Protein Tyrosine Kinase Inhibitors

Although protein kinases have been known since the discovery of protein phosphorylation in the 1950s, no one turned to them as drug targets until protein kinase C (PKC) and tyrosine phosphorylation were discovered over 20 years later. Identifying tyrosine kinase activity as being the hallmark of the oncogenic activity of pp60c-Src (and dozens of other oncoproteins) prompted researchers to investigate these proteins as novel targets for drugs. Tyrosine phosphorylation inhibitors (tyrphostins) were subsequently developed as a strategy to combat cancer and other proliferative diseases in the late 1980s [3,11,12]. Also discovered were a number of serine/threonine kinases such as cyclin-dependent kinases (Cdks), Erks, Raf, and PKB/Akt, which play a key role in cell proliferation, cell division, and anti-apoptotic signaling. In contrast to most Ser/Thr kinases known to be involved in housekeeping cellular duties, the more recently discovered protein kinases are directly involved with cellular signaling. In the human genome, we currently identify 409 Ser/Thr kinases, 59 receptor protein tyrosine kinases (RPTKs) and 32 non-RPTKs. Most of the Ser/Thr kinases

Table I Protein Kinase Inhibitors in Clinical Development






Genetech/Roche Novartis




OSI Pharmaceuticals



Ludwig Institute for Cancer Research



Eli Lilly

Herceptin(mAb) STI571/Gleevec


ZD 1839/Iressa SU 6668

Tarceva (OSI 774) CEP-1347 CEP-701 AG 1478/CDDP

BAY 43-9006 LY 333531

Breast cancer Chronic myeloid leukemia (CML) Gastrointestinal stromal tumor (GIST) Solid tumors Solid tumors Solid tumors Parkinson's disease Prostate cancer Glioblastoma multiforme

Colon cancer Diabetic retinopathy

Her-2 Marketed

Bcr-Abl Marketed

C-Kit Marketed

EGFR, Her-2 Marketed


EGFR Phase 2

Mixed lineage kinase Phase 1

NGFR Phase 2

EGFR Phase 1

Raf Phase 1

Cdk4/1 Phase 1

Protein kinase C Phase 3

are housekeeping metabolic enzymes for which only a small fraction is involved in cellular signaling, in contrast to PTKs, which are primarily involved in signaling. The activities of PTKs are associated with enhanced proliferation and strong survival signals, the two most prominent traits of cancer cells.

The development of PTK inhibitors originally known as tyrosine phosphorylation inhibitors (tyrphostins) led to the approach of signal transduction therapy aimed at eradicating cancer cells. Research in this area has demonstrated that one can generate small molecules with a high degree of selectivity against different PTKs, even closely related ones such as endothelial growth factor receptor (EGFR) and Her-2/neu. Table 1 lists PTKs and Ser/Thr kinase inhibitors currently in clinical development. It is interesting to note that, in a number of cases, the PTK inhibitor, even as a single agent, induces apoptosis in the treated cancer cell but has no such effect on normal cells and is well tolerated by the treated animal.

The first striking example is the case of the Jak-2 inhibitor AG 490 (Fig. 1) [13]. This tyrphostin was found to induce apoptosis of recurrent pre-B acute lymphoblastic leukemic (pre-B ALL), eliminating completely the pre-B ALL cells from severe combined immunodeficiency (SCID) mice engrafted with the disease; treatment of the animals began 5 to 9 weeks after disease engraftment. Furthermore, AG 490 was not inhibitory to normal B or T cells when stimulated by various means. This pioneering study validates the hypothesis discussed here [14] and is considered to be an important milestone in signal transduction therapy. The diseased Pre-B ALL cells depend for their survival and growth on the persistently active Jak-2, but normal B cells (and normal T cells) are completely oblivious to the inhibition of Jak-2. Indeed, the inhibition of Jak-2 is sufficient to induce apop-tosis in the pre-B ALL cells, whereas its inhibition in normal cells seems to have no effect whatsoever [13]. Similar results have also been obtained with interleukin-6 (IL-6)-dependent multiple myeloma cells, which are also driven by

Jak-2. Very recently, a novel class of Jak2/3 kinase inhibitors has been developed but the in vivo activity of these inhibitors has not been reported [15].

