Human immunodeficiency virus type 1 induces 1-beta-D-arabinofuranosylcytosine resistance in human H9 cell line.

We have found that chronically HIV-1(IIIB)-infected H9 cells showed 21-fold resistance to 1-beta-D-arabinofuranosylcytosine (ARA-C) compared with uninfected H9 cells. In the infected H9 cells, a 37% increase of dCTP pool and a 34% increase of dATP were observed, and no alteration of dTTP and dGTP was observed, compared with the uninfected H9 cells. A marked decrease of ARA-CTP generation was observed in the infected H9 cells after 3-h incubation with 0.1-10 microM ARA-C. The level of deoxycytidine kinase activity with ARA-C as substrate was similar in both the infected and the uninfected cells; however, a 37-fold increase of cytidine deaminase activity was observed in the infected H9 cells. These results indicate that the induction of cytidine deaminase activity by HIV-1(IIIB) infection conferred ARA-C resistance to H9 cells. This conclusion was supported by the observation that a marked reversal of ARA-C resistance in the infected H9 cells occurred after treatment with the inhibitor of cytidine deaminase, 3,4,5,6-tetrahydrouridine. The understanding of these cellular alterations in drug sensitivity may facilitate the development of effective therapeutic strategies against HIV-1-infected cells.

For the therapy of acquired immunodeficiency syndrome (AIDS),' a number of antiviral agents has now been considered to inhibit infection and proliferation of human immunodeficiency virus (HIV-1) (1). Usually, these anti-HIV agents have no specific cytotoxic effect on the HIV-1-infected cells. The CD4+ T lymphocyte is the principal target (2) and major reservoir for HIV-1 in the peripheral blood compartment of infected individuals (3,4). Other CD4+ cells including monocyte/macrophages, dendritic cells, Langerhaus cells, and thymocyte precursor cells can be infected with HIV-1 (2, 5, 6). It has been shown that HIV-infected macrophages can transmit HIV-1 through cell-cell contact to susceptible T cells, acting as a potential reservoir for HIV-1 transmission (7). To eliminate these infected cells, it is necessary to con-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1-P-D-Arabinofuranosylcytosine (ARA-C) is a nucleoside analog of deoxycytosine, and one of the most effective agents for treatment of acute leukemia. ARA-C is phosphorylated to ARA-CTP, the active metabolite for inhibition of DNA synthesis (8). We have found that the chronically HIV-l(II1B)infected H9 cells showed a 21-fold resistance to 1-8-D-arabinofuranosylcytosine, compared with the uninfected H9 cells. Our results indicated that biochemical alterations associated with pyrimidine metabolism occurred as a result of HIV-1 infection in the host cells.  University). Cells were cultured in RPMI 1640 (GIBCO) medium supplemented with 10% fetal bovine serum and kanamycin (100 gg/ ml). H9 cells were infected with HIV-l(II1B) and cultured for I month in complete medium and used in the following experiments.

Chemi~als-['~C]Deoxycytidine
Cytotoxicity of Reagents-The cytotoxic activity of the drug was measured by determining the IC50 (concentration of drug required for 50% inhibition of cell growth) using XTT assay, as described previously (11).
Determination of ARA-CTP Formation and Deoxyribonucleotide Pools-Deoxyribonucleotide pools were determined as described by Garrett and Santi (12) with slight modification. Briefly, the nucleotides were extracted from exponentially growing cells (3 X lo7) by exposure to 0.6 M triethyl chloroacetate for 10 min at 4 "C. After removal of cell debris, the supernatant was neutralized with 0.18 M trioctylamine in trichlorotrifluoroethane. The ribonucleotides in cell extracts were destroyed by treatment with sodium periodate. Nucleotides were separated and quantified by high performance liquid chromatography (HPLC) on Partisil SAX (250 X 4.6 mm)(GL Science Inc.). The column was eluted with 0.4 M ammonium phosphate (pH I.O)-acetonitrile mixture (101, v/v) at a flow rate of 2.0 ml/min.
