TrendsAlthough the oncogenic PI3K/Akt/mTORC1 cascade is well establishedin cancer, PI(4,5)P2 and PIPKI/PIPKIIfunctions are also integral parts of can-cer progression.
The phosphoinositide signaling axis incancer is collectively regulated by PI(4,5)P2 and PI(3,4,5)P3 lipid messen-gers and their generating enzymesPIPK and PI3K, both of which are over-expressed in cancer.
PI(4,5)P2 and PIPKI/PIPKII not only reg-ulate cell polarity, motility, and invasionbut also control PI3K/Akt activation incancer.
Co-targeting of the PI(4,5)P2 signalingnexus may enhance the efficacy ofcanonical anticancer drugs targetingPI3K/Akt.
1Program in Molecular and CellularPharmacology, University ofWisconsin–Madison School ofMedicine and Public Health, 1300University Avenue, Madison, WI53706, USA2Department of Oncology andMcArdle Laboratory for CancerResearch, University of Wisconsin–Madison School of Medicine andPublic Health, 1111 Highland Avenue,Madison, WI 53705, USA3Department of Human Oncology,University of Wisconsin–MadisonSchool of Medicine and Public Health,1111 Highland Avenue, Madison, WI53705, USA
ReviewThe Hidden Conundrumof Phosphoinositide Signalingin CancerNarendra Thapa,1 Xiaojun Tan,1 Suyong Choi,1 PaulF. Lambert,2 Alan C. Rapraeger,1,3 and Richard A. Anderson1,*
Phosphoinositide 3-kinase (PI3K) generation of PI(3,4,5)P3 from PI(4,5)P2 andthe subsequent activation of Akt and its downstream signaling cascades(e.g., mTORC1) dominate the landscape of the phosphoinositide signalingaxis in cancer research. However, PI(4,5)P2 is breaking its boundary as merelya substrate for PI3K and phospholipase C (PLC) and is now an establishedlipid messenger pivotal for various cellular events in cancer. Here we reviewthe phosphoinositide signaling axis in cancer, giving due weight to PI(4,5)P2
and its generating enzymes, the phosphatidylinositol (PI) phosphate(PIP) kinases (PIPKs). We highlight how PI(4,5)P2 and PIPKs serve as aproximal node in the phosphoinositide signaling axis and how interactionwith cytoskeletal proteins regulates the migratory and invasive nexus ofmetastasizing tumor cells.
Phosphoinositides and Phosphoinositide Signaling AxisLife is a delicate balance of various cellular events orchestrated in a highly regulated andcoordinated manner, such as cell cycle progression, survival, apoptosis, cell motility, and geneexpression [1,2]. These cellular events are the intricate outcome of signaling pathways thatoperate in time and space [1]. Among them, phosphoinositide (see Glossary) signaling,initiated by the generation of phosphorylated PI lipid moieties – phosphoinositides – unequivo-cally occupies a central position in health and disease [3–5]. The present and past decades haveseen a tremendous surge in the study of phosphoinositide signaling, the deregulation of which isnow one of the established culprits in cancer. The phosphorylation of PI(4,5)P2 to PI(3,4,5)P3 byPI3K, leading to the recruitment of protein kinase B (PKB)/Akt and 3-phosphoinositide-depen-dent protein kinase 1 (PDK1) to the plasmamembrane and the initiation of downstream signalingcascades (PI3K/Akt/mTORC1), dominates the landscape of the phosphoinositide signalingaxis in cancer [3,4]. Similarly, PLC hydrolysis of PI(4,5)P2 and the subsequent generation ofdiacylglycerol (DAG) and activation of various isoforms of protein kinase C (PKC) has establishedits own domain in the landscape of cancer biology [6]. As PI3K/Akt/mTORC1 and DAG/PKCcascades have dominated cancer research, PI(4,5)P2 and its generating enzymes, the PIPKs,were mostly considered regulators of basic cellular functioning [7–9]. The inability of PI(4,5)P2
to directly recruit and activate oncogenic PDK1 and Akt, despite its 10–100-fold greaterabundance than PI(3,4,5)P3, and the lack of oncogenic mutations activating PIPKs largelydampened PI(4,5)P2 signaling from the limelight of cancer research. However, PI(4,5)P2
synthesis is highly regulated at different subcellular compartments by distinct isoforms of PIPKs(PIPKI/PIPKII) (Box 1) and their diverse interacting partners from cytoskeletal proteins to adhe-sion receptors and signaling molecules, implicating PI(4,5)P2 lipid messenger functions in manycellular events that are integral parts of cancer progression [2,10–12]. PI(4,5)P2 and PIPKI/PIPKII
GlossaryPhosphatidylinositol (PI) andphosphoinositides: PI is a lipidsignaling molecule in the inner leafletof the plasma membrane. Itcomprises two fatty acid chainslinked to a glycerol moiety and awater-soluble inositol head group.The hydroxyl groups (at the 3, 4, and5 positions) in the inositol ring can bephosphorylated in variouscombinations resulting in sevendistinct and interconvertiblephosphoinositides: PI3P, PI4P, PI5P,PI(3,4)P2, PI(3,5)P2, PI(4,5)P2, and PI(3,4,5)P3. This is achieved by variousPI or phosphoinositide kinases andphosphatases that are expressed inthe cell. These molecules account forless than 1–5% of the totalphospholipid content of mammaliancells. PI phosphorylated at the fourthposition (PI4P) of the inositol ring isthe most abundant of thesephosphoinositides in the cell. It isfollowed by PI 4,5-bisphosphate [PI(4,5)P2], which is phosphorylated atthe fourth and fifth positions.Phosphatidylinositol (PI)phosphate kinases (PIPKs): lipidkinases that phosphorylatephosphoinositides (e.g., PI3P, PI4P,PI5P) in cells and are classified astype I, II, and III. Type I and type IIPIPKs generate PI(4,5)P2 in cells.Phosphatidylinositol (PI) 4,5-bisphosphate 3-kinases (PI3Ks):lipid kinases that phosphorylate the3-OH group of the inositol ring in PIto form PI3P, PI(3,4)P2, and PI(3,4,5)P3. Depending on their structure andsubstrate specificity, these arecategorized as class I, II, and III.Class I is the dominant member incell signaling and cancer biology andis activated downstream of RTKs,GPCRs, and adhesion receptors.Class IA PI3Ks are heterodimericmolecules comprising regulatory (fivevariants: p85/, p85b, p85g, p55/,and p50/) and catalytic (threevariants: p110/, p110b, and p110d)subunits. Class IB PI3K comprises ap110g catalytic subunit. PI3Kenzymes with catalytic subunitsp110/ and p110b are the mostubiquitously expressed. Class II PI3Kproduces PI3P and PI(3,4)P2
whereas class III PI3K produces onlyPI3P.
Box 1. PIPKs Synthesizing PI(4,5)P2
PIPKs are responsible for synthesizing PI(4,5)P2 in various subcellular compartments in mammalian cells, which isconsistent with the diverse cellular functions of the PI(4,5)P2 lipid messenger [9,10]. These lipid kinases are classified intotype I (PIPKI), type II (PIPKII), and type III (PIPKIII), but PIPKIII is involved in PI(3,5)P2 synthesis (Figure I). PIPKI utilizes PI4Pas a substrate for the generation of PI(4,5)P2 and is largely responsible for synthesizing the majority of PI(4,5)P2 inmammalian cells. PIPKII utilizes PI5P as a substrate to produce PI(4,5)P2 and appears to be more important in controllingcellular levels of PI5P rather than the production of PI(4,5)P2. In mammalian cells, both PIPKI and PIPKII exist in threeisoforms: /, b, and g (PIPKI/, PIPKIb, PIPKIg, PIPKII/, PIPKIIb, and PIPKIIg). These kinases display highly homologouscatalytic domains with divergent N and C termini, which is key for their interaction with specific binding partners,differential subcellular targeting, and functional divergence [10]. PIPKIg is the most complex isoform among the PIPKs,comprising several splice variants (e.g., PIPKIgi1, PIPKIgi2, PIPKIgi4, PIPKIgi5) targeted to different subcellular compart-ments and performing distinct cellular functions [9,44].
