Mechanism-based target identification and drug discovery in cancer research.

Cancer as a disease in the human population is becoming a larger health problem, and the medicines used as treatments have clear limitations. In the past 20 years, there has been a tremendous increase in our knowledge of the molecular mechanisms and pathophysiology of human cancer. Many of these mechanisms have been exploited as new targets for drug development in the hope that they will have greater antitumor activity with less toxicity to the patient than is seen with currently used medicines. The fruition of these efforts in the clinic is just now being realized with a few encouraging results.

CDC25: a component of the RAS-adenylate cyclase pathway in Saccharomyces cerevisiae.

The yeast Saccharomyces cerevisiae contains two functional homologues of the ras oncogene family, RAS1 and RAS2. These genes are required for growth, and all evidence indicates that this essential function is the activation of adenylate cyclase. In contrast, ras in mammalian cells does not appear to influence adenylate cyclase activity. To clarify the relation between ras function in yeast and in higher eukaryotes, and the role played by yeast RAS in growth control, it is necessary to identify functions acting upstream of RAS in the adenylate cyclase pathway. The evidence presented here indicates that CDC25, identified by conditional cell cycle arrest mutations, encodes such an upstream function.

Selective inhibition of ras-dependent transformation by a farnesyltransferase inhibitor.

To acquire transforming potential, the precursor of the Ras oncoprotein must undergo farnesylation of the cysteine residue located in a carboxyl-terminal tetrapeptide. Inhibitors of the enzyme that catalyzes this modification, farnesyl protein transferase (FPTase), have therefore been suggested as anticancer agents for tumors in which Ras contributes to transformation. The tetrapeptide analog L-731,735 is a potent and selective inhibitor of FPTase in vitro. A prodrug of this compound, L-731,734, inhibited Ras processing in cells transformed with v-ras. L-731,734 decreased the ability of v-ras-transformed cells to form colonies in soft agar but had no effect on the efficiency of colony formation of cells transformed by either the v-raf or v-mos oncogenes. The results demonstrate selective inhibition of ras-dependent cell transformation with a synthetic organic inhibitor of FPTase.

Transforming ras proteins accelerate hormone-induced maturation and stimulate cyclic AMP phosphodiesterase in Xenopus oocytes.

Transforming Harvey (Ha) ras oncogene products accelerated the time course of Xenopus oocyte maturation induced by insulin, insulinlike growth factor 1, or progesterone. The transforming constructs, [Val-12]Ha p21 and [Val-12, Thr-59]Ha p21, displayed equal potency and efficacy in their abilities to accelerate the growth peptide-induced response. Normal Ha p21 was only 60% as powerful and one-fifth as potent as the mutants containing valine in the 12 position. In contrast, two nontransforming constructs, [Val-12, Ala-35, Leu-36, Thr-59]Ha p21 and [Val-12, Thr-59]Ha(term-174) p21, had no effect on the time course of hormone-induced maturation. Effects of the transforming ras proteins on hormone-induced maturation correlated with their abilities to stimulate in vivo phosphodiesterase activity measured after microinjection of 200 microM cyclic [3H] AMP. When p21 injection followed 90 min of insulin treatment, there was no increase in phosphodiesterase activity over that measured after hormone treatment or p21 injection alone, but additive effects of p21 and insulin on enzyme activity were observed during the first 90 min of insulin treatment. Even though normal Ha p21 and transforming [Val-12, Thr-59]Ha p21 stimulated oocyte phosphodiesterase to equal levels when coinjected with substrate at the initiation of the in vivo assay, the transforming protein elicited a more sustained stimulation of enzyme activity. These results suggest that stimulation of a cyclic AMP phosphodiesterase activity associated with insulin-induced maturation is involved in the growth-promoting actions of ras oncogene products in Xenopus oocytes.

Dominant inhibitory Ras mutants selectively inhibit the activity of either cellular or oncogenic Ras.

