Oncogenic Transformation Induced by Membrane-Targeted Akt2 and Akt3

Oncogenic transformation induced by membrane-targeted Akt2 and Akt3

Ines Mende1,3, Scott Malstrom2, Philip N Tsichlis2, Peter K Vogt*,1 and Masahiro Aoki1

1Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, California CA 92037, USA; 2Kimmel Cancer Center, Thomas Jeerson University, Philadelphia, Pennsylvania PA 19107, USA

The kinases Akt2, Akt3 and their myristylated variants, v-P3k induces oncogenic transformation of chicken Myr-Akt2 and Myr-Akt3 were expressed by the RCAS embryo ®broblasts (CEF) in culture and hemangiosar- vector in chicken embryo ®broblasts (CEF). Myr-Akt2 comas in chickens (Chang et al., 1997). The transform- and Myr-Akt3 were strongly oncogenic, inducing multi- ing activity of the P3k proteins is routed through the layered foci of transformed cells. In contrast, wild-type serine-threonine kinase Akt1 (protein kinase Ba), Akt2 and Akt3 were only poorly transforming, their because dominant negative Akt1 interferes with P3k- eciencies of focus formation were more than 100-fold induced transformation (Aoki et al., 1998). The akt1 lower; foci appeared later and showed less multilayering. gene was originally isolated as the oncogene of the Addition of the myristylation signal not only enhanced murine lymphomagenic retrovirus AKT8 (Bellacosa et oncogenic potential but also increased kinase activities. al., 1991). Cellular Akt1 binds with its plekstrin Myr-Akt2 and Myr-Akt3 also induced hemangiosarco- homology (PH) domain to the product of PI3K, D3- mas in the animal, whereas wild type Akt2 and Akt3 PPI, and thus is translocated to the plasma membrane. were not oncogenic in vivo. Furthermore, Akt2, driven by At this location, the protein kinases PDK1 and PDK2 the lck (lymphocyte speci®c kinase) promoter in activate Akt1 by phosphorylation of threonine 308 and transgenic mice, induced lymphomas. The oncogenic serine 473 (Alessi and Cohen, 1998; Chan et al., 1999; eects of Akt2 and Akt3 described here are indis- Coer et al., 1998; Datta et al., 1999). Constitutively tinguishable from those of Akt1. The downstream targets activated and membrane-targeted Akt1 causes focus relevant to oncogenic transformation are therefore formation in CEF and hemangiosarcomas in chickens probably shared by the three Akt kinases. Oncogene identical to the tumors induced by P3k (Aoki et al., (2001) 20, 4419 ± 4423. 1998). The oncogenic activity of the PI3K-Akt path- way is also manifest in human cancer. P110a Keywords: Akt; transformation; serine/threonine kinase (PIK110CA) is ampli®ed and overexpressed in ovarian cancers (Shayesteh et al., 1999). Akt1 is ampli®ed in gastric cancer (Staal, 1987). The function of a negative Signals controlled by PI 3-kinase (PI3K) aect diverse regulator of PI3K and Akt signals, PTEN, is lost in cellular functions including response to growth factors, several types of human tumors including glioblastomas dierentiation, and cell survival (Alessi and Downes, and prostate cancer. This loss of PTEN function leads 1998; Fruman et al., 1998; Shepherd et al., 1998; to high levels of D3-PPI and activation of Akt1 (Ali et Wymann and Pirola, 1998). Oncoproteins of the al., 1999; Cantley and Neel, 1999; Di Cristofano and receptor tyrosine kinase and Src families as well as Pandol®, 2000; Maehama and Dixon, 1999). polyoma middle T antigen activate PI3K through In addition to akt1, there are two related genes in association with the regulatory subunit of PI3K, p85, mammalian genomes, akt2 and akt3. These genes code suggesting a role of PI3K in the actions of these for the Akt2 and Akt3 kinases which show a high proteins (Fukui and Hanafusa, 1989; Hu et al., 1992; degree of sequence