PIPKIα
PIPKIβ
PIPKIγ
PIPKIIα
PIPKIIβ
PIPKIIγ
Dis�nct C-tail in PIPKIγ variants (e.g. PIPKIγi1,PIPKIγi2, PIPKIγi4 and PIPKIγi5)
Kinase domain
Ac�va�on loop
1 539 aa
1 546 aa
1
1
1
1
406 aa
416 aa
421 aa
Type I PIP kinase (PIPKI)
Type II PIP kinase (PIPKII)
Figure I. Phosphatidylinositol (PI) Phosphate (PIP) Kinases (PIPKs) Synthesizing PI(4,5)P2 in MammalianCells. PI(4,5)P2 in mammalian cells is primarily synthesized through phosphorylation of the fifth hydroxyl groupon the inositol ring of PI4P (predominant substrate) by type I PIPK (PIPKI), for which there are three genes: PIPKI/,PIPKIb, and PIPKIg. Type II PIPK (PIPKII) synthesizes PI(4,5)P2 by phosphorylating the fourth hydroxyl group ofPI5P (minor substrate), which is increased by metabolic stress or oncogene expression. PIPKII is also furtherclassified as PIPKII/, PIPKIIb, and PIPKIIg. PIPKIg is the most complex isoform of PIPKI and displays variouspost-transcriptional splicing variants differing in their C tails, which specify their interactions with distinct bindingpartners.
control of PI3K/Akt signaling, metabolic stress, cytoskeletal reorganization, and the migratoryand invasive nexus are discussed in the subsequent sections of this review, although theserepresent only a fraction of the cellular functions attributed to PI(4,5)P2 signaling. We summarizethe phosphoinositide signaling axis in cancer as an integrative role of PI(4,5)P2 and PI(3,4,5)P3
lipid messengers and the enzymes generating these lipid messengers (Figure 1). We emphasizethat PI(4,5)P2 and PIPKs are an underappreciated conundrum of the phosphoinositide signalingaxis in cancer that deserve attention.
2 Trends in Cancer, Month Year, Vol. xx, No. yy
TRECAN 76 No. of Pages 13
RTK
Integrins
GPCR
Cancerprogression
Basic cellularfunc�oning
Cytoskeletalreorganiza�on
Migratory/invasive nexus
Cell metabolism
PDK1
Effectors
PIPKIαPIPKIβPIPKIγ
PI4P PI(4,5)P2 PI(3,4,5)P3
p27
Akt
PTEN
BIM
p21
Casp9
BAD
P53
MDM2
GSK3β
BAX
IKK
4EBP1
S6K1 mTORC1
Inhibi�on of apotosis,increased cell survival
Transcrip�onal-transla�on control ofan�-apopto�c and cellcycle regulator genes
PI3K
Figure 1. Collective Role of PI(4,5)P2 and PI(3,4,5)P3 Lipid Messengers and Phosphoinositide Kinases inCancer Progression. PI(4,5)P2, phosphatidylinositol (PI) phosphate (PIP) kinases (PIPKs), and PI(4,5)P2 effectors regulatevarious cellular functions including cytoskeletal reorganization, cell migration and invasion, and basic cellular functioning. PIP(4,5)P2 is predominantly synthesized by type I PIPKs (PIPKI/, PIPKIb, and PIPKIg) and functions as a substrate ofphosphoinositide 3-kinase (PI3K) for PI(3,4,5)P3 generation downstream of activated receptors. PI(3,4,5)P3 generationand activation of Akt/mTORC1 is critical for cell growth, survival, and metabolism and inhibition of the repertoire of proteinsinvolved in cell apoptosis. Various cellular events involved in cancer progression are collectively regulated by PI(4,5)P2 and PI(3,4,5)P3 lipid messengers together with their generating enzymes. Abbreviations: GPCR, G-protein coupled receptor; RTK,receptor tyrosine kinase.
Implication of Phosphoinositide Signaling in CancerThe discovery of phosphoinositide turnover by the pioneering work of the Hokins in 1953 laid thefoundation for the study of phosphoinositide signaling in mammalian cells [13]. In the early andlate 1990s, a series of key discoveries directly linked phosphoinositide signaling to cancer. Theintroduction of phosphoinositides phosphorylated at the third hydroxyl group of its inositol ring bythe Cantley group [14] and others [15] and the establishment of rapid phosphorylation of PI(4,5)P2
to PI(3,4,5)P3 by PI3K in growth factor stimulation, G protein-coupled receptor (GPCR) activation,and oncogenic transformation were seminal discoveries that unveiled a new signaling pathwayin parallel with PLC-mediated PI(4,5)P2 conversion to DAG and inositol trisphosphate (IP3) [16].This was followed by the cloning and characterization of different isoforms of PI3K [16,17] and
Trends in Cancer, Month Year, Vol. xx, No. yy 3
TRECAN 76 No. of Pages 13
the discovery of the serine/threonine kinase Akt as a downstream effector of PI(3,4,5)P3, andmTORC1 downstream of Akt [18]. Later, the tumor suppressor PTEN, which turns off PI3K/Aktsignaling by dephosphorylating PI(3,4,5)P3 back to PI(4,5)P2, was found to be commonly lost ormutated in various cancer types [19]. Furthermore, the discovery of a series of somatic mutationsin the components of the PI3K/Akt signaling pathway (e.g., catalytic subunit of PI3K, PDK1, andAkt) further consolidated the significance of the PI3K/Akt/mTORC1 signaling axis in cancer[20,21]. As activated Akt directly inhibits the repertoires of proapoptotic proteins (e.g., BAD,BAX, BIM, Caspase-9) and mTORC1 controls cell growth, activation of PI3K/Akt/mTORC1 is anindispensable signaling node in many cancers [3,4]. Similarly, responding to the availability ofoxygen and nutrients (e.g., glucose, ATP, amino acids) in the environment, cell metabolism, andautophagy are emerging areas of PI3K/Akt/mTORC1 signaling and functions [22]. All of thesejustify PI3K/Akt/mTORC1 signaling as perhaps one of the most common therapeutic targets forcancer treatment [4].
