Tumour cell responses to new fibroblast growth factor receptor tyrosine kinase inhibitors and identification of a gatekeeper mutation in FGFR3 as a mechanism of acquired resistance.

Fibroblast growth factor receptors (FGFRs) can act as driving oncoproteins in certain cancers, making them attractive drug targets. Here we have characterized tumour cell responses to two new inhibitors of FGFR1-3, AZ12908010 and the clinical candidate AZD4547, making comparisons with the well-characterized FGFR inhibitor PD173074. In a panel of 16 human tumour cell lines, the anti-proliferative activity of AZ12908010 or AZD4547 was strongly linked to the presence of deregulated FGFR signalling, indicating that addiction to deregulated FGFRs provides a therapeutic opportunity for selective intervention. Acquired resistance to targeted tyrosine kinase inhibitors is a growing problem in the clinic but has not yet been explored for FGFR inhibitors. To assess how FGFR-dependent tumour cells adapt to long-term FGFR inhibition, we generated a derivative of the KMS-11 myeloma cell line (FGFR(Y373C)) with acquired resistance to AZ12908010 (KMS-11R cells). Basal phosphorylated FGFR and FGFR-dependent downstream signalling were constitutively elevated and refractory to drug in KMS-11R cells. Sequencing of FGFR3 in KMS-11R cells revealed the presence of a heterozygous mutation at the gatekeeper residue, encoding FGFR3(V555M); consistent with this, KMS-11R cells were cross-resistant to AZD4547 and PD173074. These results define the selectivity and efficacy of two new FGFR inhibitors and identify a secondary gatekeeper mutation as a mechanism of acquired resistance to FGFR inhibitors that should be anticipated as clinical evaluation proceeds.

Tumour cell survival signalling by the ERK1/2 pathway.

Several advances in recent years have focused increasing attention on the role of the RAF-MEK-ERK1/2 pathway in promoting cell survival. The demonstration that BRAF is a human oncogene mutated at high frequency in melanoma, thyroid and colon cancer has provided a pathophysiological context, whilst the description of potent and highly selective inhibitors of BRAF or MEK has allowed a more informed and rational intervention in both normal and tumour cells. In addition, separate studies have uncovered new mechanisms by which the ERK1/2 pathway can control the activity or abundance of members of the BCL-2 protein family to promote cell survival. It is now apparent that various oncogenes co-opt ERK1/2 signalling to de-regulate these BCL-2 proteins and this contributes to, and even underpins, survival signalling in some tumours. New oncogene-targeted therapies allow direct or indirect inhibition of ERK1/2 signalling and can cause quite striking tumour cell death. In other cases, inhibition of the ERK1/2 pathway may be more effective in combination with other conventional and novel therapeutics. Here, we review recent advances in our understanding of how the ERK1/2 pathway regulates BCL-2 proteins to promote survival, how this is de-regulated in tumour cells and the opportunities this might afford with the use of new targeted therapies.

Colorectal cancer cells with the BRAF(V600E) mutation are addicted to the ERK1/2 pathway for growth factor-independent survival and repression of BIM.

The RAF-mitogen-activated protein kinase kinase 1/2-extracellular signal-regulated kinase 1/2 (RAF-MEK1/2-ERK1/2) pathway is activated in many human tumours and can protect cells against growth factor deprivation; however, most such studies have relied upon overexpression of RAF or MEK constructs that are not found in tumours. Here we show that expression of the endogenous BRAF(V600E) allele in mouse embryonic fibroblasts from conditional knock-in transgenic mice activates ERK1/2, represses the BH3-only protein BIM and protects cells from growth factor withdrawal. Human colorectal cancer (CRC) cell lines harbouring BRAF(V600E) are growth factor independent for the activation of ERK1/2 and survival. However, treatment with the MEK1/2 inhibitors U0126, PD184352 or the novel clinical candidate AZD6244 (ARRY-142886) overcomes growth factor independence, causing CRC cell death. BIM is de-phosphorylated and upregulated following MEK1/2 inhibition in all CRC cell lines studied and knockdown of BIM reduces cell death, indicating that repression of BIM is a major part of the ability of BRAF(V600E) to confer growth factor-independent survival. We conclude that a single endogenous BRAF(V600E) allele is sufficient to repress BIM and prevent death arising from growth factor withdrawal, and CRC cells with BRAF(V600E) mutations are addicted to the ERK1/2 pathway for repression of BIM and growth factor-independent survival.

