Measurement of endogenous uracil and dihydrouracil in plasma and urine of normal subjects by liquid chromatography-tandem mass spectrometry

A sensitive and specific HPLC-MS-MS method was developed for the determination of endogenous uracil (Ura) and its metabolite dihydrouracil (UH2) in human plasma and urine samples. Plasma samples were extracted with ethyl acetate-isopropanol (85:15, v/v) following added ammonium sulfate, and then separated on a Discovery Amide C16 column with 3% methanol solution as the mobile phase; urine samples were just centrifuged at 2500 g for detection. Quantitation was carried out by LC-MS-MS in the multiple reaction monitoring (MRM) mode. The limits of quantitation of the method for Ura and UH2 were 0.5 and 5 ng ml(-1) (for plasma), and 50 and 100 ng ml(-1) (for urine), respectively. This method can be useful to evaluate the activity of dihydropyrimidine dehydrogenase (DPD), a rate-limiting enzyme of the chemotherapy drug fluoropyrimidine, which will be helpful in investigating subject variation of DPD and adjusting clinical dosage in pyrimidine chemotherapy.     

High-performance liquid chromatographic assay with UV detection for measurement of dihydrouracil/uracil ratio in plasma

A rapid, robust and sensitive HPLC method for analysis of uracil (U) and dihydrouracil (UH2) in plasma was developed using solid phase extraction and ultraviolet detection. Separation was achieved with a SymmetryShield RP18 column and an Atlantis dC18 column using a 10 mM potassium phosphate buffer as mobile phase. Compounds were eluted within 15 min without interference. Recovery was 80.4 and 80.6% for U and UH2. Calibration curves were linear from 2.5 to 80 ng/mL for U and 6.75 to 200 ng/mL for UH2. The LLQ was, respectively, 2.5 ng/mL for U, and 6.75 ng/mL for UH2. Within-run and between-run precision were less than 5.94% and inaccuracy did not exceed 7.80%. The overall procedure has been applied to correlate UH2/U ratio with dihydropyrimidine dehydrogenase activity in 165 cancer patients.     

Thymidylate synthase and dihydropyrimidine dehydrogenase levels are associated with response to 5-fluorouracil in Caenorhabditis elegans

5-Fluorouracil (5-FU), a pyrimidine antagonist, has a long history in cancer treatment. The targeted pyrimidine biosynthesis pathway includes dihydropyrimidine dehydrogenase (DPD), which converts 5-FU to an inactive metabolite, and thymidylate synthase (TS), which is a major target of 5-FU. Using Caenorhabditis elegans as a model system to study the functional and resistance mechanisms of anti-cancer drugs, we examined these two genes in order to determine the extent of molecular conservation between C. elegans and humans. Overexpression of the worm DPD and TS homologs (DPYD-1 and Y110A7A.4, respectively) suppressed germ cell death following 5-FU exposure. In addition, DPYD-1 depletion by RNAi resulted in 5-FU sensitivity, while treatment with Y110A7A.4 RNAi and 5-FU resulted in similar patterns of embryonic death. Thus, the pathway of 5-FU function appears to be highly conserved between C. elegans and humans at the molecular level.     

Analysis of pyrimidine catabolism in Drosophila melanogaster using epistatic interactions with mutations of pyrimidine biosynthesis and beta-alanine metabolism

The biochemical pathway for pyrimidine catabolism links the pathways for pyrimidine biosynthesis and salvage with beta-alanine metabolism, providing an array of epistatic interactions with which to analyze mutations of these pathways. Loss-of-function mutations have been identified and characterized for each of the enzymes for pyrimidine catabolism: dihydropyrimidine dehydrogenase (DPD), su(r) mutants; dihydropyrimidinase (DHP), CRMP mutants; beta-alanine synthase (betaAS), pyd3 mutants. For all three genes, mutants are viable and fertile and manifest no obvious phenotypes, aside from a variety of epistatic interactions. Mutations of all three genes disrupt suppression by the rudimentary gain-of-function mutation (r(Su(b))) of the dark cuticle phenotype of black mutants in which beta-alanine pools are diminished; these results confirm that pyrimidines are the major source of beta-alanine in cuticle pigmentation. The truncated wing phenotype of rudimentary mutants is suppressed completely by su(r) mutations and partially by CRMP mutations; however, no suppression is exhibited by pyd3 mutations. Similarly, su(r) mutants are hypersensitive to dietary 5-fluorouracil, CRMP mutants are less sensitive, and pyd3 mutants exhibit wild-type sensitivity. These results are discussed in the context of similar consequences of 5-fluoropyrimidine toxicity and pyrimidine catabolism mutations in humans.     