Deregulated PI3K/Akt Activation in CancerVarious repertoires of genetic and epigenetic changes that tumor cells acquire contribute to theevasion of controlled regulation of PI3K/Akt/mTORC1 signaling [3,4,23,24]. The most commonmechanisms include: (i) loss of the PTEN tumor suppressor; (ii) gain of somatic mutations in thecomponents of the PI3K/Akt signaling axis; (iii) overexpression of PI3K; and (iv) overexpressionor overactivation of receptor tyrosine kinases (RTKs) leading to constitutive recruitment andactivation of PI3K. PTEN loss is commonly observed in various cancer types, resulting inaberrant activation of PI3K/Akt [25]. Besides PTEN loss, somatic mutations causing truncationof the PTEN protein and loss of its function are reported in tumor-prone germline diseases [25].The Cancer Genome Atlas (TCGA) genome-scale analysis shows PI3KCA (the gene encodingp110/, the catalytic subunit of class I PI3K) to be one of the eight most frequently mutatedgenes in cancer [26–28]. These mutations occur at the interface between the catalytic (p110/)and adaptor (p85) subunits of PI3K and generally abrogate the inhibitory effect of the adaptorsubunit on the catalytic subunit [24,26,27]. More than 75% of these activating mutations residein either the helical or the catalytic domain of P110/ and are called Hot-Spot’mutation sites [26].These activating mutations are reported only in the P110/ catalytic subunit. Additionally,overexpression or mutational activation of various tyrosine kinase receptors is responsiblefor aberrant activation of phosphoinositide signaling in cancer [3,23]. These overexpressedRTKs (e.g., PDGF, EGFR, c-MET, IGFR) generally undergo homo- or heterodimerization, evenin the absence of ligand binding. This creates the docking sites, which are phosphorylatedtyrosine residues in the context of the YXXM motif in their cytoplasmic domains that recruit thePI3K enzyme to the plasmamembrane (the SH2 domain of the adaptor subunit p85 specificallybinds to these YXXM motifs, mediating the recruitment of the catalytic subunit) [24]. Forexample, ERBB2/HER2 and ERBB3/HER3 contain multiple docking sites for PI3K and theiractivation triggers a dramatic increase in PI3K/Akt signaling [29]. Additionally, adaptor proteinssuch as insulin receptor substrate (IRS), Shc, and growth factor receptor-bound protein2 (Grb2) and the E3 ubiquitin ligase Cbl also provide docking sites for the PI3K adaptor subunitp85 [3,4,23].
Various components of the PI3K/Akt/mTORC1 signaling axis (e.g., PI3K enzyme, PDK1, Akt,mTORC1) are actively targeted for treatment of cancer [4,30,31]. More than 100 drugs (e.g.,buparlisib, duvelisib, TGR1202, copanlisib, BEZ235, RP6530, PWT33597, CUDC-907, PI-103,TG100-115I, NK1117) targeting the PI3K/Akt/mTORC1 signaling cascade are undergoing orprogressing towards various stages of clinical trials (Phase I, II, and III) [30,31]. Everolimus,temsirolimus, and idelalisib represent the handful of anticancer drugs that have so far beenapproved by the FDA for treatment of cancer. Everolimus and temsirolimus target mTORC andare used for the treatment of renal cell carcinoma, astrocytoma, and HER2-negative breastcancer, whereas idelalisib targets P110/ and is used for leukemia and lymphoma [32,33].
4 Trends in Cancer, Month Year, Vol. xx, No. yy
TRECAN 76 No. of Pages 13
Readers are referred to ClinicalTrials.gov (https://clinicaltrials.gov/) to learn more about clinicaltrials and the outcome of specific drugs in cancer treatments. It is becoming clear that targetingthe PI3K/Akt/mTORC signaling axis at multiple points (e.g., PI3K andmTOR in dual therapy) or incombination with inhibitors of TKRs (e.g., HER family tyrosine kinase inhibitors in combinationtherapies) and other, parallel nodal pathways (e.g., MAPK) is more effective in inhibiting tumorgrowth and preventing the emergence of regulatory feedback loops and development of drugresistance [34–36].
PIPKI/PIPKII Control of PI3K/Akt Activation in CancerGiven the diversity in the mechanisms of PI3K/Akt activation (e.g., from the repertoire ofoncogenic mutations in PI3K to loss of PTEN to overexpression/overactivation of tyrosinekinase receptors), identifying the common node for all of these diverse mechanisms couldpave the way for developing more effective therapeutic approaches to blocking the PI3K/Aktsignaling axis in cancer. Could the generation of PI(4,5)P2 substrate at specific subcellularcompartments provide a common regulatory node upstream of PI3K/Akt activation? Does PI3Kcollaborate with PIPKI/PIPKII for de novo synthesis of PI(4,5)P2 and PI(3,4,5)P3 or are thepreexisting pools of PI(4,5)P2 in the plasma membrane/endomembranes utilized for PI(3,4,5)P3
generation and sustenance of PI3K/Akt signaling in cancer? Although PI(4,5)P2 is the predomi-nant phosphoinositide in the plasma membrane, the availability of free PI(4,5)P2 may be ratelimiting as PLC hydrolyzes the bulk of the PI(4,5)P2 in the plasma membrane in the vicinity ofactivated growth factor receptors or adhesion receptors. This could be circumvented by spatialrecruitment of PIPKI/PIPKII along with PI3K to the plasma membrane for Akt activation.
Among PIPKI isoforms, PIPKI/ and PIPKIg appear to regulate PI3K/Akt signaling, althoughPIPKIb negatively regulates PI3K/Akt signaling [37]. Increased association of PI3K with PIPKIg ingrowth factor stimulation of cells suggests coupling of PI(4,5)P2 and PI(3,4,5)P3 synthesis for Aktactivation [37]. As upstream activators of PI3K/Akt include integrins and RTKs, PIPKIg interactionwith talin and the proto-oncogene Src facilitates the recruitment of PIPKIg to the vicinity ofintegrin-mediated adhesion complexes and activated RTKs, respectively [37]. In both condi-tions, PIPKIg potentially provides de novo PI(4,5)P2 to promote and sustain the PI3K/Aktsignaling downstream of activated integrins and RTKs [37]. Overexpression of PIPKIg alongwith Src sustains the PI3K/Akt signaling and oncogenic growth, where Src serves as a bridgingmolecule for incorporating PIPKIg and PI3K into the same complex [37] (Figure 2). Alternatively,the assembly of all of the phosphoinositide kinases PI4K, PIP5KI, and PI3K into the samecomplex by a scaffold protein could provide a self-contained mechanism to activate and sustainPI3K/Akt signaling in cancer cells. Such a mechanism is well established in the regulation of theMAPK signaling pathway, a companion of PI3K/Akt signaling [38]. IQGAP1, which scaffolds themolecules in the MAPK pathway in certain cell types, also associates with PI4P, PIPKI, and PI3Kand could potentially serve as a scaffolding molecule to streamline and self-sustain the PI3K/Aktsignaling axis in cancer [39]. A similar mechanism could exist in coupling PI(4,5)P2 generationwith PLC-mediated hydrolysis of PI(4,5)P2 and PKC activation in cancer. Unlike PIPKI, thecontribution of the PIPKII enzyme to PI3K/Akt/mTORC1 signaling appears less dominant andcounteracting [40] as the majority of PI(4,5)P2 is synthesized by PIPKI. However, studies inDrosophila still lend support to the role of PIPKII in cell growth and Akt/mTORC1 signaling [41].Additionally, phosphatidic acid, which activates various isoforms of PIPKI, inhibits endogenousinhibitor of mTORC1, suggesting that PIPKI provides another level of mechanism regulatingPI3K/Akt/mTORC1 signaling in cancer [42,43].