homology in their catalytic and PH Peles et al., 1992; Reith et al., 1991; Whitman et al., domains to Akt1 but diverge from Akt1 in other 1985). The Ras oncoprotein can activate PI3K by domains. The two regulatory phosphorylation sites binding to the catalytic subunit directly (Rodriguez- that correspond to threonine 308 and serine 473 of ONCOGENOMICS Viciana et al., 1994, 1996). We previously showed that Akt1 are conserved, and Akt2 and Akt3 also transduce the oncoprotein v-P3k of the avian retroviruses ASV PI3K signals (Altomare et al., 1998; Brodbeck et al., 16 and ASV 8905 codes for the catalytic subunit of 1999; Liu et al., 1998; Masure et al., 1999; Meier et al., PI3K, p110a (Chang et al., 1997; Aoki et al., 2000). 1997; Nakatani et al., 1999a). For both Akt2 and Akt3 there are data suggesting a role in human cancer. Akt2 is ampli®ed and overexpressed in ovarian cancer, breast cancer and pancreatic cancer and Akt3 is *Correspondence: PK Vogt, 10550 N. Torrey Pines Rd., BCC-239, overexpressed in breast and prostate cancer (Bellacosa La Jolla, CA 92037, USA et al., 1995; Cheng et al., 1992, 1996; Miwa et al., 1996; 3 Current address: Universitatsklinikum Essen, Virchowstrasse 173, Nakatani et al., 1999b; Ruggeri et al., 1998). In this 45122 Essen, Germany Received 29 November 2000; revised 20 March 2001; accepted 26 communication, we describe oncogenic transformation March 2001 by Akt2 and Akt3 in experimental systems. The in vitro Neoplastic transformation by Akt2 and Akt3 I Mende et al 4420 transformation assays have obvious utility in screens transforming, about as inecient in focus formation for Akt antagonists. as non-myristylated Akt1 (Table 1). Foci induced by Akt2 (of murine origin) and Akt3 (from rat) were R-Akt2 or R-Akt3 took about 3 weeks to develop and cloned with an HA tag in the avian retroviral showed signi®cantly reduced multilayering. Rare in- expression vector RCAS to yield the constructs R- stances of highly transformed foci in R-Akt2 or R- Akt2 and R-Akt3 (Aoki et al., 1998; Hughes et al., Akt3-infected cultures probably resulted from muta- 1987). Two additional constructs expressed the same tions acquired during retroviral replication of the tagged Akt2 or Akt3 proteins but with the myristyla- constructs. As was shown in a previous study, such tion signal of the c-Src kinase added at the amino mutations occur during RCAS replication and can termini (R-Myr-Akt2 and R-Myr-Akt3). The pre- greatly enhance focus formation mediated by a cellular viously described RCAS constructs expressing Akt1 insert in RCAS (Aoki et al., 2000). Oncogenicity of and myristylated Akt1 (R-Akt1 and R-Myr-Akt1) non-myristylated Akt2 was observed previously by served as controls (Figure 1) (Aoki et al., 1998). another group using NIH3T3 cells (Cheng et al., 1997). Separate CEF cultures were transfected with these The apparent disagreement with our results may be constructs, and supernatant growth medium containing explained by the dierence in the host cells used in the infectious RCAS virus with the respective inserts was two studies. Western blots with anti HA epitope used in assays for focus formation on CEF. R-Myr- antibody detected high levels of all Akt proteins in Akt2 and R-Myr-Akt3 were strongly transforming, infected CEF (Figure 3a). The vestigial transforming inducing foci of multilayered cells within one week as activity of R-Akt2 or R-Akt3 was therefore not caused eciently as R-Myr-Akt1 (Figure 2, Table 1). In by insucient protein expression. Kinase activities contrast, the constructs expressing the wild-type were determined by immune complex kinase assays kinases, R-Akt2 and R-Akt3 were only weakly using anti HA antibody for immunoprecipitation and