Chemistry of Tyrosine Kinase Inhibitors

Initially, a large number of natural compounds were found to be rather potent inhibitors of PTKs. Although many showed initial promise, they all were found to be highly promiscuous and toxic in that they hit many cellular targets. The first PTK inhibitors to be synthesized were benzene malononitrile tyrphostins [11,12]. These compounds (Fig. 1) are competitive with the substrate and noncompetitive with ATP. Structure-activity relationship studies generated compounds that are 1000-fold more active against the EGFR kinase as compared to insulin receptor kinase, with no measurable activity against protein kinase A (PKA) and other serine/ threonine kinases. As the number of identified tyrphostins grew, a more complex pattern of kinetics of inhibition of the EGFR kinase began to emerge. Tyrphostins competitive against either substrate or ATP were common, but so were compounds competitive with both [17]. Some compounds were found to be partially competitive (mixed competitive); their interactions with the EGFR [17] or PDGFR [18] reduce the binding affinity of ATP and substrate as well as causing a reduction in the catalytic activity of the enzyme [16]. This type of behavior suggests that the inhibitor binds to sites different than the active site and therefore qualifies as an allosteric inhibitor. As tyrphostins became cyclized (Fig. 2), incorporating nitrile nitrogen into the second ring caused most of the compounds to become ATP competitive [19-22].

Since 1994, the main thrust in the development of PTK inhibitors, especially by pharmaceutical companies, has been toward the generation of ATP mimics (ATP-competitive kinase inhibitors) [23]. Most of the inhibitors generated are based on a scaffold structured around two or more

Figure 2 Cyclized tyrphostins. Incorporating the nitrillo(cyano) nitrogen within a second ring generated two-ring tyrphostins as opposed to the one-ring system (see Fig. 1). When a second nitrogen is introduced into the second ring, selective ATP mimics emerge. Interestingly, similar compounds have been identified as PTK inhibitors by random screening rather than by semi-rational design. AG 1150 is rather inactive, whereas AG 1296 and AGL 2043 are potent and selective PDGFR kinase inhibitors.

Figure 1 Early Tyrphostins. AG 1112 and AG 957 were prepared as inhibitors of Bcr-Abl; AG 490 is an inhibitor of Jak-2, AG 825 is a Her-2/neu inhibitor, and AG 538 is a potent inhibitor of IGF1-R.

aromatic rings. These compounds were designed to be ATP mimics, but their kinetic behavior toward the substrate was not always investigated. Because the degree of conservation in the ATP binding site is not absolute, one can obtain a high degree of selectivity among closely related ATP binding domains. In 1993, the Jerusalem group [24] was able to demonstrate that ATP-competitive tyrphostins such as AG 825 (Fig. 1) can discriminate between the kinase domains of Her-2/neu and EGFR by almost two orders of magnitude in affinity, in spite of the almost 80% identity in the kinase domains of the two related PTKs. In 1994, the quinazoline ZD 1839 (Iressa; Fig. 3) was shown to be a potent EGFR kinase inhibitor with excellent bioavailability [25]. Quinazolines were originally identified by Zeneca and were shown to selectively inhibit EGFR at low nanomolar concentrations [25], whereas Her-2/neu is inhibited only at micromolar concentrations. Qunixaloines such as AG 1296 [19,26] or AGL 2043 [27] (Fig. 2) were found to block PDGFR kinase with inhibitory effects on related receptors such as c-Kit and Flt-3 and with 10- to 50-fold less efficacy against the more distantly related receptor VEGFR (unpublished data). The crystal structure of the inactive form of Hck with the Pfizer inhibitor PP1 [28] and of the active form of Lck with PP2 [29] explained why this ATP mimic binds better to the Src family kinase binding domain than does EGFR and much better than to a number of other tyrosine kinases and PKA. It was found that when threonine 338 is

Figure 2 Cyclized tyrphostins. Incorporating the nitrillo(cyano) nitrogen within a second ring generated two-ring tyrphostins as opposed to the one-ring system (see Fig. 1). When a second nitrogen is introduced into the second ring, selective ATP mimics emerge. Interestingly, similar compounds have been identified as PTK inhibitors by random screening rather than by semi-rational design. AG 1150 is rather inactive, whereas AG 1296 and AGL 2043 are potent and selective PDGFR kinase inhibitors.

substituted for methionine or alanine 402 is substituted for another amino acid, the affinity to PP1 drops markedly [28].