PI(4,5)P2 and PIPKI/PIPKII in CancerAs PI(3,4,5)P3 generation from PI(4,5)P2 and PLC-mediated PI(4,5)P2 hydrolysis/PKC activationtakes the center stage of the phosphoinositide signaling axis in cancer, PI(4,5)P2 and PIPKs havelargely gained recognition as key regulators of basic cellular functions such as ion channels and
Ac�va�on of downstreamsignaling cascadesand cellular func�ons
ECM
PIPKIγ
Figure 2. Phosphatidylinositol (PI)Phosphate (PIP) Kinase Type IGamma (PIPKIg) Assembly with Acti-vated Receptor Tyrosine Kinases(RTKs) and Adhesion Receptors forPI(4,5)P2 Synthesis and Phosphoi-nositide 3-Kinase (PI3K)/Akt Activa-tion. PIPKIg recruitment to activatedintegrins at the adhesion complex andactivated RTKs may provide the mechan-ism for de novo synthesis of PI(4,5)P2 forPI(3,4,5)P3 generation and Akt activation.PIPKIg interaction with talin provides aselective advantage for PIPKIg amongother PIPKI isoforms to be recruited inthe vicinity of activated integrins on cellstimulation with extracellular matrix pro-teins. Similarly, interaction with Src pro-motes PIPKIg recruitment to activatedRTKs for PI(4,5)P2 and PI(3,4,5)P3 gen-eration and Akt activation.
transporters, neuronal transmission, endocytosis, exocytosis, phagocytosis, vesicle trafficking,reorganization of cytoskeletal proteins, cell polarity, gene expression, and nuclear events[2,9,10,44]. This reconciles with the fact that PI(4,5)P2 is the most abundant integral lipid moietyof the plasma membrane and endomembranes, interrogating diverse protein interactomesranging from ion channels to cytoskeletal proteins and DNA polymerases, and also a substratefor the generation of other second messengers/metabolites [2]. As a result, many disorders,including channelopathies, mental retardation, bipolar diseases, schizophrenia, Alzheimer dis-ease, diabetes, ciliopathies, and Lowe syndrome, are associated with deregulation of PI(4,5)P2
signaling or PI(4,5)P2 metabolism or loss of PI(4,5)P2 regulation of protein functions [45,46].However, many cellular functions directly attributed to cancer and cancer progression, such ascytoskeletal reorganization and cell motility/invasiveness, are under the direct control of the PI(4,5)P2 lipid messenger and PIPKIs, as discussed below.Unlike PI3K, mutational activation of neither PIPKI nor PIPKII has been reported in cancer,although mutational loss of the PIPKIg kinase activity is found in congenital contracturalsyndrome type 3 (LCCS3) [47]. Similarly, different from PI3K, ectopic expression of PIPKI orPIPKII alone usually does not induce oncogenic transformation, although overexpression ofPIPKIg variants in cooperation with other oncogenes has been reported to promote oncogenicgrowth [48]. However, upregulated expression of PIPKI and PIPKII kinases and their directimplication in cancer progression is emerging, as various cellular events regulated by PI(4,5)P2
and PIPKI/PIPKII lipid kinases are integral parts of cancer progression. In tissue microarrays ofbreast cancer tissues, increased expression of PIPKIg correlates with EGFR expression in triple-negative breast cancer tissues [49]. Survival of breast cancer patients inversely correlates withPIPKIg expression. Corroborating this, xenograft studies in mice show an essential role of PIPKIgin tumor growth andmetastasis [50]. Activation of EGFR phosphorylates a tyrosine residue in theC terminus of PIPKIg (Y639) and this appears to be essential for the role of PIPKIg in tumor growthand metastasis [50]. The PIPKIg/EGFR nexus is further fine-tuned by PIPKIgi5, a splicing variantof PIPKIg, which controls the downregulation of EGFR via modulating the SNX5/Hrs lysosomal
6 Trends in Cancer, Month Year, Vol. xx, No. yy
TRECAN 76 No. of Pages 13
degradation pathway [51,52]. However, the functional role of PIPKIgi5 regulation of EGFRexpression and its impact on cancer progression and metastasis remain to be defined. TheCatalogue of Somatic Mutations in Cancer (COSMIC) database shows increased copy numbersof PIPKIA along with PI4KB, PIPK3C2B, and AKT3, molecules of the phosphoinositide signalingcascade, in breast cancer [53]. More comprehensive studies in the future will define how PIPKIsare involved in regulating the oncogenic phosphoinositide signaling nexus in cancer, as over-expression of PIPKI/PIPKII lipid kinases alone is not sufficient to activate and sustain the PI3K/Akt signaling nexus [48]. Like PIPKIs, deep transcriptome sequencing shows increased expres-sion of PIPKII in cancer cells and cancer tissues [54]. PIP4KII/ and PIP4KIIb are overexpressedin HER2-positive breast cancer tissues [55] and ACGH array shows amplification of the PIP4KIIbgene to be part of the HER2 amplicon in cancer [56]. These lipid kinases appear to be essentialfor tumor growth in the background of p53 loss or mutation. A recent study indicates a novelfunction of PIP4KIIb as a GTP sensor in the regulation of cell metabolism in cancer [57]. The useof a shRNA library targeting all knownmodulators of phosphoinositide metabolism has identifiedPIP4KII/ as a gene required for leukemia, indicating PIP4KII/ as a potential therapeutic targetfor hematological malignancies [58]. However, knockdown of PIPKIIb is associated with stronglyinduced basal and insulin-stimulated PI(3,4,5)P3 levels and Akt activation [55]. This indicatesthat, unlike PIPKI, the role of PIPKII in cancer is independent of the PI3K/Akt/mTORC1 signalingaxis. Tumor cells encounter oxidative stress as a result of oncogene expression or loss of tumorsuppressors that upregulate PI5P levels, and PIPKII lipid kinases are required for its conversionto PI(4,5)P2 by a noncanonical route [40,59]. PIP4KIIb has also been reported to regulate nuclearPI5P and gene expression [60]. This highlights PIPKII as a stress-regulated lipid kinase requiredfor cancer cells to overcome oxidative stress and maintain homeostasis of reactive oxygenspecies [61] (Figure 3).
PI(4,5)P2 and PIPKI/PIPKII Control of Cell Polarity and Cell MotilityMaintenance of cell polarity is one of the most fundamental properties of epithelial cells and lossof cell polarity is a hallmark of epithelial cancer [62,63]. Given the ability of PI(4,5)P2 to providedocking sites for myriad lipid–protein interactions on the plasma membrane and the ability ofPIPKI/PIPKII lipid kinases to interrogate a diverse array of protein interactomes, PI(4,5)P2 andPIPKI/PIPKII serve as integral parts of the epithelial cell polarity program. For example, PI(4,5)P2
is the landmark phosphoinositide entity of the apical surface and an alteration in distribution of PI(4,5)P2 from the apical to the basolateral surface in 3D culture of epithelial cells disrupts lumenformation/epithelial morphogenesis and affects the polarized secretion of basement membraneproteins [64,65]. Many other studies also support a critical role of the PI(4,5)P2 phosphoinositidemolecule in maintaining apical cell polarity [66]. By contrast, PI(3,4,5)P3 and PI3K function ascritical determinants of the basolateral domain of epithelial cells [67]. Epithelial cells lose E-cadherin/cell polarity and gain promigratory/proinvasive phenotypes as a result of oncogenictransformation, expression of E-cadherin transcriptional repressors, or activation of EMT ago-nists as seen in many cancer cells [62]. The same PI(4,5)P2 and PI(3,4,5)P3 molecules alsoparticipate in regulating the myriad cellular events essential for cell motility and the invasiveprogram. This indicates that PI(4,5)P2 and PI(3,4,5)P3 function at the conjunction of epithelial cellpolarity as well as the promigratory/proinvasive nexus of cancer cells. How are these twoseemingly opposite cellular functions regulated by phosphoinositide signaling? In this regard,PIPKIg and PI(4,5)P2 deserve special attention as they provide the molecular platform for theassembly of not only E-cadherin-mediated adherens junctions at cell–cell contact sites inepithelial cells but also the integrin-mediated adhesion complex at the interface of cell–extra-cellular matrix interaction sites [68,69].