Two dominant inhibitory Ras mutant proteins were analyzed by microinjection. One, [Asn-17]Ras, had a substitution in the putative Mg(2+)-binding site of Ha-Ras. The other, RAST, had a mutation in a yeast RAS protein that impaired its GTPase activity and increased its affinity for GAP. RAST also had a mutation that blocked its localization to the plasma membrane. In NIH 3T3 cells [Asn-17]Ras inhibited the function of normal Ras much more efficiently than that of oncogenic Ras. In contrast, RAST interfered with the transforming activity of oncogenic Ras more efficiently than that of normal Ras. These conclusions were based on two separate types of analysis. The inhibitory Ras mutant proteins were first microinjected into cells stably transformed either by oncogenic Ras or by high levels of expression of cellular Ras. Results obtained in stably transformed cells were then verified by coinjection of the inhibitory Ras mutant proteins together with transforming concentrations of either oncogenic or normal Ras protein. Whereas RAST was active in soluble form. [Asn-17]Ras required membrane localization for activity. Furthermore, mutations in the GAP/effector-binding domain reduced or eliminated the inhibitory activity of RAST but had no detectable effect on [Asn-17]Ras. These results are consistent with the possibility that [Asn-17]Ras functions by blocking the activation of endogenous Ras proteins, while RAST functions by blocking the ability of activated Ras to stimulate a downstream target within the cells. The properties of RAST suggest that interference with the GAP/effector-binding function of RAS represents a strategy for the preferential inactivation of oncogenic Ras in cells.

An adenylate cyclase from Saccharomyces cerevisiae that is stimulated by RAS proteins with effector mutations.

Conservative amino acid substitutions were introduced into the proposed effector regions of both mammalian Ha-ras (residues 32 to 40) and Saccharomyces cerevisiae RAS2 (residues 39 to 47) proteins. The RAS2[Ser 42] protein had reduced biological function in the yeast S. cerevisiae. A S. cerevisiae strain with a second-site suppressor mutation, SSR2-1, was isolated which could grow on nonfermentable carbon sources when the endogenous RAS2 protein was replaced by the RAS2[Ser 42] protein. The SSR2-1 mutation was mapped to the structural gene for adenylate cyclase (CYR1), and the gene containing SSR2-1 was cloned and sequenced. SSR2-1 corresponded to a point mutation that would create an amino acid substitution of a tyrosine residue for an aspartate residue at position 1547. The SSR2-1 gene encodes an adenylate cyclase that is dependent on ras proteins for activity, but is stimulated by Ha-ras and RAS2 mutant proteins that are unable to stimulate wild-type adenylate cyclase.

Inhibition of v-src-induced transformation by a GTPase-activating protein.

Previous work has shown that microinjection into cells of antibodies against p21ras blocks transformation by src, suggesting that oncogenic transformation by pp60v-src is dependent on p21ras. The activity of p21ras itself is regulated by its cyclic association with GDP-GTP, where p21ras-GTP is the active form and p21ras-GDP is the inactive form. A GTPase-activating protein (GAP) mediates the inactivation of p21ras by facilitating the conversion of the active p21ras-GTP to the inactive p21ras-GDP. This predicts that overexpression of GAP would inactivate p21ras and block transformation of cells by src. In this paper, we confirm this prediction. We report that overexpression of GAP in NIH 3T3 cells blocks transformation by pp60v-src but not by v-ras. Susceptibility to transformation by v-src is restored when GAP expression is lowered to levels comparable to that in control cells. These results support the suggestion that p21ras plays a central role in the signalling pathway used by pp60v-src.

Regulatory function of the Saccharomyces cerevisiae RAS C-terminus.

Activating mutations (valine 19 or leucine 68) were introduced into the Saccharomyces cerevisiae RAS1 and RAS2 genes. In addition, a deletion was introduced into the wild-type gene and into an activated RAS2 gene, removing the segment of the coding region for the unique C-terminal domain that lies between the N-terminal 174 residues and the penultimate 8-residue membrane attachment site. At low levels of expression, a dominant activated phenotype, characterized by low glycogen levels and poor sporulation efficiency, was observed for both full-length RAS1 and RAS2 variants having impaired GTP hydrolytic activity. Lethal CDC25 mutations were bypassed by the expression of mutant RAS1 or RAS2 proteins with activating amino acid substitutions, by expression of RAS2 proteins lacking the C-terminal domain, or by normal and oncogenic mammalian Harvey ras proteins. Biochemical measurements of adenylate cyclase in membrane preparations showed that the expression of RAS2 proteins lacking the C-terminal domain can restore adenylate cyclase activity to cdc25 membranes.

