Oncogene (2005) 24, 7482–7492
`& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00
`
`www.nature.com/onc
`
`The Akt/PKB pathway: molecular target for cancer drug discovery
`
`Jin Q Cheng*,1, Craig W Lindsley2, George Z Cheng3, Hua Yang1 and Santo V Nicosia1
`
`1Departments of Pathology and Interdisciplinary Oncology, H Lee Moffitt Cancer Center and Research Institute, University of South
`Florida College of Medicine, 12902 Magnolia Drive, SRB3, Tampa, FL 33612, USA; 2Department of Medicinal Chemistry, Merck &
`Co., West Point, PA 19486, USA; 3Department of Microbiology, Mount Sinai School of Medicine, New York, NY 10029, USA
`
`The serine/threonine kinase Akt/PKB pathway presents
`an exciting new target for molecular therapeutics, as it
`functions as a cardinal nodal point
`for transducing
`extracellular (growth factor and insulin) and intracellular
`(receptor
`tyrosine kinases, Ras and Src) oncogenic
`signals. In addition, alterations of the Akt pathway have
`been detected in a number of human malignancies. Ectopic
`expression of Akt, especially constitutively activated Akt,
`is sufficient to induce oncogenic transformation of cells
`and tumor formation in transgenic mice as well as
`chemoresistance. Akt has a wide range of downstream
`targets that regulate tumor-associated cell processes such
`as cell growth, cell cycle progression, survival, migration,
`epithelial–mesenchymal
`transition
`and
`angiogenesis.
`Blockage of Akt signaling results in apoptosis and growth
`inhibition of tumor cells with elevated Akt. The observed
`dependence of certain tumors on Akt signaling for survival
`and growth has wide implications for cancer therapy,
`offering the potential for preferential tumor cell killing. In
`the last several years, through combinatorial chemistry,
`high-throughput and virtual screening, and traditional
`medicinal chemistry, a number of inhibitors of the Akt
`pathway have been identified. This review focuses on
`ongoing translational efforts to therapeutically target the
`Akt pathway.
`Oncogene (2005) 24, 7482–7492. doi:10.1038/sj.onc.1209088
`
`Keywords: Akt; cancer; therapeutics; inhibitor
`
`Akt proteins: overview and rationale as antitumor targets
`
`Akt was originally discovered as an oncogene trans-
`duced by the acute transforming retrovirus (Akt-8),
`which was isolated from an AKR thymoma (Staal et al.,
`1977; Staal, 1987), and subsequently found to encode a
`serine/threonine protein kinase (Bellacosa et al., 1991).
`Akt is also known as protein kinase B (Coffer and
`Woodgett, 1991) and RAC-PK (Jones et al., 1991). Viral
`akt highly activated and oncogenic due to the fact that
`v-akt is associated with the cell membrane through a
`myristylated Gag protein sequence fused to the N-
`terminus of Akt (Bellacosa et al., 1991). The important
`
`*Correspondence: JQ Cheng; E-mail: chengjq@moffitt.usf.edu
`
`role of Akt in transformation and cancer was shortly
`thereafter strengthened by the cloning of the AKT2 gene
`(Cheng et al., 1992) and the discovery that AKT2 is
`frequently amplified and overexpressed in human
`cancers (Cheng et al., 1992, 1996; Bellacosa et al.,
`1995). To date, three Akt family members have been
`identified in mammals, designated Akt1/PKBa, Akt2/
`PKBb and Akt3/PKBg (Testa and Bellacosa, 2001). The
`members of Akt family share similar domain structure
`and are activated by various stimuli in a phosphatidy-
`linositol 3-kinase (PI3K)-dependent manner (Burgering
`and Coffer, 1995; Franke et al., 1995; Liu et al., 1998;
`Shaw et al., 1998). Activation of Akt depends on the
`integrity of
`the pleckstrin homology (PH) domain,
`which mediates its membrane translocation, and on
`the phosphorylation of Thr308 in the activation loop and
`Ser473 (Chan et al., 1999; Datta et al., 1999; Testa and
`Bellacosa, 2001; Brazil et al., 2002). Phosphoinositides,
`PtdIns-3,4-P2 and PtdIns-3,4,5-P3, produced by PI3 K
`bind directly to the PH domain of Akt, driving a
`conformational change in the molecule, which enables
`the activation loop of Akt to be phosphorylated by
`PDK1 at Thr308 (Alessi et al., 1997). Full activation of
`AKT1 is also associated with phosphorylation of Ser473
`(Alessi et al., 1996) within a C-terminal hydrophobic
`motif characteristic of kinases in the AGC kinase family.