In epithelial cells, cell polarity is maintained by E-cadherin-mediated adherens junctions betweenadjacent cells [70]. The precise regulation of targeting, recycling, and endocytosis of E-cadherinmolecules controls the integrity of adherens junctions and epithelial cell polarity [70]. PIPKIgi2, a
Trends in Cancer, Month Year, Vol. xx, No. yy 7
TRECAN 76 No. of Pages 13
RTK
Integrin
Akt
GPCR
Oncogenicinsults
ROS
Loss of tumorsuppressor
Decreasedcellularstress
PI3K
GTP-sensor formetabolism andtumorigenesis
Nuclearsignaling
Cell growth
Decreased apopto�ccell dead/senescence
PIPK
I
PIPK
II
Effectors
Transcrip�onal regula�on
Transla�onal control
Cell growth
Decreased apoptosis
Cytoskeletal reorganiza�on
Cell polarity/mo�lity/invasion
Figure 3. Distinct Roles of Phosphatidylinositol (PI) Phosphate (PIP) Kinase Type I (PIPKI) and PIPKII in CancerProgression. As PIPKI is predominantly involved in PI(4,5)P2 generation, its role in cancer is coupled with PI(3,4,5)P3
generation and Akt activation. PI(4,5)P2 and its effectors, along with PI(3,4,5)P3, activated Akt, and its downstream effectormolecules, control various cellular functions, including cell survival, cell growth, apoptosis inhibition, and cytoskeletalreorganization, all of which are required for cancer progression. PIPKII's function in cancer is associated with balancingmetabolic stress. Loss of tumor suppressor or oncogenic insult results in increased generation of PI5P, which is thenconverted into PI(4,5)P2 by PIPKII for stress alleviation. In the absence of PIPKII, increased accumulation of PI5P promotescell senescence.
specific variant of PIPKIg, integrates the clathrin adaptor protein AP1B and the evolutionarilyconserved vesicle trafficking protein complex, the exocyst, to regulate basolateral trafficking ofE-cadherin for cell polarity and epithelial morphogenesis [68,71–73] (Figure 4). In this process,PIPKIg functions as a molecular scaffold by bridging E-cadherin molecules with AP1B and theexocyst complex [68]. This facilitates targeting and trafficking of E-cadherin cargo to adherensjunctions. Tyrosine-based sorting motifs in the context of YXXQ in the C terminus of PIPKIg(YSPL and YSAQ in PIPKIgi2) recruit the adaptor protein AP1B for basolateral sorting of recyclingendosomes. However, these motifs also recruit the AP2 clathrin adaptor protein to mediateendocytosis of E-cadherin from the plasma membrane [68]. The recruitment of the AP2 clathrinadaptor protein to the YXXQ motif in PIPKIg is favored when the tyrosine residue in the motif isunphosphorylated [68,70]. However, the precise mechanism that governs the phosphorylationof these motifs and the selective recruitment of AP1B or AP2 remains poorly understood. Alongwith clathrin adaptor proteins, the exocyst complex is another direct interacting partner of PIPKIgthat mediates the basolateral targeting of E-cadherin molecules in polarized epithelial cells [73].The ability of two exocyst subunits, Sec3 and Exo70, to interact with both PI(4,5)P2 and PIPKIg
8 Trends in Cancer, Month Year, Vol. xx, No. yy
TRECAN 76 No. of Pages 13
Polarizedepithelial cells
Oncogenictransforma�on
Polarizeddelivery
PIPKIγExocyst
AP1B
Vesicle
(Adherens junc�ons assembly)
PIPKIγERM
N-WASP
VCA
Arp2/3
G-ac�n
Gelsolin
Cofilin
RhoGTPase
FAK
Talin
(Adhesionassembly)
Migratory/invasive cells
E-cad.
(Cytoskeletal reorganiza�on)
Integrins
Vinc
ulin
Ac�n
filamen
t
PIPK
IγFigure 4. Phosphatidylinositol (PI) Phosphate (PIP) Kinase Type I Gamma (PIPKIg) Plays Key Roles inEpithelial Polarity, Reorganization of the Actin Cytoskeleton, and Formation of Adhesion Complexes. PIPKIg,in association with the polarized vesicle trafficking complex (AP1B and exocyst), controls E-cadherin trafficking to adherensjunctions in polarized epithelial cells. However, in malignant cells, which have lost cell polarity, PIPKIg and PI(4,5)P2 playpivotal roles in controlling cytoskeletal reorganization and adhesion complexes, which are key for migrating and invadingtumor cells.
on the plasma membrane/endomembrane facilitates the basolateral targeting of E-cadherinmolecules in recycling or synthetic cargo [8]. The expression of an Exo70 mutant deficient in PI(4,5)P2 binding severely impairs E-cadherin targeting to developing adherens junctions [73].Further, a specific variant of PIPKIg, PIPKIgi5, in coordination with SNX5, controls thelysosomal sorting and degradation of E-cadherin molecules [74], although the involvementof the PIPKIgi5/SNX5 complex in E-cadherin degradation and its role in cancer remain poorlyunderstood.
As epithelial cells lose E-cadherin-mediated adherens junctions and cell polarity, PIPKIg and PI(4,5)P2 engage in the development of dynamic focal adhesion complexes and control themigratory and invasive nexus of various cancer cells [39,49,71,72,75] (Figure 4). One of thekey functions of the PI(4,5)P2 lipid messenger and PIPKIg is to promote the recruitment ofcytoskeletal proteins at developing adhesion complexes. Talin, vinculin, and FAK all harborpatches of basic residues that bind to PI(4,5)P2. Additionally, the spatial generation of PI(4,5)P2
at adhesion complexes promotes the recruitment and activation of these molecules by relievingthe intramolecular constraints imposed on them, thus establishing structurally and functionallycompetent adhesion complexes [76]. PIPKIgi2, the focal adhesion-targeting and talin-interactingvariant of PIPKIg, putatively provides PI(4,5)P2 at developing nascent adhesion complexes inadhering and migrating cells. However, PIPKIgi2 also competes with b1 integrin for talin binding,indicating that the assembly of PIPKIgi2, talin, and integrin at the adhesion complex is a tertiarycomplex facilitated by the PI(4,5)P2 lipid messenger. Specifically, the recruitment of talin to thecytoplasmic domain of b1 integrin at adhesion complexes is key for the initiation of ‘inside-out’and ‘outside-in’ signaling, an integral part of focal adhesion signaling in adhering and migratingcells [77]. Additionally, PIPKIg’s association with the exocyst complex and talin and PI(4,5)P2
Trends in Cancer, Month Year, Vol. xx, No. yy 9
TRECAN 76 No. of Pages 13
generation facilitate the polarized delivery of integrins required for developing nascent adhesioncomplexes at the leading edge of migrating cells [71,72]. PIPKIg also works with its interactingpartner IQGAP1 to promote actin polymerization and cell migration [39]. These illustrate PIPKIgand PI(4,5)P2 function at the conjunction of the epithelial cell polarity program and the migratory/invasive nexus of cancer cells.