EGFR Family Kinase Inhibitors

The role of EGFR in many cancers was appreciated early on and was one of the first targets for therapy. Indeed, the quinazoline Iressa (ZD 1839) (Fig. 3) [21,25] is in advanced clinical trials for treatment of cancers for which EGFR plays an important role, such as lung cancer and head and neck cancer. Similarly, the quinazoline AG 1478 (Fig. 3) [22] is in clinical development for the treatment of glioblastoma multiforme in which the EGFR and its persistently active A (2-7) EGFR are overexpressed [30,31]. This tyrphostin will be used in combination with CDDP, with which it synergizes to induce apop-tosis in glioma multiforme cells in vitro and in vivo [31]. OSI 774 (Fig. 3) is also an effective quinazoline in clinical development [32]. Over the years, we have come to realize that het-erodimer combinations of the four members of the Her family play a role in the oncogenic phenotype of many cancers; therefore, attempts are being made to generate inhibitors that inhibit both Her-1 and Her-2. The Glaxo-Welcome Her-1/Her-2 inhibitor GW 2016 (Fig. 4), blocks both receptor tyrosine kinases at 12 nM [33], and it is to be expected that more compounds aimed at the Her family will emerge in the near future.

AG 1478

OSI 774

Figure 3 Quinazoline EGFR kinase inhibitors. AG 1478, ZD 1839, and OSI 774 are reversible EGFR kinase inhibitors, whereas CI 1033 is an irreversible inhibitor.

CI 1033

Figure 3 Quinazoline EGFR kinase inhibitors. AG 1478, ZD 1839, and OSI 774 are reversible EGFR kinase inhibitors, whereas CI 1033 is an irreversible inhibitor.

Covalent EGFR Kinase Affinity Labels

The covalent attachment of a selective inhibitor to the EGFR kinase domain (or to any target PTK domain) completely abolishes the catalytic activity of the receptor and is therefore believed to possess better clinical potential. The Parke-Davis compound CI-1033 (Fig. 3) [34] is highly effective in vivo as an EGFR kinase inhibitor; the effect of CI-1033 is long lasting and seems to possess higher efficacy than its reversible analogs. The covalent label attaches to cysteine 773, close to the ATP binding domain, and most

Figure 4 Her-1/2 kinase inhibitor.

probably targets the receptor for degradation followed by cell death. Other covalent labels of the EGFR kinase site were also reported to possess strong antitumor activity in vivo at relatively low doses.

From Tyrphostins to Gleevec

In 1993, it was demonstrated that selective Bcr-Abl kinase inhibitors such as tyrphostin AG 1112 (Fig. 1) induce the terminal differentiation of K562 cells [35]. Similarly, another Bcr-Abl-selective kinase inhibitor, AG 957 [36,37], induces the purging of Ph+ cells and synergizes with anti-Fas receptor antibody to induce their demise [38]. Druker et al. [39] followed up on these studies by utilizing CGP 57148, renamed STI 571/Gleevec/Glivec and produced by Novartis (Fig. 5) [40,41]. This highly potent inhibitor inhibits PDGFR kinase as well as its homologous PTK c-Kit but is also a powerful inhibitor of Bcr-Abl kinase. Gleevec was found to induce complete remission in chronic myeloid leukemia (CML) patients which has lasted for nearly 3 years for most patients. It is interesting to note that these patients, who take about 400 to 800 mg daily, suffer only minor side effects and tolerate the drug well. This is a rather surprising, as STI 571 blocks c-Abl, PDGFR, and c-Kit, which play important roles in the function of normal cells. The most likely explanation is that normal cells that utilize c-Abl, PDGFR, or c-Kit can get by even when over 90% of these targets are blocked by utilizing the alternative pathways that all normal cells possess.