`Although the role of PDK1 in Thr308 phosphorylation is
`well established, the mechanism of Ser473 phosphoryla-
`tion is controversial. A number of candidate enzymes
`responsible for this modification have been put forward,
`including integrin-linked kinase (Persad et al., 2001),
`PDK1 when in a complex with the kinase PRK2
`(Balendran et al., 1999), Akt itself, through autopho-
`sphorylation (Toker and Newton, 2000), PKCa (Parto-
`vian and Simons, 2004), PKCbII (Kawakami et al.,
`2004), DNA-dependent kinase (Feng et al., 2004), and
`the rictor-mTOR complex (Sarbassov et al., 2005). The
`activity of Akt
`is negatively regulated by tumor
`suppressor PTEN, which is frequently mutated in
`human malignancy (Li et al., 1997; Steck et al., 1997;
`Parsons, 2004). PTEN encodes a dual-specificity protein
`and lipid phosphatase that reduces intracellular levels of
`PtdIns-3,4,5-P3 by converting them to PtdIns-4,5-P2,
`thereby inhibiting the PI3K/Akt pathway (Stambolic
`et al., 1998). Akt phosphorylates and/or interacts with a
`number of molecules
`to exert
`its normal cellular
`functions, which include roles in cell proliferation,
`survival and differentiation (Chan et al., 1999; Datta
`
`NOVARTIS EXHIBIT 2051
`Par v. Novartis, IPR 2016-01479
`Page 1 of 11
`
`

`

`7483
`
`et al., 1999; Testa and Bellacosa, 2001; Brazil et al.,
`2002). Gene knockout studies have defined the biologi-
`cal
`importance of Akt members in normal cells. In
`particular, Akt2-null mice develop typical
`type II
`diabetes (Cho et al., 2001a, b), while Akt1- and Akt3-
`deficient mice do not display a diabetic phenotype but
`exhibit a decrease in the sizes of all organs and a
`selective impairment of brain development, respectively
`(Chen et al., 2001; Cho et al., 2001a, b; Easton et al.,
`2005; Tschopp et al., 2005). Moreover, although Akt1-
`and Akt3-deficient brains are reduced in size to
`approximately the same degree, the absence of Akt1
`reduces neuronal cell number, whereas the lack of Akt3
`results in smaller and fewer cells in which mTOR
`signaling is attenuated (Easton et al., 2005).
`Several lines of evidence suggest that Akt is a critical
`target for anticancer drug discovery. First, Akt sits at
`the crossroads of multiple oncogenic and tumor
`suppressor signaling networks (see related Reviews in
`this issue). Almost all known oncogenic growth factors,
`angiogenic factors and cytokines activate Akt by
`binding to cognate receptors on cell surface. Further,
`Akt is also activated by steroid hormones, such as
`estrogen and androgen through a mechanism indepen-
`dent of their nuclear receptors (Sun et al., 2001a; Sun
`et al., 2003). In addition, Akt is shown to be activated by
`constitutively active Ras and Src (Datta et al., 1996; Liu
`et al., 1998). Second, frequent deregulations of many
`components of the Akt signaling pathway have been
`observed in human cancer (see review by Altomare and
`Testa in this issue). Of the Akt family, only the AKT2
`gene is frequently amplified in human cancer. Further,
`overexpression of AKT2 RNA and/or protein is also
`more commonly observed in human cancer than are
`AKT1 and AKT3 (Testa and Bellacosa, 2001; Kim
`et al., 2005). However, recurrent activation of the three
`Akt family members has been detected in a variety of
`types of human malignancy (Yuan et al., 2000; Sun
`et al., 2001a, b, 2003; Altomare et al., 2003, 2004, 2005;
`Balsara et al., 2004). Activation of Akt is primarily the
`result of aberrant upstream molecules of Akt, which
`include overproduction of growth factors, upregulation
`and/or mutation of receptor tyrosine kinases, Ras and
`Src as well as PIK3CA and PTEN. While a point
`mutation of AKT2 has been reported in familial diabetes
`(George et al., 2004), a dominant mutation of Akt has
`not been identified in human tumor. Third, ectopic
`expression of constitutively active Akt and even wild-
`type Akt2 results in oncogenic transformation in vitro
`and in vivo (Cheng et al., 1997; Hutchinson et al., 2001;
`Malstrom et al., 2001; Mende et al., 2001; Sun et al.,
`2001a, b; Majumder et al., 2003). Furthermore, a
`number of studies have shown that overexpression
`and/or activation of Akt render tumor cells resistant
`to chemotherapeutic drugs and signal molecule inhibi-
`tors such as Gleevec, Iressa, Herceptin and retinoid acid
`(Cheng et al., 2002; Arlt et al., 2003; Knuefermann et al.,
`2003; Yuan et al., 2003; Nagata et al., 2004). In
`addition, Akt targets many signal molecules to regulate
`tumor development-associated cell processes such as
`apoptosis, cell proliferation, differentiation, migration
`
`Akt as target for anticancer drug discovery
`JQ Cheng et al
`
`and angiogenesis. Finally, knockdown of Akt by
`antisense or siRNA significantly reduces tumor growth
`and invasiveness and induces apoptosis and cell growth
`arrest only in tumor cells overexperssing Akt (Cheng
`et al., 1996; Chen et al., 2001; Asnaghi et al., 2004; Remy
`et al., 2004; Tabellini et al., 2005).
