Inhibitors of signal transduction protein kinases as targets for cancer therapy

Theresa Mikalsen, Nancy Gerits, and Ugo Moens*

Department of Microbiology and Virology, Institute of Medical Biology, University of Troms, N-9037 Troms, Norway *Corresponding author: Ugo Moens Department of Microbiology and Virology, Institute of Medical Biology, University of Troms, N-9037 Troms, Norway Phone: +47-77644622 Fax: +47-77645350 e-mail: ugom@fagmed.uit.no

Abstract:

Cancer development requires that tumour cells attain several capabilities, including increased replicative potentials, anchorage and growth-factor independency, evasion of apoptosis, angiogenesis and metastasis. Many of these processes involve the actions of protein kinases, which have emerged as key regulators of all aspects of neoplasia. Perturbed protein kinase activity is repeatedly found to be associated with human malignancies, making these proteins attractive targets for anti-cancer therapy. The last decade has witnessed an exponential increase in the development of specific small protein kinase inhibitors. Many of them are in clinical trials in patients with different types of cancer and some are successfully used in clinic. This review describes different approaches that are currently applied to develop such specific protein kinase inhibitors and provides an overview of protein kinase inhibitors that are currently in clinical trials or are administered in the clinic. Focus is directed on inhibitors against receptor tyrosine kinases and protein kinases participating in the signalling cascades.

Keywords: anaplastic lymphoma kinase, c-Abl, clinical trials, EGFR, FLT-3, integrin-linked kinase, c-Kit, MAP kinase, mTOR, nonreceptor tyrosine kinase, PDGFR, PKC, VEGFR

Contents

• 1 Introduction
• 2 Protein kinase inhibitors and therapeutic strategies
  • 2.1 ATP-mimics
  • 2.2 Substrate mimics
  • 2.3 Monoclonal antibodies against protein kinases
  • 2.4 Antisense RNA and RNA interference
  • 2.5 Bisubstrate analogue inhibitor
  • 2.6 Other strategies to inhibit protein kinases
  
  
• 3 Clinical trials and problems facing the use of protein kinase inhibitors
• 4 Status of specific protein kinase inhibitors in clinical trials or in the clinic
  • 4.1 Inhibitors against RTK
  • 4.2 Inhibitors against nonRTK
  • 4.3 Inhibitors against serine/threonine kinases
  
  
• 5 Future perspectives

1 Introduction

Post-translational modifications are important processes in the regulation of protein activity. One of the major reversible protein modifications is phosphorylation as it is estimated that approximately 30% of all the proteins in the cell are transiently phosphorylated. The human genome project has led to the identification of approximately 520 protein kinases, underscoring the importance of phosphorylation. The phosphorylation pattern of proteins within a cell is determined by the antagonistic action of protein kinases that phosphorylate proteins and protein phosphatases that dephosphorylate proteins [1]. Protein kinases can covalently link a phosphate group from the ATP donor molecule to serine, threonine, and/or tyrosine residues on their substrates. This phosphotransfering reaction requires the presence of three specific sites within the protein kinase:

• an ATP-binding site,

• a domain catalyzing the transfer of a phosphate group of ATP and

• a substrate-binding site in the protein kinase.

Of these protein kinases, approximately 90 are tyrosine kinases, 43 are tyrosine kinase-like, while the remaining majority comprises serine/threonine kinases. Tyrosine kinases are divided into receptor tyrosine kinases (RTK) and nonreceptor tyrosine kinases (non-RTK). The former are transmembrane proteins that contain a ligand-binding extracellular domain and an intracellular catalytic domain, whereas nonreceptor tyrosine kinases are intracellular proteins that generally function downstream of the RTK.

The intrinsic tyrosine kinase activity of RTK is activated upon binding of the ligand and subsequent oligomerization of the receptor. The intracellular catalytic domain of the activated receptor autophosphorylates tyrosine residues. These phosphotyrosine residues form binding sites for proteins containing src-homology 2 (SH-2) domains. The SH-2 containing proteins further transmit the signal often via non-RTK or serine/threonine protein kinases. The cascade of phosphorylation events results in amplification and intracellular transmission of the signal [2, 3]. Protein kinases possess a wide variety of substrates including structural proteins, metabolic enzymes, protein kinases controlling the cell cycle, and transcription factors. Hence signal transduction pathways play a pivotal role in the regulation of fundamental cellular processes such as metabolism, proliferation, differentiation, survival, migration and angiogenesis. The activity state of these proteins determines the fate of the cell and aberrant expression and activities of these functional classes of enzymes result in abnormal signal transmission.

Perturbed signalling transduction provokes dysregulation of processes involved in angiogenesis, apoptosis, cell migration and cell cycle control and can therefore lead to malignant phenotype. As such, protein kinases have merged as key regulators of all aspects of neoplasia, including proliferation, invasion, angiogenesis and metastasis, hence making cancer fundamentally a disease of aberrant protein kinase activity and signal transmission. In fact, perturbed activity of more than 50% of the RTK and of several serine/threonine protein kinases have been repeatedly found to be associated with human malignancies [Table S2 in 4, 5]. Dysregulation of normal protein kinase activity may occur by :

• genomic rearrangements, including chromosomal translocations that generate hybrid proteins containing the catalytic domain of a certain protein kinase and an unrelated protein. A well-know example is the fusion of the tyrosine kinase c-Abl to the breakpoint cluster region (BCR) protein in chronic myelogenic leukaemia.
• A second important mechanism that disrupts the normal function of a protein kinase is a mutation that renders the kinase constitutive active, as exemplified by mutated c-Kit in gastrointestinal tumours.
• A third mechanism involves increased or aberrant expression of the protein kinase or the ligand for RTK amplification such as the RTK HER-2 (=ErbB2=EGFR2) in breast cancer or aberrant expression of vascular endothelial growth factor (VEGF) and its receptor (VEGFR), which enhances angiogenesis and tumour growth.
• Finally, deregulation of kinase activity by activation of oncogenes or loss of tumour suppressor functions can also contribute to tumorigenesis. For example, mutation in the GTPase Ras leads to deregulation of c-Raf, while loss of the protein phosphatase PTEN activates the serine/threonine kinase AKT [6, 7, 8].

Conventional management strategies in cancer therapy have been relying on surgery, radiation and chemotherapy. However, since the underlying molecular basis of many cancers lies in the dysfunction of protein kinases, these proteins have become attractive targets for novel anti-cancer drug design.

Here we review different therapeutic approaches in the design of specific inhibitors against protein kinases that are involved in cancer. We highlight recently developed drugs that target protein kinases engaged in signal transduction, while inhibitors against cyclin dependent kinases are beyond the scope of this review. Furthermore, we provide an overview of the status of protein kinase inhibitors that have entered the clinic or that are in clinical trials and we also focus on novel promising inhibitors. The protein kinase inhibitors such as Imatinib (STI571/Gleevec), Cetuximab (C225/Erbitux), Erlotinib (OSI774/Tarceva), Gefitinib (ZD1839/Iressa), and Trastuzumab (Herceptin) that have entered the clinic have been extensively and excellently reviewed elsewhere and will only be briefly discussed here.

2 Protein kinase inhibitors and therapeutic strategies

In order to carry out its function, a protein kinase has to bind ATP and its substrate before the catalytic activity can transfer the phosphate group. Preventing ATP binding, substrate binding or inhibiting the kinase activity or a combination may therefore form the basis for drug design, as this will block the function of the protein kinase. Interdisciplinary research has tremendously increased our knowledge on the structure, function and regulation of protein kinases. Although the 3D structure of less than 10% of the human protein kinases is known so far, X-ray structure information has facilitated the prediction and comparison of the binding sites of ATP- and substrate-related molecules to protein kinases [9]. However, the development of protein kinase drugs as anti-cancer therapy was originally hampered by the assumption that these proteins were no suitable targets. The reason for this assumption was that the high intracellular ATP concentrations would compete with the ATP-binding site inhibitors so that no ATP site-directed inhibition was obtained. Moreover, the common catalytic mechanism of the kinase domain made it unlikely to develop specific inhibitors, because blocking the activity of these kinases prevented additional physiological processes unrelated to proliferation, which resulted in severe side effects [7].

2.1 ATP-mimics

In order to develop chemical compounds that exert an inhibitory effect on protein kinases, chemical libraries that consist of hundreds of thousands of synthetic compounds are screened. The composition of the library can be based on the actual structure of the ATP-binding pocket of the protein kinase or a family member (structure-based library design) or on the structure of compounds already known to bind to the ATP pocket (ligand-based library design) [10]. In this way ATP-mimics have been identified and modified to increase their potential to compete with ATP binding at the ATP-binding site of the protein kinase, thereby inhibiting the kinase activity of the protein. The degree of conservation of the ATP-binding sites in the distinct protein kinases is not absolute, so it is possible to develop ATP-mimics with relatively high selectivity. Besides high selectivity, protein kinase inhibitors must bind with an extremely high affinity. The affinity of the inhibitor for the kinase should be several orders of magnitude higher than that of ATP, because the inhibitor will be present in M-nM range concentrations, whereas the intracellular concentration of ATP is mM [10,11].

Recently, a new approach that uses ``selectivity filters'' to increase the specificity of the drug has been described in the literature. This method selectively targets 2 determinants in the ATP-binding pocket. One is the so-called ``gatekeeper'', which is the residue that flanks a highly variable hydrophobic pocket at the rear of the ATP-binding site. The other site is a reactive cysteine within the active site. Of the 491 protein kinases, only 19 had cysteine residues within the conserved glycine loop, a region that forms close contacts with bound ATP. By combining a drug that occupied both the hydrophobic pocket and irreversibly alkylated the cysteine residue, a drug that selectivity inhibited the p90 ribosomal S6 kinases RSK1 and RSK2 was designed. Moreover, this dual targeting drug may reduce the risk of drug-resistant mutations in the kinases, as two mutations must occur in order to induce resistance [12, 13].

2.2 Substrate mimics

Another approach to interfere with the function of protein kinases is the use of substraterelated molecules that compete with the genuine substrate to become phosphorylated. The rational behind the use of this class of inhibitors was that phosphorylation of the substrate mimics, but not of the actual substrate will terminate the signalling transduction event that contributes to the neoplastic state of the cells. These inhibitors are likely to be less toxic than ATP-mimics since they bind to regions in the kinase domain that are less conserved than the ATP binding site and therefore less likely to hit many other targets (very stringent specificity). The substrate mimic ThymectacinTM (NB1011) that targets thymidylate synthase, an enzyme overexpressed in tumours, has entered clinical trials. This illustrates that substrate mimicry as a means to inhibit enzymatic activity possesses potential clinical applications [14]. However, substrate competitive inhibitors targeting protein kinases await to enter clinical trials.

2.3 Monoclonal antibodies against protein kinases

Monoclonal antibodies against protein kinases or in the case of RTK against their ligand have been developed and can be used to shut-off receptor-mediated signalling as a means of blocking (tumour) cell growth. The complexity of the immunoglobulin molecule, its murine origin, and the selection of appropriate target structures on the target antigen initially were obstacles for the production and the utilisation. The replacement of most of the mouse sequences with equivalent human sequences has helped to turn monoclonal antibodies into valuable tools in cancer therapy [15,16]. Another class of monoclonal antibodies consists of bispecific antibodies that bind a RTK and immunologic effector cell. Examples are MDX-447 and H22-EGF, which will be discussed below. The major disadvantage of therapeutic monoclonal antibodies is their production costs, leading inevitably to expensive medication. To circumvent this, researchers have developed nanobodies, which both reduces production costs and increases specificity (see section 5).