PI(4,5)P2 Control of Cytoskeletal ReorganizationAlthough PI(4,5)P2 could not shin on par with the PI(3,4,5)P3 lipid messenger in growthsignaling in cancer, PI(4,5)P2 signaling has been established as a key regulator of thecytoskeletal machinery and cytoskeletal reorganization under both physiological and patho-logical conditions [78,79]. Many excellent reviews [78] serve as great resources for under-standing PI(4,5)P2 regulation of cytoskeleton-associated and regulatory proteins. Importantly,cancer cells display upregulated expression of many cytoskeleton-associated and regulatoryproteins and these are essential for their migratory and invasive properties, as one-third ofproteins that are induced in metastatic cancers are related to adhesion or cytoskeletalreorganization [80]. Spatial and temporal reorganization of the actin cytoskeleton controlsvarious cellular events involved in metastatic cascades, such as conversion of indolentepithelial cells to the mesenchymal state, detachment from the primary tumor, and migra-tion/invasion and extravasation for secondary/tertiary growth [81]. The first evidence of anintimate association between cytoskeletal protein and phosphoinositide was demonstratedby Anderson and Marchesi [82]. Subsequent studies established PI(4,5)P2 as the dominantlipid moiety in regulating cytoskeletal organization via modulating the activities of diversearrays of cytoskeleton-associated and regulatory proteins [78]. For example, PI(4,5)P2 inhibitsthe actin-binding and -depolymerizing activity of ADF/cofilin. Similarly, PI(4,5)P2 inhibitscapping proteins like gelsolin, which prevent the addition and loss of actin monomers fromthe end of actin polymers. Besides these examples, PI(4,5)P2 also directly regulates a plethoraof proteins involved in regulating cytoskeletal reorganization such as spectrin, dynamin,myosin X, ezrin, radixin, gelsolin, profilin, actin, neural Wiskott–Aldrich syndrome protein(N-WASP), myosin, MARCKS, annexins, and /-actinin [11]. The fine-tuned activity of theseactin-binding and -regulating proteins controls the nucleation/formation of actin filaments andtheir polymerization. Furthermore, the coordinated interplay between these processes con-trols the formation as well as the geometry of actin filaments in cellular machineries like celladhesion complexes, lamellipodia, filopodia, and invadopodia, which are essential for migrat-ing and invading tumor cells [81].
PI(4,5)P2 regulation of the actin-nucleating activity of the Arp2/3 protein complex via WASP andRho family small GTPases is the most extensively studied in the context of tumor cell migration/invasion and metastasis [81,83–85]. This depends on PI(4,5)P2 along with active Cdc42 andRac1 binding to the N terminus of N-WASP and PI(4,5)P2 binding to basic motifs [84,85]. Thisleads to exposure of the intramolecularly masked VCA domain in N-WASP, which promotes thebinding of the VCA domain to Arp2/3 and G-actin and initiates the nucleation of actin polymerfrom the sides of existing actin filaments. Furthermore, Rho GTPases interact with PIPKI andactivate the synthesis of the PI(4,5)P2 lipid messenger required for activation of the Arp2/3complex, establishing the self-contained molecular complex for Arp2/3-mediated actin nucle-ation and polymerization (Figure 4). This process plays a pivotal role in the formation oflamellipodia and invadopodia, key cellular machineries employed by migrating and invadingtumor cells. Consistently, the Arp2/3 complex, along with Rho GTPases, is elevated in themajority of cancers [86], although the correlation between upregulated expression of cytoskele-ton-associated and regulatory proteins and PI(4,5)P2 and PIPKI remains poorly defined. Astubulin targeting has been a successful approach in cancer therapy [87], intervening in PI(4,5)P2
and PIPKI functions in the actin cytoskeleton could be similarly exploited in developing drugstargeting the cytoskeletal machinery for tumor therapy.
10 Trends in Cancer, Month Year, Vol. xx, No. yy
TRECAN 76 No. of Pages 13
Outstanding QuestionsHow is PI(3,4,5)P3 generation regu-lated for the sustenance of PI3K/Akt/mTORC1 signaling in cancers? Does italways need spatiotemporal produc-tion of PI(4,5)P2 as a substrate forPI3K to activate Akt? Could the spatio-temporal generation of PI(4,5)P2 byPIPKI serve as a common upstreamnode for the oncogenic PI3K/Akt/mTORC1 pathway?
Howdo PI(4,5)P2, PI(3,4,5)P3, and theirgenerating enzymes work together tocontrol the various aspects of cellularfunctions crucial for cancer progres-sion? The precise mechanisms of PI(4,5)P2 and PI(3,4,5)P3 regulation ofthe cell cycle, cell survival, apoptosis,anoikis evasion, and the stemness traitwould provide an important platformfor uncovering the collective role ofphosphoinositide lipid messengers incancer.
Does phosphatidic acid suppressesthe endogenous mTORC1 inhibitorby activating PIPKI? Could PIPKI acti-vate mTORC1 independently of PI(3,4,5)P3 generation or does it alwaysactivate mTORC1 signaling throughthe PI3K/Akt cascades?
Does constitutive activation of PI3K/Akt/mTORC1 depend on the molecu-lar assembly of PI3K and PIPKI? Whatare the factors that regulate themolec-ular assembly of PIPKI and PI3K? ArePIPKI and PI3K assembled in a uniqueway in cancers addicted to PI3K/Akt/mTORC1 signaling? How can we spe-
Concluding RemarksPhosphoinositide signaling represents a fundamental signaling nexus involved in many cellularfunctions in health and disease. Given the broad horizons of and extraordinary surges inphosphoinositide studies over the past decades, it has become difficult to emphasize particularaspects of phosphoinositide signaling and its cellular functions that are important in cancer. It isplausible that PI(4,5)P2 and PIPKIs have been overshadowed by the well-known PI3K/PI(3,4,5)P3/Akt/mTORC1 axis in cancer. However, it is also important to note the aspects of PI(4,5)P2
and PIPKI/PIPKII that are directly and indirectly implicated in cancer progression. In this reviewwe emphasized and attempted to incorporate the functional role of PI(4,5)P2 and generatingenzymes in the context of cancer and the oncogenic phosphoinositide signaling axis. Asdiscussed here, PI(4,5)P2 and PIPKI/PIPKII are companions of the PI3K/PI(3,4,5)P3/Akt/mTORC1 signaling axis in cancer. However, many other aspects of phosphoinositide signalingthat are equally important in cancer, such as cell cycle regulation, cell survival, apoptosis, anoikisevasion, cancer stem cells, autophagy, and nuclear signaling, remain uncovered in this review.Importantly, future attempts to target the phosphoinositide signaling nexus in cancer should alsoconsider the PI(4,5)P2 lipid messenger and relevant lipid kinases (see Outstanding Question).Targeting the spatiotemporal generation of the PI(4,5)P2 lipid messenger could be exploited totarget the PI3K/Akt/mTORC1 axis in cancer as this would impair the most proximal node of theoncogenic phosphoinositide signaling axis in cancer and might overcome possible variations inthe efficacy of anticancer drugs targeting PI3K and Akt due to the bewildering arrays ofmutations in these target molecules. Optimism remains high in phosphoinositide research,as drugs targeting the specific catalytic subunit of PI3K may also come into clinic within a fewyears. However, a precise understanding of the mechanism regulating spatial activation of PI3K/Akt/mTORC1 signaling in cancer would provide an important avenue for the development ofmore effective therapeutic approaches for cancer treatment.
AcknowledgmentsThe authors regret not being able to include and cite the work of many investigators in the field due to space constraints. The
laboratory of R.A.A. is supported by National Institutes of Health grants (CA104708 and GM057549). Support also comes
from the American Heart Association (AHA) to N.T. (10POST4290052) and S.C. (13PRE14690057) and the Howard
Hughes Medical Institute (HHMI) International Student Research Fellowship to X.T. The authors acknowledge the con-
structive comments and suggestions from colleagues in R.A.A.’s laboratory, Thomas Wise and Anas Rattani.
References
cifically target these unique factors forcancer treatment?
