Gene expression predicts differential capecitabine metabolism, impacting on both pharmacokinetics and antitumour activity.

Capecitabine is converted into 5'-deoxy-5-fluorocytidine (5'DFCR), 5'-deoxy-5-fluorouridine (5'DFUR) and 5-fluorouracil (5-FU) by CES1 and 2, CDD, and TP, in both liver and tumour. 5-FU is catabolised by DPD. Gene expression analysis of these enzymes was undertaken in fresh human hepatocytes, mouse liver, colorectal cancer cell lines and xenografts. Cell lines with low CDD expression (<1.5) had 5'DFCR/5'DFUR cytotoxicity ratios>2 and cell lines with TP/DPD<0.6 had 5'DFUR IC50>50 microM (SRB assay). A pharmacokinetic/pharmacodynamic study in nude mice bearing HCT 116 xenografts and treated with capecitabine by oral gavage assessed pharmacokinetic, gene expression and antitumour activity. Low liver CDD correlated with high 5'DFCR plasma concentrations in mice. CDD expression was approximately 100-fold higher in fresh human hepatocytes than mouse liver, explaining the higher plasma 5'DFUR concentrations reported previously in humans. Tumour 5-FU concentration correlated with TP/DPD and with tumour response. These studies identify the potential utility of gene expression analysis and drug monitoring in tumour in patients.

Time-dependent cytotoxicity induced by SJG-136 (NSC 694501): influence of the rate of interstrand cross-link formation on DNA damage signaling.

SJG-136 is a new pyrrolobenzodiazepine dimer inducing time-dependent cytotoxicity. HCT 116 cells were exposed to 50 nmol/L of SJG-136 for 1 hour or 1 nmol/L of SJG-136 for 24 hours to achieve similar levels of interstrand cross-links (ICL). The short exposure led to a rapid formation of ICLs (1 hour), early H2AX foci formation (4 hours), prominent S phase arrest, and greater phosphorylation of Nbs1 (on serine 343) and Chk1 (on serine 317) than a 24-hour exposure. The prolonged exposure at low concentrations of SJG-136 induced a gradual formation of ICLs (up to 24 hours) which was associated with a limited S phase arrest and delayed Nbs1 phosphorylation. Prolonged exposure was also associated with a reduced phosphorylation of p53 on serines 15 and 20, a limited and delayed phosphorylation on serine 392, and a less prominent increase in p21 levels. These data suggest that the 24-hour exposure to a low concentration of SJG-136 led to delayed and reduced DNA damage signaling compared with a higher concentration of SJG-136 for 1 hour, resulting in greater cytotoxicity and contributing to the time-dependent cytotoxic effect of SJG-136.

Validation of real-time reverse-transcription-polymerase chain reaction for quantification of capecitabine-metabolizing enzymes.

Capecitabine is an oral fluoropyrimidine carbamate activated sequentially in both liver and tumor tissues by carboxylesterases, cytidine deaminase, and thymidine phosphorylase. 5-Fluorouracil is inactivated by dihydropyrimidine dehydrogenase and targets thymidylate synthase. Here we report the validation of the real-time polymerase chain reaction assay for the quantification of the transcripts of the different enzymes involved in capecitabine activation. The method is specific, sensitive, and linear over 2-3 logs of RNA input. It is reproducible with less than 5% intraday variability and less than 10% interday variability. Five reference genes were tested for normalization. POLR2A was selected since it reduced variability between samples, demonstrated levels of expression similar to those of the genes of interest, and its expression was not modified by capecitabine treatment in samples from preclinical studies. The method was robust as the gene expression profiles from six colon cancer cell lines obtained by this method were similar to microarray data. Finally, the method was able to detect changes in gene expression in xenograft tumors treated with capecitabine. It could therefore constitute the method of choice for future correlative studies in patients receiving capecitabine.

Enhanced oxaliplatin-induced apoptosis following antisense Bcl-xl down-regulation is p53 and Bax dependent: Genetic evidence for specificity of the antisense effect.