Application of non-steroidal anti-inflammatory drugs to enhance 5-fluorouracil efficacy in experimental systems

The elevated cyclooxygenase-2 (COX-2) expression has been shown to affect the carcinogenesis and tumor progression processes, including cell proliferation, motility and angiogenesis. COX-2 is overexpressed in approximately 80% of sporadic colorectal carcinomas and COX-2 enzyme is the best defined target of non-steroidal anti-inflammatory drugs (NSAIDs). In the chemotherapy of colorectal carcinomas 5-fluorouracil (5-FU) has been the most important of the basic drugs for more than 40 years. In order to improve the effectiveness of 5-FU therapy different biological modifiers i.e. inhibitors of its catabolism or activators of anabolism have been studied recently. The rate-limiting enzyme of 5-FU catabolism is dihydropyrimidine dehydrogenase (DPD) since more than 80% of the administered 5-FU is catabolized by DPD. Tumoral DPD has become of clinical interest because elevated intratumoral DPD can decrease the tumor response to 5-FU therapy. The main purpose of our experiments was to investigate the effect of COX inhibitors on the efficacy of 5-FU on high and low COX-2 expressing HCA-7 and HT-29 human colon adenocarcinoma cell lines, respectively, and also on xenografts derived from HT-29 cells. The cytotoxic and antitumor effects of 5-FU in the presence of low doses of indomethacin (non-selective COX-2 inhibitor) and that of NS-398 (highly selective COX-2 inhibitor) on HT-29 and HCA-7 cells and also on the HT-29 xenograft were investigated. In addition, our intention was to understand the mechanism(s) by which NSAIDs could enhance the cytotoxic effect of 5-FU. Our data indicated that the elevated COX-2 expression of HCA-7, the collagen-induced HT-29-C cells and of the HT-29 xenograft were associated with reduced 5-FU sensitivity. Based on the fact that at the same time DPD activity was also increased it might be conceivable that a possible explanation for the decrease of 5-FU sensitivity is the co-existence of high COX-2 and DPD activity. Indomethacin or NS-398 enhanced in a simultaneous and significant manner the sensitivity and cytotoxic effect of 5-FU on high COX-2 expressing cells and xenografts through the modulation of DPD - decrease of its mRNA expression and/or enzyme activity. Based on our results it could be presumable that 5-FU efficacy is limited by the COX-2 associated high DPD expression and activity in patients with colorectal cancer as well, therefore further clinical studies are warranted to decide if NSAIDs in the therapeutic protocol might improve the antitumor potency of 5-FU. Réti A. Application of non-steroidal anti-inflammatory drugs to enhance 5-fluorouracil efficacy in experimental systems.     

Design and synthesis of the tumor-activated prodrug of dihydropyrimidine dehydrogenase (DPD) inhibitor,RO0094889 for combination therapy with capecitabine

A series of tumor-activated prodrugs of the inhibitors of dihydropyrimidine dehydrogenase (DPD), an enzyme catabolizing 5-fluorouracil (5-FU: 4g), has been designed and synthesized. RO0094889 (11c) is a prodrug of 5-vinyluracil (4c), a known DPD inhibitor, and was designed to generate 4c selectively in tumor tissues by sequential conversion of 11c by three enzymes: esterase, cytidine deaminase and thymidine phosphorylase, the latter two of which are known to be highly expressed in various tumor tissues. When capecitabine (1), a tumor-activated prodrug of 5-FU, was co-administered orally with 11c, 5-FU in tumor tissues was significantly increased with only a slight increase of 5-FU in plasma as compared with oral capecitabine alone.     

