Expression of deoxynucleoside kinases and 5'-nucleotidases in mouse tissues: implications for mitochondrial toxicity

Anti-HIV nucleoside therapy can result in mitochondrial toxicity affecting muscles, peripheral nerves, pancreas and adipose tissue. The cytosolic deoxycytidine kinase (dCK; EC 2.7.1.74) and thymidine kinase (TK1; EC 2.7.1.21), the mitochondrial thymidine kinase (TK2) and deoxyguanosine kinase (dGK; EC 2.7.1.113) as well as 5'-deoxynucleotidases (5'-dNT; EC 3.1.3.5) are enzymes that control rate-limiting steps in formation of intracellular and intra-mitochondrial nucleotides. The mRNA levels and activities of these enzymes were determined in mouse tissues, using real-time PCR and selective enzyme assays. The expression of mRNA for all these enzymes and the mitochondrial deoxynucleotide carrier was detected in all tissues with a 5-10-fold variation. TK1 activities were only clearly detected in spleen and testis, while TK2, dGK and dCK activities were found in all tissues. dGK activities were higher than any other dNK in all tissues, except spleen and testis. In skeletal muscle dGK activity was 5-fold lower, TK2 and dCK levels were 10-fold lower as compared with other tissues. The variation in 5'-dNT activities was about eight-fold with the highest levels in brain and lowest in brown fat. Thus, the salvage of deoxynucleosides in muscles is 5-10-fold lower as compared to other non-proliferating tissues and 100-fold lower compared to spleen. These results may help to explain tissue specific toxicity observed with nucleoside analogs used in HIV treatment as well as symptoms in inherited mitochondrial TK2 deficiencies.     

3'-Azido-2',3'-dideoxythymidine induced deficiency of thymidine kinases 1, 2 and deoxycytidine kinase in H9 T-lymphoid cells

Continuous cultivation of T-lymphoid H9 cells in the presence of 3′-azido-2′,3′-dideoxythymidine (AZT) resulted in a cell variant cross-resistant to both thymidine and deoxycytidine analogs. Cytotoxic effects of AZT, 2′,3′-didehydro-3′-deoxythymidine as well as different deoxycytidine analogs such as 2′,3′-dideoxycytidine, 2′,2′-difluoro-2′-deoxycytidine (dFdC) and 1-ß-D-arabinofuranosylcytosine (Ara-C) were strongly reduced in H9 cells continuously exposed to AZT when compared to parental cells (>8.3-, >6.6-, >9.1-, 5×10⁴-, 5×10³-fold, respectively). Moreover, anti-HIV-1 effects of AZT, d4T, ddC and 2′,3′-dideoxy-3′-thiacytidine (3TC) were significantly diminished (>222-, >25-, >400-, >200-fold, respectively) in AZT-resistant H9 cells. Study of cellular mechanisms responsible for cross-resistance to pyrimidine analogs in AZT-resistant H9 cells revealed decreased mRNA levels of thymidine kinase 1 (TK1) and lack of deoxycytidine kinase (dCK) mRNA expression. The loss of dCK gene expression was confirmed by western blot analysis of dCK protein as well as dCK enzyme activity assay. Moreover, enzyme activity of TK1 and TK2 was reduced in AZT-resistant cells. In order to determine whether lack of dCK affected the formation of the active triphosphate of the deoxycytidine analog dFdC, dFdCTP accumulation and retention was measured in H9 parental and AZT-resistant cells after exposure to 1 and 10 μM dFdC. Parental H9 cells accumulated about 30 and 100 pmol dFdCTP/10⁶ cells after 4 hr, whereas in AZT-resistant cells no dFdCTP accumulation was detected. These results demonstrate that continuous treatment of H9 cells in the presence of AZT selected for a thymidine analog resistant cell variant with cross-resistance to deoxycytidine analogs, due to deficiency in TK1, TK2, and dCK.     

