[Cancer Biology & Therapy 2:3, 222-232, May/June 2003]; ©2003 Landes Bioscience
`
`Review
`Rapamycins
`Mechanism of Action and Cellular Resistance
`Shile Huang
`ABSTRACT
`Mary-Ann Bjornsti
`Rapamycins are macrocyclic lactones that possess immunosuppressive, antifungal and
`antitumor properties. The parent compound, rapamycin, is approved as an immunosup-
`Peter J. Houghton*
`pressive agent for preventing rejection in patients receiving organ transplantation. Two
`analogues, CCI-779 and RAD001 are currently being investigated as anticancer agents.
`Rapamycins first bind a cyclophilin FKBP12, and this complex binds and inhibits the
`function of mTOR (mammalian target of rapamycin) a serine/threonine (Ser/Thr) kinase
`with homology to phosphatidylinositol 3’ kinase. Currently, as mTOR is the only identified
`target, this places rapamycins in a unique position of being the most selective kinase
`inhibitor known. Consequently these agents have been powerful tools in elucidating the
`role of mTOR in cellular growth, proliferation, survival and tumorigenesis. Increasing
`evidence suggests that mTOR acts as a central controller sensing cellular environment
`(nutritional status or mitogenic stimulation) and regulating translation initiation through the
`eukaryotic initiation factor 4E, and ribosomal p70 S6 kinase pathways. Here we review
`the conserved TOR signaling pathways, conceptual basis for tumor selectivity, and the
`mechanisms of resistance to this class of antitumor agent.
`
`Department of Molecular Pharmacology; St Jude Children's Research Hospital; 332
`N. Lauderdale; Memphis; Tennessee USA
`
`*Correspondence to: Peter J. Houghton; Department of Molecular Pharmacology; St
`Jude Children's Research Hospital; 332 N. Lauderdale; Memphis; Tennessee
`38105-2794 USA; Email: peter.houghton@stjude.org
`
`Received 04/24/03; Accepted 05/01/03
`
`Previously published online as a CB&T Paper in Press at:
`http://www.landesbioscience.com/journals/cbt/toc.php?volume=2&issue=3
`
`©2003 Landes Bioscience. Not for distribution.
`
`KEY WORDS
`
`Rapamycins, Translation initiation, Cancer,
`Resistance, Therapy, Ribosomal biogenesis, Yeast
`
`Work reported from these laboratories was supported by PHS awards CA23099,
`CA58755, CA77776, CA96996 and CA21765 (Cancer Center Support Grant) and by
`American, Lebanese, Syrian Associated Charities (ALSAC).
`
`INTRODUCTION
`Rapamycin, a macrocyclic lactone product of the soil bacteria Streptomyces hygroscopicus,
`was isolated and identified as an antifungal agent in the mid-1970’s.1-3 Rapamycin
`(sirolimus), is a structural analogue of the macrolide antibiotic FK506 (tacrolimus,
`Prograf®) (Fig. 1), and like FK506 was found to potently suppress the immune system.4-7
`The potential for rapamycin to act as an antitumor agent was recognized early in its
`development when the drug demonstrated potent inhibitory activity against numerous
`solid tumors in the NCI screening program.8-10 However, the drug was not developed
`further due to stability and solubility problems that prevented development of a parenteral
`formulation for use in clinical trials. Also at that time in the early 1980’s, the mechanism
`of action of rapamycin in blocking signal transduction was not understood.