`These observations make Akt an attractive target for
`anticancer drug discovery, and it has been postulated
`that inhibition of Akt alone or in combination with
`standard cancer chemotherapeutics will reduce the
`apoptotic threshold and preferentially kill cancer cells.
`The development of specific and potent inhibitors will
`allow this hypothesis to be tested in animals. The
`majority of small molecule inhibitors in this nascent field
`are classic ATP-competitive inhibitors, which provide
`little specificity. Phosphatidylinositol (PI) analogs have
`been reported to inhibit Akt, but these inhibitors may
`also have specificity problems with respect to other PH
`domain containing proteins and may have poor
`bioavailability. Recently, small chemical compounds
`triciribine/Akt/protein kinase B inhibitor-2 (API-2) and
`allosteric inhibitors have been reported which are PH
`domain dependent, and the latter also exhibit Akt
`isozyme selectivity.
`In addition,
`inhibitors
`toward
`upstream regulators and downstream targets of Akt
`have also been tested for their capability of reversing the
`phenotype of cancer cells expressing altered Akt. This
`review focuses on the ongoing efforts to therapeutically
`target individual components of Akt pathway including
`Akt
`itself as well as its upstream regulators and
`downstream effectors (Figure 1). Some of these efforts
`involve AKT-specific inhibition based on structure-
`based analyses (see also the Review by Kumar and
`Madison in this issue).
`
`Therapeutic targeting of upstream regulators of Akt
`
`PDK1 inhibitors
`
`PDK1 is a serine/threonine protein kinase that can
`phosphorylate and activate a number of kinases in the
`AGC kinase superfamily (Mora et al., 2004). The first
`identified and best characterized PDK1 substrates are
`the three members of the Akt family (Mitsiades et al.,
`2004). PDK1 phosphorylates the activation loop of Akt
`(also called the T-loop) on residue Thr308, which
`primarily regulates Akt activation (Alessi et al., 1997).
`Therefore, a PDK1 inhibitor should significantly block
`activation of Akt.
`Three potent PDK1 inhibitors, BX-795, BX-912 and
`BX-320 (Figure 2a), recently identified by screening of
`compound libraries, have IC50 between 11 and 30 nM
`(Feldman et al., 2005). The inhibitors blocked PDK1/
`Akt signaling in tumor cells resulting in the inhibition of
`anchorage-independent growth and the induction of
`apoptosis in a variety of tumor cell lines. A number of
`cancer cell lines with elevated Akt activity were >30-
`fold more sensitive to growth inhibition by PDK1
`inhibitors in soft agar than on tissue culture plastic,
`consistent with the cell survival function of the PDK1/
`
`Oncogene
`
`NOVARTIS EXHIBIT 2051
`Par v. Novartis, IPR 2016-01479
`Page 2 of 11
`
`

`

`7484
`
`Akt as target for anticancer drug discovery
`JQ Cheng et al
`
`PIP3
`competitor
`
`Akt
`inhibitor
`
`PI-3,4,
`
`5-P3
`
`PH
`
`Kinase
`domain
`
`Akt
`-T308-p
`
`RD
`
`-S473-p
`
`PTEN
`
`PI 3-kinase
`
`p110
`
`Src
`
`p85
`
`PI-4,
`5-P2
`
`TRK
`inhibitor
`
`RAS
`
`PDK1
`
`FTI/GGTI
`
`PI3K
`inhibitor
`
`PDK1
`inhibitor
`
`mTOR
`
`mTOR
`inhibitor
`
`p70S6K
`
`4E-BP1
`
`5’top mRNA translation
`
`Cap dependent translation
`
`Figure 1 Therapeutic targeting of the Akt pathway including Akt itself as well as its upstream regulators and downstream effectors
`
`Figure 2 Compound structures for inhibitors of PDK1 (a) and PI3K (b)
`
`Akt signaling pathway, which is particularly important
`for unattached cells. BX-320 inhibited the growth of
`LOX melanoma tumors in the lungs of nude mice after
`injection of tumor cells into the tail vein. The effect of
`BX-320 on cancer cell growth in vitro and in vivo
`indicates that PDK1 inhibitors may have clinical utility
`as anticancer agents.