2.4 Antisense RNA and RNA interference

Since anticancer drugs are dependent on a time and dose schedule, cancer cells might soon develop resistance towards treatment, thereby rendering the small molecule protein kinase inhibitors ineffective. Therefore, other therapeutic strategies have been considered. One of them uses antisense RNA (asRNA) or RNA interference (siRNA/RNAi) to prevent translation of proteins implicated in cancer. The advantage of siRNA molecules is their high specificity, i.e. siRNA can be used to target unique mRNA transcripts that are only present in cancer cells, sometimes only differing from the wild-type transcript by a single point mutation. This may circumvent unwanted side effects in healthy cells because the siRNA will not affect the expression of the healthy gene [17 and references therein]. Since cancer cells are characterized by chromosomal rearrangements, deletions, and/or alternative splicing of premRNA, they can produce unique mRNA [see e.g. Table 1 in 17]. These cancer-specific aberrant transcripts can be used as a target for highly specific hybridization with a long asRNA.

For example:

•  persistently active truncated EGFR (2-7 EGFR) is expressed in advanced glioblastoma multiform. Long dsRNA (>30 base-pairs), which only binds 2-7 EGFR but not wild-type EGFR, will activate the dsRNA-dependent protein kinase PKR, which subsequently induces cell death.

• Lentiviral vector-mediated delivery and expression of a 39 nucleotides asRNA against EGFR in glioblastoma xenograft strongly inhibited the growth of the tumour in the mouse brain [18].

Several problems that may occur with RNA interference can be imagined:

1. The asRNA may function as a classical antisense, i.e. it will inhibit the expression of the target gene, but fail to induce activation of PKR, while in other tumours, the activity of PKR is impaired.
2. Another pitfall could be that secondary structures in long dsRNA prevent its binding to target sequences. One draw back with siRNA is incomplete inhibition of expression of the target.
3. Another disadvantage is that, when the siRNA is directed against a region of the target mRNA that does not include the mutation, the inhibitory effect will also occur in healthy cells, leading to undesirable side effects.
   

siRNA vectors from which expression is induced specifically in cancer cells, e.g. by a promoter that is activated by hypoxia or glucose deprivation or by localized irradiation or drug application can solve this problem. Although it was originally assumed that the substrate specificity of individual siRNAs was very high, recent studies have indicated that siRNA can tolerate single mutations in the centre of the molecule, and up to 4 mutations are necessary for complete inactivation. Unspecificity of the siRNA may cause difficulties in knocking down the mutated transcript in the neoplastic cell. As protein kinases often carry single point mutations, it may be difficult to design siRNA that specifically reduced the expression of the mutated protein kinase in tumour cells and does not affect the wild-type transcripts in normal cells.

Probably, relatively high concentrations and continuous administration of siRNA must be given over extended periods to be beneficial for the patients. At concentration of 100 nM, siRNA non-specifically induced the expression of a significant number of genes, many of them involved in apoptosis and stress response. Reduction to 20 nM eliminated this nonspecific response. In addition, siRNA can bind to cellular proteins and induce non-specific changes in gene expression. Several siRNA sequences have been shown to activate the innate immunity response genes in freshly isolated human monocytes, resulting in the production of TNF$\alpha$ and activation of the NF$\kappa$B and mitogen-activated protein kinase (MAPK) pathways. Delivery forms another challenge in the use of siRNA in clinical treatment. Exogenous delivery by liposome-based transfection or viral-based vectors may give unwanted side effects [17, 19].

2.5 Bisubstrate analogue inhibitor

Bisubstrate inhibitors compete for both ATP and substrate binding simultaneously. The rational for bisubstrate inhibitors is promising, but their use in clinics is hampered by the difficulty of delivery. Transforming these peptides into cell-permeable molecules is absolutely required to make them successful therapeutic drugs [6].

2.6 Other strategies to inhibit protein kinases

Other approaches have been tested in the search for strategies to counteract the kinase activity of protein kinases :

• One method relies on inhibiting the expression of the regulatory subunits of certain protein kinases. The cyclic AMP (cAMP)-dependent protein kinase or protein kinase A (PKA) is a heterotetramer that consists of two regulatory subunits (R) and two catalytic subunits (C). The regulatory subunits contain two binding sites for the secondary messenger cAMP, while the catalytic subunits include the kinase domain. Both the R subunits (RI$\alpha$, RI$\beta$, RII$\alpha$, RII$\beta$) and the C subunits (C$\alpha$, C$\beta$, C$\gamma$) are differentially expressed. The R2C2 holoenzyme will dissociate into the R and C subunits upon binding of cAMP, resulting in activation of PKA [20]. Perturbed PKA activity has been implicated in tumours. For example, PKAI and/or its regulatory subunit RI$\alpha$ are generally overexpressed in human cancer cell lines and in primary tumours and inactivating mutations of RI$\alpha$ are involved in Carney complex (CNC), a multiple endocrine neoplasia syndrome [21-23]. The RI$\alpha$ subunit forms an attractive target in human cancers with aberrant PKA activity. Phase I studies with the antisense oligonucleotide GEM231 against RI$\alpha$ mRNA alone or combined with docetaxel in patients with solid tumours showed that GEM231 could be safely used, but no objective responses were noted. Stable disease (up to 16 weeks) was observed in few patients. Based on the results, additional clinical trials and phase II trials have been recommended [24-26].

• An alternative strategy relies on different conformations that activated and inactive protein kinase can acquire. Antagonists for protein kinase inhibitors can be selected that exclusively bind to the inactive form of the kinase, so as to sequester the molecule in a state that cannot participate in signal transduction. Certain classes of BCR-ABL inhibitors bind differentially to kinases in active and inactive states [7].

• Another way to restrain protein kinases is by altering their stability. Heat shock protein 90 (Hsp90) functions as a molecular chaperone by binding to various cellular proteins and regulates the folding, stability and function of the target protein. Bona fide Hsp90 target proteins include protein kinases. Thus inhibiting the normal function of Hsp90 may obstruct the proliferation of cancer cells [27]. Encouraging in vitro results with the histone deacetylase inhibitor LBH589 and an analogue of geldanamycin (17-allyl-amino-demethoxy geldanamycin or 17-AAG) support this assumption. Both LBH589 and 17-AAG disrupt the chaperone association of Hsp90 with its targets proteins BCR-ABL (see section 4.2.1) and mutant FLT-3 (see section 4.1.6), resulting in polyubiquitination and proteasomal degradation of both proteins. The combination of LBH589 and 17-AAG provoked more apoptose of imatinib mesylate-resistant primary chronic myeloid leukaemia in blast crisis cells expressing BCR-ABL and acute myeloid leukaemia cells with an activated mutant of FLT-3 than either compound alone. These data should encourage the use of both agents in clinical trials [28].

•  Phosphotyrosines on RTK function as docking sites for other proteins that will transmit the signal received by the RTK. Small molecules that inhibit this binding and prevent downstream cell signalling can also be used as thereapeutic application.

• Finally, research is aimed towards the development of inhibitors that prevent dimerization of RTK or that interfere with appropriate subcellular localization of protein kinases. However, none of these strategies have entered clinical trials so far [6].

3 Clinical trials and problems facing the use of protein kinase inhibitors

Before a drug can enter the clinic, it has to be thoroughly tested. Preclinical tests include biochemical analyses in vitro, studies in cell cultures of cancer cell lines and intact animals such as xenograft studies or animal models for cancers, and finally clinical trials. In vitro analysis, cell culture and most of the xenograft and/or syngeneic systems used for drug screening suffer from a poor predictability with respect to the history of the molecular pathophysiology of human malignancy. In that regard, transgenic and knockout animal models may provide more realistic approaches before clinical trials are initiated [6]. Clinical trials are usually divided into three phases. In phase I clinical trials, the new drug or treatment is tested in a small group of people (20-80) for the first time to evaluate its safety, determine a safe dosage range, and identify side effects. These studies also determine how the compound is absorbed, distributed, metabolized and excreted, as well as the duration of its action.

• Drugs that are used in phase I clinical trials have not been used in humans before and thus adverse reactions specific to humans are possible. That is why phase I trials are conducted stepwise starting at relatively low doses. After each dose step the trial temporarily closes and a toxicity evaluation is performed. The study only continues at a higher dose-level if no serious drugrelated toxicities have occurred.
• In phase II clinical trials, the studied drug or treatment is given to a larger group of people to see if it is effective. Phase II trials further evaluate the safety and effectiveness of an agent or intervention, and evaluate how it affects the human body. Phase II studies usually focus on a particular type of cancer, and include fewer than 100 patients.
• In Phase III trials, the drug or treatment is given to large groups of people to confirm its effectiveness, monitor side effects, compare it to commonly used treatments, and collect information that will allow the drug or treatment to be used safely. These Phase III trials compare a new agent or intervention (or new use of a standard one) with the current standard therapy (a currently accepted and widely used treatment for a certain type of cancer, based on the results of past research). Participants are randomly assigned to the standard group or the new group, usually by a computer. This method, called randomization, helps to avoid bias and ensures that human choices or other factors do not affect the study's results. In most cases, studies move into phase III testing only after they have shown promise in phases I and II. Phase III trials may include hundreds of people across the country.

Although promising results with specific protein kinase inhibitors have been obtained in clinical trials (see section 4), several problems are facing the use of these drugs.

• One obstacle is when to apply the drug during the treatment. Experiences with imatinib mesylate show that the drug is most effective in the pre-fully malignant chronic phase, indicating that this compound is probably most beneficial in the early process of tumorigenesis. However, in early stages, patients may not demonstrate clinical evidence for the disease and the clinical trials are mostly done on patients in terminal stages of the disease. Routine screening of patients may enable the detection of early stages of cancer, but will also meet ethical constrictions as the patients get to know very early that they have cancer.
• Another problem that faces the use of protein kinase inhibitors is the ubiquitous establishment that the particular protein kinase is indeed implicated in tumorigenesis. It is not sufficient simply to know whether it is activated (or inhibited) in a specific disease state, because perturbed activity can be an irrelevant consequence of the disease rather than a major contributing factor to disease pathology. Moreover, tumours rarely have a single genetic alteration driving the neoplastic process. With likely few exceptions, a single protein kinase inhibitor may not be sufficient to shut-down cancer cell activity because several protein kinases may exert perturbed activity in a specific tumour so that combinations of inhibitors may be required for positive effects. [7, 10]. Furthermore, protein kinases do not excusively regulation of cell proliferation or apoptosis, but participate in other processes too. For example, Akt/PKB plays a crucial role in nutrient metabolism so that inhibition of this kinase may have such severe unwanted side effects that the drug is toxic for humans [29].
• An additional problem facing the use of an appropriate inhibitor is defining the suitable patient population in which to test these compounds. For example, chronic myelogenous leukemia (CML) patients with the lymphoid subtype of blast crisis derive very little benefit from imatinib mesylate, while those with the myeloid subtype derive more limited benefit in comparison with chronic phase patients. This is probably because blast phase CML patients have more heterogenous karyotypes compared to chronic phase CML patients.