`Rapamycin (Rapamune®), as an immunosuppressive drug, was finally approved by the
`Food and Drug Administration (FDA) in the USA in 1999, and the European
`Commission in 2000, respectively. Results from many laboratories have demonstrated that
`rapamycin, in contrast to FK506, is not only a potent immunosuppressant, but also an
`active antitumor agent. Rapamycin can act as a cytostatic, slowing or arresting cells in G1
`phase. Under specific conditions, or in some tumor cell lines rapamycins may induce
`apoptosis in culture. To date, studies have revealed that rapamycin potently arrests growth
`of cells derived from rhabdomyosarcoma, neuroblastoma and glioblastoma, small cell lung
`cancer,11-17 osteoscarcoma,18 pancreatic cancer,19,20 breast and prostate cancer,21-23
`murine melanoma and leukemia, and B-cell lymphoma.9,24-26 In addition to broad
`spectrum activity in vitro, rapamycin and its derivatives (designated here as rapamycins)
`suppress growth of some human and murine tumor models in vivo.11-30 When combined
`with other chemotherapeutic agents, rapamycins generally show at least additive antitumor
`activity.10,12,17,31 Preliminary data from clinical trials have indicated that rapamycins are
`well tolerated and successfully suppress growth of various human tumors.32-34
`The use of rapamycin as an anticancer drug is clinically impractical, because of its poor
`water-solubility and stability in solution. Recently, rapamycin ester analogues (Fig. 1),
`CCI-779 [rapamycin-42, 2, 2-bis(hydroxymethyl)-propionic acid] (Wyeth-Ayerst, PA,
`USA) and RAD001 [everolimus, 40-O-(2-hydroxyethyl)-rapamycin] (Novartis, Basel,
`Switzerland), have entered clinical trials. These analogues have improved pharmaceutical
`properties. CCI-779 is being developed for both intravenous and oral administration,
`whereas RAD001 is only for oral administration. The antitumor activity of these analogues
`
`222
`
`Cancer Biology & Therapy
`
`2003; Vol. 2 Issue 3
`
`Ex. 1104-0001
`
`

`
`RAPAMYCINS: MECHANISM OF ACTION AND CELLULAR RESISTANCE
`
`is similar to that of rapamycin.17,21-23,27-30 RAD001 is
`in Phase I trials whereas the development of CCI-779
`is more advanced with several Phase II trials completed.
`Why should it be anticipated that rapamycins
`could exhibit tumor-selectivity, in a manner analogous
`to the activity of another kinase inhibitor, Gleevec, in
`Bcr-Abl expressing chronic lymphocytic leukemia?
`Accumulating evidence suggests that genetic mutations
`or compensatory changes in tumor cells may affect
`sensitivity to rapamycins. At least in some systems
`mutations that occur frequently in malignant trans-
`formations such as GLI amplification, or mutations
`that inactivate p53, and the dual specificity phos-
`phatase PTEN (phosphatase and tensin homolog
`deleted on chromosome ten, also known as MMAC1
`for mutated in multiple advanced cancers) or lead to
`activation of Akt appear to determine rapamycin
`sensitivity. On the other hand there may be multiple
`loci that confer intrinsic or acquired resistance. This
`review will summarize the current knowledge of the
`role of mTOR in cellular regulation, the mechanism
`of action of rapamycins, and currently understood
`resistance mechanisms.
`
`Figure 1. The structure of rapamycin and its analogues FK506, CCI-779 and RAD001. The
`FKBP12 binding face is shown by the filled bars, whereas the mTOR binding face of
`rapamycin is shown by the hatched bar.
`
`THE RAPAMYCIN TARGET (MTOR)
`The mammalian target of rapamycin, [also named
`FKBP12 and rapamycin-associated protein (FRAP),
`rapamycin and FKBP12 target 1 (RAFT1), rapamycin
`target 1 (RAPT1), or sirolimus effector protein
`(SEP)], is a 289 kDa Ser/Thr kinase orthologue of
`TOR1 and TOR2 in Saccharomyces cerevisiae.35,36
`TOR is an atypical serine/threonine kinase highly
`conserved from yeast to mammals. Human, mouse
`and rat mTOR proteins share 95% identity at the amino acid
`level.36-38 Since the C-terminus of TOR is highly homologous to the
`catalytic domain of phosphatidylinositol 3´ kinase (PI3K), mTOR is
`considered to belong to the PI3K-related protein kinase (designated
`PIKK) family, which also includes Mec1, Tel1, RAD3, MEI-41,
`DNA-PK, ATM, ATR, and TRRAP.36,37 Recently, single TOR
`homologs have also been identified in fungi (TOR1 in Cryptococcus
`neoformans), plants (AtTOR in Arabidopsis thaliana), worms
`(CeTOR in Caenorhabditis elegans), and flies (dTOR in Drosophila
`menalogaster).37
`The domain structure of mTOR is depicted in Figure 2. The
`protein consists of a catalytic kinase domain, an FKBP12-rapamycin
`binding (FRB) domain and a putative auto-inhibitory domain
`(“repressor domain”) near the C-terminus, and up to 20 tandemly
`repeated HEAT (Huntingtin, EF3, A subunit of PP2A and TOR)
`motifs at the N-terminus, as well as FAT (FRAP-ATM-TRRAP) and
`FATC (FAT C-terminus) domains. HEAT motifs may serve as
`protein-protein interaction parts, whereas FAT and FATC domains
`
`Figure 2. Schematic representation of the domains of mTOR. Structural
`domains of mTOR. HEAT :(huntingtin elongation factor A subunit of PP2A
`and TOR) repeats (positions 71-522 and 628-1147); FAT: (FRAP-ATM-
`TRAPP) domain, which is unique to PIK-related kinases located N-terminal to
`the FKBP12-rapamycin binding domain (FRB); the role of
`FAT sequences is less clear, but they are associated with
`C-terminal FAT (FATC) sequences in mTOR. Interaction
`between FAT and FATC domains may facilitate protein bind-
`ing or act as a structural scaffold; CD: Catalytic domain;
`RD: regulatory domain.