3'-Azido-3'-deoxythymidine (AZT) is a competitive inhibitor of thymidine phosphorylation in isolated rat heart and liver mitochondria

Long-term use of 3'-azido-3'-deoxythymidine (AZT) is associated with various tissue toxicities, including hepatotoxicity and cardiomyopathy, and with mitochondrial DNA depletion. AZT-5'-triphosphate (AZTTP) is a known inhibitor of the mitochondrial DNA polymerase gamma and has been targeted as the source of the mitochondrial DNA depletion. However, in previous work from this laboratory with isolated rat heart and liver mitochondria, AZT itself was shown to be a more potent inhibitor of thymidine phosphorylation (IC50 of 7.0+/-1.0 microM AZT in heart mitochondria and of 14.4+/-2.6 microM AZT in liver mitochondria) than AZTTP is of polymerase gamma (IC50 of >100 microM AZTTP), suggesting that depletion of mitochondrial stores of TTP may limit replication and could be the cause of the mitochondrial DNA depletion observed in tissues affected by AZT toxicity. The purpose of this work is to characterize the nature of AZT inhibition of thymidine phosphorylation in isolated rat heart and rat liver mitochondria. In both of these tissues, AZT was found to be a competitive inhibitor of the phosphorylation of thymidine to TMP, catalyzed by thymidine kinase 2. The inhibition constant (Ki) for heart mitochondria is 10.6+/-4.5 microM AZT, and for liver mitochondria Ki is 14.0+/-2.5 microM AZT. Since AZT is functioning as a competitive inhibitor, increasing thymidine concentrations may be one mechanism to overcome the inhibition and decrease AZT-related toxicity in these tissues.     

3'-Azido-3'-deoxythymidine (AZT) inhibits thymidine phosphorylation in isolated rat liver mitochondria: a possible mechanism of AZT hepatotoxicity

3'-azido-3'-deoxythymidine (AZT) is a staple of highly active antiretroviral therapy (HAART). Prior to HAART, long-term use of high-dosage AZT caused myopathy, cardiomyopathy, and hepatotoxicity, associated with mitochondrial DNA depletion. As a component of HARRT, AZT causes cytopenias and lipodystrophy. AZT-5'-triphosphate (AZTTP) is a known inhibitor of the mitochondrial polymerase gamma and has been targeted as the source of the mitochondrial DNA depletion. However, in previous work from this laboratory with isolated rat heart mitochondria, AZT phosphorylation beyond AZT-5'-monophosphate (AZTMP) was not detected. Rather, AZT was shown to be a more potent inhibitor of thymidine phosphorylation (50% inhibitory concentration (IC50) of 7.0+/-1.0 microM) than AZTTP is of polymerase gamma (IC50 of >100 microM), suggesting that depletion of mitochondrial stores of TTP may limit replication. This work has investigated whether an identical mechanism might account for the hepatotoxicity seen with long-term use of AZT. Isolated rat liver mitochondria were incubated with labeled thymidine or AZT, and the rate and extent of phosphorylation were determined by HPLC analysis of acid-soluble extracts of the incubated mitochondria. The results showed that in the phosphorylation of thymidine to TMP, liver mitochondria exhibit a higher Vmax and Km than heart mitochondria, but otherwise heart and liver mitochondria display similar kinetics. AZT is phosphorylated to AZTMP, but no further phosphorylated forms were detected. In addition, AZT inhibited the production of TTP, with an IC50 of 14.4+/-2.6 microM AZT. This is higher, but comparable to, the results seen in isolated rat heart mitochondria.     

Anabolism of amdoxovir: phosphorylation of dioxolane guanosine and its 5'-phosphates by mammalian phosphotransferases

Amdoxovir [(-)-beta-D-2,6-diaminopurine dioxolane, DAPD], the prodrug of dioxolane guanosine (DXG), is currently in Phase I/II clinical development for the treatment of HIV-1 infection. In this study, we examined the phosphorylation pathway of DXG using 15 purified enzymes from human (8), animal (6), and yeast (1) sources, including deoxyguanosine kinase (dGK), deoxycytidine kinase (dCK), high Km 5'-nucleotidase (5'-NT), guanylate (GMP) kinase, nucleoside monophosphate (NMP) kinase, adenylate (AMP) kinase, nucleoside diphosphate (NDP) kinase, 3-phosphoglycerate (3-PG) kinase, creatine kinase, and pyruvate kinase. In addition, the metabolism of 14C-labeled DXG was studied in CEM cells. DXG was not phosphorylated by human dCK, and was a poor substrate for human dGK with a high Km (7 mM). Human 5'-NT phosphorylated DXG with relatively high efficiency (4.2% of deoxyguanosine). DXG-MP was a substrate for porcine brain GMP kinase with a substrate specificity that was 1% of dGMP. DXG-DP was phosphorylated by all of the enzymes tested, including NDP kinase, 3-PG kinase, creatine kinase, and pyruvate kinase. The BB-isoform of human creatine kinase showed the highest relative substrate specificity (47% of dGDP) for DXG-DP. In CEM cells incubated with 5 microM DXG for 24 h, 0.015 pmole/10(6) cells (approximately 7.5 nM) of DXG-TP was detected as the primary metabolite. Our study demonstrated that 5'-nucleotidase, GMP kinase, creatine kinase, and NDP kinase could be responsible for the activation of DXG in vivo.     