`The staurosporine derivative UCN-01 (7-hydroxys-
`taurosporine), a drug now in clinical trials, has been
`shown to potently inhibit PDK1 (IC50¼ 33 nM) in vitro.
`UCN-01-induced PDK1 inhibition was also observed in
`
`human tumor xenografts (Sato et al., 2002). Over-
`expression of constitutively active Akt diminished the
`cytotoxic effects of UCN-01, suggesting that UCN-01
`may in part exert its cytotoxicity by inhibiting PDK1/
`Akt
`survival pathway. Crystal
`structure analyses
`showed that staurosporine and UCN-01 form a complex
`with the kinase domain of PDK1 (Komander et al.,
`2003). Although staurosporine and UCN-01 interact
`with the PDK1 active site in an overall similar manner,
`the UCN-01 7-hydroxy group (Figure 2a), which is not
`present in staurosporine, generates direct and water-
`
`Oncogene
`
`NOVARTIS EXHIBIT 2051
`Par v. Novartis, IPR 2016-01479
`Page 3 of 11
`
`

`

`7485
`
`mediated hydrogen bonds with active-site residues. This
`moiety is hydrogen-bonded directly to Thr222 and
`indirectly via an ordered water molecule to Gln220 of
`PDK1 (Zhao et al., 2002; Johnson and Pinto, 2002;
`Komander et al., 2003). A different water-mediated
`hydrogen-bonding network is also observed in other
`UCN-01 complexes and might serve as a starting point
`for further structure-based optimization. In addition,
`recent studies show that UCN-01 inhibits other kinases
`such as PKC and Chk1 and transcriptionally upregu-
`lates the cyclin-dependent kinase inhibitor p21waf1/cip1
`(Senderowicz, 2003a, b).
`
`PI3K inhibitors
`
`Dissection of PI3K/Akt signaling pathway has been
`aided greatly by two pharmacological PI3K inhibitors,
`wortmannin and LY294002 (Figure 2b). Wortmannin is
`a fungal metabolite and a potent inhibitor of type I
`PI3K, with an IC50 range for inhibition of PI3K from 2
`to 4 nM (Arcaro and Wymann, 1993; Vlahos et al.,
`1994). Wortmannin inhibits PI3K activity by binding
`covalently to a conserved lysine residues in the ATP-
`binding site of the enzyme (Wymann et al., 1996).
`Wortmannin has antitumor activity in vitro and in vivo,
`suggesting that it might offer a valuable approach to
`treat cancer. However, a major disadvantage of the use
`of wortmannin is its stability in an aqueous environ-
`ment. Wortmannin is soluble in organic solvents, which
`may limit its use in clinical trials. Currently, water-
`soluble wortmannin conjugates are being developed to
`circumvent this issue. LY294002 is a flavonoid deriva-
`tive and a reversible, ATP-competitive inhibitor with
`IC50 for recombinant PI3K in the low micromolar
`range. A number of in vitro studies have shown that
`LY294002 alone has antiproliferative and proapoptotic
`activities (Wetzker and Rommel, 2004). Relatively, few
`in vivo studies have been conducted to demonstrate the
`efficacy of LY294002 on the inhibition of tumor growth,
`but
`these
`studies
`showed that administration of
`LY294002 in human cancer xenografts inhibited tumor
`growth and induced apoptosis (Semba et al., 2002; Fan
`et al., 2003). Although inhibition of the PI3K/Akt
`pathway by wortmannin or LY294002 alone may inhibit
`cell proliferation, promote apoptosis and/or inhibit
`tumor growth,
`the combination of wortmannin or
`LY294002 with traditional cytotoxic drugs or radiation
`enhances the effectiveness of these treatments (Wetzker
`and Rommel, 2004).