Orotate phosphoribosyltransferase expression level in tumors is a potential determinant of the efficacy of 5-fluorouracil

Although the intratumoral expression levels of thymidylate synthase (TS) and dihydropyrimidine dehydrogenase (DPD) are known to affect the antitumor activity of 5-fluorouracil (5-FU), the importance of orotate phosphoribosyltransferase (OPRT) has remained unclear. This study investigated the relationship between intratumoral OPRT expression and the antitumor activity of 5-FU using human NCI60 cell lines with similar levels of TS and DPD messenger RNAs, as well as 31 tumor xenografts. The OPRT mRNA level was positively correlated with the 5-FU efficacy in these cell lines. In vitro, the 50% growth-inhibitory concentrations of 5-FU were closely correlated with the OPRT mRNA levels in cancer cell lines with similar levels of TS mRNAs when combined with a DPD inhibitor. Moreover, downregulation of OPRT with small-interfering RNA decreased the sensitivities of the cultured tumor cells to 5-FU. These results suggest that the OPRT expression level in tumors is an additional determinant of the efficacy of 5-FU.     

Pharmacogenetic studies on the prediction of efficacy and toxicity of fluoropyrimidine-based adjuvant therapy in colorectal cancer

The cytotoxic effect of 5-fluorouracil (5-FU) is mediated by the inhibition of thymidylate synthase (TS), however, at the same time 5-FU is catabolized by dihydropyrimidine dehydrogenase (DPD). Efficacy of 5-FU may therefore depend on the TS and DPD activity and on pharmacogenetic factors influencing these enzymes. Our aims were (1) to determine the distribution of DPD activity, the frequency of DPD deficiency and the DPD (IVS14+1G>A) mutation in the peripheral blood mononuclear cells of colorectal cancer (CRC) patients, and study the relationship between DPD deficiency and toxicity of 5-FU; (2) to investigate the influence of TS polymorphisms and DPD activity on the survival of CRC patients receiving 5-FU-based adjuvant therapy. The frequency of DPD deficiency was determined by radiochemical methods in the peripheral blood mononuclear cells (PBMCs) of 764 CRC patients treated with 5-FU. The relationship between the TS polymorphisms, DPD activity and the disease-free and overall survival was studied in 166 CRC patients receiving 5-FU-based adjuvant therapy. TS polymorphisms were determined in the DNA samples separated from the PBMCs, by PCR-PAGE and PCR-RFLP-PAGE (restriction fragment length polymorphism) methods. Low DPD values (<10 pmol/min/106 PBMCs) were demonstrated in 160/764 patients (20.9%), and of those DPD deficiency (<5 pmol/min/106 PBMCs) was verified in 38 patients (4.9%). In the latter group severe (>Gr 3) toxicity was found in 87%. The prevalence of the DPD IVS14+1G>A mutation among the 38 DPD-deficient patients was 7.8% (3/38) and was accompanied by severe Gr 4 toxic symptoms (neutropenia, mucositis, diarrhea). TS polymorphisms showed a relationship with the survival of CRC patients. It is important to mention that by combining the 3-3 genotypes of 5'-TSER and 3'-TSUTR polymorphisms the obtained 8 genotype combinations showed significantly different Kaplan-Meier survival curves. The evaluation of these curves with Cox regression analysis resulted in two prognostically different groups: "A" good prognosis (RR<1) and "B" bad prognosis (RR>1). The disease-free- and overall survival of these two groups were significantly different. DPD activity also showed correlation with the survival; patients with DPD activity <10 pmol/min/106 PBMCs showed significantly longer disease-free and overall survival. The determination of DPD activity proved to be a more valuable parameter in the evaluation of serious 5-FU-related toxicity compared to the IVS14+1G>A mutation analysis. According to the Cox multivariate analysis the combination of germline TS polymorphisms and DPD activity is/an independent prognostic marker of survival in CRC patients treated with adjuvant 5-FU therapy.     

Thymidylate synthase predicts for clinical outcome in invasive breast cancer

Thymidylate synthase (TS) is a major target of 5-fluorouracil (5-FU) and dihydropyrimidine dehydrogenase (DPD) is a rate-limiting enzyme in the degradation of 5-FU. Whether TS or DPD could be used as valuable parameters for 5-FU sensitivity in clinical patients are largely unknown. We analyzed TS and DPD expression in breast carcinomas to evaluate the clinicopathological significance of these enzymes in patients with invasive breast cancer receiving 5-FU-based chemotherapy. A total of 197 patients with invasive ductal carcinoma were included in our study. Both the TS and DPD expression were analyzed using immunohistochemical method for all the surgical samples. Sixty-three out of 197 (31.97%) patients are positive for TS expression, and 77 out of 197 (39.09%) patients are positive for DPD expression. TS expression was not correlated with DPD expression. Patients with TS-positivity had aggressive phenotype including large tumor size, low differentiation and nodal metastasis. DPD expression is not related with phenotype or prognosis. Multivariate analysis demonstrated that TS expression was an independent prognostic factor for both disease-free and overall survival. The current study demonstrated that TS but not DPD expression was associated with both progression and prognosis in breast cancer receiving 5-FU-based chemotherapy. TS expression in the primary tumor might be useful as a predictive parameter for the efficacy of 5-FU-based chemotherapy for breast cancer.     