`
`may participate in modulation of catalytic kinase activity of
`mTOR.36 The remarkable conservation of mTOR at the amino acid
`level suggests that multiple domains of this protein are essential for
`its cellular functions.
`Tor Signaling in Yeast. In yeast, Tor kinase activity is regulated
`by availability of nutrients (amino acids and glucose), whereas in
`mammalian cells, mTOR is regulated also by phosphatidic acid,
`ATP, and growth factors.37,39-42 The Tor signaling pathway in yeast
`is depicted in Figure 3, and controls translation initiation, protein
`turnover, transcription, and actin cytoskeleton organization. In yeast
`these pathways have been rigorously established,37 and are at least in
`part maintained in mammalian cells. The Tor1/2 complex (desig-
`nated TORC1) comprising Kog1 (the yeast homologue of the
`mammalian protein raptor) and Lst8 controls translation, protein
`stability and transcription,43-45 whereas the TORC2 complex
`controls actin organization. As TORC2 is not a rapamycin target it
`will not be considered further. The evolutionarily conserved TORC1
`complex controls translation initiation probably through activation
`of eIF4E, and transcription of ribosomal genes, stress response
`genes, ribosomal biogenesis and tRNA synthesis.
`Tor, through the TORC1 complex, controls protein turnover and
`some aspects of transcription through regulation of protein phos-
`
`www.landesbioscience.com
`
`Cancer Biology & Therapy
`
`223
`
`Ex. 1104-0002
`
`

`
`RAPAMYCINS: MECHANISM OF ACTION AND CELLULAR RESISTANCE
`
`Figure 3. Nutrient signaling in yeast. (Adapted from Jacinto and Hall.37)
`
`phatases. Tor positively regulates Tap42, which binds to and inacti-
`vates the Sit4 and Pph protein phosphatases.46,47 Pph21 and Pph22
`are calaytic subunits of PP2A, and Sit4 is the catalytic subunit of a
`PP2A-related phosphatase in yeast. Nutrient deficiency or rapamycin
`leads to dissociation of these phosphatases from Tap42, resulting in
`increased phosphatase activity. This leads to dephosphorylation of
`Npr1 and Gln3 involved in protein turnover and transcription,
`respectively. Control of protein turnover and some aspects of tran-
`scription appear to be regulated through Tap42 binding and inactivat-
`ing Sit4, whereas pathways regulated by release of Pph phosphatases are
`less well defined. In the presence of nutrients Tor signaling represses
`autophagy and leads to stabilization of proteins by suppressing
`ubiquitin-dependent degradation.48 For example, Tor signaling
`prevents ubiquitylation, vacuolar targeting and degradation of the
`tryptophan transporter Tat2 by maintaining phosphorylation and
`inactivation of Npr1 a putative Tat2 kinase.49,50 Under conditions of
`starvation Sit4 becomes activated leading to dephosphorylation of
`Npr1 and degradation of Tat2. The Tap42-Sit4 pathway also controls
`the Gln3 transcription factor. Under nutrient replete conditions
`Gln3 is phosphorylated and is bound to the Ure2 protein in the
`cytoplasm. Inhibition of Tor by rapamycin or nitrogen starvation
`leads to dephosphorylation of Gln3, its nuclear translocation and
`transcription of genes required for the use of secondary nitrogen
`sources.43,44 The pathway(s) downstream of Tap42 involving Pph
`are less clear. Similarly, Tor negatively regulates the heterodimeric
`transcription factor Rtg1-Rtg3 through an unknown mechanism.51
`The Tor pathway also is important in control of stress-responses
`through modulation of transcription. The TORC1 complex nega-
`
`tively controls transcription of stress-responsive genes through the cyto-
`plasmic sequestration of the general stress transcription factors Msn2
`and Msn4. Although the mechanism is not fully understood this
`may occur through Tor signaling promoting the binding of these
`transcription factors to the 14-3-3 homologues Bmh1 and Bmh2.50
`Proximal Signaling in Mammalian Cells. In mammalian cells,
`mTOR is regulated not only by nutrients but also by growth fac-
`tors.37,39-42 It appears that growth factors regulate mTOR signaling
`through both PI3K and Akt pathways, whereas proximal activators
`regulated by nutrients and ATP are less well characterized. In mam-
`malian cells mTOR is activated as a consequence of ligand binding
`to various growth factor receptors that result in activation of PI3K,
`Figure 4. Activated PI3K catalyzes the conversion of phostidylinosi-
`tol (4,5)-bisphosphate (PIP2) to phosphatidylinositol-3,4,5-trispho-
`sphate (PIP3). This pathway is negatively regulated by a dual-speci-
`ficity protein and lipid phosphatase PTEN. Full activation of Akt,
`downstream of PI3K requires binding of PIP3 to the pleckstrin
`homology domain of Akt, and phosphorylation by phosphoinositide-
`dependent kinases 1/2 (PDK1/2) and other unidentified kinases.