`It
`is noteworthy that neither wortmannin nor
`LY294002 displays selectivity for different members of
`the class I PI3K (Finan and Thomas, 2004). Wortman-
`nin also inhibits class III PI3K to reduce autophagy, a
`class II programmed cell death (Shintani and Klionsky,
`2004). LY294002 also inhibits casein kinase 2 with
`similar potency to PI3K. At higher concentrations,
`wortmannin inhibits PI3K-related enzymes, such as
`mTOR, ATM and PI4-kinase b (Finan and Thomas,
`2004). Moreover, methylxanthines, such as caffeine and
`theophylline, inhibit p110d, even though their activity is
`rather weak (Arcaro and Wymann, 1993; Abraham, 2004).
`
`Akt as target for anticancer drug discovery
`JQ Cheng et al
`
`Recently, several new compounds have been de-
`scribed to have some selectivity for individual members
`of PI3K (Figure 2b). PIramed have described several
`imidazopyridine derivatives that exhibit excellent PI3K
`inhibitory activities, especially against p110a. A series of
`morpholino-substituted compounds related LY294002
`have shown isoform selectivity. Quinolone and pyrido-
`pyrimidine (Kinacia) are approximately 100-fold more
`potent against p110a/b as compared to p110g (Finan
`and Thomas, 2004). ICOS Corporation has recently
`claimed a new PI3K inhibitor, IC87114, which selec-
`tively inhibits p110d with IC50¼ 0.5 mM and >50-fold
`selectivity over the other class PI3K isoforms (Sadhu
`et al., 2003). In addition, Novartis has described 5-
`phenylthiazole derivatives as PI3K inhibitors (Finan
`and Thomas, 2004). However, antitumor activity of
`these compounds needs further investigation both in
`vitro and in vivo. It is noteworthy that use of PI3K
`inhibitors may be associated with undesirable side
`effects because of the many important cellular targets
`of this lipid kinase.
`
`Inhibitors of the prenylation
`
`Protein prenylation, including farnesylation and geranyl-
`geranylation,
`is a lipid post-translational modification
`required for the cancer-causing activity of proteins such as
`the GTPase Ras. Farnesyltransferase and geranylgeranyl-
`transferase I inhibitors (FTIs/GGTIs) represent a new
`class of anticancer drugs that exhibit a remarkable ability
`to inhibit malignant transformation without significant
`toxicity to normal cells, especially FTIs are currently in
`clinical trials. However, the mechanism of FTI and GGTI
`antitumor activity remains elusive. It has been shown that
`FTIs inhibit PI3K/Akt-mediated growth factor- and
`adhesion-dependent survival pathways and induce apop-
`tosis in human cancer cells that overexpress Akt (Jiang
`et al., 2000; Prendergast, 2000; Sebti and Der, 2003).
`Furthermore, overexpression of Akt, but not oncogenic H-
`Ras, sensitizes NIH3T3 cells to FTI-277, and a high serum
`level prevents FTI-277-induced apoptosis in H-Ras- but
`not Akt-transformed NIH3T3 cells. A constitutively active
`form of Akt rescues human cancer cells from FTI-277-
`induced apoptosis. Integrin-dependent activation of Akt is
`also blocked by FTI-277. In addition, GGTI-298 and
`GGTI-2166 have also been shown to inhibit PI3K/Akt
`pathway, resulting in apoptosis in human cancer cells
`(Dan et al., 2004). Thus, a mechanism for FTIs and
`GGTIs inhibition of human tumor growth is by inducing
`apoptosis through inhibition of the PI3K/Akt pathway.
`However, neither FTIs nor GGTIs directly inhibits PI3K/
`Akt, suggesting that the unidentified prenylated proteins
`that activate PI3K/Akt are the targets of FTIs and/or
`GGTIs. Recent studies suggest that RhoB could be a
`candidate to mediate this action (Liu and Prendergast,
`2000; Adini et al., 2003; Jiang et al., 2004).