Clinical pharmacokinetics of 5-fluorouracil with consideration of chronopharmacokinetics

Even though 5-fluorouracil (FU) is one of the oldest anticancer drugs, its use in cancer chemotherapy continues to increase. Fluorouracil is a pro-drug that requires intracellular activation to exert its effects. This makes it difficult to associate blood drug concentration with cell toxicity directly, although data from the literature show the existence of such a relationship. The relationship between FU pharmacokinetics and patient response has been explored extensively and reports attest a link between systemic drug exposure and response and survival. This has led to the concept of maximal tolerated exposure, and strategies to achieve this rely on pharmacokinetic follow-up and individual dose adjustment. More than 80% of the administered FU dose is eliminated by catabolism through dihydropyrimidine dehydrogenase (DPD), the rate-limiting enzyme. Dihydropyrimidine dehydrogenase activity is found in most tissues but is highest in the liver. Peripheral blood mononuclear cells (PBMC) are used to monitor clinically DPD activity. A significant, but weak correlation between PBMC and liver DPD activity has been observed. The relationship between PBMC-DPD activity and FU systemic clearance is weak (r²=0.10); thus, simply determining PBMC-DPD is not sufficient to predict accurately FU clearance. Population pharmacokinetic analysis identified patient co-variables that influence FU clearance; drug kinetics is significantly reduced by increased age, high serum alkaline phosphatase, length of drug infusion, and low PBMC-DPD. Autoregulation of FU metabolism also is suggested; inhibition of DPD activity was observed after FU administration in both colorectal cancer patients and an animal model. Circadian rhythmicity in DPD activity is suggested from both human and animal investigations. In patients receiving protracted low dose 5-FU infusion, the circadian rhythm in FU plasma concentration peaks at 11:00h and is lowest at 23:00h, on average. The inverse relationship observed between the circadian profile of FU plasma concentration and PBMC-DP activity in these same patients suggests a link between DPD activity and FU pharmacokinetics. The impact of the biological time of drug administration was also studied with short venous infusions; clearance was 70% greater at 13:00h than at 01:00h. Similarly, peak drug concentration occurred in the first half of the night in patients receiving constant rate 5-FU infusion for 2-5 d. Several studies describe wide interindividual variation in the timing of the peak and trough of the 24h rhythm in DPD activity. The rational for FU chronomodulated therapy has been the circadian rhythm in host drug tolerance, which is greatest during the night time when the proliferation of normal target tissue is least. A randomized study of chronomodulated FU therapy with maximal delivery rate at 04:00h was shown clearly to be significantly more effective and less toxic than control flat FU therapy. Future research must focus on easy-to-obtain markers of specific rhythms to individualize the chronomodulated FU delivery.     

Clinical pharmacokinetics of 5-fluorouracil with consideration of chronopharmacokinetics