`Pharmacological studies with albeit relatively non-specific inhibitors
`of PI3K (wortmannin and LY294002) indicate that mTOR is down-
`stream of PI3K. How mTOR is regulated by PI3K or Akt, however,
`is still not well understood. Akt can phosphorylate mTOR
`(Ser2448) directly, although the significance remains to be deter-
`mined.37,38 Recent studies have placed the tuberous sclerosis
`(TSC1/2) complex as a modulator between PI3K/Akt and mTOR.52-54
`The TSC1/2 complex comprises harmartin (TSC1) and tuberin
`(TSC2). These proteins form a physical and functional complex in
`
`224
`
`Cancer Biology & Therapy
`
`2003; Vol. 2 Issue 3
`
`Ex. 1104-0003
`
`

`
`RAPAMYCINS: MECHANISM OF ACTION AND CELLULAR RESISTANCE
`
`Figure 4. Growth factor signaling in mammalian cells.
`
`vivo, which binds and inhibits mTOR.52-54 Loss of TSC1/2 results
`in mTOR-dependent increase of ribosomal p70 S6 kinase (S6K1)
`activity, and confers resistance of cells to amino acid starvation.53
`Conversely co-expression of TSC1 and TSC2 inhibits amino
`acid-induced activation of S6K1 in nutrient-deprived cells.54
`Mitogenic stimuli, such as insulin or serum, activate Akt which can
`directly phosphorylate TSC2 on multiple sites in vitro and in
`vivo.52,55,56 Phosphorylation of TSC2 at Ser939 and Thr1462 is
`PI3K-dependent.52,56 Akt-mediated phosphorylation of TSC2
`destabilizes TSC2 and thereby inhibits the formation of TSC1/2
`complex, leading to de-repression of mTOR, and consequently
`increasing the kinase activity of mTOR. By contrast, treatment with
`rapamycin does not influence the phosphorylation of TSC2,52 and
`together with other data,53,54 suggests that mTOR lies downstream
`of TSC2. Other results57,58 imply that the TSC1/2 complex may
`mediate S6K1 activation through a pathway parallel to mTOR.
`However, further studies will be required to address the relationship
`of mTOR and TSC1/2.
`mTOR Associated Proteins. mTOR forms a scaffold complex
`with other proteins, such as raptor (regulatory associated protein of
`mTOR) and mLST8,59,60 which are the mammalian counterparts of
`yeast Kog1 (kontroller of growth 1) and Lst8,61 respectively. The
`exact role of raptor remains unclear.38,59-61 Raptor may act as a scaffold
`protein linking mTOR to S6K1 and 4E-BP1.60 Alternatively, it may
`have a dual function, inhibiting mTOR under nutrient-deficient
`conditions and stimulating mTOR in a nutrient-replete environment.59
`
`Thus TSC1/2, raptor and possibly mLST8 act as potential modulators
`of mTOR function in response to availability of nutrients.