1. Kolch, W. et al. (2015) The dynamic control of signal transduc-tion networks in cancer cells. Nat. Rev. Cancer 15,515–527
2. Balla, T. (2013) Phosphoinositides: tiny lipids with giant impact oncell regulation. Physiol. Rev. 93, 1019–1137
3. Thorpe, L.M. et al. (2015) PI3K in cancer: divergent roles of iso-forms, modes of activation and therapeutic targeting. Nat. Rev.Cancer 15, 7–24
4. Vanhaesebroeck, B. et al. (2012) PI3K signalling: the pathto discovery and understanding. Nat. Rev. Mol. Cell Biol. 13,195–203
5. Davis, W.J. et al. (2015) Nuclear PI3K signaling in cell growth andtumorigenesis. Front. Cell Dev. Biol. 3, 24
6. Follo, M.Y. et al. (2015) PLC and PI3K/Akt/mTOR signalling indisease and cancer. Adv. Biol. Regul. 57, 10–16
7. Posor, Y. et al. (2015) Phosphoinositides in endocytosis. Biochim.Biophys. Acta 1851, 794–804
9. Choi, S. et al. (2015) PIP kinases define PI4,5P2 signaling speci-ficity by association with effectors. Biochim. Biophys. Acta. 1851,711–723
10. Heck, J.N. et al. (2007) A conspicuous connection: structuredefines function for the phosphatidylinositol-phosphate kinasefamily. Crit. Rev. Biochem. Mol. Biol. 42, 15–39
11. Tsujita, K. and Itoh, T. (2015) Phosphoinositides in the regulationof actin cortex and cell migration. Biochim. Biophys. Acta 1851,824–831
12. Ling, K. et al. (2006) Movin’ on up: the role of PtdIns(4,5)P2 in cellmigration. Trends Cell Biol. 16, 276–284
13. Hokin, M.R. and Hokin, L.E. (1953) Enzyme secretion and theincorporation of P32 into phospholipides of pancreas slices. J.Biol. Chem. 203, 967–977
14. Auger, K.R. et al. (1989) PDGF-dependent tyrosine phosphory-lation stimulates production of novel polyphosphoinositides inintact cells. Cell 57, 167–175
15. Traynor-Kaplan, A.E. et al. (1989) Transient increase in phospha-tidylinositol 3,4-bisphosphate and phosphatidylinositol trisphos-phate during activation of human neutrophils. J. Biol. Chem. 264,15668–15673
16. Carpenter, C.L. et al. (1990) Purification and characterization ofphosphoinositide 3-kinase from rat liver. J. Biol. Chem. 265,19704–19711
17. Shibasaki, F. et al. (1991) Two types of phosphatidylinositol3-kinase from bovine thymus. Monomer and heterodimer form.J. Biol. Chem. 266, 8108–8114
18. Nave, B.T. et al. (1999) Mammalian target of rapamycin is a directtarget for protein kinase B: identification of a convergence point foropposing effects of insulin and amino-acid deficiency on proteintranslation. Biochem. J. 344, 427–431
19. Maehama, T. andDixon, J.E. (1998) The tumor suppressor, PTEN/MMAC1, dephosphorylates the lipid second messenger, phos-phatidylinositol 3,4,5-trisphosphate. J. Biol. Chem. 273, 13375–13378
20. Alessi, D.R. et al. (1998) 3-Phosphoinositide-dependent proteinkinase 1 (PDK1) phosphorylates and activates the p70 S6 kinasein vivo and in vitro. Curr. Biol. 8, 69–81
21. Alessi, D.R. et al. (1997) 3-Phosphoinositide-dependent proteinkinase-1 (PDK1): structural and functional homology with theDrosophila DSTPK61 kinase. Curr. Biol. 7, 776–789
22. Dunlop, E.A. and Tee, A.R. (2014) mTOR and autophagy: adynamic relationship governed by nutrients and energy. Semin.Cell Dev. Biol. 36, 121–129
23. Fruman, D.A. and Rommel, C. (2014) PI3K and cancer: lessons,challenges and opportunities.Nat. Rev. Drug Discov. 13, 140–156
24. Bunney, T.D. and Katan, M. (2010) Phosphoinositide signalling incancer: beyond PI3K and PTEN. Nat. Rev. Cancer 10, 342–352
25. Song, M.S. et al. (2012) The functions and regulation of the PTENtumour suppressor. Nat. Rev. Mol. Cell Biol. 13, 283–296
26. Bader, A.G. et al. (2005) Oncogenic PI3K deregulates transcriptionand translation. Nat. Rev. Cancer 5, 921–929
27. Zhao, L. and Vogt, P.K. (2008) Class I PI3K in oncogenic cellulartransformation. Oncogene 27, 5486–5496
28. Samuels, Y. et al. (2004) High frequency of mutations of thePIK3CA gene in human cancers. Science 304, 554
29. Freeman, M.R. (2004) HER2/HER3 heterodimers in prostate can-cer: whither HER1/EGFR? Cancer Cell 6, 427–428
30. Bauer, T.M. et al. (2015) Targeting PI3 kinase in cancer. Pharma-col. Ther. 146, 53–60
31. Yap, T.A. et al. (2015) Drugging PI3K in cancer: refining targets andtherapeutic strategies. Curr. Opin. Pharmacol. 23, 98–107
32. Patel, S.B. et al. (2016) Everolimus versus temsirolimus in meta-static renal cell carcinoma after progression with previous sys-temic therapies. Clin. Genitourin. Cancer 14, 153–159
33. Shah, A. and Mangaonkar, A. (2015) Idelalisib: a novel PI3Kdinhibitor for chronic lymphocytic leukemia. Ann. Pharmacother.49, 1162–1170
34. Zanardi, E. et al. (2015) Better together: targeted combinationtherapies in breast cancer. Semin. Oncol. 42, 887–895
35. Massacesi, C. et al. (2016) PI3K inhibitors as new cancer thera-peutics: implications for clinical trial design. Onco Targets Ther. 9,203–210
36. Yang, S.X. et al. (2016) New insights on PI3K/AKT pathwayalterations and clinical outcomes in breast cancer. Cancer Treat.Rev. 45, 87–96
37. Thapa, N. et al. (2015) Phosphatidylinositol phosphate 5-kinase Igand phosphoinositide 3-kinase/Akt signaling couple to promoteoncogenic growth. J. Biol. Chem. 290, 18843–18854
38. Good, M.C. et al. (2011) Scaffold proteins: hubs for controlling theflow of cellular information. Science 332, 680–686
39. Choi, S. et al. (2013) IQGAP1 is a novel phosphatidylinositol 4,5bisphosphate effector in regulation of directional cell migration.EMBO J. 32, 2617–2630
40. Fiume, R. et al. (2015) PIP4K and the role of nuclear phosphoi-nositides in tumour suppression. Biochim. Biophys. Acta 1851,898–910
41. Gupta, A. et al. (2013) Phosphatidylinositol 5-phosphate 4-kinase (PIP4K) regulates TOR signaling and cell growth duringDrosophila development. Proc. Natl Acad. Sci. U.S.A. 110,5963–5968
42. Jenkins, G.H. et al. (1994) Type I phosphatidylinositol 4-phosphate5-kinase isoforms are specifically stimulated by phosphatidic acid.J. Biol. Chem. 269, 11547–11554
43. Yoon, M.S. et al. (2015) Rapid mitogenic regulation of themTORC1 inhibitor, DEPTOR, by phosphatidic acid. Mol. Cell58, 549–556
44. Sun, Y. et al. (2013) Phosphatidylinositol 4,5-bisphosphate: tar-geted production and signaling. Bioessays 35, 513–522
45. McCrea, H.J. and De Camilli, P. (2009) Mutations in phosphoino-sitide metabolizing enzymes and human disease. Physiology(Bethesda) 24, 8–16
12 Trends in Cancer, Month Year, Vol. xx, No. yy
46. Wymann, M.P. and Schneiter, R. (2008) Lipid signalling in disease.Nat. Rev. Mol. Cell Biol. 9, 162–176