Regulation of cell cycle re-entry by growth, survival and stress signalling pathways.

The mitogen-activated and stress-activated protein kinases transduce signals from plasma membrane signalling machinery into the nucleus to modulate gene expression. By regulating the genomic response to environmental cues (growth factors, stresses) these pathways determine whether a cell re-enters the cell cycle, undergoes cell cycle arrest, senescence or apoptosis. We are particularly interested in how these pathways integrate with each other, and interact with the cell cycle machinery to achieve these discrete biological responses.

Sustained MAP kinase activation is required for the expression of cyclin D1, p21Cip1 and a subset of AP-1 proteins in CCL39 cells.

In CCL39 cells thrombin is a potent growth factor which requires sustained activation of mitogen activated protein kinases (MAPKs) to promote DNA synthesis. Some of the effects of thrombin can be mimicked by peptides based on the new amino terminus of the cleaved receptor; however, these thrombin receptor peptides (TRPs) fail to induce sustained activation of MAPK or DNA synthesis. We have used thrombin, TRP-7 and other agonists which elicit sustained or transient MAPK activation to identify immediate-early and delayed-early genes which are only expressed under conditions of sustained MAPK activation focusing on cyclin D1, p21CiP1 and the AP-1 transcription factor. Of the stimuli tested only FBS and thrombin were able to stimulate a sustained activation of MAPK, expression of cyclin D1, p21Cip1 and cell cycle re-entry. The expression of cyclin D1 was strongly, though not completely, inhibited by the MEK1 inhibitor PD098059. Thrombin stimulated a rapid but transient accumulation of c-Fos whereas the expression of Fra-1, Fra-2, c-Jun and JunB was sustained throughout the G1 phase of the cell cycle. We focussed our analysis on c-Fos (typical of AP-1 genes which are expressed rapidly and transiently) and Fra-1 and JunB (typical of AP-1 genes expressed after a delay but in a sustained manner). The expression of c-Fos, Fra-1 and JunB was dependent upon the activation of MAPK since these responses were inhibited by PD098059. However, a comparison of responses to FBS, thrombin, TRPs, LPA and EGF revealed that Fra-1 and JunB expression required sustained activation of MAPK whereas c-Fos expression was strongly induced even by non-mitogenic stimuli which elicited only transient MAPK activation. The expression of c-Fos (in response to thrombin, TRP or LPA) or Fra-1, JunB and cyclin D1 (thrombin only) was also inhibited by pertussis toxin. This suggests that both early and late AP-1 gene expression is regulated by the same Gi-mediated, MEK-dependent MAPK signalling pathway but that expression of late AP-1 genes and cyclin D1 requires that this pathway be persistently activated. The results suggest that the duration of receptor signalling and therefore MAPK activation is a key determinant of qualitative changes in gene expression during cell cycle re-entry.

The relationship between intrinsic thymidylate synthase expression and sensitivity to THYMITAQ in human leukaemia and colorectal carcinoma cell lines.

Thymidylate synthase (TS) expression has been characterized for a panel of eight human colorectal carcinoma and five human leukaemia cell lines, to relate differences in intrinsic TS activity, protein and mRNA levels to growth inhibition caused by continuous exposure to THYMITAQ, a specific non-classical antifolate TS inhibitor. Although a 20-fold variation in sensitivity to THYMITAQ was found within the colorectal cell line panel (IC50 0.12-2.7 microM), sensitivity was not related to TS activity, TS protein or TS mRNA levels. For the leukaemic cell lines, only a twofold range in sensitivity to THYMITAQ was observed (IC50 0.87-2.3 microM), and this did not correlate with TS activity, TS protein or TS mRNA levels. Across all of the cell lines, TS activity was linearly related to TS protein levels (r2 = 0.87, P < 0.0001). However, for both the colorectal and leukaemia cell line panels, no relationship was found between TS mRNA/18S rRNA ratios and either TS activity or TS protein, consistent with the importance of post-transcriptional mechanisms in regulating TS activity. Two of the colorectal cell lines (BE and HCT116) and one of the human leukaemic cell lines (HL60), were intrinsically resistant to THYMITAQ (IC50 > 2 microM) in the absence of TS overexpression, suggesting that, subsequent to TS inhibition, events such as DNA repair and tolerance to apoptotic stimuli are also important determinants of sensitivity to THYMITAQ.