`Signaling Distal to mTOR. mTOR controls translation initiation
`through two pathways, S6K1 and eukaryotic initiation factor 4E
`(eIF4E) binding proteins (4E-BPs). mTOR either directly phospho-
`rylates Thr389 of S6K1, or suppresses a phosphatase responsible for
`maintaining the hypophosphoryalted state of this residue. Activation
`of S6K1 enhances translation of mRNAs bearing a 5´ terminal
`oligopyrimidine tract (5´ TOP).62,63 Inactivation of S6K1 decreases
`synthesis of ribosomal proteins and translation factors.62,63
`Activation of S6K1 is complex. The process involves interplay
`between four different domains and at least seven specific sites medi-
`ated by multiple upstream kinases.64 At least 12 sites (Ser17,
`Thr229, Thr367, Thr371, Thr389, Ser404, Ser411, Ser418, Tr421,
`Ser424, Ser429, and Thr447) can be phosphorylated in response to
`serum stimulation.65 However, the responsible kinases have not been
`fully characterized. Phosphoinositide-dependent protein kinase 1
`(PDK1) phosphorylates Thr229 in vitro and in vivo.66 Atypical
`PKC isoforms and the Rho family of small G proteins (cdc42 and
`Rac1) may contribute to phosphorylation of S6K1,65 although the
`specific sites regulated by these kinases remain to be determined. In
`vitro, mTOR phosphorylates only Thr389 in the regulatory
`domain.67-69 However, whether this phosphorylation is directly or
`indirectly regulated by mTOR is in question, since recent data suggest
`that mTOR may regulate S6K1 activation by inhibiting phos-
`phatases rather than directly phosphorylating S6K1.64,70
`
`www.landesbioscience.com
`
`Cancer Biology & Therapy
`
`225
`
`Ex. 1104-0004
`
`

`
`RAPAMYCINS: MECHANISM OF ACTION AND CELLULAR RESISTANCE
`
`mTOR phosphorylates and inactivates the 4E-BP suppressor
`proteins causing their dissociation from the RNA cap-binding
`protein eIF4E. In response to mitogens, six sites (Thr37, Thr46,
`Ser65, Thr70, Ser83, and Ser112) of 4E-BP1 can be phosphorylated.65
`So far, only mTOR and ATM have been identified to be involved in
`phosphorylation of 4E-BP1.71-74 Other kinases that phosphorylate
`4E-BP1 remain to be characterized. ATM phosphorylates 4E-BP1 at
`Ser112,74 however the physiological significance of this remains to
`be elucidated. In vitro mTOR phosphorylates 4E-BP1 at two sites
`(Thr37, Thr46) and possibly at two additional Ser/Thr residues
`(Thr70 and Ser65) in the N-terminal region.73,75 Phosphorylation
`is a hierarchical process.73,75-77 Phosphorylation of Thr37/Thr46 is
`followed by Thr70 phosphorylation. Ser65 is phosphorylated last77
`and is dependent on phosphorylation of all three Thr/Pro sites.75,76
`Mutation of Thr37 and/or Thr46 to alanine(s) prevents phosphory-
`lation of Ser65 and Thr70, indicating that phosphorylation of
`Thr37 and Thr46 serves as a requisite “priming” event.36 Single
`phosphorylation of these residues is not sufficient to dissociate
`4E-BP1 from eIF4E, indicating the requirement of combined
`phosphorylation of at least Thr37, Thr46, Ser65, and Thr70 in
`4E-BP1 to suppress association with eIF4E.36,78 Inhibition of
`mTOR leads to rapid hypophosphorylation of 4E-BP1 which then
`tightly binds to eIF4E. This prevents formation of eIF4F complex
`that contains eIF4E, eIF4G, eIF4A and eIF3, and inhibits
`cap-dependent translation initiation.36 Once 4E-BP1 is hyperphos-
`phorylated, it releases eIF4E, facilitating eIF4F complex formation
`and promoting cap-dependent protein synthesis.36 Overall inhibition
`of mTOR by rapamycin leads to a decrease in protein synthesis of
`15 to 20 percent. However, as the eIF4E pathway is required for
`translation of mRNA’s encoding cyclin D1,79,80 and ornithine
`decarboxylase81 inhibition of mTOR leads to slowing or arrest of
`cells in G1 phase of the cell cycle. However, the exact mechanism(s)
`by which mTOR regulates cell cycle progression are complex, poor-
`ly understood, and potentially context specific. Although
`rapamycins are highly specific inhibitors, the TOR pathway regu-
`lates multiple cellular processes. The mTOR pathway regulates trans-
`initiation of survival factors such as c-MYC82 and
`lation
`hypoxia-inducible factor 1α, and consequently vascular endothelial
`growth factor.83,84 In addition, mTOR is involved in the regulation
`of cyclin A, cyclin dependent kinases (cdk1/2), cdk inhibitors
`(p21Cip1 and p27Kip1), retinoblastoma protein, RNA polymerase
`I/II/III-transcription and translation of rRNA and tRNA, protein
`phosphatases (PP2A, PP4 and PP6), and CLIP-170.36,37,85-91
`
`TUMOR SELECTIVITY OF RAPAMYCINS
`To date there are no reports suggesting that activating mutations
`of mTOR, or overexpression occur as primary events in malignant
`transformation. However, activation of signaling pathways both
`proximal and distal to mTOR appear to occur frequently in human
`cancer. Loss of the phosphatase PTEN by deletion, silencing or
`mutation leads to constitutive activation of Akt,29,30,92 and upregu-
`lation of mTOR-dependent pathways. In PTEN-deficient tumor
`cells or mouse embryo fibroblasts (MEFs), activated Akt is associat-
`ed with enhanced activity of S6K1 and hyperphosphorylation of
`4E-BP1,29,30 or increased levels of c-MYC.92 It is speculated that
`Akt-activated cells become dependent on upregulated mTOR
`signaling for proliferation, hence become more susceptible to
`rapamycin or CCI-779. Increased sensitivity to rapamycins has been
`demonstrated in a panel of brain, prostate, and breast cancer cells,
`multiple myeloma cells and in MEFs.23,29 The association of PTEN
`deficiency and sensitivity to rapamycin is further supported by the
`
`activity of CCI-779 against the growth of human tumors implanted
`in athymic nude mice.23,30 There are, however, some exceptions;
`cells with functional PTEN and low constitutive activation of Akt
`are equally sensitive to inhibition of proliferation by rapamycins.