47. Narkis, G. et al. (2007) Lethal contractural syndrome type 3(LCCS3) is caused by a mutation in PIP5K1C, which encodesPIPKIg of the phosphatidylinositol pathway. Am. J. Hum. Genet.81, 530–539
48. Thapa, N. et al. (2013) Phosphatidylinositol phosphate 5-kinaseIgi2 in association with Src controls anchorage-independentgrowth of tumor cells. J. Biol. Chem. 288, 34707–34718
49. Sun, Y. et al. (2010) Type I gamma phosphatidylinositol phosphatekinase modulates invasion and proliferation and its expressioncorrelates with poor prognosis in breast cancer. Breast CancerRes. 12, R6
50. Chen, C. et al. (2015) Targeting type Ig phosphatidylinositol phos-phate kinase inhibits breast cancer metastasis. Oncogene 34,4635–4646
51. Sun, Y. et al. (2013) Endosomal type Ig PIP 5-kinase controls EGFreceptor lysosomal sorting. Dev. Cell 25, 144–155
52. Tan, X. et al. (2015) LAPTM4B is a PtdIns(4,5)P2 effector thatregulates EGFR signaling, lysosomal sorting, and degradation.EMBO J. 34, 475–490
53. Waugh, M.G. (2014) Amplification of chromosome 1q genesencoding the phosphoinositide signalling enzymes PI4KB,AKT3, PIP5K1A and PI3KC2B in breast cancer. J. Cancer 5,790–796
54. Nagaraj, N. et al. (2011) Deep proteome and transcriptome map-ping of a human cancer cell line. Mol. Syst. Biol. 7, 548
55. Emerling, B.M. et al. (2013) Depletion of a putatively druggableclass of phosphatidylinositol kinases inhibits growth of p53-nulltumors. Cell 155, 844–857
56. Keune, W.J. et al. (2013) Low PIP4K2B expression in humanbreast tumors correlates with reduced patient survival: a role forPIP4K2B in the regulation of E-cadherin expression. Cancer Res.73, 6913–6925
57. Sumita, K. et al. (2016) The lipid kinase PI5P4Kb is an intracellularGTP sensor for metabolism and tumorigenesis. Mol. Cell 61,187–198
58. Jude, J.G. et al. (2015) A targeted knockdown screen of genescoding for phosphoinositide modulators identifies PIP4K2A asrequired for acute myeloid leukemia cell proliferation and survival.Oncogene 34, 1253–1262
59. Glasauer, A. and Chandel, N.S. (2014) Targeting antioxidants forcancer therapy. Biochem. Pharmacol. 92, 90–101
60. Stijf-Bultsma, Y. et al. (2015) The basal transcription complexcomponent TAF3 transduces changes in nuclear phosphoinosi-tides into transcriptional output. Mol. Cell 58, 453–467
61. Jones, D.R. et al. (2013) PtdIns5P is an oxidative stress-inducedsecond messenger that regulates PKB activation. FASEB J. 27,1644–1656
62. Lamouille, S. et al. (2014) Molecular mechanisms of epithelial–mesenchymal transition. Nat. Rev. Mol. Cell Biol. 15, 178–196
63. Jung, H.Y. et al. (2015) Molecular pathways: linking tumor micro-environment to epithelial–mesenchymal transition in metastasis.Clin. Cancer Res. 21, 962–968
64. Devergne, O. et al. (2014) Polarized deposition of basementmembrane proteins depends on phosphatidylinositol synthaseand the levels of phosphatidylinositol 4,5-bisphosphate. Proc. NatlAcad. Sci. U.S.A. 111, 7689–7694
65. Martin-Belmonte, F. et al. (2007) PTEN-mediated apical segrega-tion of phosphoinositides controls epithelial morphogenesisthrough Cdc42. Cell 128, 383–397
66. Claret, S. et al. (2014) PI(4,5)P2 produced by the PI4P5K SKTLcontrols apical size by tethering PAR-3 in Drosophila epithelialcells. Curr. Biol. 24, 1071–1079
67. Gassama-Diagne, A. et al. (2006) Phosphatidylinositol-3,4,5-tri-sphosphate regulates the formation of the basolateral plasmamembrane in epithelial cells. Nat. Cell Biol. 8, 963–970
68. Ling, K. et al. (2007) Type I gamma phosphatidylinositol phosphatekinase modulates adherens junction and E-cadherin trafficking viaa direct interaction with mu 1B adaptin. J. Cell Biol. 176, 343–353
69. Ling, K. et al. (2002) Type I gamma phosphatidylinositol phosphatekinase targets and regulates focal adhesions. Nature 420, 89–93
70. Schill, N.J. and Anderson, R.A. (2009) Out, in and back again:PtdIns(4,5)P2 regulates cadherin trafficking in epithelial morpho-genesis. Biochem. J. 418, 247–260
71. Thapa, N. et al. (2012) Phosphoinositide signaling regulates theexocyst complex and polarized integrin trafficking in directionallymigrating cells. Dev. Cell 22, 116–130
72. Thapa, N. and Anderson, R.A. (2012) PIP2 signaling, an integratorof cell polarity and vesicle trafficking in directionally migrating cells.Cell Adh. Migr. 6, 409–412
73. Xiong, X. et al. (2012) An association between type Ig PI4P 5-kinase and Exo70 directs E-cadherin clustering and epithelialpolarization. Mol. Biol. Cell 23, 87–98
74. Schill, N.J. et al. (2014) Isoform 5 of PIPKIg regulates the endo-somal trafficking and degradation of E-cadherin. J. Cell Sci. 127,2189–2203
75. Sun, Y. et al. (2007) Type I gamma phosphatidylinositol phosphatekinase is required for EGF-stimulated directional cell migration. J.Cell Biol. 178, 297–308
76. Wehrle-Haller, B. (2012) Structure and function of focal adhesions.Curr. Opin. Cell Biol. 24, 116–124
77. Hu, P. and Luo, B.H. (2013) Integrin bi-directional signaling acrossthe plasma membrane. J. Cell. Physiol. 228, 306–312
78. Zhang, L. et al. (2012) Phosphatidylinositol 4, 5 bisphosphate andthe actin cytoskeleton. Subcell. Biochem. 59, 177–215
79. Sechi, A.S. and Wehland, J. (2000) The actin cytoskeleton andplasmamembrane connection: PtdIns(4,5)P2 influences cytoskeletal
protein activity at the plasma membrane. J. Cell Sci. 113, 3685–3695
80. Clark, E.A. et al. (2000) Genomic analysis of metastasis reveals anessential role for RhoC. Nature 406, 532–535
81. Gross, S.R. (2013) Actin binding proteins: their ups and downs inmetastatic life. Cell Adh. Migr. 7, 199–213
82. Anderson, R.A. and Marchesi, V.T. (1985) Regulation of the asso-ciation of membrane skeletal protein 4.1 with glycophorin by apolyphosphoinositide. Nature 318, 295–298
83. Iwaya, K. et al. (2007) Coexpression of Arp2 andWAVE2predictspoor outcome in invasive breast carcinoma. Mod. Pathol. 20,339–343
84. Tomasevic, N. et al. (2007) Differential regulation of WASP and N-WASP by Cdc42, Rac1, Nck, and PI(4,5)P2. Biochemistry 46,3494–3502
85. Prehoda, K.E. et al. (2000) Integration of multiple signals throughcooperative regulation of the N-WASP–Arp2/3 complex. Science290, 801–806
86. Lane, J. et al. (2014) Structure and role of WASP andWAVE in RhoGTPase signalling in cancer. Cancer Genomics Proteomics 11,155–165
87. Dumontet, C. and Jordan, M.A. (2010) Microtubule-bindingagents: a dynamic field of cancer therapeutics. Nat. Rev. DrugDiscov. 9, 790–803