Ras GTPase-activating protein physically associates with mitogenically active phospholipids.

The physical interaction between GTPase-activating protein (GAP) and lipids has been characterized by two separate analyses. First, bacterially synthesized GAP molecules were found to associate with detergent-mixed micelles containing arachidonic but not with those containing arachidic acid. This association was detected by a faster elution time during molecular exclusion chromatography. Second, GAP molecules within a crude cellular lysate were specifically retained by a column on which certain lipids had been immobilized. The lipids able to retain GAP on such columns were identical to those which were shown previously to be most active in blocking GAP activity. The association between lipids and GAP was dependent upon magnesium ions. Lipids unable to inhibit GAP activity were also unable to physically associate with GAP. The tight association of GAP with these lipids was predicted by and helps to rationalize their ability to inhibit GAP activity.

Transformation by pp60src or stimulation of cells with epidermal growth factor induces the stable association of tyrosine-phosphorylated cellular proteins with GTPase-activating protein.

GTPase-activating protein (GAP) is a cytosolic protein that stimulates the rate of hydrolysis of GTP (GTP to GDP) bound to normal p21ras, but does not catalyze the hydrolysis of GTP bound to oncogenic, activated forms of the ras protein. Transformation of cells with v-src or activated transforming variants of c-src or stimulation of cells with epidermal growth factor resulted in the stable association of GAP with two tyrosine-phosphorylated cellular proteins of 64 kDa (p64) and 190 kDa (p190). Analysis of GAP immune complexes isolated from extracts of metabolically labeled src-transformed cells and epidermal growth factor-stimulated cells indicated that tyrosine phosphorylation of p64 and p190 appeared to be coincident with the stable association of these proteins with GAP. Quantitation of the amount of p64 associated with GAP in v-src-transformed cells, however, indicated that only 15 to 25% of tyrosine-phosphorylated p64 was found in complex with GAP. Mutations within the SH2 region of pp60src that render activated pp60src defective for transformation inhibited the efficient formation of complexes between GAP and the tyrosine-phosphorylated forms of p64 and p190. From these data, we suggest that tyrosine phosphorylation and stable association of p64 with GAP is an important step in mediating cellular signaling through the p21ras-GAP pathway.

Kinase-deficient Pak1 mutants inhibit Ras transformation of Rat-1 fibroblasts.

Among the mechanisms by which the Ras oncogene induces cellular transformation, Ras activates the mitogen-activated protein kinase (MAPK or ERK) cascade and a related cascade leading to activation of Jun kinase (JNK or SAPK). JNK is additionally regulated by the Ras-related G proteins Rac and Cdc42. Ras also regulates the actin cytoskeleton through an incompletely elucidated Rac-dependent mechanism. A candidate for the physiological effector for both JNK and actin regulation by Rac and Cdc42 is the serine/threonine kinase Pak (p65pak). We show here that expression of a catalytically inactive mutant Pak, Pak1(R299), inhibits Ras transformation of Rat-1 fibroblasts but not of NIH 3T3 cells. Typically, 90 to 95% fewer transformed colonies were observed in cotransfection assays with Rat-1 cells. Pak1(R299) did not inhibit transformation by the Raf oncogene, indicating that inhibition was specific for Ras. Furthermore, Rat-1 cell lines expressing Pak1(R299) were highly resistant to Ras transformation, while cells expressing wild-type Pak1 were efficiently transformed by Ras. Pak1(L83,L86,R299), a mutant that fails to bind either Rac or Cdc42, also inhibited Ras transformation. Rac and Ras activation of JNK was inhibited by Pak1(R299) but not by Pak1(L83,L86,R299). Ras activation of ERK was inhibited by both Pak1(R299) and Pak1(L83,L86,R299), while neither mutant inhibited Raf activation of ERK. These results suggest that Pak1 interacts with components essential for Ras transformation and that inhibition can be uncoupled from JNK but not ERK signaling.

Farnesyltransferase inhibition causes morphological reversion of ras-transformed cells by a complex mechanism that involves regulation of the actin cytoskeleton.

