Hydroxyurea Enhances the Activities of Didanosine, 9-[2-(Phosphonylmethoxy)ethyl]adenine, and 9-[2-(Phosphonylmethoxy)propyl]adenine against Drug-Susceptible and Drug-Resistant Human Immunodeficiency Virus Isolates

Abstract

We assessed the effects of hydroxyurea (HU) at a concentration of 50 μM on the in vitro activities of 2′,3′-dideoxyinosine (ddI), 9-[2-(phosphonylmethoxy)ethyl]adenine (PMEA), and 9-[2-(phosphonylmethoxy)propyl]adenine (PMPA) against a wild-type human immunodeficiency virus (HIV) type 1 (HIV-1) laboratory isolate and a panel of five well-characterized drug-resistant HIV isolates. Fifty micromolar HU significantly increased the activities of ddI, PMEA, and PMPA against both the wild-type and the drug-resistant HIV-1 isolates. In fixed combinations, both ddI and PMEA were synergistic with HU against wild-type and drug-resistant viruses.

The experimental human immunodeficiency virus (HIV) reverse transcriptase (RT) inhibitors 9-[2-(phosphonylmethoxy)ethyl]adenine (PMEA) and 9-[2-(phosphonylmethoxy)propyl]adenine (PMPA) are acyclic phosphonate analogs of AMP (3, 26, 32). Although they already contain a single phosphate, PMEA and PMPA, like other nucleoside analogs, rely on intracellular kinases for phosphorylation to their active diphosphate forms (27, 28). The diphosphates of PMEA and PMPA and the triphosphate of 2′,3′-dideoxyinosine (ddI) (ddATP) compete with the cellular nucleotide dATP for the active binding sites on the RT enzyme (1, 8). Therefore, the antiretroviral activities of PMEA, PMPA, and ddI are dependent on two factors: (i) the activities of intracellular phosphorylating enzymes and (ii) the ratio of the amount of phosphorylated drug to the amount of competing intracellular nucleoside triphosphate pools.

The anticancer agent hydroxyurea (HU) is used for the treatment of myleoproliferative disorders (9, 34). HU is a potent inhibitor of the cellular enzyme ribonucleotide reductase, which catalyzes the reduction of ribonucleotides to deoxyribonucleotides (14). Cells exposed to HU show measurable reductions in several deoxynucleotide pools, with the reduction of dATP pools being the most pronounced (4, 10–12, 24). These decreases in deoxynucleotide pools effectively block cellular DNA synthesis (4).

HU increases the anti-HIV activities of ddI and 2′-β-fluoro-2′,3′-dideoxyadenosine, probably due to the favorable shift in the ratio of adenosine drug triphosphates versus competing cellular dATP pools which favors the binding of drug triphosphates to RT (4, 10–13, 18, 24). Due to these promising in vitro results, several clinical trials of ddI in combination with HU have been initiated (5–7, 17, 35, 36).

In the present study, we investigated the effects of HU on the anti-HIV activities of the three adenosine analogs PMEA, PMPA, and ddI. We assessed the interaction of HU with these drugs against wild-type HIV and versus a panel of drug-resistant HIV strains. We also analyzed the cytotoxicity of HU alone and in combination with PMEA, PMPA, or ddI.

HIV-1 strains.

The antiviral activities of the drugs and drug combinations were assessed against six different HIV type 1 (HIV-1) strains: a wild-type laboratory isolate (HIV_NL4-3), three recombinant isolates containing ddI resistance mutations (HIV_K65R, HIV_L74V, and HIV_{L74V, M184V}), one molecularly constructed multinucleoside-resistant strain (HIV_{V75I, F77L,F116Y, Q151M}) (15), and a recently reported multidrug-resistant clinical isolate containing six major RT mutations (HIV_{M41L, D67N, M184V, L210W, T215Y, K219N}) (30).

Sequence analysis of HIV-1 strains.