INTRODUCTION: Oxaliplatin, licensed for colorectal cancer chemotherapy, damages DNA by generating intrastrand and interstrand cross-links and can induce apoptosis via a Bax-dependent pathway. Bcl-xl, an antiapoptotic Bcl-2 family member, regulates apoptosis and chemoresistance in several cancer models. Bcl-xl expression correlates with invasiveness in primary colorectal cancer. Bcl-xl may therefore represent a therapeutic target in this disease. We used the mismatch repair-deficient HCT116 colorectal cancer cell line (wild-type HCT116) and p53 null, Bax null, or p21/WAF1 null derivatives to identify genetic determinants of the response to oxaliplatin and tested the hypothesis that antisense-mediated Bcl-xl down-regulation would enhance the apoptotic response in a p53- or Bax-dependent manner. RESULTS: At clinically relevant concentrations, oxaliplatin induced p53 and p53-dependent Bax, Bcl-xl, and p21/WAF1 protein accumulation. A minor degree of apoptosis resulted via a p53- and Bax-dependent pathway. The major response was a transient mixed G(1) and G(2) growth arrest. The G(1) arrest was p53 and p21/WAF1 dependent. A 2'-O-ribose methoxyethyl phosphorothioate antisense oligonucleotide reduced Bcl-xl protein expression by approximately 90% in HCT116 (Bcl-xl knockdown). Missense controls were inactive. Prior Bcl-xl knockdown enhanced the apoptotic and the global cytotoxic effect of oxaliplatin. The extent of enhancement of apoptosis depended on the integrity of the p53- and Bax-mediated apoptotic pathway, providing genetic evidence that the desired proapoptotic antisense effect is due to specific down-regulation of the Bcl-xl target. CONCLUSIONS: The combination of oxaliplatin and Bcl-xl antisense merits testing in models of colorectal cancer in vivo.

Glucuronidation as a mechanism of intrinsic drug resistance in human colon cancer: reversal of resistance by food additives.

Colon cancer exhibits inherent insensitivity to chemotherapy by mechanisms that are poorly characterized. We have shown that human colon cancer cells are efficient in drug conjugation catalyzed by UDP-glucuronosyltransferases (UGTs) and now report on the role of glucuronidation in de novo resistance to two topoisomerase I inhibitors. Identification of the UGT responsible for glucuronidation of SN-38 and the anthraquinone NU/ICRF 505 was achieved by first using a panel of human cDNA-expressed isozymes to measure conjugating activity. HT29 colon cancer cells were then probed by reverse transcriptase-PCR, Western Blot analysis, and liquid chromatography with mass spectrometry for their profile and activity of UGT isozymes and screened for effective inhibitors of glucuronidation. Expression analysis was also conducted in colon cancer biopsies and paired adjacent normal colon specimens. UGT1A9 was identified as the isozyme catalyzing biotransformation of the two compounds in HT29 cells and propofol as an effective competitive inhibitor of this metabolism. Inhibition of glucuronidation resulted in up to a 5-fold enhancement in drug activity. The majority of colon cancer biopsies studies expressed UGT protein at levels greater than in HT29 cells but with marked interpatient variations and proficiently glucuronidated the two anticancer drugs. A range of UGT aglycones were capable of modulating glucuronidation in the biopies with octylgallate being 10-fold more potent (ID(50) 24 microM) than propofol. In a subset of tumors (33%), UGT protein levels and activity exceeded that of paired normal colon. Glucuronidation may represent a mechanism of intrinsic drug resistance in colon cancer open to modulation by a range of food additives and proprietary medicines.

Antisense Bcl-xl down-regulation switches the response to topoisomerase I inhibition from senescence to apoptosis in colorectal cancer cells, enhancing global cytotoxicity.

