Targeting Microtubules by Natural Agents for Cancer Therapy

Vinca alkaloids and related drugs

Vinca alkaloids, among the earliest developed for disruption of microtubule, were isolated from periwinkle leaves, Catharanthus roseus (L.) G. Don (also known as Vinca rosea). These compounds have been successful in the clinic not only for the treatment of diabetes, high blood pressure, and as disinfectants, but also for being cancer fighters (26). Vinca alkaloids were discovered in the 1950s by Canadian scientists, Robert Noble and Charles Beer, for their antimitotic and, therefore, cancer chemotherapeutic potential. They were first used as single agent in the treatment of childhood hematologic malignancies, then widely spread into use for solid malignancies and, shortly after, for adult hematologic malignancies (Table 1; ref. 27). Because of their great success in the treatment of childhood leukemia, they were considered “wonder drugs.” Further development of the Vinca alkaloids has been motivated by the increased understanding of their mechanisms of action, their synergy in several combination therapies, and the desire to develop orally available analogs. This has led to the development of various novel semisynthetic analogs, including vindesine, vinorelbine, and vinflunine to overcome the side effects of peripheral neuropathy and reversible myelosuppression that commonly occur with these drugs (28).

Although the causes of the neurotoxicity are undefined, they definitely involve the effects of the drugs on microtubules, which are a key component of neurons (29). Neuropathy might result from disruption of axonal flow by bundling of microtubules by paclitaxel (30), from steric hindrance of motor-protein binding to microtubules, or from the effects of altered microtubule dynamics in axonal processes or on transport in growth cones. Another cause of neurotoxicity might result from microtubule destabilization or from suppression of microtubule dynamics due to neuronal retraction and reduced arborization (30); also, from reduced responsiveness of neurons to incoming signals; or from demyelinization (29). Myelosuppression, also called bone marrow suppression or myelotoxicity, is a result of blockage of mitosis and proliferation of the rapidly cycling bone marrow cells. This is a common side effect associated with Vinca alkaloids chemotherapy.

Jordan and colleagues reported that Vinca alkaloids destroy mitotic spindles at high concentrations (e.g., 10–100 nmol/L in HeLa cells) and depolymerize microtubules (4), therefore leaving the dividing cancer cells blocked in mitosis with condensed chromosomes. Also, the same authors stated that at low but clinically relevant concentrations, vinblastine does not depolymerize spindle microtubules, yet it powerfully blocks mitosis (e.g., IC₅₀ 0.8 nmol/L in HeLa cells) and cells die by apoptosis (4). Studies from the same laboratory revealed that the mitotic-blocking action of low concentrations of Vinca alkaloid agent in living cancer cells is due to suppression of microtubule dynamics rather than microtubule depolymerization (12).

Vinblastine binds to the β-subunit of tubulin dimers at a distinct region called the Vinca-binding domain (31). Various other novel chemotherapeutic drugs also bind at this domain. The binding of vinblastine to soluble tubulin is rapid and reversible (12). This induces a conformational change in tubulin in connection with tubulin self-association (32).

In vitro studies revealed that vinblastine also binds with high affinity to tubulin at the extreme microtubule ends; however, it binds with low affinity to tubulin that is buried in the tubulin lattice (12, 33). These studies show that binding of one or two molecules of vinblastine per microtubule plus end is sufficient to reduce both treadmilling and dynamic instability by approximately 50%, without significant microtubule depolymerization. This suppression of dynamics prevents the mitotic spindle from assembling normally and it reduces the tension at the kinetochores of the chromosomes. Mitotic progress is delayed with chromosomes often stuck at the spindle poles not capable of assembly at the spindle equator. The cell-cycle signal to the anaphase-promoting complex to pass from metaphase into anaphase is blocked and the cells eventually die by apoptosis (12, 33).

In addition, this group includes a large number of compounds that have not undergone clinical development for cancer therapy, including the antitussive noscapine (15), maytansine, rhizoxin, spongistatins, podophyllotoxin, steganacins, and curacins (34); several herbicides that inhibit microtubule polymerization; antifungal, and antihelmintic agents (35); and some psychoactive drugs (ref. 32; Table 1).

Colchicine

Colchicine was first isolated in 1820 by French chemists P.S. Pelletier and J.B. Caventou from autumn crocus (Colchicum autumnale), and described for treatment of rheumatism and swelling (36). Although colchicine has been used clinically in the treatment of gout, neither colchicine nor compounds that bind to the colchicine site on tubulin have been found to be of significant use in cancer treatment because of their potent toxicity (Table 1; ref. 37).