`
`RTK inhibitors
`
`Among RTKs, EGFR and Her2/Neu/ErbB2 are fre-
`quently altered in human cancer and primarily activate
`
`Oncogene
`
`NOVARTIS EXHIBIT 2051
`Par v. Novartis, IPR 2016-01479
`Page 4 of 11
`
`

`

`7486
`
`Akt as target for anticancer drug discovery
`JQ Cheng et al
`
`the PI3K/Akt pathway. Two major approaches have
`been used to target the ErbB family, that is, small-
`molecule tyrosine kinase inhibitors and humanized
`antibodies against the receptor extracellular domains
`(Yu and Hung, 2000; Chen et al., 2003). In general,
`antibodies bind to the extracellular domain of
`the
`receptors,
`inhibiting their activation by ligand, and
`promoting receptor internalization and downregulation,
`whereas small molecules competitively inhibit ATP
`binding to the receptor, thereby hindering autopho-
`sphorylation and kinase activation. At present, the most
`advanced of the newer therapies in clinical development
`are anti-EGFR monoclonal antibody IMC-C225 (ce-
`tuximab, Erbitux; Imclone), anti-ErbB2 and the rever-
`sible small-molecule inhibitors of EGFR, ZD-1839
`(gefitinib, Iressa; AstraZeneca) and OSI-774 (erlotinib,
`Tarceva; OSI Pharmaceuticals). Both ZD-1839 and
`OSI-774 have been through phase I and phase II trials.
`Promising single-agent clinical antitumor activity has
`been reported in advanced NSCLC, head and neck
`cancer and prostate carcinoma. Furthermore, huma-
`nized monoclonal antibodies, IMC-C225 against EGFR
`and trastuzumab (Herceptin) targeted ErbB2, are also in
`phase II and phase III trials. These agents potently
`inhibit EGFR and ErbB2 resulting in the reduction of
`Akt kinase activity (Yakes et al., 2002; Mitsiades et al.,
`2004). A recent study has shown that patients who
`become resistance to Herceptin have elevated levels of
`Akt due to loss of PTEN, suggesting that elevated Akt
`activation is responsible for Herceptin resistance (Na-
`gata et al., 2004).
`
`Akt inhibitors
`
`Akt antibody
`
`Antibodies and antibody-based reagents have been used
`for the treatment of cancer. As described above, the
`humanized IgG1 trastuzamab (Herceptin) is an effective
`treatment for breast cancers that overexpress ErbB2 (Yu
`and Hung, 2000). Genetic engineering of antibodies can
`be used to modify and enhance antibody efficacy. For
`example, mouse monoclonal antibodies can be chimer-
`ized by such approaches to prevent the production of
`human antimurine antibodies when administered to
`immune-competent patients (Clark, 2000). An alterna-
`tive strategy is to replace the antibody gene present in
`mouse B cells with human antibody genes. These
`modified B cells can then be used to produce hybridoma
`cell
`lines that express fully humanized monoclonal
`antibodies that avoid cross-species immune response
`(Fishwild et al., 1996). Over a decade ago, McCafferty
`et al. (1990) demonstrated that recombinant antibody
`fragments could be displayed on the tip of M13
`bacteriophage. Some of
`the advantages of phage-
`displayed recombinant antibodies over the conventional
`polyclonal or monoclonal antibodies are quick genera-
`tion time, cheap production cost, and,
`importantly,
`accessibility to the antibody DNA for further genetic
`manipulations. Recently, Shin et al. (2005) developed
`
`novel recombinant anti-Akt single-chain antibodies by
`panning a mouse phage-displayed scFv recombinant
`antibody library using GST-Akt1 fusion protein. To
`generate a membrane-permeable version of the anti-
`Akt1-scFv, the scFv gene was subcloned into a GST
`expression vector carrying a membrane-translocating
`sequence (MTS) from Kaposi fibroblast growth factor.
`A purified GST-anti-Akt1-MTS fusion protein accumu-
`lates intracellularly and inhibits activation of all three
`Akt family members. Interestingly, in vitro kinase assay
`shows that GST-anti-Akt1-MTS also inhibits constitu-
`tively active forms of Myr-Akt1, -Akt2 and -Akt3 as
`well as phosphomimetic mutant of Akt-DD, where
`Thr308 and Ser473 are replaced with aspartic acid.
`Furthermore, GST-anti-Akt1-MTS induces apoptosis in
`cancer cell lines that express constitutively active Akt. In
`addition, anti-Akt scFv exhibits antitumor activity in
`PyVmT-expressing transgenic tumors
`implanted in
`mouse dorsal window chambers (Shin et al., 2005).
`These data indicate that GST-anti-Akt1-MTS is a cell-
`permeable inhibitor of Akt and that this approach can
`be used to generate compounds that target tumor cells
`dependent on aberrant Akt for their growth.