PURPOSE: To identify determinants of the effect of antisense-mediated Bcl-xl down-regulation (Bcl-xl knockdown) on the response of colorectal cancer cells to SN38, the active metabolite of irinotecan, a topoisomerase I inhibitor licensed for colorectal cancer chemotherapy. EXPERIMENTAL DESIGN: Using wild-type HCT116, p53 null, Bax null, or p21/WAF1 null isogenic derivatives, we measured expression of regulators of cellular response, and associated growth arrests or apoptosis, after SN38 treatment, with or without antisense-mediated Bcl-xl knockdown. RESULTS: A modified phosphorothioate antisense oligonucleotide (ISIS15999) reduced Bcl-xl protein expression by approximately 90%. SN38 induced p53, Bax, Bcl-xl, and p53-dependent p21/WAF1 protein accumulation. The Bax:Bcl-xl ratio changed little. In wild-type HCT116, but not in Bax null cells, Bcl-xl knockdown induced a shift in response from drug-induced senescence to apoptosis, and enhanced the global cytotoxicity of SN38. In p53 null or p21/WAF1 null cells marked apoptosis occurred after SN38 alone, and was additionally enhanced by Bcl-xl knockdown in p21/WAF1 null cells but not in p53 null cells. CONCLUSIONS: Drug-induced senescence is associated with late relapse after therapy in transgenic models of cancer in vivo. We have shown that abolition of p21/WAF1-mediated drug-induced senescence or antisense-mediated Bcl-xl knockdown can both, independently, enhance the apoptotic response of colorectal cancer cells to SN38 in vitro. The growth arrest suppresses a p53-independent apoptotic pathway, whereas Bcl-xl induction suppresses a p53 and Bax-dependent apoptotic pathway. The combination of irinotecan and Bcl-xL antisense merits testing in models of colorectal cancer in vivo.

Enhanced clearance of topoisomerase I inhibitors from human colon cancer cells by glucuronidation.

As part of a program to identify novel mechanisms of resistance to topoisomerase I (topo I) inhibitors, the cellular pharmacology of 7-ethyl-10-hydroxycamptothecin (SN-38), the active metabolite of clinically used irinotecan (CPT-11) and NU/ICRF 505, an anthraquinone-tyrosine conjugate, has been investigated in two human colorectal cancer (CRC) cell lines. Two novel metabolites of NU/ICRF 505 (M1 and M2) and a single metabolite of SN-38 (M1) were detected by high performance liquid chromatography in the culture medium of HT29 cells but were absent in HCT116 cells. Identities of all three metabolites were established by a combination of biochemical and physicochemical techniques. M1 of SN-38 was the C10-(beta)-glucuronide of the parent lactone while M1 of NU/ICRF 505 was the C4-O-glucuronide and M2 the tyrosine-O-glucuronide, both of the parent compound. Drug transport studies revealed that by 24hr HT29 cells had effectively cleared 82.5% of NU/ICRF 505 (10 microM) into the culture medium as the two glucuronides. In contrast, intracellular concentrations of NU/ICRF 505 were maintained in HCT116 cells in the absence of glucuronidation at a level 550 times greater than in HT29 cells. HT29 cells cleared 40.9% of SN-38 (1 microM) as the glucuronide to the culture medium, while the parent drug was maintained at a level 2-fold greater in HCT116 cells. Enhanced drug clearance due to glucuronidation may contribute to intrinsic drug resistance of human CRC.

Factors influencing the cellular accumulation of SN-38 and camptothecin.

PURPOSE: The influence of biophysical factors (drug metabolism, transport proteins, and chemical stability) on the cellular accumulation of camptothecin (CPT) and SN-38 was examined. METHODS: Drug transporter RNA transcript levels were measured by real-time reverse transcriptase polymerase chain reaction (RT-PCR). Intracellular drug accumulation, metabolism, and drug stability studies were all performed by HPLC. RESULTS: A panel of three human cell lines exhibiting different drug resistant phenotypes was investigated. HT29 colon cells glucuronidated SN-38 but did not express P-gp or MRP1 or 2. HCT116 colon cells expressed P-gp and MRP2 but did not catalyse conjugation. A2780 ovarian cells neither catalysed drug metabolism nor contained these drug transporters. In all lines, SN-38 lactone was rapidly taken up achieving peak concentrations at the earliest time point studied (5 min, 3.3-4.1 ng/10(6) cells). Subsequently, a fall in intracellular lactone concentration occurred, stabilising after 4 h at 0.48-1.18 ng/10(6) cells. No significant differences in intracellular levels of lactone were observed between the three cell lines with one exception: a twofold increase in HCT116 cells at 24 h. Stability studies in culture medium revealed that SN-38 lactone concentrations disappeared at the same rate regardless of whether cells were present, initially falling to reach equilibrium with the hydroxy acid by 4 h. Indeed, changes in intracellular lactone concentrations followed closely chemical stability profiles in media. Similar patterns of cellular retention and chemical degradation were observed with CPT. CONCLUSION: The major determinant of drug accumulation in three diverse cell line phenotypes was lactone chemical stability in culture medium.