Thus, it is not merely the expression of a mutated protein kinase in a tumour that determines success for protein kinase-directed therapies, the context in which the mutation occurs is also important [7]. Recently, mutation analyses of all protein kinases were performed in breast cancer and colon cancer tumour samples [30]. This strategy allowed mapping the mutation spectrum in single cancer samples and the design of a protein kinase inhibitor cocktail therapy adapted to individual patients. In this regard, it is important to consider that some protein kinase inhibitors have an additive or synergistic effect when combined with chemotherapy, while others fail to work or have severe side effects. For example, the combination of UCN-01, which prevents AKT activation, and the MEK inhibitor CI-1044 (PD184352) worked synergistically to induce apoptosis in human leukaemia cells. However, VEGFR inhibitors in combination with cytotoxic agents has been associated with thrombosis or haemorrhagic side effects [7]. Other examples of combined clinical trials are given in section 4.

4 Status of specific protein kinase inhibitors in clinical trials or in the clinic

More than 40 specific protein kinase inhibitor drugs are currently in clinic or have entered clinical trials and their number is still expanding. The drugs discussed in this review are summarized in Table 1.

4.1 Inhibitors against RTK

4.1.1 Epidermal growth factor receptor (EGFR)

Figure: Schematic presentation of the epidermal growth factor receptor (EGFR) signalling pathway. The EGFR family consists of the family members ErbB1/EGFR, ErbB2/HER-2, ErbB3/HER-3 and ErbB4/HER-4, which can form homo- and heterodimers. After activation, the intrinsic tyrosine kinase activity will phosphorylate and thereby activate different signalling pathways, including the Ras-Raf-MAP kinase, the PI3K/Akt, the Crk/c-Abl and the PKC pathways. This leads to activation of transcription factors and modulation of cellular processes such as proliferation, cell survival and protein synthesis. Perturbed activation of the EGFR signalling pathway may contribute to tumorigenesis. Several drugs and strategies have been developed to specifically inhibit protein kinases in this pathway. The compounds discussed in this article are boxed. The red box represents antisense oligonucleotides, the blue box contains small inhibitors that interfere with the kinase activity of the enzyme, while the drugs in the green box encompass antibodies against members of the EGFR family.

The family of epidermal growth factor receptors consists of four family members known as

1. EGFR (HER-1, ErbB-1),
2. HER-2 (ErbB-2, neu),
3. HER-3 (ErbB-3) and
4. HER-4 (ErbB-4).

Activation of the receptor requires binding of growth factors, which normally leads to oligomerization and tyrosine-autophosphorylation by the intracellular kinase domain of the receptor (Figure 1). At least 12 different ligands that bind to HER-1, HER-3, and HER-4 receptors have been identified, whereas there is no known ligand for HER-2 so far. Instead, this protein functions as a co-receptor through heterodimerization with the other members in the EGFR family [31]. All members of the EGFR family are implicated in the development and progression of numerous human tumours, including breast, cervical, lung, colorectal, ovarian, glioma, non small cell lung cancer (NSCLC), prostate, oesophageal, bladder, endometrial, and head and neck cancer [32,33]. RTK are frequently overactive in cancer cells, even in the absence of ligand. The presence of such an overactive EGFR, due to overexpression or truncated forms, is often correlated with advanced disease or poor prognosis. The most common mutant form of EGFR is EGFRvIII, which lacks important parts of the extracellular domain and is unable to bind a ligand, but displays constitutive kinase activity [8, 34, 35]. Anti-EGFR therapy includes small ATP-mimicking molecules and monoclonal antibodies against the receptor and several of them have reached clinical trials or the clinics. However, results from early phase clinical trials suggest that monotherapy targeting EGFR alone may not sufficient to effectively fight established tumours. Combining EGFR inhibitors with chemotherapy or incorporating EGFR inhibition into cancer prevention have been proven much more efficient [35].

4.1.1.1 Erlotinib, TarcevaTM

Erlotinib (TarcevaTM) is a small molecule ATP-competitive inhibitor of HER1/EGFR. Its inhibiting effect has been studied in a whole range of cancers, including NSCLC, pancreatic cancer, ovarian cancer, cancer of head and neck. The Food and Drug Administration (FDA) recently approved Erlotinib for treatment of patients with locally advanced or metastatic NSCLC after failure of at least one chemotherapy regime [3].

4.1.1.2 Gefitinib, Iressa

Gefitinib (Iressa or ZD1839) is also an ATP-competitive inhibitor of EGFR that has been studied in NSCLC, colon cancer, breast cancer, cancer of head and neck, and skin cancer. Recently, a phase III trial revealed that compared to placebo, IressaTM showed no improved survival in NSCLC patients and the drug was retained for new patients [36].

4.1.1.3 CI-1033

CI-1033, an abbreviation for Canertinib dihydrochloride belongs to the anilinoquinazoline class compound. This inhibitor, also known as PD183805, is a water-soluble analogue of PD169414. It not only irreversibly inhibits pan-erbB kinase, but its activity extends to blocking all four members of the erbB family. ErbB1, ErbB2 and ErbB4 are directly blocked, whereas ErbB3 is indirectly affected because of the lack of catalytically active heterodimeric partners. Additionally, the highly tumorigenic constitutively activated variants of the ErbB1, like EGFRvIII, can be blocked as well. This inhibition prevents further activation of the downstream signalling pathways through PI3kinase/Akt and Ras/MAP kinase (see section 4.3.1 and4.3.3, respectively). The activity of CI-1033 lies in the formation of a covalent bond between the acrylamide side chain at position six and the cysteine 773 (784 and 778 in erbB2 and erbB4 respectively), thereby permanently preventing autophosphorylation and thus activity of the tyrosine kinase. Since the interaction with the kinase is irreversible, the effects last longer, making the receptor more prone to ubiquitin tagging and subsequent degradation. In this way, new receptor synthesis is required rather than pharmacological clearance of the drug. Other pharmacological properties of the highly protein bound CI-1033 include rapid absorption, billiary excrection and good tissue distribution, except for the brain. However, toxicity seemed to be a problem due to the highly reactive acryloyl group [11]. By reducing this reactivity and increasing its affinity for the receptor, annoying side effects could be decreased [11, 37, 38].

In preclinical xenograft models of breast, colon, lung, pancreatic, ovarian cancers and cancers of glial origin, CI-1033 delayed tumour growth or induces tumour stasis. Especially in human vulvae carcinoma xenografts, near maximum suppression of ErbB1 phosphorylation (even at the lowest dose level) was acquired. The disadvantages of this treatment are related to toxicities that concern primarily the gastrointestinal tract in rats and monkeys and dermatological toxicities in rats only. Subsequent phase I studies have been carried out in advanced stages of colon cancer, NSCLC, head and neck and breast cancers. Here, mild to moderate adverse effects were noted (thrombocytopenia and rare hypersensitivity reactions). The overall result of the heavy CI-1033 treatment schedule indicated a 40-50% decrease in the phosphorylated receptor after seven days of treatment. Therefore, a less intensive treatment schedule was evaluated in another phase I study with patients suffering from advanced solid malignancies. The patients were submitted to treatment for seven days, and were then untreated for the subsequent seven days. This schedule seemed to be better tolerated by the patients, yet with the same type of toxicities. However, a disappointing 6 out of 31 cases responded to the treatment with stable disease, and no better results were obtained at higher doses. Therefore, a third phase I was conducted to evaluate the responses of patients treated for 14 days combined with one week without treatment. Although the patients in this study could tolerate higher doses, the side effects remained. Some combination studies have also been conducted and these learned that the combination of CI-1033 with topotecan, gemcitabine and SN38 (active metabolite of irinotecan), and cisplatin lead to synergistic activities against the ErbB receptors. CI-1033 will be further evaluated in subsequent phase II/III trials and as in many anti-cancer drugs combinatorial studies with other agents are on the program as well [37, 39, 40].

4.1.1.4 TheraCIM h-R3

TheraCIM h-R3 is a humanized monoclonal antibody with specificity against EGFR. Phase I studies have shown that it is well tolerated and that it may enhance radiotherapy. Phase I/II trials in 24 patients with advanced carcinomas of the head and neck showed that 14 of 16 evaluable patients had an objective response and 7 had complete response [41]. Phase I/II trials in pancreatic cancer, breast cancer and paediatric gliomas have been initiated or are planned. Derivatives of h-R3 labeled with radioisotopes are also being investigated [42]. A phase II trial including 17 children with gliomas demonstrated that of six patients that were fully evaluable all demonstrated either stable disease or partial response [43]. In another study with 24 evaluable glioma tumour patients 4 had complete response, while 5 had partial response. In total, 20 patients had stable disease [44]. HKI-272/Compound25o HKI-272 (compound 25o) is an EKB-569 derivative with improved efficacy against HER-2, but it also inhibits EGFR and blocks HER4. It binds within the ATP binding site of these protein kinases [45, 46]. HKI-272 is currently being tested in phase I and II clinical trials with breast cancer and NSCLC patients [38, 47-49].

4.1.1.5 MDX-447

MDX-477 represents an example of a bispecific antibody. This monoclonal antibody was constructed by cross-linking the F/ab' fragment of monoclonal antibody H22 (directed against CD64) and monoclonal antibody H425 (=anti-EGFR). The CD64 receptor (or Fc$\gamma$RI), which is present on key cytotoxic effector cells, has been implicated in the destruction of tumour cells, while overexpression of EGFR correlates with a worse clinical outcome in several cancers, including NSCLC and cancer of the prostate, breast, stomach, colon, ovary, and head and neck [50,51].

Phase I studies have been or are being performed with patients with solid tumours and adult glioblastoma [52,53]. Of the 62 evaluable patients with solid tumours, 21 experienced a stable disease for 12 weeks, while one patient with mucoepidermoid parotid tumour had tumour reduction [52]. A phaseI/II trial with renal cell carcinoma, head and neck, bladder, ovarian, prostate, and skin cancer showed that MDX-447 was well tolerated and that of the 36 evaluable patients, 9 had a stable disease of 3 to 6 months [50]. Phase II studies with squamous cell carcinoma of the head and neck are ongoing [54]. Panitumumab, ABX-EGF This monoclonal antibody blocks activation of EGFR and prevents tumour growth and perhaps shrinks tumours. It has reached phase II clinical trial in treatment of advanced NSCLC, and metastatic colorectal cancer with relapsed or progressive disease following treatment with fluoropyrimidine, irinotecan and oxaliplatin chemotherapy [53, 55].

4.1.1.6 Cetuximab

Cetuximab, also known as Erbitux or IMC-C225, is a chimeric EGFR inhibitor that was developed from a panel of antibodies by immunizing mice with human A431 epidermoid carcinoma cells that express high levels of EGFR. Subsequently, the antibodies that bound specifically to the extracellular part of the EGFR were selected and screened for those that inhibit EGF binding and receptor phosphorylation. One of those antibodies was MAb225, for which a human:murine chimeric version was produced. The antibody binds via its heavy and light chains CDR region (rich in tyrosine and tryptophane) to a site exclusively on domain III that covers an epitope that partially overlaps the ligand-binding site. Thereby, it not only prevents the substrate from binding to the receptor but it also prevents the EGFR extracellular region from adopting the dimerization-competent extended configuration [56]. The inhibition of the EGFR activity leads to a decreased cell proliferation, angiogenesis, metastasis and increased apopotosis. Furthermore an indirect anti-angiogenic effect of Cetuximab has been recorded due to dose-dependent downregulation of the expression of important angiogenic factors (VEGF, IL8, bFGF).