A 1.3-kb fragment of cDNA encompassing HIV-1 protease and the first 300 codons of RT was sequenced from each cultured supernatant as described previously (38). Briefly, purified viral RNA (Qiagen Viral RNA Extraction Kits Qiagen, Chatsworth, Calif.) was reverse transcribed and amplified by PCR with the Superscript-One-Step-RT-PCR Reagent (Life Technologies, Gaithersburg, Md.) and two primers, MAW-26 and RT21 (23). A 5-μl aliquot of the first PCR product was used for a second-round nested PCR with primers PRO-1 (29) and RT20 (23). Approximately 70 ng of the 1.3-kb product was sequenced by dye-labelled dideoxyterminator cycle sequencing (Applied Biosystems, Foster City, Calif.). Isolate sequences were compared to both patient plasma sequences and the consensus B sequence from the Los Alamos HIV Sequence Database (21).

Drug susceptibility assays.

In vitro drug susceptibility assays were performed by a modified AIDS Clinical Trials Group–U.S. Department of Defense consensus method (virology manual for ACTG HIV laboratories, 1997). Peripheral blood mononuclear cells (PBMCs) were preinfected with titrated viral stocks for 4 h at 37°C in a humidified atmosphere of 5% CO₂. Each microtiter plate well contained 100,000 preinfected PBMCs and eight serial drug dilutions in cell media of ddI, PMEA, PMPA, 3′-azido-3′-deoxythymidine (AZT), 2′-deoxy-3′-thiacytidine (3TC), or indinavir (IDV) in the presence or absence of 50 μM HU. A 50 μM concentration of HU was used since it is in the range of the average steady-state HU concentration in serum during HIV treatment (35 to 56 μM) (37). An 8:1 series of combinations of ddI and HU or PMEA and HU was also analyzed. The drug dilutions were chosen to span the 50% inhibitory concentration (IC₅₀) of each single drug (2, 3, 25, 26, 32). The drugs were combined in fixed clinically achievable ratios, based on the relative potencies of the drugs, by the median-effect method of analyzing drug interactions. Control wells containing cells and virus were coincubated on each plate.

To enable assay standardization and comparison, the 50% tissue culture infective dose of each isolate was maintained at between 30 and 100. After a 7-day incubation at 37°C and in a humidified atmosphere of 5% CO₂, viral growth was determined by a p24 antigen assay with supernatants (Dupont Pharmaceuticals, Wilmington, Del.). The percent inhibition of viral growth compared to the viral growth in the control wells without drugs was calculated. Results were expressed as the mean IC₅₀ of four to six values obtained in two to three different experiments per isolate.

The results for the two-drug combinations were calculated by using a computer program that follows the median-effect principle. The computer constructs a median-effect plot of log fraction affected/fraction unaffected against the log of the dose of the two separate drugs and the dose of the combination. A combination index (CI), which compares the amount of drug which gives a 50% effect when used in combination with that which gives a 50% effect when the drug is used alone, is calculated. A CI of <1 indicates synergy, and a CI of >1 indicates antagonism. However, in accordance with variation in raw data, CIs of between 0.8 and 1.2 were considered to represent additivism.

Cytotoxicity assays.

Thymidine uptake analyses were used to assess the effects of the drugs on cellular DNA synthesis. Phytohemagglutinin-stimulated PBMCs were plated at 100,000 cells per well and exposed to 3.1, 12.5, and 50 μM concentrations of ddI, PMEA, or PMPA in the presence or absence of HU at concentrations of 25 to 500 μM. Control cells, without drugs, were coincubated on each plate and were used for comparison when measuring inhibition of cellular DNA synthesis by the drugs. The plates were incubated at 37°C for 7 days. Sixteen hours prior to cell harvest, 50 μCi of [³H]thymidine was introduced into all wells. The cells were harvested onto preprinted filter paper with rinsings of water and 95% ethanol. After drying at 37°C, scintillant was added and the counts on the filters were determined with a Wallac beta counter (LKB Wallac, Turku, Finland).

The presence of 50 μM HU decreased the IC₅₀s of ddI in vitro, which enhanced the anti-HIV activity of ddI against all viral strains analyzed (Table 1). The IC₅₀ of ddI for many of the ddI-resistant viral strains was reduced to less than the range for the wild type in the presence of HU. These observations are consistent with previous studies showing that HU at concentrations of 50 to 100 μM increased the activity of ddI (11, 12). Moreover, recent clinical studies have shown that patients who respond to ddI and HU therapy may harbor ddI-resistant viral strains (7). Further in vitro analysis found that these resistant strains are phenotypically sensitive to inhibition by ddI and HU (19). Similar to ddI, the IC₅₀s of the acyclic adenosine derivatives PMEA and PMPA were decreased by the presence of 50 μM HU for all viruses analyzed, including five drug-resistant strains: HIV_K65R, HIV_L74V, HIV_{L74V, M184V}; HIV_{V75I, F77L, F116Y, Q151M}; and HIV_{M41L, D67N, M184V, L210W, T215Y, K219N}.