Activation of deoxycytidine kinase by UV-C-irradiation in chronic lymphocytic leukemia B-lymphocytes

Deoxycytidine kinase (dCK), a key enzyme of the deoxynucleoside salvage pathway, might have a preponderant role in DNA synthesis in resting chronic lymphocytic leukemia B-lymphocytes. In these cells, two important enzymes in deoxynucleoside triphosphate production, ribonucleotide reductase and thymidine kinase (TK), both cell-cycle regulated, are indeed very weakly expressed. This study investigated the regulation of dCK activity in response to UV-C light, a condition which causes DNA lesions and DNA repair synthesis. We observed that activity of dCK in B-CLL cells was upregulated up to 3-fold, 30 min after irradiation with 30 J/m(2) UV-C, whereas TK activity was unchanged. Activation of dCK by UV-C light was caused neither by a change in concentration of a low molecular weight metabolite nor by an increase in the amount of dCK protein. Activation of dCK by UV-C was mimicked by H(2)O(2), markedly counteracted by N-acetylcysteine, a general antioxidant, and completely abolished by the growth factor receptor inhibitor suramin. Taken together, these results indicate that dCK activity is upregulated by UV-C light through a postranslational modification that may be initiated at the cell surface through oxidative mechanisms. Suramin also suppressed the increase in DNA repair synthesis elicited by UV-C irradiation, suggesting that upregulation of dCK activity could contribute to the normal completion of DNA repair synthesis elicited by UV light.     

Substrate specificity of feline and canine herpesvirus thymidine kinase

The thymidine kinases from feline herpesvirus (FHV TK) and canine herpesvirus (CHV TK) were cloned and characterized. The two proteins are closely sequence-related to each other and also to the herpes simplex virus type 1 thymidine kinase (HSV-1 TK). Although FHV TK and CHV TK have a level of identity of 31 and 35%, respectively, with HSV-1 TK, and a general amino acid similarity of approximately 54% with HSV-1 TK, they do not recognize the same broad range of substrates as HSV-1 TK does. Instead the substrate recognition is restricted to dThd and pyrimidine analogs such as 1-beta-d-arabinofuranosylthymine (araT), 3'-azido-2',3'-dideoxythymidine (AZT) and (E)-5-(2-bromovinyl)-2'-deoxyuridine (BVDU). FHV TK and CHV TK differ in substrate recognition from mammalian cytosolic thymidine kinase 1 (TK1) in that TK1 does not phosphorylate BVDU and they also differ from mammalian mitochondrial thymidine kinase 2 (TK2), which, in addition to thymidine and thymidine analogs also phosphorylates dCyd. Although the nucleoside analog BVDU was a good substrate for FHV and CHV TK, the compound was poorly inhibitory to virus-induced cytopathic effect in FHV- and CHV-infected cells. The reason is likely the poor, if any, thymidylate kinase activity of FHV and CHV TK, which in HSV-1 TK-expressing cells convert BVDU-MP to its 5'-diphosphate derivative.     

The pattern of deoxycytidine- and deoxyguanosine kinase activity in relation to messenger RNA expression in blood cells from untreated patients with B-cell chronic lymphocytic leukemia

Deoxycytidine kinase (dCK) and deoxyguanosine kinase (dGK) catalyze the first step in the intracellular cascade of fludarabine (2-fluoroadenine-beta-D-arabinofuranoside) and cladribine (2-chlorodeoxyadenosine) phosphorylation, which leads to activation of these prodrugs, commonly used for treatment of chronic lymphocytic leukemia (CLL). Thus, resistance to nucleoside analogues may primarily be due to low levels of deoxynucleoside kinase activity. The purpose of this study was to investigate the activity profiles of dCK and dGK and characterize the possible relationship between the levels of dCK enzymatic activities and mRNA levels in B-CLL cells from untreated patient samples in an attempt to determine the best approach for predicting sensitivity to nucleoside analogues and thereby optimizing treatment of CLL. For this purpose, dCK and dGK analyses were done in blood cells from 59 untreated symptomatic patients with CLL. The dGK activity towards 2-chlorodeoxyadenosine was significantly lower than of dCK (median 73 pmol/mg protein/min (85-121, 95% CI) versus 353 pmol/mg protein/min (331-421)). The median dCK mRNA level was 0.107 (0.096-0.120, 95% CI). There was a lack of correlation between the activities of dCK and dGK, which indicates that these proteins are regulated independently. We also found that the dCK and dGK activity measurement towards their endogenous substrates were comparable to the nucleoside analogues tested. Such variations in enzyme activities and mRNA levels may well explain differences in clinical responses to treatment. There was no correlation between the levels of dCK mRNAs and enzymatic activities using a quantitative real-time PCR procedure. Sequencing of dCK mRNA did not reveal alternate splicing or mutations in the coding region. The relation between activity and mRNA levels was studied by short interfering RNA (siRNA) method, which showed that in the siRNA treated cells the down-regulation of dCK expression, and activity followed each other. However, in control cells the mRNA levels remained stable but the protein activity markedly decreased. These data demonstrate that the dCK activity is not reflected by dCK mRNA expression that indicates a post-translational mechanism(s).     