A potent and specific small molecule inhibitor of farnesyl-protein transferase, L-739,749, caused rapid morphological reversion and growth inhibition of ras-transformed fibroblasts (Rat1/ras cells). Morphological reversion occurred within 18 h of L-739,749 addition. The reverted phenotype was stable for several days in the absence of inhibitor before the transformed phenotype reappeared. Cell enlargement and actin stress fiber formation accompanied treatment of both Rat1/ras and normal Rat1 cells. Significantly, inhibition of Ras processing did not correlate with the initiation or maintenance of the reverted phenotype. While a single treatment with L-739,749 was sufficient to morphologically revert Rat1/ras cells, repetitive inhibitor treatment was required to significantly reduce cell growth rate. Thus, the effects of L-739,749 on transformed cell morphology and cytoskeletal actin organization could be separated from effects on cell growth, depending on whether exposure to a farnesyl-protein transferase inhibitor was transient or repetitive. In contrast, L-739,749 had no effect on the growth, morphology, or actin organization of v-raf-transformed cells. Taken together, the results suggest that the mechanism of morphological reversion is complex and may involve farnesylated proteins that control the organization of cytoskeletal actin.

Identification of amino acid residues required for Ras p21 target activation.

The Krev-1 gene has been shown to suppress ras-mediated transformation in vitro. Both ras and Krev-1 proteins have identical effector domains (ras residues 32 to 40), which are required for biological activity and for the interaction of Ras p21 with Ras GTPase-activating protein (GAP). In this study, five amino acid residues flanking the ras effector domain, which are not conserved with the Krev-1 protein, were shown to be required for normal protein-protein interactions and biological activity. The substitution of Krev-1 p21 residues 26, 27, 30, 31, and 45 with the corresponding amino acid residues from Ras p21 resulted in a Krev-1 protein which had ras function in both mammalian and yeast biological assays. Replacement of these residues in Ras p21 with the corresponding Krev-1 p21 amino acids resulted in ras proteins which were impaired biologically or reduced in their affinity for in vitro GAP binding. Evaluation of these mutant ras proteins have implications for Ras p21-GAP interactions in vivo.

A farnesyltransferase inhibitor induces tumor regression in transgenic mice harboring multiple oncogenic mutations by mediating alterations in both cell cycle control and apoptosis.

The farnesyltransferase inhibitor L-744,832 selectively blocks the transformed phenotype of cultured cells expressing a mutated H-ras gene and induces dramatic regression of mammary and salivary carcinomas in mouse mammary tumor virus (MMTV)-v-Ha-ras transgenic mice. To better understand how the farnesyltransferase inhibitors might be used in the treatment of human tumors, we have further explored the mechanisms by which L-744,832 induces tumor regression in a variety of transgenic mouse tumor models. We assessed whether L-744,832 induces apoptosis or alterations in cell cycle distribution and found that the tumor regression in MMTV-v-Ha-ras mice could be attributed entirely to elevation of apoptosis levels. In contrast, treatment with doxorubicin, which induces apoptosis in many tumor types, had a minimal effect on apoptosis in these tumors and resulted in a less dramatic tumor response. To determine whether functional p53 is required for L-744,832-induced apoptosis and the resultant tumor regression, MMTV-v-Ha-ras mice were interbred with p53(-/-) mice. Tumors in ras/p53(-/-) mice treated with L-744,832 regressed as efficiently as MMTV-v-Ha-ras tumors, although this response was found to be mediated by both the induction of apoptosis and an increase in G1 with a corresponding decrease in the S-phase fraction. MMTV-v-Ha-ras mice were also interbred with MMTV-c-myc mice to determine whether ras/myc tumors, which possess high levels of spontaneous apoptosis, have the potential to regress through a further increase in apoptosis levels. The ras/myc tumors were found to respond nearly as efficiently to L-744,832 treatment as the MMTV-v-Ha-ras tumors, although no induction of apoptosis was observed. Rather, the tumor regression in the ras/myc mice was found to be mediated by a large reduction in the S-phase fraction. In contrast, treatment of transgenic mice harboring an activated MMTV-c-neu gene did not result in tumor regression. These results demonstrate that a farnesyltransferase inhibitor can induce regression of v-Ha-ras-bearing tumors by multiple mechanisms, including the activation of a suppressed apoptotic pathway, which is largely p53 independent, or by cell cycle alterations, depending upon the presence of various other oncogenic genetic alterations.

A peptidomimetic inhibitor of farnesyl:protein transferase blocks the anchorage-dependent and -independent growth of human tumor cell lines.