As with Vinca alkaloids, Colchicine depolymerizes microtubules at high concentrations and stabilizes microtubule dynamics at low concentrations. Colchicine inhibits microtubule polymerization by binding to microtubule ends rather than to the soluble tubulin pool. However, free colchicine itself probably does not bind directly to microtubule ends. Instead, it first binds to soluble tubulin, induces slow conformational changes in the tubulin, and ultimately forms a poorly reversible tubulin–colchicine complex (38), which then copolymerize into the microtubule ends in small numbers along with large numbers of free tubulin molecules (39).

So, despite the differences between the effects at high concentrations of the Vinca/colchicine-like drugs and the taxane-like drugs, almost all of the microtubule-targeted antimitotic drugs stabilize microtubule dynamics at their lowest effective concentrations (4).

Several natural products that bind in the colchicine domain and disrupt microtubule assembly in vitro and in vivo are known, such as podophyllotoxin, from the mandrake herb Podophllum peltatum, and combretastatins (40).

Combretastatins

Combretastatins are a class of natural cis-stilbenes (phenols) isolated from the South African tree Combretum caffrum (Combretacae). Combretastatins exhibit cytotoxic properties and tubulin polymerization inhibition in cancer cells in vitro (41). The most potent member of this group is combretastatin A-4 with IC₅₀ of 7 nmol/L (42). The combination of this compound with a phosphate prodrug, called combretastatin A4 phosphate, also known as Zybrestat, fosbretabulin (9), progressed into clinical trials for the treatment of solid tumors (Table 1; refs. 41, 42). The advantages of this class of compounds over other antimitotic agents is that beside their antimitotic properties, they are also angiogenesis inhibitors and known as vascular disrupting agents (42, 43). In addition, they are being evaluated for their efficacy in the treatment of diabetic retinopathy which is the biggest single cause of blindness (44).

Paclitaxel and related drugs

The taxanes (paclitaxel, docetaxel), a novel class of antimicrotubule agents, are the most important addition to the chemotherapeutic armamentarium against cancer over the past several decades. Paclitaxel was discovered as part of a National Cancer Institute (Bethesda, MD) screening program which extracted thousands of plants for anticancer activity (45). Paclitaxel, Taxus brevifolia, was extracted from the stem bark of the Pacific yew tree in 1966 (46). The structure of paclitaxel was reported in 1971; however, the microtubule-stabilizing properties of this compound were discovered later by Schiff and colleagues in 1979 (ref. 47; Table 1). Although taxanes bind poorly to soluble tubulin, they bind to tubulin along the length of the microtubule with high affinity (48). Paclitaxel binds to the β-tubulin subunit, and promotes microtubule stabilization by inducing conformational changes of the M-loop of β-tubulin that result in more stable lateral interactions between adjacent protofilaments (48). In contrast with a high concentration of taxane required to increase microtubule polymerization, a study found that the binding of a very small number of paclitaxel molecules powerfully stabilizes the dynamics of the microtubules without increasing microtubule polymerization (11). Suppression of microtubule dynamics by paclitaxel leads to mitotic blocking in the absence of significant microtubule bundling (11). Jordan and colleagues show that half of HeLa cells in mitosis are blocked at 8 nmol/L paclitaxel, whereas there is no increase in microtubule-polymer mass below 10 nmol/L paclitaxel (12). The polymer mass of paclitaxel is half increased at 80 nmol/L (12). The suppression of spindle-microtubule dynamics by taxanes prevents the dividing cancer cells from progressing from metaphase into anaphase and the cells eventually die by apoptosis (11, 12, 49).

Because of the limited quantities of paclitaxel from the bark of Taxus brevifolia, alternate sources, particularly an approved semisynthesis of 10-deacetylbaccatin III, were obtained to cover the need (4). The U.S. Food and Drug Administration (FDA) approved taxol for the treatment of metastatic ovarian cancer in 1992. Today, taxol is one of the most important anticancer drugs and is widely used clinically for the treatment of ovarian, breast, and non–small cell lung cancer, either as monotherapy or in combination with cis-platin. It is also used for the treatment of acquired immunodeficiency syndrome-related Kaposi sarcoma.

Paclitaxel and its semisynthetic analogue docetaxel are widely prescribed antineoplastic agents for a broad range of malignancies (50). The side effects are neurotoxicity and myelosuppression similar to those associated with Vinca alkaloids (51). Treatment of prostate cancer with taxanes also inhibits androgen receptor signaling by inhibiting the androgen-dependent nuclear translocation of the androgen receptor (AR; ref. 52). In the absence of ligand dihydrotestosterone, AR binds with Hsp90 chaperone in the cytoskeleton. However, in the presence of dihydrotestosterone, the ligand binding results in AR homodimerization and translocation in the nucleus, where it binds with specific androgen-dependent genes, resulting in proliferative and trophic effects (53). This findings is an evidence that in addition to blocking cell division, microtubule-targeting drugs also impair AR signaling.

