INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The reverse transcriptase (RT) of human immunodeficiency virus (HIV) type 1 (HIV-1) is a multifunctional enzyme which catalyzes the conversion of viral genomic RNA into double-stranded proviral DNA. RT is an important target for antiviral chemotherapy, and two classes of inhibitors are currently being used in treatment. Nonnucleoside RT inhibitors (NNRTIs) bind directly to a hydrophobic pocket adjacent to the polymerase active site of RT. Nucleoside RT inhibitors (NRTIs) competitively inhibit reverse transcription by competing with native deoxynucleotide triphosphates (dNTPs) and cause chain termination when incorporated into nascent proviral DNA due to a lack of a 3' hydroxyl group (3, 15).

The emergence of drug-resistant variants of HIV-1, observed in patients undergoing prolonged antiviral therapy, is a factor implicated in treatment failure and can be selected for in tissue culture in vitro (12, 17, 27, 36, 37, 41). RT exhibits lower fidelity and processivity than cellular DNA polymerases and lacks a 3'-to-5' proofreading activity, resulting in a high mutation rate and heterogeneous viral populations (21, 22, 30, 32, 42). Sequence data have shown that point mutations that cause single amino acid substitutions within RT are responsible for the drug resistance phenotypes (12, 27, 36, 37). Mutations in RT which confer resistance to NRTIs map to the fingers and palm subdomains of RT, including both the dNTP binding site and the polymerase active site.

Different NRTIs select for specific mutations in RT, resulting in different mechanisms of resistance to individual compounds (38). For example, single mutations are sufficient to generate resistance to 2',3'-dideoxycytidine (ddC), 2',3'-dideoxyinosine (ddI), and the () enantiomer of 2'-deoxy-3'-thiacytidine (3TC) through decreased incorporation of the respective dNTPs. 3TC-resistant variants containing the substitutions M184I and M184V at the polymerase active site have been isolated both in cell culture and from patients undergoing therapy with this drug (25, 36). The M184V mutation is associated with high-level resistance to 3TC and lower-level cross-resistance to both ddC and ddI (25, 36). In contrast, the occurrence of high-level resistance to 3'-azido-3'-deoxythymidine (AZT) commonly requires the accumulation of mutations M41L, D67N, K70R, L210W, T215Y/F, and K219Q (6, 16, 18, 24, 27). Biochemical studies have shown that the D67N and K70R mutations increase the pyrophosphorolysis of AZT triphosphate (AZT-TP), the reverse reaction of polymerization. The T215Y and K219Q mutations increase the processivity of RT and decrease the dissociation rate of template-primer from RT (1, 2). In combination, the mutations generate a free 3' hydroxyl group from AZT-TP-terminated products and increase the rate of polymerization.

The (+) and () enantiomers of a novel nucleoside inhibitor, 2'-deoxy-3'-oxa-4'-thiocytidine (dOTC), and its fluorinated derivative (dOTFC) are structurally similar to those of 3TC but contain differences in the structure of the sugar moiety and a fluorinated base. The structures of the compounds are illustrated in Fig. 1. We have previously shown that they possess antiviral activity in vitro (28, 29, 31). The geometry of the sugar in 3TC is important for the development of resistance. Therefore, we were interested in determining how the structural differences between dOTFC, dOTC, and 3TC would affect the mutation patterns of viruses resistant to each of the compounds.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Chemical structures of the (+) and () enantiomers of dOTC and its fluorinated derivative dOTFC. (A) ()dOTC. (B) (+)dOTC. (C) ()dOTFC. (D) (+)dOTFC.

(This work was performed by N. Richard in partial fulfillment of the Ph.D degree, Faculty of Graduate Studies and Research, McGill University, Montréal, Quebec, Canada).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Viruses and cells. The human T-cell line MT-4 was used to grow both wild-type and resistant variants of HIV-1 and was cultured in RPMI 1640 medium (Gibco-BRL, Mississauga, Ontario, Canada) with 10% heat-inactivated fetal calf serum (Flow Laboratories, Toronto, Ontario, Canada), L-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 µg/ml) as described previously (12). Human cord blood mononuclear cells (CBMCs; obtained from the Department of Obstetrics of the Jewish General Hospital) were isolated by Ficoll-Hypaque centrifugation and were stimulated with 0.1% phytohemagglutinin for 3 days and cultured in RPMI 1640 medium supplemented with interleukin 2 as described previously (11). The infectious HXB2D molecular clone of HIV-1 was used in these studies (10), and viral stocks were generated by transfection of cloned plasmid DNA into MT-4 cells by electroporation (12). Several clinical viral isolates were used in this study, and these were obtained by coculture of patient peripheral blood mononuclear cells (PBMCs) with healthy CBMCs as described previously (11).