 Cetuximab has gone through phase III trials in stage III colon cancer patients [53]. Since the outcome was promising, it has recently been appointed as a second and third-line treatment for the metastatic type of colon cancer of patients that showed disease progression while they were under Irinotecan treatment. Currently clinical trials are being set up to evaluate the combination of Cetuximab with other anticancer agents [3,40,57].

4.1.1.7 PKI166

The pyrolo-pyrimidine compound PKI166 (CGP-75166) potently inhibits the EGFR and HER-2 tyrosine kinase activities, but possesses good selectivity against the tyrosine kinases Met, Kit and TCK as well [34].

Phase I studies in patients with solid tumours, including colorectal, renal, thyroid, head and neck and NSCLC demonstrated a stable disease in some of the patients and a partial remission in one of the NSCLC patient [58, 59]. Hence, PKI166 reached phase II trial, but Novartis announced in October 2002 that the development of PKI166 was discontinued [34].

4.1.1.8 AEE788

AEE788 is a potent EGFR, HER2 and VEGFR inhibitor that is currently, alone or combined with RAD001, in phase I clinical trials with patients with glioblastoma or solid tumours [53]. A phase II trial in NSCLC patients has been initiated in 2005 [60].

4.1.1.9 EMD72000

EMD-72000 (or Matuzumab) is a monoclonal antibody directed against the EGFR. It is the human descendant of the murine precursor antibody EMD55900. It consists of the human IgG1 heavy and light chains with some remaining murine amino acids within the CDRs. The antibody binds with high affinity and specificity (Kd=3,4x10M) and competes with the natural ligand by blocking the receptor's binding site, which abrogates receptor-mediated downstream signalling.

The antitumour activities of both compounds, EMD55900 and EMD72000 have been illustrated in xenographic rodent models. However, the former elicited the HAMA (Human Anti-Mouse Antibody) response and was therefore developed into a human variant. Matuzumab displays significant anti-tumour activity in several solid tumours. Several phase I studies were carried out in advanced EGFR-positive solid malignancies unresponsive to therapy, and although there was no clear dose response relationship, sufficient inhibition of the EGFR effector network was achieved at doses way below the Maximum Tolerated Dose (MTD). The overall response rate was 23%, and the disease control rate 50%, demonstrating that single agent activity can occur in these patients. A similar study in combination with Paclitaxel in heavily pre-treated NSCLC patients indicated that the treatment was well tolerated, with moderate toxicity of Matuzumab, but with a partial clinical response (one complete response, several partial and stable disease cases). A phase II clinical trial was carried out in heavily pre-treated subjects (in primary peritoneal and platinum resistant carcinoma cases), now showing the absence of significant clinical activity. This could be due to the far advanced stages of the cancers. Therefore, currently other phase II studies are ongoing in e.g. recurrent ovarian cancer [61-63].

4.1.1.10 Trastuzumab, Herceptin

One of the success stories of protein kinase inhibitors is Trastuzumab (=Herceptin) and its approval in 1998 was considered a milestone in EGFR-directed therapy. The history and clinical applications of Trastuzumab have been extensively discussed in excellent reviews and will therefore be briefly summarized here [64, 65]. Trastuzumab is a monoclonal antibody against HER2/ErbB2. This member of the EGFR family is overexpressed in approximately 30% of invasive human breast cancers. Trastuzumab inhibits HER2 activity by prevention of ligand-induced ErbB receptor activation. Patients with HER2-overexpressing tumours are measured using a scale from 0 (negative) to 3+ (strongly positive), and the stronger the overexpression, the more likely the patient is to benefit from the drug. The FDA approved the use of Trastuzumab to treat metastatic breast cancer. Trastuzumab is now also being studied in clinical trials phase II for osteosarcoma and endometrium cancer [53].

4.1.1.11 TAK165

Another HER-2/ErbB-2 inhibitor is TAK165, also called Mubritinib. Phase I trials with TAK165 are in progress in breast cancer patients [66].

4.1.1.12 Lapatinib (GW2016)

 Lapatinib is a dual inhibitor of the receptors ErbB1 and ErbB2, which have been implicated in various tumours such as breast cancer and lung cancer. A phase II trial is currently testing Lapatinib as a first-line therapy for breast cancer patients with tumours expressing large amounts of ErbB2 [67].

4.1.1.12 Pertuzumab

Pertuzumab, also known as Omnitarg, is a recombinant humanized monoclonal antibody that also binds to the dimerization domain of the HER2 receptor. This occlusion prevents the contact of a hairpin from the dimerization partner and consequently blocks downstream propagation of the signal.

A phase I clinical study in patients with advanced malignancies revealed its clinically activity and indicated it was well tolerated on a 3-week dosing schedule. Currently, Pertuzumab undergoes evaluation in several phase II studies for the treatment of prostate, ovarian, NSCLC and metastatic breast cancer. It has been suggested that this antibody may also be effective against androgen independent prostate cancer [68-70].

4.1.2 EKB-569

EKB-569 or EGFR kinase inhibitor 86 irreversible inhibits ErB1 and ErbB2 by forming a covalent bond with Cys773 of the ATP-pocket. Since EKB-569 is very specific and water soluble, it has a good bioavailability and specific reactivity towards its target, and therefore exerts potent anti-tumour effects and causes few side effects. Apart from inhibiting the EGFR kinase, it also seems to contain, although less efficiently, activity towards HER2 in BT474 cells.

Clinical trials with this component are steadily growing. Reports of its use in solid tumours indicated that EKB-569 is well tolerated and has an acceptable pharmacokinetic safety profile. Toxicities associated with EKB-569 treatment were of gastrointestinal and sometimes of dermatological origin. A phase I-II dose escalation study of EKB-569 in combination with chemotherapy FOLFOX4 and FOLFIRI pointed to some additional toxicities, including thrombocytopenia, and in cases of high doses haematological toxicities and neuropathy. But the overall responses were good as in the majority of the cases either complete and partial responses or stable disease was noted, although a minority showed signs of progressive disease as well. Currently, phase II studies in advanced colorectal cancers and combination studies of CCI-779 and Celecoxib in combination with EKB-569 are being set up to evaluate the potency of EKB569 at a larger scale [38, 46, 53, 71-73].

4.1.3 Vascular endothelial growth factor receptor (VEGFR)

Figure: Signalling through class III receptor tyrosine kinases (RTKs). These RTKs are characterized by an insertion between the two conserved elements of the tyrosine kinase domain and include the vascular endothelial growth factor receptor family (VEGFR), c-KIT and FLT-3. The VEGFR family consists VEGFR1 (= FLT-1), VEGFR2 (=KDR=FLK-1) and VEGFR3 (=FLT-4). The class III RTK pathways modulate processes such as proliferation, cell migration and survival. The compounds discussed in this article are boxed. The red box represents antisense oligonucleotides, the blue box contains small inhibitors that interfere with the kinase activity of the enzyme, while the drugs in the green box encompass antibodies against members of the class III RTK family.

Angiogenesis is the process of new capillary blood vessels formation. Solid tumours, regardless of their type and origin, cannot grow beyond a certain size and therefore depend on the establishment of new blood vessels to ensure the survival and growth of the tumours. These vessels extend from existing blood vessels and they facilitate the delivery of nutrients to the tumour, as well as the removal of waste products. The vascular endothelial growth factor family (VEGF), mitogens specific for vascular endothelial cells, play a pivotal role in the angiogenic process. VEGF are secreted by tumour cells and macrophages and bind the VEGF receptors (VEGFR), a family of RTK [8].

There are three members of VEGFR (see 2):

1. VEGFR1 (also known as fms-like tyrosine kinase, FLT-1),
2. VEGFR2 (kinase-insert domain receptor KDR or FLK-1) and
3. VEGFR3 (also known as fms-like tyrosine kinase 4, FLT-4).
4. A naturally occurring splice variant of VEGFR1, sVEGFR1 (sflt1) exists, while two isoforms that differ in their C-termini have been described in humans. The soluble sVEGFR1 variant lacks the intracellular tyrosine kinase domain.
   

VEGFR1 and VEGFR2 mediate angiogenesis, while VEGFR2 and VEGFR3 are involved in lymphangiogenesis, i.e. the growth of new lymphatic vessels. Lymphangiogenesis often accompanies angiogenesis and may contribute to tumour metastasis. Expression of the VEGFR is cell-specific. VEGFR1 and VEGFR2 are located on e.g. activated vascular endothelial cells, dendritic cells, osteoblasts and some tumour cells, while VEGFR1 can also be present on haematopoietic stem cells. VEGFR2 is in addition expressed on circulating endothelial cells, while VEGFR3 is exclusively found on lymphatic endothelial cells in adults [74-77]. The VEGF ligands are excessively expressed in virtually all types of cancer and not only stimulate angiogenesis, but they can inhibit the antitumour immune response as well, thereby decreasing the host's ability to eradicate tumour cells. This makes anti-angiogenesis an attractive strategy to treat cancer.

Anti-angiogenesis has a number of potential advantages over conventional cancer treatment. The risk of resistance is lower presumably because tumour-associated endothelial cells are more genetically stable than cancer cells. Moreover, inhibition of angiogenesis may prove to be more specific and less toxic because the vasculature is normally quiescent in adults. Several therapeutic strategies targeting angiogenesis in cancer are being developed. They include drugs that inhibit the VEGF ligand or the tyrosine kinase activity of their receptors, but also antibodies that inhibit the action of VEGFR, ribozymes that degrade VEGFR mRNA, and soluble decoy receptors that trap and inactivate VEGF are under development or have entered clinical trials or the clinic. Problems that may arise with anti-angiogenesis therapy are unexpected haemorrhage and that this strategy is only useful in combinational therapy, but is insufficient alone [75, 77]. Some of the anti-angiogenic drugs are discussed below.

4.1.3.1 Avastin

Avastin or Bevacizumab is a monoclonal antibody against the VEGF. Treatment with Avastin blocks the VEGF and prevents angiogenesis [78]. Phase I clinical trials in refractory solid tumours (renal cell carcinoma, breast cancer, sarcoma and lung cancer) showed safe Avastin administration without dose-limiting toxicities even in combination with chemotherapy. However, patients run an increased risk of encountering thromboembolic events. Therefore, Avastin should be administered in combination with anticoagulatory agents [79, 80]. Phase II studies with Avastin were carried out in several solid tumours as well (androgen-independent prostate cancer, in peritoneal and epithelial ovary carcinoma, in non-metastatic hepatocellular cancer, relapsed metastatic breast cancer and IL-2 refractory renal cell carcinoma and studies in combination with first line treatment in NSCLC, metastatic colorectal cancer). The most encouraging results were obtained in renal cell cancer, NSCLC and colorectal cancers. Consequently, the effect of Avastin treatment in colorectal cancers and NSCLC (with paclitaxel and carboplatin) was further evaluated in a phase III trial. Since treatment of the former cancers lead to increased survival, the FDA approved Avastin as first line treatment for colorectal cancers. This has also been done for metastatic breast cancer and advanced renal cell carcinomas in combination with respectively Paclitaxel and interferon [81]. However, in breast cancer patients with poor prognosis, the combined activity of Avastin with capecitabine resulted in a less effective activity, indicating that Avastin should not be used in more advanced stages of breast cancers (e.g. third line treatment) [80]. Currently several combined clinical trials, including paclitaxel (phase I and II), Sorafenib (phase I), Rituximab (phase II), Tarceva (phase II), Erlotinib (phase II trials) are ongoing. Additionally, some trials investigate the use of Avastin as an adjuvant in combination with Irinotecan (also called Camptosar, a topoisomerase inhibitor often used in colon and rectal cancers) and capecitabine [53]. As indicated by this wide range of studies, the applicability of Avastin in several cancers seems to expand rapidly. However, for these different cancers, other optimal doses are of concern. This needs to be carefully investigated [81].