Phase I and pharmacokinetic study of a water-soluble etoposide prodrug, etoposide phosphate (BMY-40481).

Etoposide phosphate is a water-soluble prodrug of etoposide. A phase I and pharmacokinetic study has been performed over the dose range 25-110 mg/m2/day for 5 days (etoposide equivalent doses). The maximum tolerated dose (MTD) was 110 mg/m2/day for 5 days every 3 weeks and the dose-limiting toxicity was neutropenia. Other toxicities were mild, with the exception of 2 patients who displayed significant hypersensitivity reactions. The etoposide phosphate:etoposide area under the plasma concentration versus time curve (AUC) ratio was < 1% and the pharmacokinetic parameters for etoposide were within previously reported ranges. Pharmacodynamic analyses demonstrated that etoposide AUC and baseline white blood cell count were significant determinants of leucopenia (model r2 = 0.51).

Clinical pharmacokinetic and pharmacodynamic studies with the nonclassical antifolate thymidylate synthase inhibitor 3, 4-dihydro-2-amino-6-methyl-4-oxo-5-(4-pyridylthio)-quinazolone dihydrochloride (AG337) given by 24-hour continuous intravenous infusion.

3,4-Dihydro-2-amino-6-methyl-4-oxo-5-(4-pyridylthio)-quinazolon e dihydrochloride (AG337) is a nonclassical inhibitor of thymidylate synthase (TS) designed to avoid potential resistance mechanisms that can limit the activity of classical antifolate antimetabolites. A clinical pharmacokinetic and pharmacodynamic study of AG337 given as a 24-h i.v. infusion was performed. Thirteen patients received 27 courses over the dose range 75-1350 mg/m2. Plasma AG337 concentrations were achieved which, in preclinical models, were associated with antitumor effects. AG337 clearance was saturable, and the pharmacokinetics of the drug at doses above 300 mg/m2 was best described by a one-compartment model with saturable elimination (median Km = 6.5 microgram/ml; range, 4.1-13 microgram/ml; median Vmax = 2.0 microgram/ml/h/m2; range, 0.96-5.6 microgram/ml/h/m2). Following the end of the infusion, AG337 was cleared rapidly (t1/2, 53-193 min), and levels were less than 0.2 microgram/ml in all patients by 48 h. Plasma protein binding was 96-98%, and the urinary excretion of AG337 as unchanged drug did not exceed 30% of the dose administered. Measurements of plasma deoxyuridine (dUrd) concentrations showed that doses of 600 mg/m2 and above of AG337 produced a consistent elevation in plasma dUrd levels (60-290%), suggesting that TS inhibition was being achieved in patients. However, in all cases dUrd concentrations had returned to pretreatment levels 24 h after the end of the infusion, suggesting that TS inhibition was not maintained. Local toxicity, probably due to the infusate pH, was the only significant adverse effect observed. These studies have shown that cytotoxic AG337 plasma concentrations can be readily achieved without acute toxicity and that these concentrations are associated with elevations in plasma dUrd levels. The lack of prolonged dUrd elevations indicates that extended administration should be explored using central line or p.o. administration to avoid local toxicity.

Pharmacokinetics and pharmacodynamics of prolonged oral etoposide in women with metastatic breast cancer.