For measurement of ARA-CTP formation, exponentially growing cells (3 X 10') were preexposed to ARA-C for 3 h, and the nucleotides were extracted as described above. After treatment of the cell extract with 0.6 M triethyl chloroacetate, an aliquot of the fraction was submitted to HPLC with a Partisil-10 SAX column equilibrated with buffer A (7 mM ammonium phosphate, pH 3.8). After elution with buffer A for 10 min, a linear gradient was run from 100% buffer A to 100% buffer B (250 mM ammonium phosphate, 0.5 M potassium chloride, pH 4.5) over 30 min, followed by isocratic elution for the next 10 min at a flow rate of 3 ml/min. Uptake of [3H/ARA-C-Cell suspensions of HIV-l(II1B)-infected and uninfected cells (2 X IO6 cells/ml) in the growth medium were incubated at 37 "C with [3H]ARA-C. After incubation for 3 h, the cells were collected and treated with 0.5 M KOH for 1 h at 60 "C, and the radioactivity in the cells was determined by liquid scintillation counting as described previously (13). Enzyme Assay-For estimation of the dCyd kinase activity, 5 X IO7 cells were resuspended in 350 pl of extraction buffer (composed of 20 mM Tris-HC1 (pH 7.5), 100 mM KCl, 20 mM NaC1,4 mM MgC12, 1 mM CaC12, and 10% glycerol). The cells were disrupted by freezing and thawing. Cell debris was removed by centrifugation, and samples were stored at -80 "C. The dCyd kinase assay was performed according to the method described by Cheng et al. (14). The dCyd kinase activity was measured with either deoxycytidine or ARA-C as substrate. The crude cell extract was incubated with the substrate for 20, 40, and 60 min. dCyd kinase activity was calculated by linear regression analysis of the time-response curves.
For the estimation of cytidine deaminase activity, 5 X lo7 cells were resuspended in 350 hl of extraction buffer (50 mM Tris-HC1 pH 7.5, 2 mM dithiothreitol, and 10% glycerol). The cells were disrupted by freezing and thawing. Cell debris was removed by centrifugation, and samples were stored at -80 "C. The cytidine deaminase assay was performed according to the method described by Chabner et al. (15). The crude cell extract was incubated with the substrate for 20, 40, and 60 min. Cytidine deaminase activity was calculated by linear regression analysis of the time-response curves.

RESULTS AND DISCUSSION
Comparisons of the concentration of drug required to inhibit cell growth by 50% (ICso) on chronically HIV-l(II1B)infected and uninfected H9 cells are shown in Table I. The of ARA-C in the chronically HIV-1-infected H9 cells was 21-fold higher than that in the uninfected H9 cells. The infected H9 cells showed 4.1-and 3.3-fold resistance to ARA-C derivatives, N4-behenoyl-ARA-C and N4-palmitoyl-ARA-C, respectively, but did not show resistance to anti-HIV-1 agents including 3'-azido-3'-deoxythymidine and 2',3'-deoxycytidine, and other inhibitors for DNA synthesis including 5-fluorouracil, methotrexate, 6-mercaptopurine, and aphidi-Colin. Chronically HIV-l(II1B)-infected human cell lines, MOLT-4 clone 8 and U937, also showed 23-and 21-fold resistance to ARA-C, respectively (data not shown). These results indicated that biochemical alterations associated with deoxycytidine metabolism occurred through HIV-1 infection in the host cells.
There was no statistical difference in accumulation of ARA-C between the infected and the uninfected H9 cells (Table   TABLE I Growth inhibitory effect of drugs on HIV-l(IIIB)-infected and uninfected H 9 cells Cells (2 X 103/well) were cultured for 5 h in a 96-well plate, and graded concentrations of drugs were added. After 72 h of continuous drug exposure, growth of the cells was measured by the XTT-assay (111, and concentrations of drug producing 50% cell growth inhibition (IC60) were determined. Relative resistance was calculated as the ratio of the IC,, obtained for HS/IIIB to the ICso for the uninfected H9 cells. Ratios based on significantly different ICso values ( p < 0.05 paired t test) are indicated with an asterisk.  Table 11. A 37% increase of the dCTP pool was observed in the infected cells, but the increase could not explain the induction of 21-fold resistance to ARA-C relative to the uninfected H9 cells. A 34% increase in the dATP pool was also observed, but there was no alteration in the dTTP or dGTP pools. ARA-C is phosphorylated to ARA-CTP by kinases of the salvage pathway for nucleotide synthesis (8). In the infected H9 cells, a marked diminution of ARA-CTP generation was observed, compared with the uninfected H9 cells (Fig. 1); after 3 h of incubation with ARA-C at concentrations of 0.1 p~, detectable ARA-CTP levels were not produced in the infected cells, whereas 32 k 5.0 pmol/107 cells of ARA-CTP were produced in the uninfected cells. At 1 and 10 p~, ARA-CTP levels in the infected H9 cells were only 4 and 18% of that found in the uninfected H9 cells.