Paclitaxel was the first microtubule-stabilizing agent that was investigated in an animal model of neurodegenerative tauopathies, the T44 tau Tg mouse. The ability of paclitaxel and docetaxel to cross the blood–brain barrier (BBB) is limited. This is believed to be caused at least in part by the P-glycoprotein (Pgp) efflux pump, which is highly expressed in the BBB (54). Thus, taxane analogs that are capable of overcoming Pgp-mediated transport may result in improving the effectiveness of the drugs. Several compounds of this type have been reported, such as cabazitaxel, an FDA-approved semisynthetic taxane with weak Pgp substrates that can saturate the active transporter (55). Furthermore, pharmacokinetic studies with cabazitaxel show that administering the compound via rapid infusions in the brain significantly enhances the drug concentration, which results in plasma drug levels that are above the threshold needed to saturate Pgp (55). Other examples of taxanes that circumvent Pgp-mediated efflux are orally active BMS-275183 and milataxel (MAC-321; ref. 56). In addition, other approaches with brain-targeted delivery have been reported with favorable results, such as paclitaxel–peptide conjugate GRN1005, a Pgp-insensitive prodrug that exploits the low density lipoprotein receptor-related protein 1. Albumin nanoparticle carriers proposed for taxanes not only enhance the microtubule inhibition, but also prevent the use of solubilizing cremophors, which sometimes induce severe hypersensitivity (15). Other nanoparticle-based carriers, such as nanoporous silicon particles, for chemotherapeutic shutting, have been recently proposed.

Search for other drugs that enhance microtubule polymerization has yielded several promising compounds, including the epothilones, discodermolide, the sarcodictyins, eleutherobin, and laulimalide. Some of these compounds compete with paclitaxel for binding to microtubules and bind at or near the taxane site (epothilones, discodermolide, eleutherobins, and sarcodictyins), but others, such as laulimalide, seem to bind to unique sites on microtubules (25).

Epothilones

Epothilones are a family of novel MTAs, stabilizing microtubules and inhibiting microtubule dynamic behavior at mitotic spindle and, therefore, preventing cancer cells from mitosis. Epothilones A and B were discovered by Gerth and colleagues as antifungal agents produced by the soil bacterium Sorangium cellulolus (57). These compounds compete with paclitaxel for the taxane-binding site on β-tubulin, suggesting that this class of compounds may act on microtubules in a paclitaxel-like manner and promote microtubule assembly (58). This observation led to the suggestion that epothilones, taxanes, and other classes of microtubule-stabilizing natural products may share a similar pharmacophore (ref. 59; Table 1). Many studies support this common pharmacophore model (60); however, other studies evaluated the complex of epothilone A with tubulin polymerized in zinc-stabilized sheets by electron crystallography and demonstrated that epothilone A and paclitaxel interact in substantially different ways within the same binding pocket in β-tubulin (61). Such differences in the binding modes of β-tubulin may provide a possible explanation of how the epothilones maintain their high antimitotic activity in cell lines that are resistant to taxanes because of point mutations in the β-tubulin subunit (62). Another unique feature of the epothilones is that unlike docetaxel and paclitaxel, epothilones maintain their cytotoxic performance even in cells overexpressing Pgp.

Currently there are five epothilone or epothilone-derived compounds undergoing clinical evaluation. One of the most advanced of these compounds is epothilone B (Patupilone; EP0-906). This drug is being evaluated by Novartis and is in phase III clinical trials for a number of cancer types (63). Similarly, lactam analog (Ixabepilone; BMS-247550) of epothilone B is in phase III trial targeting breast cancer being developed by Bristol-Myers Squibb (BMS).

In addition to epothilones A and B, epothilone D has been isolated as a minor component of fermentation of myxobacteria (64). This compound revealed a greater therapeutic index as a chemotherapeutic agent, compared with epothilones B (65); however, severe side effects including central nervous system toxicities were reported from clinical trials (66).

Toxicity and side effects

Although MTAs have proven to be highly effective for the treatment of cancer, the adverse side effects associated with their long- and short-term use, affect their applicability in the clinical setting. DNA damage and apoptosis are the main causes of drug-induced cytotoxicity (67). Neurologic and hematologic side effects are the main and often dose-limiting toxicities, but several other side effects also occur during the treatment with each individual drug (68).