Intracellular metabolism of CycloSaligenyl 3'-azido-2', 3'-dideoxythymidine monophosphate, a prodrug of 3'-azido-2', 3'-dideoxythymidine (zidovudine)

The administration of CycloSaligenyl 3'-azido-2',3'-dideoxythymidine monophosphate (CycloSal-AZTMP) to CEM cells resulted in a concentration- and time-dependent conversion to the 5'-monophosphate (AZTMP), 5'-diphosphate (AZTDP), and 5'-triphosphate (AZTTP) derivatives. High ratios of AZTMP/AZTTP were found in the CEM cell cultures treated with CycloSal-AZTMP. The intracellular T(1/2) of AZTTP in CEM cell cultures treated with either AZT and CycloSal-AZTMP was approximately 3 h. A variety of human T- and B-lymphocyte cell lines efficiently converted the prodrug to the AZT metabolites, whereas peripheral blood lymphocytes and primary monocyte/macrophages showed at least 10-fold lower metabolic conversion of the prodrug. CycloSal-AZTMP failed to generate marked levels of AZT metabolites in thymidine kinase-deficient CEM/TK(-) cells, an observation that is in agreement with the substantial loss of antiviral activity of CycloSal-AZTMP in CEM/TK(-) cells. The inability of CycloSal-AZTMP to generate AZTMP in CEM/TK(-) cells is presumably due to a relatively high hydrolysis rate of AZTMP to the parent nucleoside AZT, combined with the inability of CEM/TK(-) cells to phosphorylate AZT to AZTMP through the cytosolic salvage enzyme thymidine kinase.     

Expression of human mitochondrial thymidine kinase in Escherichia coli: correlation between the enzymatic activity of pyrimidine nucleoside analogues and their inhibitory effect on bacterial growth

Mitochondrial thymidine kinase (TK2) phosphorylates pyrimidine nucleosides to monophosphates and is expressed constitutively through the cell cycle in all cells. Because of the overlap of its substrate specificity with that of the cytosolic thymidine kinase (TK1) and deoxycytidine kinase (dCK), it has been difficult to determine the role of TK2 in activating nucleosides used in chemotherapy. In this report, we described the construction of a recombinant Escherichia coli strain which could be used to test if TK2 activity is limiting for the toxicity of nucleosides. Enzymes of bacterial origin which are involved in thymidine and deoxyuridine anabolism and catabolism were eliminated, and the cDNA for human TK2 was introduced. In the crude extract of the engineered E. coli, the level of thymidine kinase was, after induction of TK2 expression, several hundred fold higher than in the control strain. Several pharmacologically interesting nucleoside analogues, including 3'-azidothymidine, 2',3'-didehydro-2',3'-dideoxythymidine, and 2', 3'-dideoxy-beta-L-3'-thiacytidine, were tested for their effects on the growth of this recombinant strain. For a comparison, the phosphorylation of these compounds was determined with purified recombinant TK1, TK2, and dCK. A correlation was observed between the phosphorylation of several of these compounds by TK2 and their effects on bacterial growth. These results demonstrate that activation of growth-inhibiting pyrimidine nucleosides can be catalyzed by TK2, and together with recombinant E. coli strains expressing other cellular nucleoside kinases, this whole-cell bacterial system may serve as a tool to predict the efficacy and side effects of chemotherapeutic nucleosides.     

Deoxynucleoside phosphorylating enzymes in monkey and human tissues show great similarities, while mouse deoxycytidine kinase has a different substrate specificity.