Drugs. 3TC and the (+) and () enantiomers of each of dOTC and dOTFC were obtained from BioChem Therapeutic Inc., Laval, Quebec, Canada.

Selection of drug-resistant variants. MT-4 cells (2.5 × 10⁵) were preincubated in the presence or absence of subinhibitory concentrations of drug for 2 h. The cells were washed and infected with 10⁶ RT units of HIV-1 for 2 h. The cells were then washed and maintained in the presence or absence of drug. Cells were monitored for cytopathic effects (CPEs), and culture supernatants were clarified once 50% of cells displayed CPEs. A new round of infection was initiated by infecting fresh MT-4 cells with 0.5 ml of clarified culture supernatant. The concentration of drug was increased gradually over 12 passages from 1 to 200 µM. Drug susceptibility assays were then performed to determine the 50% inhibitory concentration (IC₅₀) of each drug by measuring the RT activity in culture fluids through the incorporation of [³H]dTTP (ICN, Montreal, Quebec, Canada) by using poly(rA) · oligo(dT)_(12-18) (Pharmacia, Mississauga, Ontario, Canada) as template-primer, as described previously (7).

Cloning and sequencing. Cellular DNA was extracted with phenol-chloroform from uninfected MT-4 cells and cells infected with the drug-selected variants following lysis in 50 mM Tris-HCl-1 mM EDTA (pH 8.0)-0.5% sodium dodecyl sulfate and overnight incubation with 1 mg of proteinase K per ml at 4°C. The DNA was precipitated with ethanol and was resuspended in TE buffer (50 mM Tris-HCl, 1 mM EDTA [pH 8.0]). The complete RT-coding regions of the drug-selected variants were amplified by PCR with the primers 5'-GTAGAATTCTGTTGACTCAGATTGG-3' and 5'-GATAAGCTTGGGCCTTATCTATTCCAT-3' and were cloned by using the TA Cloning kit (Invitrogen, Carlsbad, Calif.). The sequence of the RT region was determined by using the double-stranded DNA Cycle Sequencing System (Gibco-BRL) with six different primers as described previously (15).

Site-directed mutagenesis and generation of mutant HXB2D viral stocks. The D67G mutation was introduced into plasmid pHIVpol, which contains the RT-coding region, by using the Quickchange mutagenesis kit (Stratagene, La Jolla, Calif.) and the primers 5'-CCATTTAGTACTGCCTTTTTTCTTTATGGC-3' and 5'-CAAAAGTGCCACCTGAATTCTAAGAAACCAT-3'. The mutation Met-184Ile was introduced into pHIVpol by cloning a 1.0-kb EcoRV-PflMI fragment from a positive TA clone. Full-length HXB2D was generated by cloning a 1.9-kb MscI fragment from pHIVpol into HXB2D as described previously (14). HXB2D-M184V was generated as described previously (15). Viral stocks of the mutated HXB2D clones were generated by transfecting MT-4 cells with plasmids encoding mutated proviral DNA by electroporation, and culture fluids containing virions were harvested and clarified.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Selection for resistance with () and (+) dOTFC. The HXB2D wild-type molecular clone of HIV-1 was used to select for resistance to the nucleoside analogues ()dOTFC and (+)dOTFC in MT-4 cells. A final concentration of 200 µM was achieved for each compound over 12 passages over 3 months, and the susceptibilities of the drug-selected viruses are summarized in Table 1. A 10-fold decrease in susceptibility compared to that of the wild type was observed for the virus selected with ()dOTFC. Cloning and sequencing of the complete RT-coding region of this virus revealed that 14 of 14 clones contained the substitution MetVal at position 184, located within the YMDD motif of the polymerase active site of RT. HXB2D virus selected with (+)dOTFC was also found to be 10-fold more resistant to the compound than the wild type, and the novel mutation D67G, which is located in the fingers domain of RT, was identified in five of five RT clones.