4.1.3.2 IMC-1C11

IMC-1C11 is a chimeric monoclonal antibody that prevents VEGFR2 activation by specific binding to the extracellular domain of this RTK. The conclusions from a phase I study in 14 patients with colorectal carcinoma and hepatic metastasis are that IMC-1C11 is both safe and well tolerated and one patient showed prolonged stable disease for 28 weeks [82].

4.1.3.3 CP547632

CP547632 (3-(4-bromo-2,6-difluorobenzyloxy)-5-[4-(pyrrolidin-1-yl) butylaminocarbonylamino] isothiazole-4-carboxamide hydrobromide salt) is a potent, selective inhibitor of the tyrosine kinase activity of VEGFR2 [83]. This drug has been used in phase I clinical trials in patients with advanced NSCLC and patients with advanced solid tumours. Positive effects were observed on evaluable patients [84, 85]. This has advanced CP547632 into phase II trials with ovarian cancer patients with minimal disease and NSCLC [86, 87].

4.1.3.4 ZD-6474/ZactimaTM

ZD6474 is an anilinoquinazoline derivative that possesses potent inhibitory activity against VEGFR2, but also EGFR and Ret. The VEGF pathway is a key mediator of tumour angiogenesis, while the EGF pathway plays a pivotal role in cell proliferation. Hence, combined targeting of both pathways may provide additive or even synergistic benefit on inhibiting tumour growth [88,89].

The drug can be orally administered and phase I studies have shown that ZD6474 is well tolerated by patients. One study showed that 4 out of nine NSCLC patients exhibited a positive response [88, 89]. Monotherapy and combined phase II studies have been evaluated in patients with NSCLC and breast cancer patients. A phase II trial in 44 evaluable breast cancer patients with ZD6474 alone reported no objective response, although one patients has stable disease for >24 weeks [90]. In another study with breast cancer patients, a 6-7% response rate was observed [91]. A combined phase II study in NSCLC patients with ZD6474 and docetaxel indicated that combined therapy improved time to cancer progression more effectively than docetaxel alone [89]. ZD6474 plus paclitaxel/carboplatin regimes showed partial responses in 7 out of 18 evaluated patients with NSCLC. Another two patients had stable disease for >12 weeks [92]. Phase II comparative studies between ZD6474 and Gefitinib are in progress [89], while phase II studies with SCLC patients, multiple myeloid, and thyroid cancer patients are planned/ongoing [53]. No objective responses, however, were measured in multiple myeloma patients thus far [93]. Recruitment for phase III trials in NSCLC patients has begun.

4.1.3.5 AZD2171

The indole-ether quinazoline AZD2171 is a highly potent inhibitor of the tyrosine kinase activity of VEGFR1, -2, and 3, but it does not posses activity against EGFR. A phase I trial is studying the side effects and best dose of AZD2171 when given together with chemotherapy in treating patients with advanced NSCLC or colorectal cancer. This compound has also been tested in a phase I clinical trials including patients with advanced solid tumours or liver metastases [53, 88, 94].

4.1.3.6 AG-013736

The inhibitor AG-013736 targets both the VEGFR and PDGFR. This inhibition prevents downstream signalling and receptor activation, which leads to inhibition of proliferation and angiogenesis. Therefore, AG-013736 could be used in metastatic melanomas, renal cell cancer, thyroid carcinoma and NSCLC. A Phase I study in advanced solid tumours indicated that there was a considerable difference between administrations to patients in a variable nutritional state. In the fasting state, two patients who had suggestions of a tumour response, suffered from an episode of haemoptysis. This reminded the researchers to the same type of episodes under treatment of Bevacizumab/Carboplatin/Paclitaxel, which resulted in 4 fatal cases. In the study with AG-013736, however, there was less interpatient variability in the fasted state and within the dose range evaluated and kinetics was linear. The dose limiting toxicities like hypertension, seizures, stomatitis, pancreatitis and abnormal liver function were noted as well. In a phase II study with kidney cancer patients, 46% of the patients showed a partial response [95-97].

4.1.3.7 CEP5214 and CEP7055

CEP5214, alias compound 21, consists of a C3 (isopropylmethoxy) part fused to a pyrrolocarbazole part and functions as a potent low-nanomolar pan-inhibitor of all three human VEGF receptors by preventing autophosphorylation. CEP5214 also seems to contain a limited selectivity against PKC, Tie2, TrkA, CDK1, p38 and IRK.

CEP5214 demonstrated in vitro activity in human umbilical vein endothelial cells, seemingly in the absence of toxicity. Furthermore, it establishes a significant in vivo anti-tumour activity in xenographic tumour models (prostate carcinomas and renal cell carcinomas) where it reduced metastasis by increasing tumour apoptosis and decreasing microvessel density. Therefore, it was advanced into Phase I clinical trials as the water soluble N,N dimethylglycine ester prodrug 40 (CEP-u) for the treatment of a variety of solid tumours. First signs of possible clinical activity in a sarcoma and prostate cancer patient were obtained, but no treatment-related toxicities were established although at higher doses CEP5214 seemed to delay wound angiogenesis. Currently, other phase I studies are ongoing to evaluate increased dosage [98-100].

4.1.3.8 KRN-951

KRN-951 is a urea-based compound that specifically inhibits VEGF RTKs of class III and V. It potently suppresses VEGF-driven mitogenesis and capillary tube formation of the endothelial cells in vitro. However, no suppression of proliferation was discovered even at 1M. On the other hand xenograft models have indicated growth inhibition during KRN-951 treatment and thus currently phase I clinical trials are in progress to explore the biological activity of this compound [101].

4.1.3.9 SU-type inhibitors

Since often ATP-competitive inhibitors favour mutation in a specific residue of the ATP binding site, conferring resistance of the tumour to therapy, researchers sought to develop drugs that simultaneously target multiple protein kinases. The first such compound synthesised was SU5416 or Semaxanib. This highly lipophilic protein-bound small compound inhibited primarily autophosphorylation by blocking the conserved ATP binding site within the kinase domain and thus activation of the VEGFR2 (KDR). Additionally, it had some affinity for the bFGFR, PDGFR and the VEGFR1 (Flt1) as well. The compound's high lipophilicity made it difficult to solubilise and therefore it had to be administered with Cremophor via i.v. injection. In preclinical studies of xenografts (malignant melanoma, glioma, fibrosarcoma, carcinomas of the lung, breast, prostate, colon and skin) tumour associated microvessel formation and proliferation were inhibited. Furthermore, SU5416 did not seem to exert cytotoxic properties, although a significant dose-dependent tumour regression occurred. Phase I clinical studies in head and neck cancer yielded a benefit for 36% of the patients, but increased the risk of thromboembolic events. Other phase I-II clinical trials with SU5416 alone or combined with radiotherapy were initiated. Unfortunately, only minimal signs of clinical activity could be detected. Due to the formulation with cremophor and i.v. administration, patients suffered from skin sensitivities, tromboembolic events as well as gastrointestinal toxicities. Finally, in a Phase III study with metastatic colorectal cancer patients in combination with 5'fluoroacil and leucovorin failed to show survival benefit. Therefore, the development of this inhibitor was terminated [96, 102, 103].

The follow-up drug of SU5416 was SU6668. This indolino type of inhibitor had increased pharmacological properties and an increased potency as was shown by its simultaneous inhibition of three receptor-tyrosine kinases (Flk1, platelet-derived growth factor receptor, basic fibroblast growth factor receptor) involved in neovascularisation. Xenograft models of primary patient tumours showed inhibition of phosphorylation of the targeted receptors and subsequent tumour regression. However, again clinical phase I trials indicated problems with toxicity and no objective responses were observed (although 4 patients acquired stable disease). Despite these toxicities, hints of anti-neoplastic activity included stable disease in patients with melanomas, soft tissue sarcomas, renal cell carcinoma and breast cancer, but these were generally weaker activities than those indicated by the xenograft models [95, 104106].

SU11248 or Sunitib Malate (Sutent) represents the new generation of drugs. This broad spectrum of competitive ATP inhibitor prevents activation of the VEGFR, PDGFR$\alpha$, c-KIT and FLT-3 by targeting signalling through the class III RTK. In cell culture, SU11248 showed a rather weak activity, although in combination with Cytarabine synergistic inhibition of proliferation of primary AML myoblasts with Flt3-ITD mutations was obtained. In contrast to cell cultures, very strong and broad potency was observed in xenograft models where this compound inhibited both tumour growth and angiogenesis. A phase I study in AML patients showed that Sutent was generally well tolerated and partial remission for short durations in time was noticed. Patients with metastatic kidney cancer, breast cancer and gastrointestinal stromal tumours (GIST), that were previously unresponsive to therapy manifested partial responses and stable disease in a series of phase II studies. Phase III studies in renal cell cancer, imatinib resistant GIST, and others are ongoing. In some of those studies, tumour growth delayed considerably, and the studies were unblinded for the control group so they could benefit from the treatment as well [38, 78, 103, 107-110]. SU014813, which inhibits the members of the VEGFR family, PDGFR, KIT, RET and FLT3, is in phase I clinical trials [111].

4.1.3.10 Vatalanib

Vatalanib, also called ZK-222584, ZK-22854 and PTK-787, selectively, potently and reversibly inhibits the VEGFR2 (and the VEGFR1 in a less efficient way). When applied in higher concentrations, other tyrosine kinases like PDGFR-$\beta$, c-kit and c-FMS are inhibited as well. However, Vatalanib contains no affinity towards the EGFR, the bFGFR-1, c-Met, Tie2 or intracellular kinases like c-Src, c-Abl and PKC$\alpha$.

Xenographic models treated with Vatalanib showed a reduced microvessel growth, density and vascular permeability. In multiple myeloma cells and bone marrow cells it inhibits cell growth, survival and drug resistance. The combination of Vatalanib with the histone deacetylase inhibitor NVP-LAQ824 resulted in synergistic activity against VEGFR in prostate, breast and renal cell carcinoma cell lines. Based on these findings phase I studies were set up. These clinical trials indicated that Vatalanib was well tolerated and caused mild but frequent side effects. Minor regressions in renal cell cancer and stable disease in colorectal cancer, prostate and renal cell cancer and liver metastasis was observed. In haematological malignancies, the drug can be tolerated at higher doses, however in that case patients suffer from more side effects. Subsequently phase III studies were carried out in patients with solid tumours. In general, few side effects (hypertension) are noted, which is in contrast to other broad range VEGFR inhibitors that produce rashes and other allergic reactions. Ongoing combination studies indicate that Vatalanib can be safely administered with 5'FU and leucovorin/oxiplatin [38, 103, 112-115].