Three key enzymes in the anabolic phosphorylation of deoxyribonucleosides and deoxyribonucleoside analogs were purified i.e. cytoplasmic thymidine kinase (TK1), mitochondrial thymidine kinase (TK2) and cytoplasmic deoxycytidine kinase (dCK) from human, mouse and monkey liver and spleen. Their subunit structure and substrate specificities were compared. Extensive purification of TK1 and dCK from mouse spleen and TK2 from mouse and monkey livers revealed major polypeptide bands of 25, 30 and 28 kD, respectively, on sodium dodecyl sulphate-polyacrylamide gel electrophoresis which are very similar to the subunit molecular weights of the corresponding human enzymes. Affinity purified polyclonal antibodies against human dCK also cross-reacted with 30 kD bands in extracts from both mouse and monkey spleen. Thus, the molecular weights of the subunits of these three enzymes appeared to be very similar in all three species. TK1 and TK2 from these different sources appeared to have similar substrate specificities against several deoxyribonucleoside analogs. However, mouse dCK differed significantly from monkey and human dCK in its capacity to phosphorylate dAdo and 2',3'-dideoxycytidine (ddCyd) with a Vmax approximately 10-fold lower than that of the two latter enzymes. The Km and Vmax values for dCyd and arabinocytosine appeared to be very similar with the enzymes from all three species. The fact that mouse dCK shows low activity with dAdo and ddCyd explains differences reported previously in the metabolism of dAdo and ddCyd in mouse compared to that in human lymphocytes. These results argue against the use of mice as model systems for human deoxynucleoside metabolism.     

Relationships between DNA incorporation, mutant frequency, and loss of heterozygosity at the TK locus in human lymphoblastoid cells exposed to 3'-azido-3'-deoxythymidine.

3'-Azido-3'-deoxythymidine (AZT), a thymidine analogue widely used in the treatment of AIDS patients and for prevention of the onset of AIDS in HIV-seropositive individuals, causes tumors in mice exposed as adults or in utero. The purpose of this study was to investigate the potential mechanisms of AZT mutagenicity and carcinogenicity by quantifying the incorporation of AZT into cellular DNA, measuring AZT-induced thymidine kinase (TK) mutant frequencies (Mfs), and determining the percentage of loss of heterozygosity (LOH) in spontaneous or AZT-induced TK mutants in the human lymphoblastoid cell line, TK6. Cells were exposed to 300 microM AZT for 0, 1, 3, or 6 days, or to 0, 33, 100, 300, or 900 microM AZT for 3 days (n = 5 flasks/group). The effects of exposure concentration on incorporation of AZT into cellular DNA were evaluated by an AZT radioimmunoassay, and the effects of duration and concentration of AZT exposure on the TK Mfs were assessed by a cell-cloning assay. AZT was incorporated into DNA in a dose-related manner at concentrations up to 300 microM, above which no further increase was observed. TK Mf increased with the extended duration and with incremental concentrations of AZT exposure. There was a positive correlation (P = 0.036, coefficient = 0.903) between AZT-DNA incorporation and AZT-induced TK Mfs, suggesting that AZT incorporation into cellular DNA has a direct role in the genotoxicity of AZT. Southern blot analyses indicated that 84% (6.2 x 10(-6)/7.4 x 10(-6)) of AZT-induced mutants were attributable to LOH, consistent with the known mechanism of AZT as a DNA chain terminator. Considering the importance of LOH in human carcinogenesis, AZT-induced LOH warrants further study.     

Characterization of the activation pathway of phosphoramidate triester prodrugs of stavudine and zidovudine.

The phosphoramidate triester prodrugs of anti-human HIV 2', 3'-dideoxynucleoside analogs (ddN) represent a convenient approach to bypass the first phosphorylation to ddN 5'-monophosphate (ddNMP), resulting in an improved formation of ddN 5'-triphosphate and, hence, higher antiviral efficacy. Although phosphoramidate derivatization markedly increases the anti-HIV activity of 2',3'-didehydro-2', 3'-dideoxythymidine (d4T) in both wild-type and thymidine kinase-deficient CEM cells, the concept is far less successful for the 3'-azido-2',3'-dideoxythymidine (AZT) triesters. We now investigated the metabolism of triester prodrugs of d4T and AZT using pure enzymes or different biological media. The efficiency of the first activation step, mediated by carboxylesterases, consists of the formation of the amino acyl ddNMP metabolite. The efficiency of this step was shown to be dependent on the amino acid, alkyl ester, and ddN moiety. Triesters that showed no conversion to the amino acyl ddNMP accumulated as the phenyl-containing intermediate and had poor, if any, anti-HIV activity. In contrast to the relative stability of the triesters in human serum, carboxylesterase-mediated cleavage of the prodrugs was found to be remarkably high in mouse serum. The subsequent conversion of the amino acyl ddNMP metabolite to ddNMP or ddN was highest in rat liver cytosolic enzyme preparations. Although L-alaninyl-d4TMP was efficiently converted to d4TMP, the main metabolite formed from L-alaninyl-AZTMP was the free nucleoside (AZT), thus explaining why d4T prodrugs, but not AZT prodrugs, retain anti-HIV activity in HIV-infected thymidine kinase-deficient cell cultures. The rat liver phosphoramidase responsible for the formation of ddNMP was shown to be distinct from creatine kinase, alkaline phosphatase, and phosphodiesterase.     