`
`PI analog inhibitors
`
`As PtdIns(3,4,5)P3 directly binds to the PH domain of
`Akt and PDK1 and is required for activation of Akt, the
`development of a PtdIns(3,4,5)P3 analog would be a
`reasonable approach to develop an Akt inhibitor. This
`mode of inhibition would prevent Akt translocation to
`the plasma membrane and activation. The feasibility of
`this approach was suggested by the demonstration that
`D-3-deoxy-myo-inositols inhibited the growth of trans-
`formed cells (Powis et al., 1991). It was subsequently
`found that the inositol derivative DPI had an IC50 of
`35 mM against H-29 colon cancer cell growth (Kozi-
`kowski et al., 1995). A recent study examined 24
`modified phosphatidylinositol ether
`lipid analogues
`(PIAs) and found that five of them, PIA5, 6, 23, 24,
`and 25 (Figure 3), with modifications at two sites on the
`inositol ring, inhibited Akt with IC50o5 mM (Castillo
`et al., 2004a, b). PIAs decreased phosphorylation of
`many downstream targets of Akt without affecting
`upstream kinases, such as PI3K or PDK1. Importantly,
`PIAs selectively induced apoptosis in cancer cell lines
`with high levels of endogenous Akt activity. These
`findings identify PIAs as effective Akt inhibitors, and
`provide proof of principle for targeting the PH domain
`of Akt. However, whether PIAs are effective in vivo and
`whether PIAs affect other PH-domain containing
`proteins are currently unknown.
`Perifosine is a novel orally bioavailable alkylpho-
`spholipid and structurally resembles naturally occurring
`phospholipids (Figure 3). Perifosine is known to be a
`CDK inhibitor and has displayed significant antiproli-
`ferative activity in vitro and in vivo in several human
`tumor model systems (Senderowicz, 2003a, b; Vink
`et al., 2005). It has been shown that perifosine can
`cause cell cycle arrest with induction of p21WAF1/CIP1 in a
`p53-independent fashion. By searching for the under-
`
`Oncogene
`
`NOVARTIS EXHIBIT 2051
`Par v. Novartis, IPR 2016-01479
`Page 5 of 11
`
`

`

`Akt as target for anticancer drug discovery
`JQ Cheng et al
`
`7487
`
`Figure 3 Compound structures for representive Akt inhibitors
`
`lying mechanism, Kondapaka et al. (2003) has demon-
`strated that perifosine inhibits Akt phosphoylation by
`decreasing the plasma membrane localization of Akt,
`and this is substantially relieved by Myr-Akt along with
`relief of downstream drug effect on induction of
`p21WAF1/CIP1. While perifosine does not directly affect
`PI3K, PDK1 or Akt activity, its antitumor activity is, at
`least in part, mediated by inhibition of the Akt pathway
`(Kondapaka et al., 2003). In a recent phase I clinical
`trial, 1 patient had a partial response to perifosine and 2
`patients had stable disease among 42 patients with
`incurable solid malignancies, thus justifying additional
`investigation of this agent in a phase II trial (Van
`Ummersen et al., 2004).
`
`ATP-competitive inhibitors
`
`Through the elucidation of the sequence and structural
`composition of kinase active sites, coupled with the
`solution of numerous ATP-competitive ligand complex
`structures, significant advances have been made in
`developing inhibitors that are highly selective. This has
`been the case not only for kinases that are divergent in
`primary structure, but also for isoforms with highly
`conserved structure and ATP-binding sites. However,
`due to the fact that there is a high degree of homology in
`the ATP-binding pocket between Akt, PKA and PKC
`(Yang et al., 2002), many typical PKA and PKC
`inhibitors have been identified as inhibitors of Akt
`(Reuveni et al., 2002). So far, no ATP-competitive
`inhibitors specific against Akt have been identified.
`While a recent report described the design and synthesis
`of three Akt-selective inhibitors, their ability to inhibit
`Akt has not been tested in either cell culture or animal
`models (Breitenlechner et al., 2005).
`
`Pseudosubstrate inhibitors
`
`In contrast to ATP-competitive inhibitors, pseudosub-
`strate peptide inhibitors bind to the peptide/protein
`substrate sites of
`the catalytic domain and have
`been proven to be selective and potent for many kinases
`(Luo et al., 2004). This selectivity derives from the
`much larger peptide-kinase contact region that has
`evolved to discriminate between various protein sub-
`strates. A 14-mer peptide (AKTide-2T) was identified
`that binds to the substrate binding site of Akt1 and
`inhibits Akt1 activity (Luo et al., 2004). A hybrid
`peptide was recently constructed between this sequence
`and a sequence (amino acids 16–24) of the forkhead
`transcription factor FOXO3. The resulting 20-mer
`peptide construct inhibits Akt1 with a Ki of 1.1 mM.