4.1.3.11 Angiozyme (RPI4610)

Angiozyme or RPI4610 is a specific ribozyme that binds to and cleaves the mRNA encoding VEGFR1. A phase II trial in 83 patients with metastatic colorectal cancer with angiozyme combined with chemotherapy (Irinotecan/5-fluorouracil/leucoviron) reported objective responses in some of the patients [116]. A phase Ib combined study of angiozyme with carboplatin and paclitaxel in 12 patients with solid tumours revealed a complete response for more than 7 months in one bladder cancer patient and a partial response lasting longer than 3 months in an oesophageal cancer patient. Stable disease lasting from 2 up to more than 12 months was observed in 3 other patients [117]. In another phase I study with 28 evaluable patients with solid tumours, two patients (nasopharyngeal carcinoma and melanoma) showed minor responses, while 7 patients had stable disease for at least 6 months, with the longest 16 months [117]. Phase II studies with breast, lung, colorectal, melanoma, and renal cancer patients have begun or are planned.

4.1.4 Platelet-derived growth factor receptors (PDGFR)

Figure: Signalling through the platelet-derived growth factor receptors (PDGFR) controls cellular processes such as proliferation, cell survival and migration. The PDGFR family consists of two members, PDGFR$\alpha$ and PDGFR$\beta$ that bind particular homo- and heterodimers combinations of the four ligands PDGF-A, PDGF-B, PDGF-C and PDGF-D described so far. PDGFR signalling can be mediated by the PI3K/Akt, the PKC and the Ras/Raf/MAPK pathways. The PDFGR inhibitors discussed in this article are shown. Activation and overexpression of the nonRTK Src is strongly associated with cancer progression. The Scr inhibitor AZD0530 has recently entered phase I clinical trials. The colour code of the inhibitors is as described in the legend of 1.

Platelet-derived growth factor receptors (PDGFR) form another family of RTK. They are socalled class III RTK and are characterized by an insertion between the two conserved elements of the tyrosine kinase domain. This family consists of two members, PDGFR$\alpha$ and PDGFR$\beta$.
Four ligands have been described:

1. PDGF-A,
2. PDGF-B,
3. PDGF-C and
4. PDGF-D.

PDGF-A and PDGF-B can form homo- and heterodimers, while little is known about PDGFC and PDGF-D. PDGFR$\alpha$ binds PDGF-AA, -BB,-AB, and CC ligand dimers with similar affinity, while PDGFR$\beta$ possesses the highest affinity for PDGF-BB and PDGF-DD homodimers, but does not seem to bind PDGF-AA or PDGF-AB dimers. The phosphorylated intracellular domain of PDGFR can bind more than 10 different SH2-domain containing proteins, including the signal transduction molecules tyrosine kinase Src, PI3-K, PLC$\gamma$, SHP2 and GAP, and the transcription factor STAT. The diversity of the PDGFR-interacting proteins indicates that activated PDGFR can stimulate a wide variety of signalling pathways and participate in several cellular processes (see figure 3). Indeed, PDGFR and their ligands play critical roles in mesenchymal cell migration and proliferation. Abnormalities of PDGF/PDGFR are thought to contribute to a number of human diseases, especially malignancy. Abnormalities of PDGFR activity or expression is implicated in myeloid leukaemia and many solid tumours, including breast cancer, prostate cancer, ovarian cancer, lung cancer, melanoma, meningioma, osteoblastoma, glioblastoma, medullablastoma and astrocytoma.

 For example:  PDGFR$\beta$ is constitutively activated by enforced dimerization mediated by fusion with the TEL transcription factor (TEL-PDGFR$\beta$) in chronic myelogenous monocytic leukaemia patients. Approximately half of the KIT (see section 4.1.4) mutation negative cases in GIST patients have activating mutations of PDGFR$\alpha$ [reviewed in 119]. Several RTK inhibitors not only target PDGFR, but also other members of the class III RTK. They will be discussed below.

4.1.4.1 CEP-751 and CEP-2563

CEP751 or KT-6587 is the biologically active compound of the prodrug CEP-2563, a soluble lysinyl-$\gamma$-alanyl ester of CEP-751. It belongs to the same class of molecules as CEP-701 and can be converted to the latter by O-desmethylation. CEP751 inhibits tyrosine kinase receptors like the PDGFR, by competing with ATP for binding to the kinase domain of the receptor, thereby preventing autophosphorylation and reducing downstream signalling. Apparently, CEP-751 can also inhibit Trka, Trkb and Trkc to some extent.

When administered in preclinical xenograft models both CEP-2563 and CEP-751 possess inhibitory activity against a variety of tumours (e.g. medullary thyroid carcinoma, which harbours a mutation activating the RET allele). Even extended exposures to the inhibitor did not elicit neurological damage or side effects. However, application of doses above the maximum tolerated dose resulted in cardiovascular dysfunction, gastrointestinal effects, hypersensitivity and neurological reactions in rats and dogs. CEP-2563 has been evaluated in a phase I clinical trial in patients with advanced solid tumours refractory to standard therapy. The study indicated that CEP2563 was reliably converted to CEP-751 and that even with rapid dose escalation studies in single patient cohorts, toxicity levels remained acceptable. Generally CEP-2563 and CEP-751 seem to be safe and establish the same efficiency profiles with most common dose limiting toxicities being hypotension and urticaria in grade III. Since the majority of the responses occurred within 80-120% of the recommended phase II dose, no further dose escalations were explored for reasons of patient's safety [120, 121].

4.1.4.2 MLN-518

MLN 518, formerly known as CT-53518 is a small molecule inhibitor developed to inhibit type III RTKs (Flt3, PDGFR and c-Kit). These RTKs are often mutated in cases of GIST, AML (acute myelogenous leukaemia) and SM (systemic mastocytosis) and contain mutations in the KIT juxtamembrane region and the kinase domain. Imatinib can only inhibit the juxtaform but not the kinase mutations. Therefore, MLN-518 and PD-180970 were developed, the latter compound also contains activity against Src, Abl and Kit and is used in many imantinib resistant active site mutations. However, it is not active against malignancies expressing KIT active site mutations. MLN-518 inhibits both the juxtaform variant and the kinase mutant forms. The results of this inhibition cause inactivation of c-KIT and STAT3, which translates in the repression of cell proliferation and the induction of apoptosis.

Phase I clinical trials have proven that MLN-518 effectively inhibits activation loop mutants of c-KIT and that different c-KIT mutants showed different sensitivities towards treatment with this compound. Furthermore, MLN-518 elicited low toxicity in ALM cases, which indicates a possible new treatment replacing Imatinib in these cancers. Currently, MLN-518 undergoes preclinical evaluation in gliomas, in which PDGF-PDGFR are believed to play a role as well. It is also being evaluated in phaseI/II trial for refractory AML [38, 122].

4.1.4.3 AMG706

AMG706 selectively targets the intrinsic tyrosine kinase activity of all known VEGF, PDGF, Kit and Ret receptors. Results of a Phase I trial showed that this drug was well tolerated and at least half of the subjects with advanced metastatic cancer experienced disease stabilization while on AMG706 therapy. Partial/minor responses were measured in 6 of the 34 evaluable patients [123]. Phase II trial to determine the safety and effectiveness of AMG 706 in patients with advanced GIST is ongoing.

4.1.5 c-Kit

The viral oncogene v-kit was originally isolated from the Hardy-Zuckerman 4 feline sarcoma virus. This RTK is structurally similar to PDGFR and Flt-3 and therefore classified as a class III RTK (see figure 2). Gain-of-function mutations of KIT result in development of tumours from mast cells, germ cells and GIST. It took only three years from the first successful administration of Gleevec to a GIST patient and the identification of the mutated c-kit gene as a cause of GIST [124]. Another inhibitor of the tyrosine kinase activity of KIT is E-7080, a compound similar to KRN-951 (see section 4.1.3). This molecule inhibits specifically c-kit in SCLC lines. Xenograft models also seem to confirm its biological activity towards cKit [101]. The compound PKC412 (see section 4.1.6) was also used in a patient with mast cell leukaemia with the imatinib-resistent D816V mutation in receptor tyrosine kinase KIT. The patient showed a partial response, suggesting that resistance of the D816V KIT mutant to imatinib can be circumvented by another inhibitor (e.g. PKC412), which targets the same protein kinase [125].

4.1.6 FLT-3 2

The FMS-like tyrosine kinase-3 or FLT-3 (the v-fms oncogene was originally isolated from the feline sarcoma virus SM-FeSV) belongs to the class III RTK and plays a role in normal haematopoiesis. Moreover, FLT-3 is involved in regulating the cell proliferation and survival because FLT-3 transduces signals through the phosphotidylinositol-3 kinase/AKT, the Ras/MEK/ERK and the STAT-5 pathways. Furthermore, FLT-3 participates in the induction of the expression of the anti-apoptotic bcl-2 gene and inhibition of the pro-apoptotic bax gene. Mutations in FLT-3 resulting in constitutive active tyrosine kinase activity have been detected in acute myelogenous leukaemia (AML) and acute lymphoblastic leukaemia (ALL). In fact, flt-3 is the most commonly mutated gene in AML, where an internal tandem duplication in the juxtamembrane region of FLT-3 is most frequent. This internal tandem duplication can vary in length, but it causes ligand-independent dimerization and constitutive activation of the receptor. This insertion occurs in 30-40% of all AMLs and in 1-3% of ALL cases, and is associated with a more aggressive form of the disease. In 5-10% of AML patients, the gainof-function Asp-835 to Tyr substitution in the catalytic domain of FLT-3 is detected. Other substitutions have been identified as well. Mutations in FLT-3 have rarely been observed in adult patients with ALL, but seem to be more common in paediatric ALL patients [8,126,128].

4.1.6.1 CEP-701

Because of the close structural homology between FLT-3 and other class III RTK (KIT, FMS and PDGFR), several inhibitors also target these tyrosine kinases. The indolocarbazole alkaloid CEP-701 or KT-5555, which resembles CEP-751 (see section 4.1.4), is a derivative of K252a and a fermentation product of Nonomurea longicatena. This small molecule inhibitor of the flt3 gene product also contains selectivity towards the VEGFR2 and to a lesser extent to the PDGFR-$\beta$. CEP-701 competes with ATP in the kinase domain of its target receptor, and thereby inhibits autophosphorylation and downstream signalling.

In 5 out of 6 preclinical xenograft mouse models of the human pancreatic ductal adenocarcinoma tested, CEP-701 inhibited both tumour growth and invasiveness. Other reports (e.g. androgen (in)dependent prostate carcinomas, FLT3/ITD leukaemia) have indicated the same potency for CEP-701 as a single agent or in combination with other anti-cancer compounds. Consequently, clinical phase I studies have been carried out in patients with advanced malignancies. These patients tolerated CEP-701 well and showed no particular toxicities. Although a recommended dose level for phase II studies was given, CEP-701 elicited no convincing clinical activity. Another study carried out in patients with AML-expressing FLT3 activating mutations indicated clinical activity in older and heavily pre-treated patients. In this study, some cases of drug resistance were found, suggesting that CEP-701 should maybe be combined with other kinds of treatment to eliminate this problem. In Phase II studies carried out thus far, the same type of cancer has been used, and some clinical activity was detected in a few patients. Furthermore, due to its low water solubility, a lot of interindividual pharmacokinetic variability was demonstrated. Other phase II studies should help determine the clinical activity and chemotherapeutic utility of the oral administration of CEP701 with a focus on prostate and pancreatic malignancies. Currently phase I combinatorial studies are ongoing with CEP701 and Gemcitabine [129-131].