Neurotoxicity such as peripheral neuropathy, which is characterized by the loss of deep tendon reflex at the ankle, numbness, and motor weakness, is very common. Cranial neuropathy may also occur after treatment with the Vinca alkaloids and taxanes and may cause different symptoms such as jaw pain and vocal cord dysfunction. Another common symptom is autonomic neuropathy causing constipation, abdominal cramping, and urinary retention. In addition, headache, dizziness, and mental depression are other symptom of neurotoxicity. Neurotoxicity typically occurs after prolonged treatment with MTAs. The inhibition of axonal microtubules, which are crucial for axonal transport in neurons, is at least in part the reason for such toxicity (29).

Hematologic toxicity is another major side effect as a result of treatment with MTAs, and is referred to as myelosuppression (68). Severe neutropenia, in particular, occurs early after treatment. Myelosuppression may result from the inhibition of the rapidly dividing hematopoietic cells. Nausea, vomiting, and diarrhea are other common symptoms besides the neurologic and hematologic toxicities (ref. 68; Table 1).

Drug resistance

Drug resistance refers to the status of poor responsiveness of tumor cells to chemotherapeutic drugs. Multiple mechanisms of drug resistance in cancer have been described, but typically involve efflux mechanisms and membrane-associated changes to prevent drug accumulation within tumor cells (4). Acquired drug resistance may be a result of the cellular response to MTAs, which modulate by adaptive changes of the cell. This is a common theme of anticancer drugs. Otherwise, because of the presence of specific resistance proteins, cells may be inherently protected from the antiproliferative effects of growth inhibitors even upon first exposure (4). Drug efflux by ATP-binding cassette (ABC) transporters such as Pgp, is one of the most frequent mechanisms of drug resistance encountered in cancer cells (4). Although Pgp-mediated drug efflux is a major resistance mechanism for taxol, some of the newer MTAs have been found to be less vulnerable to Pgp action, and thus may offer more benefits over taxol (4). Concrete efforts to reveal the specific resistance mechanisms to MTAs have been investigated over the last several years. Three major mechanisms have been proposed: first, overexpression of specific tubulin isotypes, especially of βIII-tubulin; second, inherit tubulin mutations that may affect microtubule stability and dynamics; third, the emergence of tubulin mutations, which lead to impaired ligand binding. Several groups have reported that the overexpression of class III β-tubulin is the cause of taxol resistance in cancer cell lines and human tumors as well (32). In addition, in vitro studies reported that microtubules that assemble solely from αβIII dimers exhibit different assembly properties and significantly enhanced dynamicity over microtubules that assemble from αβII or αβIV dimers or unfractionated tubulin (63). In paclitaxel resistance studies, taxol-resistant A549 lung carcinoma cells, which overexpress βIII-tubulin, exhibited inherently increased microtubule dynamics. This phenomenon does not allow the cells to proceed from metaphase to anaphase (4). However, upon treatment with low doses of taxol, the normal cell-cycle progression is restored. These results support the theory that these cells are dependent on taxol to normalize their microtubule dynamics and to successfully proceed through mitosis. Another study showed the effects of βIII-tubulin overexpression on microtubule dynamics and taxol sensitivity by inducing either βI- or βIII-tubulin overexpression in CHO cells (51). Cells that harbor specific tubulin mutations may be resistant to MTAs as a result of microtubule stability having been reduced and/or increased microtubule dynamics. This concept is supported by the observation that these cells often become dependent on the selecting agent and are hypersensitive to microtubule-destabilizing agents, thus indicating inherently reduced microtubule stability (69). None of these mechanisms, however, have been proven to be the cause of drug resistance in clinical settings (70). The importance of β-tubulin mutations in the clinical resistance to taxol in patients with non–small cell lung cancer by DNA sequence of the β-tubulin gene was proposed (71). However, these findings were highly controversial and could not be confirmed in a number of subsequent prospective studies. Berrieman and colleagues investigated the possible role of β-tubulin mutations in the resistance to cancer chemotherapy and concluded that β-tubulin mutations in clinical samples are rare and unlikely to contribute to drug resistance (72). It should be noted that β-tubulin pseudogenes belong to a family of homologous DNA sequences that need further investigation (73).

Enhanced DNA repair by O6-alkylguanine DNA alkyltransferase has been known as a common mechanism of resistance to drugs (74). Another mechanism of chemoresistance has been proposed by which the alteration in death-inducing signaling pathway may result in inhibition or prevention of apoptosis (75). Overexpression of antiapoptotic genes such as Bcl-2, besides mutation or reduced expression of proapoptotic genes such as p53, has been shown to induce cancer cells to become resistant to anticancer treatment. The way that Pgp can protect tumor cells from apoptosis has recently been discussed in view of observations that inhibition of Pgp by chemosensitizing agents can restore the normal apoptotic cascade in cells with defective signaling pathways (75).

Overall, there are multiple, complex and, possibly, interrelated mechanisms of drug resistance. These mechanisms are often found in combination, which can further complicate the chemotherapy of tumors (Table 1).