View this table: [in this window] [in a new window]   TABLE 1.   Selection of drug-resistant HIV HXB2D variants with (+)dOTFC and ()dOTFC in MT-4 cells

Resistance of recombinant mutated HXB2D-D67G to nucleoside analogues. In order to confirm the biological significance of the D67G substitution, recombinant HXB2D containing the mutation was generated by site-directed mutagenesis as described in Materials and Methods. Mutation D67G is novel and has not been reported previously; however, mutation D67N, in combination with other mutations, is implicated in AZT resistance (6, 24, 27). The susceptibilities of wild-type HXB2D and HXB2D-D67G to the novel nucleosides were determined in MT-4 cells and are reported in Table 2. The D67G substitution conferred approximately fivefold resistance to (+)dOTFC, confirming the tissue culture selection data. D67G also conferred fivefold resistance to ()dOTFC, but viruses with this mutation displayed wild-type sensitivity to both (+)dOTC and ()dOTC as well as to 3TC. These data suggest that the D67G mutation is important for conferring resistance to the fluorinated derivatives of the dOTC compounds.

Resistance of recombinant mutated HXB2D-M184I and HXB2D-M184V to dOTFC compounds. In order to confirm the significance of mutations at position 184, recombinant HXB2D clones containing the M184I or M184V mutation were generated as described in Materials and Methods. These mutations have previously been reported to confer moderate (M184I) and high-level (M184V) resistance to 3TC. In addition, M184V has been shown to confer low-level cross-resistance to ddI, ddC, dOTC, and abacavir (5, 11, 23, 31, 35, 36, 39, 40). Table 4 summarizes the sensitivities of wild-type HXB2D and mutated clones HXB2D-M184I and HXB2D-M184V to the dOTFC compounds in both MT-4 cells and CBMCs. Since MT-4 cells were used for the drug resistance selections, these were the first cells studied. The M184I mutation had little effect on the susceptibility to either ()dOTFC or (+)dOTFC in MT-4 cells, as indicated by an increase in the IC₅₀ of only fourfold, although it did confer resistance to 3TC (Table 4). The M184V mutation, however, conferred greater than 13-fold greater resistance to ()dOTFC and greater than 15-fold greater resistance to (+)dOTFC, and viruses with this mutation were highly resistant to 3TC; i.e., they had >600-fold greater resistance. Although the M184V mutation was originally selected in those experiments with ()dOTFC, it also confers resistance to (+)dOTFC.

Susceptibilities of clinical isolates to dOTFC. Clinical isolates which display wild-type, 3TC resistance, or AZT resistance phenotypes were obtained from patients. The sensitivities of the clinical isolates are summarized in Table 5. Isolates 4242 and 4246 were obtained from drug-naive patients and were sensitive to all the compounds tested. These viruses did not contain any drug resistance-associated mutations. Isolates 3350, 3887, and 4205 were obtained from patients who had undergone 12, 12, and 52 weeks of 3TC monotherapy, respectively, and the IC₅₀s for the viruses ranged from 10 to greater than 100 µM for 3TC, whereas they were 0.2 and 0.025 µM for wild-type isolates 4242 and 4246, respectively. Isolates 3887 and 3350 contained the M184V substitution, while isolate 4205 contained both the M184V and the T69D mutations. Isolate 3887 was sensitive to ()dOTFC but was moderately resistant to (+)dOTFC, whereas isolates 3350 and 4205 were moderately resistant to both ()dOTFC and (+)dOTFC. All of the isolates were sensitive to AZT. Isolates 1075 and 4170 were obtained from patients who had undergone more than 1 year of AZT monotherapy. The isolates were AZT resistant and displayed low-level cross-resistance to (+)dOTFC. Isolate 4170 contained the K70R mutation, and isolate 1075 contained both the M41L and the T215Y substitutions.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the studies described here we have selected for HIV-1 resistance to (+)dOTFC and ()dOTFC in MT-4 cells. Selection with ()dOTFC generated the M184V mutation following 12 passages with ()dOTFC over 3 months. HXB2D-M184V was also found to be cross-resistant to (+)dOTFC. M184V also confers high-level resistance to 3TC and low-level cross-resistance to ddC, dOTC, and ddI (25, 31, 36). Amino acid position 184 lies within the polymerase active site 183YMDD186 of RT. Crystallographic and modeling studies show that the methionine side chain at position 184 is in contact with the base and sugar of the terminal dNTP. The branched side chains of isoleucine and valine are also in contact with the sugar ring and cause interference with the modified oxathiolane ring of 3TC (20, 34). Since the dOTFC compounds also contain an oxathiolane ring, the mechanism of resistance is likely the same as that for 3TC.