Figure 1 Schematic representation of the Akt constructs used in this study. The pleckstrin homology (PH) domain, protein kinase domain, the two regulatory phosphorylation sites, the myristylation signal, and the HA epitope tag are indicated. Constructs of Akt1 and the Myr-Akt1 mutant were described previously. Akt2, and Myr-Akt2 with the myristylation signal of c-Src proteins were subcloned into the avian retroviral vector RCAS.S® via the adapter vector pBSFI (Aoki et al., 1998). The incomplete carboxyl terminal sequences of the rat akt3 cDNA (Konishi et al., 1995) was reconstructed by PCR with a degenerate primer 5'- GCTCTCTAGATTATTC(T/C)CGTCC(A/G)CTTGCAGAGTAG-3' and the internal primer 5'-CAGGGCTCTTGATAAAG- GATCC-3' using rat brain phage cDNA library (lZAP, Stratagene) as a template (Brodbeck et al., 1999; Masure et al., 1999; Nakatani et al., 1999a). After con®rming the sequences, the wild-type and the myristylated version of Akt3 were subcloned into RCAS.S®

Oncogene Neoplastic transformation by Akt2 and Akt3 I Mende et al 4421 a

Figure 3 Expression and in vitro kinase activity of the three Akt Figure 2 Transformation of CEF induced by Akt constructs. proteins. (a) Western blot analysis of the Akt proteins. Cells were Fertilized chicken eggs were obtained from SPAFAS. Primary lysed in Akt lysis buer supplemented with 1 mM microcystin CEF cultures, focus assays and DNA transfections using DMSO- (Calbiochem) (Aoki et al., 1998). Lysates containing 40 mgof polybrene method were described previously (Aoki et al., 1998). protein were separated by SDS ± PAGE and transferred to an Assay plates were evaluated after staining with crystal violet. Immobilon-P membrane (Millipore). The membrane was probed Tests for tumor formation in young chickens followed published with anti HA monoclonal antibody HA-11 (BabCo) followed by procedures (Aoki et al., 1998). Cell culture supernatants of horseradish peroxidase conjugated secondary antibody (Amer- transfected, RCAS construct-releasing CEF were serially diluted sham). Akt proteins were visualized by incubation with a and added to fresh secondary cultures of CEF in 35 mm plates. chemiluminescent substrate (Renaissance Plus, NEN). (b) Immune The cultures were overlaid with nutrient agar for 3 weeks and complex kinase assay. The Akt proteins were immunoprecipitated stained with crystal violet (Aoki et al., 2000). See Table 1 for the from the lysates containing 80 mg of protein with anti HA titer of individual constructs antibody and protein G sepharose beads (Pharmacia). Immune complex kinase assay was performed according to the protocol of Upstate Biotechnology with the Crosstide peptide as a substrate Table 1 Transforming activities of three Akt kinases in vitro and in vivo Infecting constructs FFU/ml Tumor incidence The myristylated Akt1 and Akt2 proteins showed RCAS vector 0 0/3 signi®cantly higher kinase activity than their non- R-Akt1 16103 0/3 myristylated counterparts which did not exceed back- R-Myr-Akt1 66106 3/3 ground with the empty vector alone (Figure 3b). Akt3 R-Akt2 76102 0/4 was unusual because of its high kinase activity 6 R-Myr-Akt2 4610 4/4 (Nakatani et al., 1999a). Even the non-myristylated R-Akt3 86102 0/5 R-Myr-Akt3 46106 5/5 form of Akt3 exceeded the kinase activities of myristylated Akt1 or myristylated Akt2, and that FFU=focus forming units indigenous activity of Akt3 was further elevated by myristylation. Akt3 may therefore possess stronger intrinsic kinase activity either on all or merely on Crosstide, a peptide representing the Akt phosphoryla- selected substrates. Substrate preference between Akt tion site in glycogen synthase kinase 3, as a substrate. kinases has been observed. In the present study, Akt2