Even though 5-fluorouracil (FU) is one of the oldest anticancer drugs, its use in cancer chemotherapy continues to increase. Fluorouracil is a pro-drug that requires intracellular activation to exert its effects. This makes it difficult to associate blood drug concentration with cell toxicity directly, although data from the literature show the existence of such a relationship. The relationship between FU pharmacokinetics and patient response has been explored extensively and reports attest a link between systemic drug exposure and response and survival. This has led to the concept of maximal tolerated exposure, and strategies to achieve this rely on pharmacokinetic follow-up and individual dose adjustment. More than 80% of the administered FU dose is eliminated by catabolism through dihydropyrimidine dehydrogenase (DPD), the rate-limiting enzyme. Dihydropyrimidine dehydrogenase activity is found in most tissues but is highest in the liver. Peripheral blood mononuclear cells (PBMC) are used to monitor clinically DPD activity. A significant, but weak correlation between PBMC and liver DPD activity has been observed. The relationship between PBMC–DPD activity and FU systemic clearance is weak (r²=0.10); thus, simply determining PBMC–DPD is not sufficient to predict accurately FU clearance. Population pharmacokinetic analysis identified patient co-variables that influence FU clearance; drug kinetics is significantly reduced by increased age, high serum alkaline phosphatase, length of drug infusion, and low PBMC–DPD. Autoregulation of FU metabolism also is suggested; inhibition of DPD activity was observed after FU administration in both colorectal cancer patients and an animal model. Circadian rhythmicity in DPD activity is suggested from both human and animal investigations. In patients receiving protracted low dose 5-FU infusion, the circadian rhythm in FU plasma concentration peaks at 11:00h and is lowest at 23:00h, on average. The inverse relationship observed between the circadian profile of FU plasma concentration and PBMC–DP activity in these same patients suggests a link between DPD activity and FU pharmacokinetics. The impact of the biological time of drug administration was also studied with short venous infusions; clearance was 70% greater at 13:00h than at 01:00h. Similarly, peak drug concentration occurred in the first half of the night in patients receiving constant rate 5-FU infusion for 2–5 d. Several studies describe wide interindividual variation in the timing of the peak and trough of the 24h rhythm in DPD activity. The rational for FU chronomodulated therapy has been the circadian rhythm in host drug tolerance, which is greatest during the night time when the proliferation of normal target tissue is least. A randomized study of chronomodulated FU therapy with maximal delivery rate at 04:00h was shown clearly to be significantly more effective and less toxic than control flat FU therapy. Future research must focus on easy-to-obtain markers of specific rhythms to individualize the chronomodulated FU delivery.     

Mutational analysis of the human dihydropyrimidine dehydrogenase gene by denaturing high-performance liquid chromatography

Mutations in the DPYD gene, which encodes dihydropyrimidine dehydrogenase (DPD), the rate-limiting enzyme in the catabolism of pyrimidines, are responsible for an inborn error of metabolism associated with thymine-uraciluria and neurological symptoms. Because the antimetabolite 5-fluorouracil (5-FU) is metabolized by the same enzyme, deficient DPYD alleles may also constitute a risk factor for severe toxicity following treatment with this anticancer drug. The aim of this study was to develop a comprehensive and rapid method to detect sequence variations within the DPYD gene. Using polymerase chain reaction (PCR) amplification and denaturing high-performance liquid chromatography (DHPLC), we established a protocol that makes it possible to screen all 23 exons of the DPYD gene and their exon-intron boundaries for both known and unknown mutations under identical conditions. A novel one-step PCR mutagenesis procedure was developed to generate heterozygous mutant amplicons as positive controls to optimize DHPLC detection of any sequence variation. DHPLC analysis was shown to result in mutation-specific elution profiles and to be able to distinguish different base changes within the same exon or different heterozygous combinations of mutations within the same exon. By analyzing the DPYD gene in 16 affected individuals, a total of 47 base changes were detected, representing eight known mutations and three novel intronic base changes. Sequence analysis confirmed all base changes detected. This method will be useful in identifying patients at risk for toxicity prior to 5-FU treatment, as well as in the analysis of individual patients with thymine-uraciluria.     

Identification of three novel mutations in the dihydropyrimidine dehydrogenase gene associated with altered pre-mRNA splicing or protein function

Dihydropyrimidine dehydrogenase (DPD) is the initial and rate-limiting enzyme in the catabolism of the pyrimidine bases uracil and thymine, as well as of the widely used chemotherapeutic drug 5-fluorouracil (5FU). Analysis of the DPD gene ( DPYD ) in two patients presenting with complete DPD deficiency and the parents of an affected child showed the presence of three novel mutations, including one splice site mutation IVS11 + 1G-->T and the missense mutations 731A-->C (E244V) and 1651G-->A (A551T). The G-->T mutation in the invariant GT splice donor site flanking exon 11 (IVS11 + 1G-->T) created a cryptic splice site within exon 11. As a consequence, a 141-bp fragment encoding the aminoacid residues 400-446 of the primary sequence of the DPD protein was missing in the mature DPD mRNA. Analysis of the crystal structure of pig DPD suggested that the E244V mutation might interfere with the electron flow between NADPH and the pyrimidine binding site of DPD. The A551T point mutation might prevent binding of the prosthetic group FMN and affect folding of the DPD protein. The identification of these novel mutations in DPYD will allow the identification of patients with an increased risk of developing severe 5FU-associated toxicity.     