In patients undergoing 3TC monotherapy, the M184I mutation appears first and is eventually supplanted by M184V. In primary cells, the two mutations are generated independently from the wild-type codon by ATGATA in the case of M184I and ATGGTG in the case of M184V (25, 36). It has been shown in vitro that the frequency of GA substitutions is greater than that of AG substitutions, resulting in the M184I mutation occurring first (19, 25). However, the M184V mutation yields viruses with greater fitness than does the M184I mutation, and viruses containing this mutation will eventually dominate (4). We did not identify mutation M184I in our virus selections; however, HXB2D-M184I displayed decreased susceptibility to both ()dOTFC and (+)dOTFC. M184I may have transiently appeared in earlier passages and may have been replaced by M184V by passage 12.

Different phenotypes for HXB2D-M184I and HXB2D-M184V were observed in CBMCs and MT-4 cells. Neither mutated virus conferred resistance to ()dOTFC or (+)dOTFC in CBMCs, although both were resistant to 3TC. The same phenotype has previously been observed with ddI and ddC, and the level of resistance was also greater in MT-4 cells (33). Differences in cellular metabolism may partially account for the different resistance phenotypes and affect the size of the dNTP pools, which is important for reverse transcription and/or the rate of viral replication (3, 4, 13, 33). Differential viral replication rates in different cell types may be a factor since it has been shown that viruses that contain the M184I and M184V mutations display a greater growth defect in PBMCs than in a T-cell line compared to the growth of the wild type (4). Viral genotypes may also play an important role since clinical isolates resistant to 3TC were less susceptible to ()dOTFC and (+)dOTFC in CBMCs, even though recombinant HXB2D-M184V was not. There was a correlation between the level of 3TC resistance and the degree of cross-resistance to (+)dOTFC and ()dOTFC with the clinical isolates that contained the M184V substitution in RT. These data support our findings from studies with MT-4 cells that mutations at position 184 are important for resistance to the dOTFC compounds. However, the M184V mutation alone is not sufficient to confer high-level resistance to (+)dOTFC or ()dOTFC in CBMCs.

Selection with (+)dOTFC over 12 passages yielded the novel mutation D67G. HXB2D-D67G was also cross-resistant to ()dOTFC but not to the dOTC compounds, suggesting that the mutation is specific for the fluorinated analogues of dOTC. D67G did not confer cross-resistance to 3TC or the other RT inhibitors tested. Although D67G is a novel mutation in drug resistance selection studies, the polymorphisms D67G/E/S have been observed in combination with amino acid insertions between codons 67 and 70 that are involved in multinucleoside resistance (26). Since M184V conferred greater resistance to (+)dOTFC than D67G did, HXB2D-D67G was further selected with (+)dOTFC, and three of six variants sequenced contained both the D67G and the M184V mutations. The other three clones maintained the D67G mutation alone. Both mutations were generated by an AG substitution. The reason that the D67G mutation would appear first is unclear. Studies of the growth kinetics and fitness of viruses containing these mutations and the characterization of purified mutated RTs will be important in understanding the mechanism of resistance to (+)dOTFC. Studies are in progress to determine the earliest passage number at which each of the D67G and M184V mutations occur in culture as a consequence of the selective pressure imposed by (+)dOTFC and ()dOTFC. These studies represent an extensive evaluation that uses each of (+)dOTFC and ()dOTFC alone and in combination with other nucleosides. Preliminary findings based on CalcuSyn software analysis indicate that both (+)dOTFC and ()dOTFC have additive effects in culture with AZT; these results will be reported separately.

The mechanism of AZT resistance is complex and until recently was poorly understood. Amino acid position 67 is in the 3-4 connecting loop in the fingers subdomain of RT (20, 38). Recent studies suggest that the 3-4 connecting loop is in closer proximity with the incoming nucleotide than was previously thought and may affect dNTP binding (20). D67N alone is not sufficient for conferring significant resistance to AZT and is found in combination with the M41L, K70R, L210W, T215Y/F, and K219Q mutations when resistance to this compound is present (6, 16, 18, 24, 27). Biochemical studies have shown that the D67N and the K70R mutations increase the rate of pyrophosphorolysis of AZT-TP, the reverse reaction of polymerization which removes the terminal dNTP from the primer (1, 2). Increased pyrophosphorolysis is compensated by the mutations T215Y and K219Q, which increase the processivity of RT and decrease the rate of dissociation of AZT chain-terminated DNA from mutated RT (8, 9). In combination, the mutations allow efficient reverse transcription and confer high-level resistance to AZT. Studies to determine the effects of the D67G mutation on the rate of pyrophosphorolysis are under way.