, Duplication of AKT3 is associated with macrocephaly and speech delay, Am. J. Med. Genet. A (2014), http://dx. doi.org/10.1002/ajmg.a.36521 (Epub ahead of print). [19] V. Conti, M. Pantaleo, C. Barba, et al., Focal dysplasia of the cerebral cortex and infantile spasms associated with somatic 1q21.1-q44 duplication including the AKT3 gene, Clin. Genet. (2014), http://dx.doi.org/10.1111/cge.12476 (Epub ahead of print). [20] M. Nellist, O. Sancak, M.A. Goedbloed, et al., Distinct effects of single amino acid changes to tuberin on the function of the tuberin–hamartin complex, Eur. J. Hum. Genet. 13 (2005) 59–68. [21] M. Hoogeveen-Westerveld, M. Wentink, D. van den Heuvel, et al., Functional assessment of variants in the TSC1 and TSC2 genes identified in individuals with tuberous sclerosis complex, Hum. Mutat. 32 (2011) 424–435. [22] M. Hoogeveen-Westerveld, L. van Unen, A. van den Ouweland, D. Halley, A. Hoogeveen, M. Nellist, The TSC1–TSC2 complex consists of multiple TSC1 and TSC2 subunits, BMC Biochem. 13 (2012) e18. [23] A. Poduri, G.D. Evrony, X. Cai, et al., Somatic activation of AKT3 causes hemispheric developmental brain malformations, Neuron 74 (2012) 41–48. [24] C. Greenman, P. Stephens, R. Smith, et al., Patterns of somatic mutation in human cancer genomes, Nature 446 (2007) 153–158. [25] S. Ramaswamy, N. Nakamura, F. Vazquez, et al., Regulation of G1 progression by the PTEN tumor suppressor protein is linked to inhibition of the phosphatidylinositol 3kinase/AKT pathway, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 2110–2115. [26] W.I. Wu, W.C. Voegtli, H.L. Sturgis, F.P. Dizon, G.P. Vigers, B.J. Brandhuber, Crystal structure of human AKT1 with an allosteric inhibitor reveals a new mode of kinase inhibition, PLoS ONE 5 (2010) e12913.

473

[27] P.C. Ng, S. Henikoff, Predicting the effects of amino acid substitutions on protein function, Annu. Rev. Genomics Hum. Genet. 7 (2006) 61–80. [28] I.A. Adzhubei, S. Schmidt, L. Peshkin, et al., A method and server for predicting damaging missense mutations, Nat. Methods 7 (2010) 248–249. [29] M.A. Davies, K. Stemke-Hale, C. Tellez, et al., A novel AKT3 mutation in melanoma tumours and cell lines, Br. J. Cancer 99 (2008) 1265–1268. [30] S. Menon, C.C. Dibble, G. Talbott, et al., Spatial control of the TSC complex integrates insulin and nutrient regulation of mTORC1 at the lysosome, Cell 156 (2014) 771–785. [31] C. Parikh, V. Janakiraman, W.I. Wu, et al., Disruption of PH-kinase domain interactions leads to oncogenic activation of AKT in human cancers, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 19368–19373. [32] S.E. Flanagan, A.-M. Patch, S. Ellard, Using SIFT and PolyPhen to predict loss-offunction and gain-of-function mutations, Genet. Test. Mol. Biomarkers 14 (2010) 533–537. [33] B. Dummler, O. Tschopp, D. Hynx, Z.Z. Yang, S. Dirnhofer, B.A. Hemmings, Life with a single isoform of AKT: mice lacking AKT2 and AKT3 are viable but display impaired glucose homeostasis and growth deficiencies, Mol. Cell. Biol. 26 (2006) 8042–8051. [34] A.F. Rowland, D.J. Fazakerley, D.E. James, Mapping insulin/GLUT4 circuitry, Traffic 12 (2011) 672–681. [35] N.N. Danial, A.L. Hartman, C.E. Stafstrom, L.L. Thio, How does the ketogenic diet work? Four potential mechanisms, J. Child Neurol. 28 (2013) 1027–1033. [36] S.S. McDaniel, N.R. Rensing, L.L. Thio, K.A. Yamada, M. Wong, The ketogenic diet inhibits the mammalian target of rapamycin (mTOR) pathway, Epilepsia 52 (2011) e7–e11.