Oncogene Neoplastic transformation by Akt2 and Akt3 I Mende et al 4422 proved inecient on histone H2B which is a good substrate for Akt1, whereas both Akt2 and Akt1 eciently phosphorylated myelin basic protein (data not shown). As Akt3 is poorly transforming, high kinase activity cannot be the sole determinant of oncogenicity. This conclusion is also supported by previous observations on an Akt1 mutant in which the activation phosphorylation sites threonine 308 and serine 473 were replaced by aspartic acid (Akt1- T308DS473D). This mutant showed constitutive kinase activity equivalent to that of Myr-Akt1, yet it failed to induce oncogenic transformation (Aoki et al., 1998). The myristylated Akt2 or myristylated Akt3 proteins were also oncogenic in the animal. Injection of R-Myr- Akt2- or R-Myr-Akt3-producing CEF into the wing web of young chickens caused aggressively growing hemangiosarcomas in all inoculated birds within one Figure 4 (a) Schematic representation of the Myr-Akt2 transgene used in this study. The myristylation signal and HA epitope week. The histopathology of these tumors was identical tag are indicated. Myr-Akt2 was subcloned from pcDNA3-Myr- to the hemangiosarcomas induced by R-Myr-Akt1 or Akt2 into HindIII ± XbaI sites of CMV5 plasmid, and then the myristylated p110 of PI3K. The RCAS constructs BglII ± BamHI fragment of Myr-Akt2 was transferred to pLCK expressing the non-myristylated forms of the Akt vector digested with BamHI (Abraham et al., 1991; Andersson et al., 1989). The pLCK/Myr-Akt2 construct was sequenced to kinases did not induce tumors in vivo. These in vitro ensure proper orientation of the insert. The insert containing the and in vivo observations suggest that besides kinase lck (lymphocyte speci®c kinase) promoter, Myr-Akt2, and the activity, constitutive membrane localization is a human growth hormone gene polyadenylation signal was isolated prerequisite for the oncogenicity of Akt kinases. from the vector using NotI, puri®ed, and then microinjected into To test the oncogenic potential of Myr-Akt2 in C57Bl/6J X C3H/HeJ F2 mouse oocytes as described previously (Ceci et al., 1989, 1991; Osborn et al., 1987). To identify mammals, we generated Myr-Akt2 transgenic mice transgenic founder mice, genomic DNA was isolated from the under the control of the lck gene promoter (Figure 4a). tails of live-born weanling mice (Siracusa et al., 1987), DNA was The lck promoter was chosen because it has been used digested to completion with SacI and analysed by Southern to direct transgene expression in T lymphocytes, and blotting using a 7700 bp probe obtained by KprI±SacI digestion of pCMV/Akt2. (b) Western blot analysis of the Myr-Akt2 AKT8 virus containing the akt1 oncogene speci®cally protein in a Myr-Akt2 transgenic mouse and its non transgenic induces thymomas in mice. We followed the survival of littermate these mice for 600 days. Twenty-three of the thirty- three transgenics died during this period, 18 with a diagnosis of lymphoma and ®ve of undetermined are a loss of function in the tumor suppressor PTEN or causes. In the same period, six of the 30 controls died; mutations in one of the downstream signaling two had splenomegaly, the other four showed no components that are responsive to Akt. An elevated evidence of neoplastic disease. Expression of the Myr- basal activity of Akt3 kinase was found in a prostate Akt2 transgene in thymoma cells was con®rmed by cancer cell line PC-3, which over-expresses Akt3 and Western blot using anti Akt2 antibody (Figure 4b). lacks PTEN (Nakatani et al., 1999b). Akt1, Akt2 and Akt3 are functionally similar in their The Akt kinases phosphorylate directly or indirectly eects on cellular growth; membrane-targeted versions diverse substrates. Examples are the Fox-O transcrip- of these proteins are oncogenic in vitro and in vivo. All tion factors FKHR, FKHR-L1 and AFX, nitric oxide three induce hemangiosarcomas in chickens; Akt1 and synthase, the eukaryotic initiation factor 4E-binding Akt2 are also lymphomagenic in mice. This dierence protein (4E-BP1), the pro-apoptotic protein BAD, the in tumor spectrum between avian and mammalian kinases IKK, mTOR, GSK-3b and p70S6k (IkB species may be more apparent than real: Injected kinase, mammalian target of rapamycin, glycogen chickens develop hemangiosarcomas so rapidly that the synthase kinase 3b, p70 S6 kinase (Alessi and Cohen, typically slower arising lymphomas would be missed, 1998; Chan et al., 1999; Coer et al., 1998; Datta et al., and the murine injections are biased toward expression 1999). Substrate selection may be aected by cellular in lymphocytes by the lck promoter and probably also localization (Dufner et al., 1999). In CEF, myristylated by the LTR of the AKT8 retrovirus. Akt1 but not the nonmyristylated yet kinase-active The oncogenic potential of Akt2 and of Akt3 mutant Akt1-T308DS473D induces phosphorylation of uncovered in this study suggests that the frequent p70 S6k and 4E-BP1 (Aoki et al., 2001). Since the Akt ampli®cation and upregulation of these kinases in kinases are indistinguishable in their oncogenic action human tumors may play an important role in the in vitro and in vivo, it is likely that they share essential induction and maintenance of the neoplastic pheno- oncogenic targets which now must be identi®ed in type. However, since overexpression of any wild-type order to understand the mechanisms of Akt and of Akt protein is insucient to cause transformation, PI3K-induced transformation. Phosphorylation of a set other genetic changes aecting PI3K signaling probably of these Akt targets will be both necessary and occur in these tumors. Candidates for such mutations sucient for oncogenic transformation.