Validation of an isocratic HPLC method to detect 2-fluoro-β-alanine for the analysis of dihydropyrimidine dehydrogenase activity

The efficacy of the chemotherapeutic drug 5′-fluorouracil is reduced by catabolism to 2′-fluoro-β-alanine (FBAL), a three-step reaction in which dihydropyrimidine dehydrogenase (DPD) catalyzes the rate-limiting step. To study in vitro DPD activity, we developed and validated an isocratic, reverse-phase HPLC method to detect and quantify FBAL without using multiple columns or radiolabeled substrates. Pre-column derivatization of FBAL was performed using o-phthalaldehyde in the presence of two sulfur donors, ethanthiol or β-mercaptoethanol, and the resulting products assayed. Calibration curves were linear over a range of 10–200 μg/ml and the method was successfully applied to the examination of DPD activity in cultured cells.     

Circadian rhythm in dihydropyrimidine dehydrogenase activity and reduced glutathione content in peripheral blood of nasopharyngeal carcinoma patients

Dihydropyrimidine dehydrogenase (DPD) is a rate-limiting enzyme of 5-fluorouracil (5-FU) catabolism. Glutathione (GSH) is a tripeptide involved in platinum complex detoxification. This study explored the circadian rhythms of DPD activity and GSH concentration in the peripheral blood of 16 patients with histologically proven nasopharyngeal carcinoma (NPC) in order to guide the establishment of chronotherapeutic schedules for this cancer. DPD activity and GSH concentration were determined by high performance liquid chromatography (HPLC). Both variables displayed significant circadian rhythms (Cosinor analysis: p=0.009 and 0.012, respectively). Peak DPD activity occurred at about 02:30 h; whereas, peak GSH concentration occurred around 12:40 h. The differences between the peak and nadir mean values were 25.5% and 38.7%, respectively. The study showed that the circadian rhythms in DPD activity and GSH concentration in Chinese NPC are similar to those reported for western patients with colorectal cancer, despite the differences in race and kinds of cancer. These findings imply that the chronotherapeutic schedule of 5-FU and platinum used to treat European colorectal cancer patients probably is applicable to Chinese NPC patients.     

A high-throughput denaturing high-performance liquid chromatography method for the identification of variant alleles associated with dihydropyrimidine dehydrogenase deficiency

Dihydropyrimidine dehydrogenase (DPD) is the initial, rate-limiting enzyme in the catabolism of 5-fluorouracil (5-FU). A pharmacogenetic syndrome has been described in which DPD-deficient patients are at risk for toxicity following administration of 5-FU. To date, there are at least 21 previously described mutations and/or polymorphisms that have been associated with DPD deficiency. In this study we describe the development of a highly specific, sensitive, inexpensive, and robust denaturing HPLC (DHPLC) method for rapidly identifying sequence variations (mutations and/or polymorphisms) in the gene (DPYD) that codes for the DPD enzyme. DHPLC conditions were optimized at three temperatures for analysis of the 23 exons of the DPYD gene using 25 amplicons representing the entire coding sequence, including all intron/exon boundaries (splice sites). Resolution of all 25 amplicons at the optimized temperature can be performed in 4.2 h. All 21 previously described sequence variations (mutations and/or polymorphisms) were prepared using site-directed mutagenesis from the wild-type DPYD gene, confirmed by sequence analysis, and subsequently resolved by DHPLC using the optimized conditions. These analyses generated reference chromatogram patterns for all known sequence variations previously encountered in DPD-deficient patients. In order to examine the utility and sensitivity of this approach, samples from patients with known sequence variations in the DPYD gene were analyzed. This DHPLC technique resolved 100% of the known DPYD sequence variations and differentiated between homozygous and heterozygous genotypes. We conclude that this DHPLC method is a highly specific and sensitive technique for rapidly detecting known sequence variations in the DPYD gene. In addition, this approach can be used to identify currently unrecognized unknown sequence variations in the DPYD gene and should be useful in future pharmacogenetic studies examining DPD deficiency.