Chronic CML cells are highly dependent on Bcr-Abl for their survival and therefore die when the target is blocked, thus validating the principle of enhanced sensitivity of the cancer cell to an inhibitor that targets the element whose signaling is enhanced and on which the cancer cell depends for its survival. The findings on Bcr-Abl are reinforced by the remarkable activity of Gleevec on a subpopulation of gastrointestinal stromal tumor (GIST) patients [42]. The common denominator of the patients who respond to the drug with complete remission or almost complete remission is that their tumors express Kit receptors carrying mutations in exon 11, which converts the receptor to a persistently active kinase. As is true for chronic CML, it seems that the survival of the tumor cells is highly dependent on the signaling of mutated

Figure 5 STI 571/Gleevec.

Ji STI 571/Gleevec

Figure 5 STI 571/Gleevec.

Kit receptor kinase, whereas normal cells can sustain and compensate for c-Kit blockade. Again, as for chronic CML patients, STI 571 has little side effects, probably because normal cells utilize alternative pathways when c-Kit is blocked or can get by even when a high fraction of the normal c-Kit molecules are blocked by utilizing alternative pathways.

ATP Mimics and Substrate Mimics

We have suggested that the optimal PTK inhibitors would be compounds that compete for the substrate binding site within the kinase binding domain. Such agents would be less toxic than ATP mimics because they bind to those domains at the kinase site that are less conserved than the substrate binding domains. Indeed, tyrphostins such as AG 490 (which blocks Jak-2 [13]) and AG 556 (which possesses antiinflammatory properties) have been shown to be highly nontoxic in vivo [43-46]. The main problem with these compounds is that they possess hydroxyl groups, which are metabolized relatively quickly, although this characteristic has not eliminated DOPA as an anti-Parkinsonian drug. Recently, we have developed substrate mimics in which the hydroxyls are replaced with bioisosteres. Half of the hydroxyl groups in AG 538 [47] were replaced with a type of bioisostere without losing much of the potency against IGF-1R and still retaining the substrate-competitive nature of the compound [48]. The double bond, which is present in many tyrphostins, may be a substrate for the Michael addition, which shortens the half-life of these compounds. Although this is rarely a problem, as most tyrphostins are stable in tissue culture medium for many hours [13,24], higher chemical stability can be achieved by eliminating this double bond. Work in progress indeed suggests that one can probably design substrate-competitive inhibitors devoid of double bonds.

A strong argument for developing substrate-competitive protein kinase inhibitors is that they are likely to offer higher selectivity, as the portion of the kinase site outside the ATP binding domain is less conserved among protein kinases. This in principle should enable one to discover highly selective kinase inhibitors, which even at high doses will exhibit minimal toxicity in vivo. All of the kinase inhibitors currently in development (Table 1) are used at high doses, between 20 and 100 mg/kg. These high doses reflect the relatively low efficacy of these compounds in vivo, despite the fact that their IC50 values for their molecular targets, such as the EGFR, VEGFR-2/Flk-1 etc., are in nanomolar concentrations.

When one examines the efficacy of the ATP-competitive inhibitors in cellular assays, it is observed that these nanomolar compounds act on cells at micromolar concentrations. For example, quinazolines that bind to the EGFR with Ki of a few nanomolars [7,20] inhibit EGFR autophos-phorylation in intact cells at micromolar concentrations [22,49]. Similarly, PP1 and PP2 that inhibit Src family kinases with IC50 values of 20 to 170 nM in cell-free assays [50] block Src activity in cells in the range of 5 to 40 |M [49,51]. It seems, therefore, that the high doses required in vivo probably at least partly reflect the competitive relationship between the intracellular millimolar concentrations of ATP and the administered drug. It is noteworthy that drugs such as P-adrenergic blockers are administered at doses that are lower by about 100-fold. In this case, a drug possessing an affinity at nanomolar concentrations must compete with the up to 200-nM concentration of the endogenous ligand adrenaline or noradrenaline. Thus, P-blockers can be administered to patients at doses of 1.0 mg/kg or less with great efficacy. The same is true for other receptor-directed drugs for which the endogenous ligand is present at low concentrations within body fluids.