Oncogene Neoplastic transformation by Akt2 and Akt3 I Mende et al 4423 Acknowledgments thank Dr Ushio Kikkawa for generously providing the Supported by US Public Health Service Research Grants akt3 construct. This is manuscript number 13437 of the CA 78230, CA 42564 and CA 79616. Douglas Geerdes and Department of Molecular and Experimental Medicine, The Jeery Ludwig provided valuable technical assistance. We Scripps Research Institute.

References

Abraham KM, Levin SD, Marth JD, Forbush KA and Fukui Y and Hanafusa H. (1989). Mol. Cell. Biol., 9, 1651 ± Perlmutter RM. (1991). Proc. Natl. Acad. Sci. USA, 88, 1658. 3977 ± 3981. Hu P, Margolis B, Skolnik EY, Lammers R, Ullrich A and Alessi DR and Cohen P. (1998). Curr.Opin.Genet.Dev.,8, Schlessinger J. (1992). Mol. Cell. Biol., 12, 981 ± 990. 55 ± 62. Hughes SH, Greenhouse JJ, Petropoulos CJ and Sutrave P. Alessi DR and Downes CP. (1998). Biochim. Biophys. Acta, (1987). J. Virol., 61, 3004 ± 3012. 1436, 151 ± 164. KonishiH,KurodaS,TanakaM,MatsuzakiH,OnoY, Ali IU, Schrimi LM and Dean M. (1999). J. Natl. Cancer Kameyama K, Haga T and Kikkawa U. (1995). Biochem. Inst., 91, 1922 ± 1932. Biophys. Res. Commun., 216, 526 ± 534. Altomare DA, Lyons GE, Mitsuuchi Y, Cheng JQ and Testa Liu AX, Testa JR, Hamilton TC, Jove R, Nicosia SV and JR. (1998). Oncogene, 16, 2407 ± 2411. Cheng JQ. (1998). Cancer Res., 58, 2973 ± 2977. Andersson S, Davis DL, Dahlback H, Jornvall H and Russell Maehama T and Dixon JE. (1999). Trends Cell. Biol., 9, DW. (1989). J. Biol. Chem., 264, 8222 ± 8229. 125 ± 128. Aoki M, Batista O, Bellacosa A, Tsichlis P and Vogt PK. Masure S, Haefner B, Wesselink JJ, Hoefnagel E, Mortier E, (1998). Proc. Natl. Acad. Sci. USA, 95, 14950 ± 14955. Verhasselt P, Tuytelaars A, Gordon R and Richardson A. Aoki M, Blazek E and Vogt PK. (2001). Proc. Natl. Acad. (1999). Eur. J. Biochem., 265, 353 ± 360. Sci. USA, 98, 136 ± 141. Meier R, Alessi DR, Cron P, Andjelkovic M and Hemmings Aoki M, Schetter C, Himly M, Batista O, Chang HW and BA. (1997). J. Biol. Chem., 272, 30491 ± 30497. Vogt PK. (2000). J. Biol. Chem., 275, 6267 ± 6275. Miwa W, Yasuda J, Murakami Y, Yashima K, Sugano K, Bellacosa A, de Feo D, Godwin AK, Bell DW, Cheng JQ, Sekine T, Kono A, Egawa S, Yamaguchi K, Hayashiazaki Altomare DA, Wan M, Dubeau L, Scambia G, Masciullo Y and Sekiya T. (1996). Biochem. Biophys. Res. Commun., V et al. (1995). Int. J. Cancer, 64, 280 ± 285. 