The rate-limiting step in ARA-C activation is the process of phosphorylation catalyzed by dCyd kinase. With deoxycytidine as substrate, dCyd kinase activity in the infected H9 cells showed a 12-fold increase compared with the uninfected H9 cells (Table 11). With ARA-C as substrate, however, there was no difference in dCyd kinase activity between the infected and the uninfected cells, indicating that dCyd kinase without specificity for ARA-C was induced in the infected cells. These results indicate that phosphorylation of ARA-C was not the critical step for ARA-C resistance.
ARA-C is inactivated to 1-P-D-arabinofuranosyluridine by cytidine deaminase. We next determined cytidine deaminase activity in the infected H9 cells. Marked increase (37-fold) of cytidine deaminase activity was observed in the infected cells (Table II), indicating that induction of ARA-C resistance in the infected cells was due to increase of cytidine deaminase activity. Virus infection was immediately followed by an in-  Deoxyribonucleotide pools were determined as described by Garrett and Santi (12). Deoxyribonucleotides were separated and quantified by HPLC on Partisil SAX (250 x 4.6 mm). Values represent the mean k S.D. obtained from three separately treated aliquots of the cell extract.
The dCyd kinase assay was performed according to the method described by Cheng et al. (14).
Ratios based on significantly different values for H9/IIIB versus the infected cells ( p < 0.005, paired t test).
'The cytidine deaminase assay was performed according to the method described by Chabner et al. (15).   infection in H9 cells. H9 cells were exposed to HIVl(I1IB) at multiplicity of infection of 0.1. From day 2 postinfection, an equal volume of fresh medium was added every day to replace supernatants which were sampled. p24 gag of HIV-1 in the supernatants was measured using antigen capture enzyme-linked immunosorbent assay (0). The cytidine deaminase activity was measured according to the method described by Chabner et al. (15). Relative cytidine deaminase activity compared to chronically infected H9 cells was plotted (0). crease of cytidine deaminase activity in H9 cells (Fig. 2). This result indicates that the increase of cytidine deaminase activity in the H9 cells was induced as a result of the infection, rather than as a result of selection of the cell population during establishment of the chronically infected cell line. The infected H9 cells showed relatively lower (5.0and 3.3-fold) resistance to the ARA-C derivatives, N4-behenoyl-ARA-C, and N'-palmitoyl-ARA-C, respectively (Table I). These results are consistent with previous reports that N4-behenoyl-ARA-C and N4-palmitoyl-ARA-C were resistant to deamination compared with ARA-C (16,17). To confirm the resistance mechanism, the effect of a cytidine deaminase inhibitor on cytotoxicity of ARA-C was examined. It has been shown that 3,4,5,64etrahydrouridine (THU) is an effective competitive inhibitor of cytidine deaminase (18). At concentrations of 10 PM THU, ARA-C resistance decreased to 4.6-fold and THU at 100 ~L M could completely reverse ARA-C resistance in the infected cells (Fig. 3). There was no detectable growth inhib- itory effect on the infected and uninfected H9 cells by 3-day exposure to 10-600 PM THU (data not shown). These results indicate that THU enhanced the cytotoxicity of ARA-C in the infected cells by inhibition of cytidine deaminase. While additional study will be required to fully understand mechanism of cytidine deaminase induction by HIV-1 infection, our results show that infection with human retrovirus can be a significant drug resistance factor.
During progression of AIDS, HIV-1 is harbored in CD4+ T cells which act as the primary reservoir for virus transmission (3, 4). HIV-1-infected monocyte/macrophages were involved in HIV-1 infection in specific body compartments such as brain and lung (19,20). One approach to the treatment of patients with AIDS could be the eradication of these infected cells by chemotherapy. The understanding of the cellular alterations in drug sensitivity may facilitate the development of effective therapeutic strategies against HIV-1-infected cells.