`Consequently the role of PTEN as an independent variable predicting
`for rapamycin sensitivity remains to be demonstrated. Disruption of
`the TSC1 or TSC2 gene leads to the development of tumors in
`multiple organs, notably kidney brain heart and lung. Even low
`doses of rapamycin causes apoptosis and necrosis of spontaneous
`renal tumors in Eker rats with germline mutation in TSC2 and
`results in tumor regression.93 These studies suggest that tumors
`developing in patients with tuberous sclerosis may be sensitive to
`rapamycins. Oncogene expression may also regulate the response to
`rapamycin. For example, in RK3E cells transformed with c-MYC or
`Ras rapamycin treatment increased global protein synthesis. In contrast
`rapamycin inhibited global protein synthesis and turnover in GLI
`transformed isogenic lines leading to inhibition of proliferation.94
`Although generally considered to be cytostatic agents, rapamycins
`can induce apoptosis in some cell systems. Rapamycins induce
`apoptosis of B-cells, rhabdomyosarcoma cells, renal tubular cells and
`dendritic cells.16,95-97 Rapamycin enhances transforming growth
`factor-β induced cell cycle arrest,98 and through blocking survival
`factor signaling99 rapamycins enhance cell death. Our results suggest
`that the functional status of the p53 tumor suppressor may dictate
`the cellular fate of rapamycin treated cells, as depicted in Figure 5.
`For example, under serum free conditions, the response to
`rapamycin in cells lacking functional p53 is apoptosis, suggesting
`that only in the absence of p53/p21 inhibition of mTOR becomes
`lethal (so-called synthetic lethality). Ectopic expression of p53 or
`p21Cip1 protects cells from rapamycin-induced apoptosis. The
`implication is that rapamycins may have potential tumor-selective
`therapeutic effects.16 Recent results show that inhibition of mTOR
`by rapamycin induces a cellular stress response characterized by rapid
`and sustained activation of ASK1 (apoptosis signal-regulating kinase
`1) signaling in p53-mutant cells.100 This leads to sustained phos-
`phorylation of c-JUN (Ser63) that appears to be responsible for
`inducing apoptosis. Rapamycin-induced stress appears distinct from
`that induced by ultra violet radiation in that MEKK1 is not activated,
`and from cytotoxic agents such as DNA damaging agents where
`other stress pathways (p38 or ERK1/2) are also activated.101,102 In
`contrast, cells expressing wild type p53, (or constitutive p21Cip1
`expression) there is only transient activation of ASK1.100
`Suppression of ASK1 is associated with binding of p21Cip1 in
`rapamycin treated cells, and protection from apoptosis. Since the
`rapamycin-induced apoptosis is specifically prevented by insulin-like
`growth factors (IGF-I/II) and insulin,96 combination of IGF receptor
`inhibitors with rapamycins may be selectively cytotoxic and induce
`regression of tumors with p53 mutations. Whether such anticipation
`is justified requires vigorous testing.
`
`MECHANISMS OF RESISTANCE TO RAPAMYCINS
`Intrinsic sensitivity to rapamycins between cell lines may vary by
`several orders of magnitude. For example, rhabdomyosarcoma cells
`in vitro are very sensitive (concentration for 50% inhibition (IC50)
`of proliferation ~ 1 nM) compared to colon carcinoma cells (IC50 >
`5000 nM).11 Mechanisms of intrinsic and acquired resistance may
`have either a genetic or epigenetic basis.