Adenosine triphosphate competitors suffer from another potential problem—the selectivity of newly developed compounds is only tested against a limited number of PTKs and Ser/Thr kinases. The number of PTKs is around 150 and the number of Ser/ kinases ranges from 500 to 600. It has already been observed that the so-called selective Src family kinase inhibitor PP1 is equipotent as a PDGFR kinase inhibitor [52]. Similarly, the Novartis Bcr-Abl kinase inhibitor, STI 571/Gleevec, used to treat CML, is as potent against PDGFR kinase and c-Kit [42]. An inhibitor that competes for both ATP and substrate simultaneously may actually be highly useful. Such tyrphostins have been generated [17], but their toxicity and efficacy in vivo have not been evaluated. It has been generally assumed that one should aim to treat patients for short periods in order to diminish side effects, a goal that is particularly applicable to cytotoxic drugs because of their severe side effects. Perhaps not surprisingly, Gleevec has been given to patients for many months with very minor side effects. In fact, the absence of toxicity may allow the prolonged use of protein kinase inhibitors and other signal transduction inhibitors, maximizing their therapeutic effect.


Angiogenesis is mediated by the activity of various receptors, primarily VEGFR but also PDGFR and FGFR. Following development of AG 1433 (Fig. 6) [10], Sugen developed SU 6668 (Fig. 6) as a PTK inhibitor for VEGFR, PDGFR, and FGFR simultaneously. This very interesting and promising agent (Table 1) is currently in clinical studies [53].

Figure 6 AG 1433 and SU 6668. AG 1433 was found to be a potent PDFR and EGFR kinase inhibitor. This compound was found to inhibit angio-genesis [10] but was not suitable for development. SU 6668 [53] with a new scaffold was developed and is currently in clinical development (see Table 1).

PTK Inhibitors Synergize with Pro-Apoptotic Agents

It has been observed that transformed cells possess a heightened state of sensitivity to stress/apoptotic signals as compared to their parental nonmalignant cells [2,54]. This sensitized state renders the cancer cells higher sensitivity to stress/apoptotic agents such as ds-Platin (CDDP) and pro-apoptotic ligands such as FasL. As the cancer progresses, the potentiated state of sensitivity to stress is covered up by a massive shield of anti-apoptotic signaling networks. Thus, when one applies an anti-apoptotic agent, the potentiated state of the cell is uncovered and the cancer cell becomes hypersensitive to the pro-apoptotic stress agent. This is probably why PTK inhibitors synergize with cytotoxic drugs or pro-apoptotic proteins such as FasL. The first example demonstrating this principle was reported for Her-2/neu-expressing lung cancer cell lines. It was shown that the degree of synergism between a Her-2 kinase inhibitor (AG 825) (Fig. 1) and the cytotoxic agents CDDP, etoposide, or doxorubicin increases with the level of expression of Her-2/ neu in a series of isogenic patient-derived NSLC cell lines [55]. It seems that the degree of anti-apoptotic signaling of Her-2 is proportional to the latent potentiated sensitivity of the cancer cell; thus, the degree of synergism is proportional to the level of Her-2/neu signaling. The fact that PTK inhibitors can synergize with pro-apoptotic agents is already being implemented in the clinic. ZD 1839 (Iressa), a potent EGFR kinase inhibitor developed by AstraZeneca, is currently in clinical trials as a single agent and in combination with cytotoxic agents such as Taxol® [56]. Another interesting case is advanced glioblastoma multiforme (GM), where the truncated EGFR A (2-7 EGFR) is responsible for its resistance to chemotherapy and radiation therapy. Blockade of the EGFR by the EGFR kinase inhibitor AG 1478 sensitizes the tumor to CDDP and enhances the survival of tumor-bearing nude mice treated with the two agents [30,31]. This combination is in clinical development for the treatment of GM. As mentioned previously, STI 571/Gleevec shows remarkable activity against chronic CML, and patients on STI 571 have been in complete remission for nearly 3 years, but it is much less effective as a single agent in blast crisis CML, where the disease recurs in 80% of the patients within a few months. These patients are likely to be given a combination of STI 571 with pro-apoptotic agents early in the treatment. So far, PTK inhibitors have not been tried in humans in combination with pro-apoptotic proteins or antibodies. In tissue culture, the synergy between PTK inhibitors, cytotoxic agents [55,56], FasL, or anti-Fas receptor antibody [38,57] was established.