Rottlerin inhibits multiple steps involved in insulin-induced glucose uptake in 3T3-L1 adipocytes

Recently, it was shown that rottlerin inhibits insulin-stimulated glucose uptake and reduces intracellular adenosine triphosphate (ATP) levels in 3T3-L1 adipocytes, suggesting that these two events are causally linked. However, several other reports show that ATP-depletion induces glucose uptake in both muscle cells and adipocytes. In the present study, the mechanism of inhibition by rottlerin was studied in detail, in order to resolve this apparent discrepancy. It was found that rottlerin strongly reduces insulin-stimulated 2-deoxyglucose (2-DOG) uptake in 3T3-L1 adipocytes by a partial inhibition of the translocation of the insulin-responsive GLUT4 glucose transporter towards the plasma membrane (PM). Whereas the insulin-induced phosphatidyl-inositol-3' (PI-3') kinase signaling pathway is unaffected by rottlerin, Cbl tyrosine phosphorylation, which provides an essential, PI-3' kinase-independent signal towards GLUT4 translocation, is markedly attenuated. Furthermore, we also observed a direct inhibitory effect of rottlerin on insulin-induced glucose uptake in 3T3-L1 adipocytes. The direct inhibition of insulin-stimulated 2-DOG uptake by rottlerin displayed characteristics of uncompetitive inhibition: with the K(m(app)) of glucose uptake reduced from 1.6 to 0.9 mM and the V(max(app)) reduced from 5.2 to 1.0 nmol/minmg in the presence of rottlerin. In conclusion, rottlerin inhibits multiple steps involved in insulin-stimulated 2-DOG uptake in 3T3-L1 adipocytes. The observed reduction in GLUT4 translocation towards the PM and the uncompetitive inhibition of the glucose transport process provide alternative explanations for the inhibitory effects of rottlerin aside from the effects of rottlerin on intracellular levels of ATP.     

Cytotoxic activity of gemcitabine in cultured cell lines derived from histologically different types of bladder cancer: Role of thymidine kinase 2

World wide incidence of bladder cancer is rising with nearly 13,760 deaths attributed to bladder cancer in 2007 in the USA. Tumor types of the urothelium include transitional cell carcinomas, squamous cell carcinomas, and adenocarcinomas. This study was undertaken to determine gemcitabine's efficacy against bladder cancer cell lines of different origins (HTB2, a papilloma; HTB3, a squamous cell carcinoma; and HTB4, a transitional cell carcinoma). Roles of nucleoside transporters and key enzymes in gemcitabine pharmacology were examined on the premise that cells originating from different types of bladder cancer exhibit different levels and/or types of nucleoside transporters and enzymes and thus may respond differently to gemcitabine. HTB2 cells had the highest transport efficiency and were also most responsive to gemcitabine. HTB3 and HTB4 cells had similar transport efficiencies, but exhibited different sensitivities to gemcitabine (HTB4 > HTB3). The highest accumulation of [³H]gemcitabine was in HTB2 cells and the lowest was in HTB3 cells. Sequencing experiments revealed no mutations either in coding exons or intron-exon boundaries of the hENT1 genes of the three cell lines. HTB3 cells exhibited high thymidine kinase 2 (TK2) activity whereas HTB2 and HTB4 cells lacked detectable TK2 activity and pretreatment of HTB3 but not of HTB2 and HTB4 cells with extracellular thymidine resulted in enhanced sensitivity to gemcitabine. Our results highlight the importance of hENT1 and TK2 activities in response to gemcitabine. Elevated TK expression in squamous cell carcinomas warrants further study and offers new insights into rational treatment strategies based on bladder cancer phenotype.