4.1.6.2 PKC412

PKC412 (4'-N-Benzoyl-staurosporine or midostaurin; CGP41251) is not only a broad spectrum PKC inhibitor, it also inhibits the tyrosine kinase activities of FLT3. Phase I trials in AML patients with PKC412 combined with 5-fluorouracil or with cisplatin and gemcitabine showed partial responses and tumour stabilization [132, 133]. A phase II trial in 20 AML patients with FLT3 mutations achieved 7 significant responders. One patient remained stable for 11 months, while one patient attained a near complete remission [134].

4.1.6.3 CT53518

CT53518 is a selective inhibitor for FLT3, PDGFR and KIT in vitro and was found to inhibit ectopically expressed mutated FLT-3, as well as to induce apoptosis in human AML cell lines with gain-of-function mutations in the FLT3 gene. Positive responses and increase in survival were observed in two mouse models of mutant FLT3-mediated AML when CT53518 was administered [135]. These encouraging results have opened for phase I and II clinical studies in AML patients [53, 136].

4.1.7 ALK

The receptor tyrosine kinase anaplastic lymphoma kinase (ALK) was originally identified as part of the chimeric nucleophosmin (NPM)-ALK protein associated with anaplastic large cell lymphoma (ALCL). Although its expression in normal tissues is fairly restricted, ALK expression has been described in cell lines derived from neuroblastomas and in neuroectodermal tumours [137]. Like many other RTKs, ALK is implicated in oncogenesis due to genetic abnormalities. Translocation of the ALK gene at 2p23 is shown to produce oncogenic ALK fusion proteins, resulting in ALCL and inflammatory myofibroblastic tumour (IMT). This makes ALK one of the few examples of a RTK to be involved in both nonhaematopoietic and haematopoietic oncogenesis [reviewed in 138 and 139]. NPM-ALK has been linked to several apoptotic and cell proliferation signalling pathways, thus making it a potential target for alternative therapeutical approaches [138]. CD30 is a member of the tumor necrosis factor (TNF) receptor family and found to physically interact with NPM-ALK. Although the exact functional relevance of CD30 in ALK-positive ALCL is still unclear, both conjugated to saporin and unconjugated anti-CD30 antibodies possess in vivo anti-tumour activity in human ALCL cells. Anti-CD30 mAb (or SGN-30) have reached phase I/II clinical trials [53, 140]. The 7-hydroxystaurosporine (UCN-01) is an ATP30 competitive small molecule inhibitor that blocks the activity of numerous kinases and a Phase I study suggested an effect in treatment of ALK-positive malignancies [141]. Herbimycin A inhibits NPM-ALK kinase activity in cell models, but so far this tyrosine kinase inhibitor has not tested clinically on ALCL patients [142]. Ribozyme-mediated therapeutic approaches have been tried, but cleavage of the NPM-ALK fusion transcript by anti-ALK ribozymes failed in human ALCL cell lines due to the prolonged half-life of NPM-ALK [138]. However, ribozyme-mediated degradation against full-length ALK is more likely to succeed due to the frequent low levels of ALK expression in tumours.

4.2 Inhibitors against nonRTK

4.2.1 c-Abl

The proto-oncogene c-abl was originally identified because of its homology with the viral oncogene v-abl, whose gene product was able to induce acute neoplastic transformation in the mouse. The cellular c-abl encodes a nonRTK that plays key functions in the signalling pathways regulating growth factor-induced proliferation and in the regulation of cell growth and cell cycle (see figure 1). Moreover, this protein is engaged in the mechanisms that regulate the variations of the cellular morphology and the intercellular adhesion. In addition c-ABL has been shown to inhibit migration of fibroblasts. Finally, c-Abl may affect gene expression as it possesses a DNA-binding domain in its C-terminus and c-Abl can interact with the transcription factor cAMP-response element binding protein [143, 144]. c-Abl is strongly implicated in chronic myeloid leukaemia (CML), which is tightly associated with a chromosomal abnormality known as the Philadelphia chromosome. This reciprocal translocation involves the long arms of chromosomes 9 and 22 and results in the juxtaposition of parts of the BCR and c-abl genes to form a hybrid bcr-abl gene that encodes a 210 kDa fusion protein BCR-ABL. This protein has a causative role in the neoplastic transformation of stem cells leading to CML. The oncogenic action of the BCR-ABL chimaeric protein remains elusive, but the tyrosine kinase activity of the fusion protein is substantially increased compared to wild-type ABL [144, 145].

4.2.1.1 Imatinib mesylate (STI571, Glivec, Gleevec)

Imatinib mesylate was among the first anticancer agents to be developed and is probably the best proof of the therapeutic potentials of rational anticancer drug design. The history, mode of action and resistance problems of imatinib mesylate have been extensively and excellently reviewed by others. Therefore imatinib mesylate and its follow-up drugs are only briefly discussed here. For further information, the reader is referred to recent reviews [3, 146]. Imatinib mesylate inhibits the tyrosine kinase activity of the fusion protein BCR-ABL, an oncogenic hybrid found in 90% of all patients with chronic myelogenic leukaemia (CML) and 5 25% of all patients with acute lymphoblastic leukaemia (ALL). Imatinib also inhibits the tyrosine kinase activities of PDGFR-$\alpha$, PDGFR-$\beta$ and the c-KIT receptor, and is therefore used in patients with metastatic GIST [3, 147].

4.2.1.2 AMN-107

AMN-107 was designed as a follow-up of Gleevec as 95% of all patients treated with Gleevec achieve remission. The rational was to develop a drug that binds tighter to BCR-ABL to increase its potency and overcome resistance due to mutations in BCR-ABL. AMN-107 is approximately 20-fold more potent than imatinib and displays improved inhibitory activity against most of the common BCR-ABL mutations. It has reached clinical trial phase II for patients with CML and other blood-related cancers [148].

4.2.1.3 BMS-354825

BMS-354825 was designed for Philadelphia chromosome CML (Ph+CML) and CML patients that were resistant or intolerant to Gleevec. This is an oral agent that inhibits five tyrosine kinases, including BCR-ABL and the nonRTK Scr. The latter kinase may be activated in rare patients with Imatinib resistance and play a role in signal transduction downstream of BCRABL. The drug has undergone phase I trial [149].

4.2.1.4 ON-012380

A reoccurring problem with Imatinib treatment is that a significant portion of the treated patients develops mutations in the kinase domain of BCR-ABL resulting in resistance to the drug. New drugs that bind outside the ATP-binding site and thereby circumvent the problem of resistance are therefore needed. The small-molecule inhibitor ON012380 has this potential. It inhibits BCR-ABL by substrate competition and induces cell death in Ph+ CML cells, and induces apoptosis in Imatinib-resistant mutants in vivo. In mice expressing T315I, the most common BCR-ABL mutant, ON-012380 was able to induce a regression in the leukaemia [150]. Clinical trials are being conducted [53].

4.2.1.5 SKI-606

SKI-606 is an inhibitor of the nonRTK Abl and Src that can be orally administered. Phase I studies in subjects with advanced malignant solid tumours, including breast, colorectal, pancreatic, CML and non-small cell lung cancers are in progress [47, 151, 152].

4.2.2 Src

Src was originally identified as the transforming gene in the genome of the Rous sarcoma virus (v-Src) causing sarcoma in chickens. The v-src gene was a transduced form of the cellular gene c-src. The Src protein is the founding member of a family comprised of 8 other members: Fyn, Yes, Lck, Hck, Blk, Fgr, Lyn and Yrk [153]. The nonRTK Src has been found both over-expressed and highly activated in a number of human cancers, and the relationship between c-Src activation and cancer progression is significant. Furthermore, c-Src may play a role in the acquisition of the invasive and metastatic phenotype. The Scr inhibitor AZD0530 has entered phase I clinical trials [154].

4.2.3 General nonRTK inhibitors

Genistein, an isoflavone that is present at high levels in soy, has been shown to alter the cellular levels of tyrosine phosphorylation of proteins. At high concentrations, genistein acts as a non-specific protein-tyrosine kinase inhibitor, whereas at lower concentrations, inhibition of tyrosine-specific protein phosphatases appears to predominate. A phase I analysis in 13 cancer patients (11 prostate and 2 colon) showed that of twelve evaluable patients, all except one, experienced disease progression [155, 156]. Phase II trials in metastatic breast cancer, bladder cancer and localized prostate cancer have been approved or are in progress [53, 157]. However, it has been reported that genistein can cause genetic mutation by chromosomal aberrations in human lymphoblastoid cells and peripheral blood lymphocytes in in vitro experiments [158, 159].

4.3 Inhibitors against serine/threonine kinases

4.3.1 PI3K/Akt

AKT/PKB (protein kinase B) is a family of serine/threonine kinases. In humans, 3 Akt/PKB genes have been identified encoding PKB$\alpha$/AKT1, PKB$\beta$/AKT2 and PKB$\gamma$/AKT3, respectively. A splice variant of PKB$\gamma$, referred to as PKB$\gamma$1 has been identified. AKT/PKB proteins are central mediators of signal transduction pathways in response to growth factors and other extracellular stimuli and they contribute to several cellular functions such as nutrient metabolism, cell growth, transcriptional regulation and cell survival. AKT/PKB act as downstream effectors of phosphoinositide-3 kinase (PI3-K). PI3-K can be activated by RTK and G-protein coupled receptors and activated PI3-K will generate phosphatidyl inositol3,4,5-triphosphate (PIP3) from phosphatidyl inositol-3,4-diphosphate (PIP2). PIP3 does not activate AKT/PKB directly, but recruits AKT/PKB to the plasma membrane where it is subsequently phosphorylated (Figure 1). The PKB$\alpha$, $\beta$ and $\gamma$ isoenzymes, except the PKB$\gamma$1 splice variant, contain two regulatory phosphorylation sites, one in the activation loop within the kinase domain (Thr-308, Thr-309 and Thr-305, respectively) and one in the C-terminal regulatory domain (Ser-473, Ser-474 and Ser-472, respectively). Phosphorylation of threonine residue alone partially activates AKT, while phosphorylation of serine alone has little effect on AKT activity. For complete activation, AKT requires phosphorylation at both sites. Phosphorylation of the threonine phosphoacceptor site is mediated by the 63 kDa serine/threonine kinase phosphoinositide-dependent kinase (PDK1).

The mechanism of phosphorylation of the serine residue in the C-terminal regulatory domain remains controversial. Autophosphorylation, indirect phosphorylation by PDK1 and phosphorylation by other kinases such as PDK2 and integrin-linked kinase has been suggested. Recently, it was demonstrated that mTOR in complex with rictor can phosphorylate Ser-473 in vitro and in vivo (see section 4.3.2; [160]). Furthermore, phosphorylation of tyrosine residues and PI3-Kindependent phosphorylation (by cAMP-dependent protein kinase or by Ca2+/calmodulindependent kinase) have been suggested as alternative mechanism to activate AKT/PKB, but the biological significance remains to be determined [29, 161, 162].