Farnesyl protein transferase (FPTase) catalyzes the first of a series of posttranslational modifications of Ras required for full biological activity. Peptidomimetic inhibitors of FPTase have been designed that selectively block farnesylation in vivo and in vitro. These inhibitors prevent Ras processing and membrane localization and are effective in reversing the transformed phenotype of Rat1-v-ras cells but not that of cells transformed by v-raf or v-mos. We have tested the effect of the FPTase inhibitor L-744,832 (FTI) on the anchorage-dependent and -independent growth of human tumor cell lines. The growth of over 70% of all tumor cell lines tested was inhibited by 2-20 microM of the FTI, whereas the anchorage-dependent growth of nontransformed epithelial cells was less sensitive to the effects of the compound. No correlation was observed between response to drug and the origin of the tumor cell or whether it contained mutationally activated ras. In fact, cell lines with wild-type ras and active protein tyrosine kinases in which the transformed phenotype may depend on upstream activation of the ras pathway were especially sensitive to the drug. To define the important targets of FTI action, the mechanism of cellular drug resistance was examined. It was not a function of altered drug accumulation or of FPTase insensitivity since, in all cell lines tested, FPTase activity was readily inhibited within 1 h of treatment with the inhibitor. Furthermore, the general pattern of inhibition of cellular protein farnesylation and the specific inhibition of lamin B processing were the same in sensitive and resistant cells. In addition, functional activation of Ras was inhibited to the same degree in sensitive and resistant cell lines. However, the FTI inhibited the epidermal growth factor-induced activation of mitogen-activated protein kinases in sensitive cells but not in two resistant cell lines. These data suggest that the drug does inhibit ras function and that resistance in some cells is associated with the presence of Ras-independent pathways for mitogen-activated protein kinase activation by tyrosine kinases. We conclude that FPTase inhibitors are potent antitumor agents with activity against many types of human cancer cell lines, including those with wild-type ras.

Antitumor effect of a farnesyl protein transferase inhibitor in mammary and lymphoid tumors overexpressing N-ras in transgenic mice.

We tested the antineoplastic effect of the farnesyltransferase inhibitor L-744,832 in mammary and lymphoid tumors overexpressing the N-ras proto-oncogene in transgenic mice. Mice bearing mammary tumors were randomly assigned to receive daily 40 mg/kg s.c. injections of this compound (experimental group, n = 6) or vehicle (control group, n = 6) per day for 5.5 weeks. Treatment with the compound significantly reduced the mammary tumor mean growth rate in the experimental group (-0.7 mm3/day), as compared with the control group (+28.2 mm3/day; P < 0.001). There was a significant difference in lymphoma incidence at the end of the treatment between the experimental (0 of 6) and the control (3 of 6) groups (P < 0.05). Therefore, this compound is effective in treating in vivo mammary carcinomas and lymphomas in which an activated N-Ras pathway drives tumorigenesis. The number of apoptotic figures in mammary tumors was significantly higher (P = 0.04) in the experimental (14.7 +/- 8.1) than it was in the control (5.7 +/- 3.5) group, indicating that apoptotic induction could contribute to the mechanism of antitumor activity of this compound. We analyzed the level of processing of N-Ras and H-Ras after immunoprecipitation and Western blotting of protein extracts obtained from mammary tumors treated with L-744,832 or vehicle, either in vivo or in vitro (after primary culture of the same tumors), and from several in vitro treated control cell lines. In all compound-treated mammary tumors and cell lines, H-Ras was mostly unprocessed (more so after in vitro than after in vivo treatment), whereas N-Ras remained mostly processed. Both H-Ras and N-Ras remained fully processed in all vehicle-treated samples. These findings are consistent with a less intense antineoplastic effect of the treatment with the compound in our N-ras model than the effect previously reported for the same compound in H-ras transgenics. In addition, the finding that, in compound-treated mammary tumors, the N-Ras protein remains mainly processed suggests that, in our model, other proteins in addition to Ras may be a target for the compound. Our results and the previous findings of frequent N-ras activation in human hematopoietic malignancies support a role for L-744,832 in the treatment of lymphomas and of mammary carcinomas with an activated N-Ras pathway, as well as the testing of a farnesyl protein transferase inhibitor in humans to establish its clinical relevance.