`Replacement of a putative phosphorylation site serine
`(Ser) in the sequence with an alanine (Ala) moiety
`resulted in a further 10-fold improvement of potency.
`Efforts to truncate the peptide sequence showed that the
`20-mer could be truncated to a 17-mer with only slight
`loss of potency but further chain shortening resulted in
`larger losses of potency. These peptides are highly
`selective versus p70S6K, p90S6K PKA, Cdc2, Src, and
`PKC. Interestingly, the original 20-mer was equally
`active against SGK but the subsequent Ala for Ser
`replacement resulted in a highly selective inhibitor.
`Fusion peptides were also constructed to allow cell
`uptake and demonstrated dose-dependent inhibition of
`GSK3b phosphorylation. While these peptides are
`potent and selective inhibitors of Akt, their size makes
`them poor leads for small molecule inhibitor develop-
`ment. Significant
`truncation of
`the peptides and
`incorporation of peptidomimetic functionality would
`likely be needed to lead to cell-permeable molecules with
`good pharmacokinetics.
`
`Oncogene
`
`NOVARTIS EXHIBIT 2051
`Par v. Novartis, IPR 2016-01479
`Page 6 of 11
`
`

`

`7488
`
`API-2/triciribine
`
`Akt as target for anticancer drug discovery
`JQ Cheng et al
`
`By screening of the National Cancer Institute Diversity
`Set 2000-compound library, which are derived from
`approximate 140 000 compounds, a small molecule Akt
`pathway inhibitor, API-2, has recently been identified
`(Yang et al., 2004). API-2 suppressed the kinase activity
`and phosphorylation level of all
`three Akt
`family
`members. The inhibition of Akt kinase resulted in
`suppression of cell growth and induction of apoptosis in
`human cancer cells that harbor constitutively activated
`Akt due to overexpression of Akt or other genetic
`alterations such as PTEN mutation. API-2 is highly
`selective for Akt and does not inhibit PI3K, PDK1,
`PKC, SGK, PKA, Stat3, Erk-1/2 or JNK. Furthermore,
`API-2 potently inhibited tumor growth in nude mice of
`human cancer cells in which Akt is aberrantly expressed/
`activated but not of those cancer cells in which it is not.
`Recent data suggest
`that API-2 inhibition of Akt
`depends on the interaction with the PH domain of Akt.
`API-2 is a synthetic small molecule compound
`identified previously and named triciribine (TCN,
`NSC-154020; Figure 3) or tricyclic nucleoside (Chung
`et al., 1980; Wotring et al., 1986). Its 5’phosphate ester,
`triciribine monophosphate (TCN-P, NSC-280594) is the
`chemical entity advanced into clinical trials because it is
`more soluble than the parent drug (Wotring et al., 1986).
`TCN and TCN-P have antiviral and antineoplastic
`activity at low micromolar or submicromolar concen-
`trations (Migawa et al., 2005). By the early 1980s, TCN-
`P had been identified as an inhibitor of DNA and
`protein synthesis, and shown preclinical activity against
`leukemias and carcinomas. A number of phase I and II
`clinical trials of TCN/API-2 have been conducted in
`patients with advanced tumors, including carcinomas of
`the breast, colon, bladder, ovary, pancreas and lung
`(Cobb et al., 1983; Mittelman et al., 1983; Feun et al.,
`1984, 1993; Powis et al., 1986; Schilcher et al., 1986;
`Hoffman et al., 1996). Owing to the fact that TCN-p
`was used as a cytotoxic drug, the majority of clinical
`trials were carried out with high doses of the drug in
`order to achieve maximal clinical efficacy. While it
`exhibited antitumor activity in some patients, TCN/
`API-2 had significant
`side effects at high doses,
`including hepatotoxicity, hypertriglyceridemia, throm-
`bocytopenia, and hyperglycemia, which hampers its
`application in the clinic.
`As TCN/API-2 specifically targets Akt, it is reason-
`able to speculate that a given tumor with high levels of
`Akt activity will be more sensitive/responsive to TCN/
`API-2 treatment and that lower doses of TCN/API-2
`should achieve clinical efficacy without significant side
`effects in patients whose tumors exhibit elevated Akt
`levels. In fact, a recent study has shown that TCN/API-2
`effectively induces apoptosis and cell growth arrest in
`tumor cells with activation of Akt at a concentration of
`10 mM. In xenograft experiments, no detectable side
`effects were observed in 50 mice treated with TCN/API-
`2 at a concentration of l mg/kg/day, which significantly
`inhibited tumor growth in cancer cells overexpressing