AKT/PKB seems to play a causative role in several human cancers. Overexpression of AKT2 due to gene amplification has been observed in 30-40% of ovarian carcinomas and pancreatic cancers, while amplification of the Akt1 gene was reported in gastric cancers. Increased AKT1 kinase activity has been described in prostate cancer (in >50% of the tumours), breast and ovarian cancers (~40% of the cancers), while AKT2 activity was especially upregulated in hepatocellular carcinomas and in colorectal cancers. Increased AKT3 activity was monitored in oestrogen receptor-deficient breast cancer and in androgen-insensitive prostate cancer cell lines [161]. The implication of AKT/PKB in human cancer has urged the development of inhibitors of the PI3-K/AKT signalling pathway.

4.3.1.1 Perifosine

Perifosine (KRX0401; NSC639966; D-21266) is a synthetic heterocyclic alkylphosphocholine analogue derived from the miltefosine type of molecules. It differs from miltefosine in that it contains a longer alkyl chain and a piperidine head group. These features improve the inhibitors bioavailability and compared to miltefosine, Perifosine causes less nausea.

In vitro experiments with cancer cell lines from squamous carcinoma, leukaemia, lung, prostate, larynx and others indicated inhibitory activity of the PI3K-AKT/PKB pathway, which is often associated with tumour survival and growth. Preclinical xenograft models with breast cancer showed activity upon administration of high doses (and the high doses being correlated with the onset of the response). Three phase I studies in patients with solid tumours resulted in only minor positive responses, but mainly stable diseases was recognized. Toxicities associated with this therapy comprise gastroinstestinal complaints, which increased with increasing doses. Seemingly, variations in metabolism also contributed to the fact that patients only showed stable disease. Additionally, prophylactic 5-HT3 and steroid anti-emetics need to be supplied to the patient whilst under treatment. Perifosine is now being explored in phase II and in combinational studies with other compounds on the market [163, 164].

4.3.2 mTOR

Growth and proliferation of cancer cells are dependent on external signals like growth factors and signals indicating the availability of sufficient nutrient and blood supply. Many of these signals are conveyed by pathways engaging mammalian target of rapamycin (mTOR). The serine/threonine kinase mTOR is a 289 kDa protein that belongs to the family of PI3K and related kinases (PIKK). The kinase domain of these members resembles the catalytic domain of phosphoinositide 3-kinase (PI3K). The PIKKs include the subfamilies: TOR (target of rapamycin), ataxia telangiectasia mutated (ATM), ataxia telangiectasia and Rad3-related (ATR), and DNA-dependent protein kinase.

The kinase mTOR seems to act as a master switch of cellular metabolism, signalling cells to expend, grow and proliferate. Moreover, a role for mTOR in cell survival has also been suggested. mTOR, which forms part of the PI3K/AKT/mTOR signalling pathway, regulates the response of tumour cells to nutrients and growth factors, and controls tumour blood supply and angiogenesis through effects on vascular endothelial growth factor (VEGF) in tumour and endothelial cells (Figure 1). This pathway is activated in many cancers, especially those with elevated PI3-K signalling or those harbouring mutations in the tumour suppressor PTEN (phosphatase and tensin homologue deleted on chromosome 10). It is estimated that the mTOR signalling pathway is activated in 30-50% of prostatic cancer, 30-60% of malignant gliomas, 30-50% of endometrial carcinoma, >50% of melanoma, >30% of renal cell carcinoma and ~10% of breast cancer. This makes components of this pathway very attractive targets for therapeutic inhibitors. [165-167].

However, recent findings have shown that mTOR exists at least in two distinct complexes: a rapamycin-sensitive complex defined by its interaction with the raptor protein (regulatoryassociated protein of mTOR) and a rapamycin-insensitive complex including the rictor protein (rapamycin-insensitive companion of mTOR). The latter complex cannot bind the rapamycinFKBP12 and is therefore refractory to rapamycin. The rapamycin-rictor complex is suggested to regulate the cytoskeleton through PKC$\alpha$ and was recently shown to phosphorylate Ser-473 of AKT/PKB (see section 4.3.1) [160]. Despite this rapamycin-insensitive complex, rapamycin analogues have entered clinical trials to inhibit the hyperactivated rapamycin-sensitive mTOR pathway in human cancers.

The natural antibiotic rapamycin (or sirolimus) produced by Streptomyces hygroscopicus interacts with the cytoplasmic receptor FK506-binding protein12 (FKBP12) and the rapamycin-FKBP12 complex specifically interacts with mTOR to inhibit mTOR signalling. Although originally used as immunosuppressant drug for organ transplantation, rapamycin derivates are now entering clinical trial in cancer therapy [165, 168]. The development of specific mTOR-rictor pathway inhibitors may form a novel anticancer therapeutic strategy in cancer with elevated AKT/PKB phosphorylation.

4.3.2.1 AP23573

The rapamycin analogue AP23573 was obtained by modification of the C-43 secondary alcohol moiety of the cyclohexyl group of rapamycin with substituted phosphonate and phosphinate groups [167]. AP23573 starves cancer cells and shrinks tumours by inhibiting the critical cell-signalling protein and is currently used in clinical trials.

 Phase I clinical trials indicated that AP23573 was well tolerated with no serious adverse events. In a phase I clinical trial, AP23573 completely blocked in vivo mTOR activity in peripheral blood mononuclear cells. Of 8 evaluable patients, one partial response has been observed in a patient with metastatic renal cell cancer and 1 patient with metastatic sarcoma obtained stable disease for more than 6 months [169]. In another report of a phase I clinical trial, one of 5 evaluable patients had stable medullary thyroid cancer for more than 2 months [170]. In another study, nineteen out of 52 patients (37%) of patients treated with AP23573 and evaluable for at least four months demonstrated sustained anti-tumour responses including three patients with partial responses (confirmed tumour regression >30%) and 16 patients with stable disease for at least four months. Five of these patients continue on trial with stable disease for at least six months. Twenty-three of the 32 (72%) sarcoma patients who entered the trial with tumourrelated symptoms (e.g., pain, shortness of breath and cough) demonstrated clinically beneficial symptomatic relief early during AP23573 treatment. Another phase II study (12.5 mg i.v. daily for 5 days every two weeks) in 12 patients with haematological malignancies (leukaemia and lymphomas) revealed that 6 of 11 evaluable patients had minor improvement or stable disease [171]. Chawla and co-workers reported their observations of a phase II clinical trial on advanced sarcoma patients. Of the 23 evaluable patients, 13 gave symptomatic clinical improvement [172]. Phase II study and phase Ib clinical trials with AP23573 as single agent or in combination with other anticancer therapies in solid tumours (sarcomas, breast, ovarian, prostate cancer, NSCLC and glioblastoma multiforme) are in progress [173]. Interestingly, AP23573 may also be used to prevent reblockage of injured vessels following stent-assisted angioplasty, a common non-surgical procedure for dilating or opening narrowed arteries because AP23573 has also been shown to potently block the growth, proliferation and migration of vascular smooth muscle cells, the primary cause of narrowing and blockage of injured vessels.

4.3.2.2 CCI-779 or temsirolimus

Cell cycle inhibitor-779 (CCI-779), developed by Wyeth, is a more water-soluble ester derivative of rapamycin and can be administered i.v. Minor to major responses were obtained in phase I clinical trials involving patients with lung, renal cell, breast carcinaoma, neuroendocrine tumours, soft tissue sarcoma, endometrial and cervical carcinomas. A phase I combined trial with CCI-779 and 5-fluorouracil increased the cytotoxic effects in patients with advanced solid tumours [167]. Subsequently, phase II studies have been initiated in patients with advanced breast cancer or advanced renal cell carcinoma. Clinical benefit was noted in 26-37% of the breast cancer patients. However, none of the 32 Her-2 negative tumours showed any significant response to the treatment with CCI-779. Combining CCI-779 with letrozole did not seem to have an additional effect. The overall response in renal cell carcinoma patients was 7%. Another phase II study in recurrent glioblastoma multiforme showed that the time of tumour progression had significantly increased in 36% of the patients [174,175]. A ~40% response was observed in a phase II clinical trial in patients with mantle cell lymphoma [176]. A phase II trial with CCI-7999 plus the antibody Rituximab is in progress [53]. Phase III trials with renal cell carcinoma and mantle cell lymphoma patients have been initiated [165, 167, 177, 178].

4.3.2.3 RAD001 or everolimus

RAD001 (40-O-(2-hydroxyrthyl)-rapamycin; Everolimus) is another analogue of rapamycin that can be administered orally. This mTOR inhibitor has been/is being tested in several phase I trials in patients with solid tumours, brain tumours and leukaemia. Out of 16 evaluable patients with solid tumours, four (one with hepatocellular cancer, one with fibrosarcoma and two with NSCLC) had a stable disease for more than 16 weeks [179]. Phase I combination studies with imatinib mesylate (patients with GIST [180]), gemcitabine (patients with solid tumours, [181]), letrozole (Femara; breast cancer patients, [182]), Bevacizumab and Terlotinib [183]), Gefitinib (NSCLC patients, [184]) or Terlotinib (NSCLC patients, [185]) are ongoing. Of 14 evaluable patients with GIST treated with a combination of RAD001 and imatinib mesylate, 8 had stable disease for more than 4 months, while two patients displayed partial response [180]. Phase II studies in patients with chronic lymphocytic leukaemia, mantle cell lymphoma and endometrial cancers are in progress. The rational for the latter study is that 40-50% of endometrial tumours have mutations in PTEN, which leads to constitutive activation of AKT and hence of mTOR [186]. Patients with refractory rhabdomyosarcoma or non-rhabdomyosarcomatous soft tissue sarcoma (aged 3-21 years) are being enrolled for a phase I trial at the St. Jude Children¡Çs Research Hospital [187], while Phase III studies in breast cancer patients are planned [53].

4.3.3 Mitogen-activated protein kinases

The mitogen-activated protein kinases (MAPK) regulate diverse cellular activities including gene expression, mitosis, metabolism, motility, cell survival, apoptosis, and differentiation. The MAPK pathways consist basically of a three-component module in which a cascade of phosphorylation events successively activates a MAP kinase kinase kinase (MKKK), a MAP kinase kinase (MKK), and a MAP kinase (MAPK). MAPK can phosphorylate substrates directly, or may phosphorylate yet another kinase (MAPK activated protein kinase or MAPKAPK), which then phosphorylates substrates [188]. The role of MAPK in cancer is well documented and has been recently reviewed [189, 190]. The Raf kinases and the mitogen-activated or extracellular signal-regulated protein kinases 1 and 2 have been the main focus for anticancer therapy therapeutic and will be discussed further.

4.3.4 The serine/threonine kinase Raf

The Raf kinase family consists of the members c-Raf-1, B-Raf and A-Raf in humans. c-Raf-1 is the human homologue of viral Raf, which was isolated more than 20 years ago from the acutely transforming murine sarcoma virus 3311-MSV. The name Raf is derived from rapidly transforming fibrosarcoma and this protein turned out to be the first oncoprotein with serine/threonine kinase activity. Activation of the Raf members, especially c-Raf-1 is complex, incompletely understood and beyond the scope of this review. In brief, binding of extracellular ligands to their cognate receptors activates the small G protein Ras, which in its GTP-bound form recruits and activates Raf. Activated Raf, in turn activates the dual specificity protein k