Mouse mammary tumor virus-Ki-rasB transgenic mice develop mammary carcinomas that can be growth-inhibited by a farnesyl:protein transferase inhibitor.

For Ras oncoproteins to transform mammalian cells, they must be posttranslationally modified with a farnesyl group in a reaction catalyzed by the enzyme farnesyl:protein transferase (FPTase). Inhibitors of FPTase have therefore been developed as potential anticancer agents. These compounds reverse many of the malignant phenotypes of Ras-transformed cells in culture and inhibit the growth of tumor xenografts in nude mice. Furthermore, the FPTase inhibitor (FTI) L-744,832 causes tumor regression in mouse mammary tumor virus (MMTV)-v-Ha-ras transgenic mice and tumor stasis in MMTV-N-ras mice. Although these data support the further development of FTIs, it should be noted that Ki-ras is the ras gene most frequently mutated in human cancers. Moreover, Ki-RasB binds more tightly to FPTase than either Ha- or N-Ras, and thus higher concentrations of FTIs that are competitive with the protein substrate may be required to inhibit Ki-Ras processing. Given the unique biochemical and biological features of Ki-RasB, it is important to evaluate the efficacy of FTIs or any other modulator of oncogenic Ras function in model systems expressing this Ras oncoprotein. We have developed strains of transgenic mice carrying the human Ki-rasB cDNA with an activating mutation (G12V) under the control of the MMTV enhancer/promoter. The predominant pathological feature that develops in these mice is the stochastic appearance of mammary adenocarcinomas. High levels of the Ki-rasB transgene RNA are detected in these tumors. Treatment of MMTV-Ki-rasB mice with L-744,832 caused inhibition of tumor growth in the absence of systemic toxicity. Although FPTase activity was inhibited in tumors from the treated mice, unprocessed Ki-RasB was not detected. These results demonstrate the utility of the MMTV-Ki-rasB transgenic mice for testing potential anticancer agents. Additionally, the data suggest that although the FTI L-744,832 can inhibit tumor growth in this model, Ki-Ras may not be the sole mediator of the biological effects of the FTI.

Dominant inhibitory Ras mutants demonstrate the requirement for Ras activity in the action of tyrosine kinase oncogenes.

Two inhibitory Ras mutant proteins [(Asn 17) Ras and RAST] were microinjected into NIH3T3 cells in order to compare their inhibitory activity with that of a neutralizing anti-ras antibody. Both mutants were able to block efficiently the mitogenic effects of serum added to quiescent NIH3T3 cells. Furthermore, each of the inhibitors blocked cell cycle progression at the same point as the injected anti-ras antibody, just prior to the initiation of a new round of DNA synthesis. Finally, as with the injected anti-ras antibody, each of the inhibitors was efficiently able to block proliferation and reverse the transformed morphology of cells transformed by tyrosine kinase oncogenes, while cells transformed by serine kinase oncogenes were unaffected. Therefore, results with all three reagents clearly indicate that cellular Ras activity is required in the late G1 phase of the cell cycle and is essential for the maintenance of the transformed phenotype induced by tyrosine but not serine kinase oncogenes. These studies demonstrate the utility of dominant inhibitory mutants as a means of interfering with the activity of cellular oncogenes.

Dissociation of RNA synthesis from the calcium requirement for serum-increased ornithine decarboxylase activity in rat glioma cells.

When C6-2B rat glioma cells were stimulated with calf serum in the presence of calcium, ornithine decarboxylase activity increased maximally in 6-8 h after an initial 2-3 h lag period wherein RNA synthesis occurred. The increase of ornithine decarboxylase activity in serum-stimulated C6-2B cells was prevented by the calcium chelator EGTA, but EGTA had no effect upon RNA synthesis as judged by [3H]uridine incorporation into RNA. In addition, the calcium requirement for increased ornithine decarboxylase activity was temporally distal to the lag period. EGTA appeared to inhibit the synthesis of ornithine decarboxylase, because the half-life values of ornithine decarboxylase activity were similar (37-47 min) in the presence of EGTA or protein synthesis inhibitors such as cycloheximide or emetine. Also, calcium readdition rapidly reversed EGTA inhibition of ornithine decarboxylase activity by a mechanism which could be blocked by cycloheximide.