The pharmacokinetics and pharmacodynamics of prolonged oral etoposide chemotherapy were investigated in 15 women with metastatic breast cancer who received oral etoposide 100 mg as a single daily dose for up to 15 days. There was considerable interpatient variability in the day 1 pharmacokinetic parameters: area under the plasma concentration time curve (AUC) (0-24 h) 1.95 +/- 0.87 mg/ml per min (mean +/- SD), apparent oral clearance 60.9 +/- 21.7 ml/min per 1.73 m2, peak plasma concentration 5.6 +/- 2.5 micrograms/ml, time to peak concentration 73 +/- 35 min and half-life 220 +/- 83 min. However, intrapatient variability in systemic exposure to etoposide was much less with repeated doses. The intrapatient coefficient of variation (CV) of AUC for day 8 relative to day 1 was 20% and for day 15 relative to day 1 was 15%, compared to the day 1 interpatient CV of 45%. Neutropenia was the principal toxicity. Day 1 pharmacokinetic parameters were related to the percentage decrease in absolute neutrophil count using the sigmoidal Emax equation. A good fit was found between day 1 AUC and neutrophil toxicity (R2 = 0.77). All patients who had a day 1 AUC > 2.0 mg/ml per min had WHO grade III or IV neutropenia. The predictive performance of the models for neutrophil toxicity was better for AUC (percentage mean predictive error 5%, percentage root mean square error 18.1%) than apparent oral clearance, peak plasma concentration, or daily dose (mg/m2). A limited sampling strategy was developed to predict AUC using a linear regression model incorporating a patient effect. Data sets were divided into training and test sets. The AUC could be estimated using a model utilizing plasma etoposide concentration at only two time points, 4 h and 6 h after oral dosing (R2 = 98.9%). The equation AUCpr = -0.376 + 0.631 x C4h + 0.336 x C6h was validated on the test set with a relative mean predictive error of -0.88% and relative root mean square error of 6.4%. These results suggest monitoring of AUC to predict subsequent myelosuppression as a strategy for future trials with oral etoposide.

Carboplatin pharmacokinetics in children: the development of a pediatric dosing formula. The United Kingdom Children's Cancer Study Group.

PURPOSE: The aim of this study was to define the pharmacokinetics of carboplatin in children and use the data to develop a pediatric dose formula. It was anticipated that renal function would be a major determinant of carboplatin disposition and the relationship between carboplatin clearance and glomerular filtration rate (GFR) was examined in detail. PATIENTS AND METHODS: Plasma carboplatin pharmacokinetics were measured as ultrafiltrable platinum in 22 patients (5 to 63 kg) following 200 to 1,000 mg/m2 of carboplatin. GFR was measured by the plasma clearance of chromium 51-edathamil (51Cr-EDTA). RESULTS: Carboplatin pharmacokinetics in children were best described in most patients (16 of 22) by a two-compartment model. The dose-normalized area under the plasma carboplatin concentration versus time curve (AUC) ranged from 3.1 to 9.6 mg/mL.min/400 mg/m2 and there was only a weak linear relationship between carboplatin dose and AUC (R2 = .31). There was a significant relationship between absolute carboplatin and 51Cr-EDTA clearances (R2 = .56), but the relationship was weaker (R2 = .28) when both clearances were normalized for body surface area. Carboplatin plasma clearance was predicted by the equation: clearance = GFR (mL/min) + 0.36 x body weight (BW; kg), and a modified form of the adult carboplatin dose formula is proposed: dose (mg) = target AUC x (GFR [mL/min] + [0.36 x BW(kg)]). Two further equations were developed that use the 51Cr-EDTA half-life (t1/2) to calculate the GFR and these may reduce errors resulting from inaccurate measurement of the volume of distribution for 51Cr-EDTA. In patients treated with single-agent carboplatin or carboplatin plus vincristine, there was a significant sigmoidal relationship between AUC and thrombocytopenia (R2 = .56). CONCLUSION: GFR-based carboplatin dosing in children should be feasible and will be evaluated prospectively.

Phase II trial of trimelamol in refractory ovarian cancer.

Trimelamol is an analogue of hexamethymelamine which exhibited activity against refractory ovarian cancer in phase I clinical trial. The dose limiting toxicity was leukopenia. In a phase II study, 42 patients with recurrent, or platinum-complex resistant, advanced ovarian cancer were treated using the dose schedule 800 mg m-2 i.v. daily for 3 days. There were one complete, three partial and five minor responses, objective response rate: 9.5%. The main toxicity observed was nausea and vomiting, myelosuppression was minor. The role of Trimelamol in the treatment of ovarian cancer remains to be defined, but its activity is limited in refractory disease.