`Mutations in FKBP12 and mTOR. Rapamycins first bind to the
`cyclophilin FKBP12 in mammalian cells, forming the FKBP-
`rapamycin complex. This complex then interacts with the FRB
`domain in mTOR (Fig. 2), and inhibits the function of mTOR.
`
`226
`
`Cancer Biology & Therapy
`
`2003; Vol. 2 Issue 3
`
`Ex. 1104-0005
`
`

`
`RAPAMYCINS: MECHANISM OF ACTION AND CELLULAR RESISTANCE
`
`Rapamycin resistance may be conferred by
`mutations in FKBP12 that prevent the forma-
`tion of FKBP-rapamycin complex, or mutations
`in the FRB domain of mTOR that block binding
`of FKBP-rapamycin complex to mTOR. Such
`mutations were first identified in budding yeast
`S. cerevisiae in which treatment with rapamycin
`irreversibly arrests cells in the G1 phase. In the
`yeast S. cerevisiae, deletion of the RBP1 gene, a
`homologue of mammalian FKBP-12, results in a
`recessive rapamycin resistance, whereas expres-
`sion of RBP1 restores rapamycin sensitivity.103
`This observation has been further confirmed by
`RBP1 disruption experiments using the patho-
`genic yeast Candida albicans. Wild-type
`RBP1/RBP1 parental strain and the rbp1/RBP1
`heterozygous mutant were sensitive to rapamycin
`inhibition, whereas rbp1/rbp1 homozygous
`mutant was rapamycin resistant.104 In addition, in
`S. cerevisiae mutation of a specific residue (Tyr89)
`which is conserved in RBP1 or FKBPs, also
`resulted in decreased binding of rapamycin and
`conferred a recessive resistance phenotype.105 In
`murine mast cells, two distinct point mutations
`in FKBP12 confer resistance. By altering a
`hydrophobic residue within the drug-binding
`pocket (Trp59"Leu) or changing a charged sur-
`face residue (Arg49"Gln), the binding affinity for
`rapamycin decreases substantially.106
`A genetic screen identified rapamycin-resist-
`ant alleles with mutations in genes designated
`TOR1 and TOR2. Strains with mutated to tor1-1
`(Ser1972"Arg) and tor2-1 (Ser1975"Arg), were
`completely resistant to the growth- inhibitory
`effect of rapamycin. These resistant alleles encode mutant Tor proteins
`that do not bind the FKBP-rapamycin complex.103,107-111 This
`result suggests that a conserved serine residue (Ser1972 in Tor1;
`Ser1975 in Tor2) in Tor proteins is critical for FKBP-rapamycin
`binding. In mammalian cells mutations in the FRB domain confer
`a dominant resistant phenotype consistent with decreased affinity
`for binding of the FKBP- rapamycin complex.111 Expression of a
`mutant mTOR (Ser2035"Ile), having greatly reduced binding affin-
`ity for the FKBP-rapamycin complex, confers high level resist-
`ance.14,112,113
`Deregulation of eIF4E. mTOR phosphorylates and regulates the
`function of 4E-BP1, the suppressor of eIF4E.36 Recently, our group
`has found that acquired resistance to rapamycin was associated with
`decreased levels of 4E-BP1.121 In the absence of selective pressure
`(rapamycin), resistance was unstable and cells reverted to being
`sensitive to growth inhibition of rapamycin within ten weeks. In
`resistant cells the levels of 4E-BP1 were reduced significantly
`(~10-fold), whereas in rapamycin-sensitive revertants the 4E-BP1
`levels increased to those in wild type cells. Levels of 4E-BP1 transcripts
`were unaltered in rapamycin resistant clones suggesting post-tran-
`scriptional regulation. Further studies indicate that the synthesis of
`4E-BP1 significantly decreased in rapamycin-resistant clones.
`Whether the steady state level of 4E-BP1 is also regulated by
`increased degradation remains to be determined. Thus, the changes
`in 4E-BP1 levels are reminiscent of those reported for p27Kip1 in
`rapamycin resistant BC3H cells.115 In contrast, no consistent
`changes were detected in the level or activity of S6K1 between parental
`
`Figure 5. Loss of p53 function alters cellular response to rapamycin from cytostasis to apoptosis
`in murine embryo fibroblasts (MEFs). Left: schematic representation of synthetic lethality. Right:
`Wild type, and p53-/- MEFs, and p53-/- MEFs infected withAd-p53 (MOI of 100) were grown
`without or with rapamycin (100 ng/ml). Cells were harvested after 5 days and apoptosis deter-
`mined by quantitative FACs analysis (ApoAlert) assay. The per cent distribution of cells in each
`quadrant is presented (from Huang et al. 200116).