225, 968 ± 974. Bellacosa A, Testa JR, Staal SP and Tsichlis PN. (1991). Nakatani K, Sakaue H, Thompson DA, Weigel RJ and Roth Science, 254, 274 ± 277. RA. (1999a). Biochem. Biophys. Res. Commun., 257, 906 ± Brodbeck D, Cron P and Hemmings BA. (1999). J. Biol. 910. Chem., 274, 9133 ± 9136. Nakatani K, Thompson DA, Barthel A, Sakaue H, Liu W, Cantley LC and Neel BG. (1999). Proc. Natl. Acad. Sci. Weigel RJ and Roth RA. (1999b). J. Biol. Chem., 274, USA, 96, 4240 ± 4245. 21528 ± 21532. Ceci JD, Kovatch RM, Swing DA, Jones JM, Snow CM, Osborn L, Rosenberg MP, Keller SA and Meisler MH. Rosenberg MP, Jenkins NA, Copeland NG and Meisler (1987). Mol. Cell. Biol., 7, 326 ± 334. MH. (1991). Oncogene, 6, 323 ± 332. Peles E, Lamprecht R, Ben-Levy R, Tzahar E and Yarden Y. Ceci JD, Siracusa LD, Jenkins NA and Copeland NG. (1992). J. Biol. Chem., 267, 12266 ± 12274. (1989). Genomics, 5, 699 ± 709. Reith AD, Ellis C, Lyman SD, Anderson DM, Williams DE, Chan TO, Rittenhouse SE and Tsichlis PN. (1999). Ann. Rev. Bernstein A and Pawson T. (1991). EMBO J., 10, 2451 ± Biochem., 68, 965 ± 1014. 2459. Chang HW, Aoki M, Fruman D, Auger KR, Bellacosa A, Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck Tsichlis PN, Cantley LC, Roberts TM and Vogt PK. B, Gout I, Fry MJ, Water®eld MD and Downward J. (1997). Science, 276, 1848 ± 1850. (1994). Nature, 370, 527 ± 532. ChengJQ,AltomareDA,KleinMA,LeeWC,KruhGD, Rodriguez-Viciana P, Warne PH, Vanhaesebroeck B, Water- Lissy NA and Testa JR. (1997). Oncogene, 14, 2793 ± 2801. ®eld MD and Downward J. (1996). EMBO J., 15, 2442 ± Cheng JQ, Godwin AK, Bellacosa A, Taguchi T, Franke TF, 2451. Hamilton TC, Tsichlis PN and Testa JR. (1992). Proc. Ruggeri BA, Huang L, Wood M, Cheng JQ and Testa JR. Natl. Acad. Sci. USA, 89, 9267 ± 9271. (1998). Mol. Carcinog., 21, 81 ± 86. ChengJQ,RuggeriB,KleinWM,SonodaG,AltomareDA, Shayesteh L, Lu Y, Kuo WL, Baldocchi R, Godfrey T, Watson DK and Testa JR. (1996). Proc. Natl. Acad. Sci. Collins C, Pinkel D, Powell B, Mills GB and Gray JW. USA, 93, 3636 ± 3641. (1999). Nat. Genet., 21, 99 ± 102. Coer PJ, Jin J and Woodgett JR. (1998). Biochem. J., 335, Shepherd PR, Withers DJ and Siddle K. (1998). Biochem. J., 1±13. 333, 471 ± 490. Datta SR, Brunet A and Greenberg ME. (1999). Genes Dev., Siracusa LD, Russell LB, Jenkins NA and Copeland NG. 13, 2905 ± 2927. (1987). Genetics, 117, 85 ± 92. Di Cristofano A and Pandol® PP. (2000). Cell, 100, 387 ± Staal SP. (1987). Proc. Natl. Aad. Sci. USA, 84, 5034 ± 5037. 390. WhitmanM,KaplanDR,SchahausenB,CantleyLand Dufner A, Andjelkovic M, Burgering BM, Hemmings BA Roberts TM. (1985). Nature, 315, 239 ± 242. and Thomas G. (1999). Mol. Cell. Biol., 19, 4525 ± 4534. Wymann MP and Pirola L. (1998). Biochim. Biophys. Acta, Fruman DA, Meyers RE and Cantley LC. (1998). Ann. Rev. 1436, 127 ± 150. Biochem., 67, 481 ± 507.

Oncogene