PTK Inhibitors for the Treatment of Non-Cancer Diseases

The enhanced activity of PTKs has been related to diseases other than cancer. The enhanced activity of EGFR is the hallmark of psoriasis and papilloma. In both instances, EGFR is overexpressed and the diseased cells produce EGFR-stimulating ligands. This is the reason why EGFR kinase inhibitors inhibit the growth of both psoriatic ker-atinocytes [58-61] and keratinocytes immortalized with human papillomavirus 16 (HPV-16) [62,63]. It has been suggested that EGFR kinase inhibitors be used as topical agents to treat these conditions. PDGFR is the key player in restenosis following balloon angioplasty; therefore, PDGFR kinase inhibitors have been tested as inhibitors of balloon-induced stenosis. Local application of AG 1295 [64] and AGL 2043 [27] during balloon angioplasty has indeed been shown to be highly effective against the development of stenosis after balloon angioplasty. One can envisage the utilization of PTK inhibitors for other indications, such as pulmonary fibrosis, in which the enhanced activity of PTKs plays an important role in the pathophysiology of the disease.

PTK Inhibitors as Cancer Prevention Agents

Over the past few years, interest has increased in agents that are natural components of foods and have anticancer properties [65]. It has been suggested that males from the Far East suffer less from prostate cancer because they consume food rich in genistein, which is a nonselective PTK inhibitor. Studies suggest that genistein prevents the growth of metastatic cancer [66-68]. Dietary genistein downregu-lates the expression of the androgen receptor and estrogen receptor-a and -P in the rat prostate, at concentrations comparable to those found in humans on a soy diet [69]. Thus, downregulated sex steroid receptor expression may be responsible for the lower incidence of prostate cancer in

Figure 7 Key signaling Ser/Thr kinases. A small fraction of Ser/Thr kinases are valid targets for drug development; these are indicated by bold type in boxes. Crosses appear on tumor suppressors for which activities are compromised in cancer and therefore augment the activation of various kinases.

populations living on a diet containing high levels of phytoestrogens. Clinical trials are therefore planned for patients with metastatic prostate cancer who will receive genistein to examine whether this agent does, indeed, have any antimeta-static effects. Also, it has been suggested that polyphenols, which are components of wine, possess antineoplastic effects. These effects are largely attributed to their antioxidant properties, but it is possible that the inhibitory effects on PTKs are also responsible.

PTK Inhibitors for Diagnostic Purposes

It is always essential to establish a treatment modality based on the presence of the drug target in the diseased tissue. For example, it is extremely important to know ahead of treatment with an EGFR kinase inhibitor if the tumor does, indeed, overexpress the receptor. This can be achieved by imaging the tumor utilizing an EGFR kinase inhibitor, which is a positron emitter. Recent studies show that a reversible EGFR kinase inhibitor [70] is inferior to an irreversible one [71] for achieving this goal. Positron emission tomography (PET) imaging of the EGFR is extremely important for the determining which patients with non-small-cell lung carcinoma are eligible for treatment with Iressa. Because many of the tumors express either EGFR or Her-2, or both, a dual PET imager based on GW 2016 (Fig. 4) may also be useful.

Kicking the Habit

Kicking the Habit

Stop Thinking About How You're Going To Quit Smoking When You Can Instantly Stomp It In Less Than 30 Days With A Proven Set of Techniques. When your

Get My Free Ebook

Post a comment