`
`and resistant clones. Rapamycin also inhibited growth factor activa-
`tion of S6K1 equally in parental and rapamycin-resistant clones.
`Intrinsic resistance to rapamycin has been shown in glioblastoma
`cells and colon adenocarcinoma that have very low 4E-BP1.114 For
`example, 4E-BP1 is barely detected in HCT8 colon carcinoma cells
`that are highly resistant to rapamycin (IC50 > 10,000 ng/ml). When
`4E-BP1 is overexpressed, these cells become sensitive (IC50 < 10
`ng/ml) to rapamycin, Figure 6.114 These data suggest that low levels
`of 4E-BP1 results in de-regulation of eIF4E, conferring rapamycin
`resistance.
`These results suggest that rapamycin-regulation of the eIF4E
`pathway is crucial in inducing growth arrest. Further de-regulation
`of eIF4E may facilitate a malignant phenotype. Of interest is that
`both rapamycin-resistant and -revertant cells exhibited elevated
`c-MYC levels, and increased anchorage-independent growth. That
`deregulation of the eIF4E pathway is associated with increased
`malignancy is supported by certain clinical observations that dereg-
`ulation of the eIF4E pathway does promote tumor progression.116
`In addition to decreased 4E-BP1 expression, as described above,
`increased eIF4E levels may also cause de-regulation of eIF4E. In
`advanced head and neck carcinoma,117 breast carcinoma118 gas-
`trointestinal carcinoma,119 and peripheral carcinoma of the lung120
`eIF4E levels are elevated. However, levels of 4E-BP1 suppressor
`proteins have not been reported in a consistent manner. Potentially,
`the ratio of 4E-BP:eIF4E may determine whether inhibition of
`mTOR elicits a biologically significant tumor response. Further
`studies will be necessary to determine if this ratio has predictive
`value for drug sensitivity of tumors.
`
`www.landesbioscience.com
`
`Cancer Biology & Therapy
`
`227
`
`Ex. 1104-0006
`
`

`
`RAPAMYCINS: MECHANISM OF ACTION AND CELLULAR RESISTANCE
`
`A
`
`B
`
`C
`
`Figure 6. Overexpression of 4E-BP abrogates resistance to rapamycin. (A)
`Western blot analysis of 4E-BP, eIF4E, and tubulin (loading control) in cell
`lines that have different intrinsic sensitivities to rapamycin. Colon carcinoma
`cell lines CaCo2, GC3/c1, HCT8, HCT29, HCT116, and VRC5/c1 are
`intrinsically resistant to rapamycin, with IC50 concentrations > 1200 ng/ml.
`Pediatric solid tumor lines SJ-G2 (glioblastoma) and Rh18 and Rh30 (rhab-
`domyosarcoma) are sensitive to rapamycin (IC50 < 1 ng/ml). (B) Expression
`of 4E-BP and eIF4E in HCT8 clones stably transfected with a 4E-BP expres-
`sion plasmid (pcDNA3-PHAS-I). Expression of 4E-BP was greater in clones
`C2, C4, and C5 than in parental HCT8 cells, but expression of was similar
`in parental and C1 and C3 transfected clones. (C) Sensitivity to rapamycin.
`Cells were plated at low density in increasing concentrations of rapamycin,
`and colonies were counted after 7 days of exposure to rapamycin. Symbols:
`Parental HCT8 (#) and clones C1 ($), C2 (%), C3(&), C4(▲), and C5( ∆).
`(Adapted from Dilling et al. 2002114).
`
`Mutations in S6K1. Ribosomal S6K1 is the other principal
`downstream effector of mTOR. Inhibition of mTOR by rapamycin
`primarily inhibits phosphorylation of Thr389 in the regulatory
`domain.64 However, since phosphorylation of S6K1 is hierarchical
`with phosphorylation of several other sites dependent on phospho-
`rylation of Thr389,64 rapamycin in vivo influences phosphorylation
`of other sites, including Thr229 in the catalytic domain, and S404
`in the regulatory domain.67 Mutation of Thr389"Glu abrogates the
`ability of rapamycin to inhibit S6K1 activation.67,121 Similarly,
`substitution of Thr229 by either a