TABLE 1.

Effect of HU on antiviral activities of ddI, PMEA, and PMPA against a wild-type laboratory isolate and drug-resistant HIV-1 strains

Viral strain or isolate and presence of HU	IC50 (μM)a in culture
	ddI	PMEA	PMPA
HIV-1NL4-3
Without HU	0.38 ± 0.06	3.1 ± 0.77	1.3 ± 0.02
With 50 μM HU	<0.05 (>8)b	0.37 ± 0.12 (8)	<0.05 (>26)
HIV-1K65Rc
Without HU	2.3 ± 0.8	13.4 ± 1.3	4.9 ± 0.9
With 50 μM HU	<0.05 (>46)	0.10 ± 0.1 (134)	0.08 ± 0.1 (61)
HIV-1L74V
Without HU	1.2 ± 0.3	4.0 ± 0.2	1.7 ± 0.5
With 50 μM HU	<0.05 (>24)	0.29 ± 0.03 (14)	<0.05 (>34)
HIV-1L74V, M184V
Without HU	3.9 ± 0.06	1.6 ± 0.5	1.1 ± 0.1
With 50 μM HU	<0.05 (>78)	<0.05 (>32)	<0.05 (>22)
HIV-1V75I, F77L, F116Y, Q151M
Without HU	38 ± 3.9	8.7 ± 0.3	6.4 ± 0.58
With 50 μM HU	1.7 ± 0.45 (22)	1.9 ± 0.39 (5)	0.2 ± 0.06 (32)
HIV-1M41L, D67N, M184V, L210W, T215Y, K219N
Without HU	3.5 ± 0.6	5.1 ± 0.64	2.6 ± 0.72
With 50 μM HU	0.2 ± 0.15 (18)	1.4 ± 0.42 (4)	0.4 ± 0.02 (7)

^aResults are expressed as the means ± standard deviations of four to six values obtained in two to three different experiments. 

^bValues in parentheses are the fold decrease in IC₅₀ of each compound in the presence of HU. The HU-induced decrease in drug IC₅₀s were statistically significant (P < 0.001; two tailed t test). 

^cSubscripts are resistance mutations of the RT gene. 

The HU-induced reductions in IC₅₀s of ddI, PMEA, and PMPA for the wild-type isolate were from 8- to >26-fold (Table 1), whereas in parallel experiments the reduction in IC₅₀s of AZT, 3TC, and IDV for the wild type were between 2- and 5-fold (data not shown). The HU-induced fold decrease in IC₅₀s of these drugs for the wild-type strain could be ranked as PMPA ≥ ddI > PMEA > AZT > 3TC > IDV. The differences of these drug IC₅₀s may be attributed to the effects of HU, a known inhibitor of ribonucleotide reductase, upon intracellular nucleotide pools (4, 10–12, 24). Cells exposed to HU experience a severe loss in dATP pools. After 5 days of continuous exposure to HU, the levels of dATP pools remain lower than those in control cells not exposed to HU (11). In contrast, studies show that natural dTTP and dCTP pools and the thymidine and deoxycytidine phosphorylating enzymes are elevated in cells exposed to HU (4, 10–12, 24). Consequently, in HU-treated cells, the ratio of phosphorylated adenosine analogs (ddATP, PMEA diphosphosphate, or PMPA diphosphate) to natural dATP may be substantially higher than the ratios of AZT-triphosphate/dTTP or 3TC-triphosphate/dCTP. The shift of the phosphorylated adenosine analogs/dATP ratio favors the binding of the analog to RT and is the probable cause of the more pronounced effect of HU on the anti-HIV activities of the adenosine analogs (ddI, PMEA, and PMPA) versus AZT and 3TC. The anti-HIV activity of IDV is independent of intracellular nucleotide levels, which may explain the limited effect of HU upon the activity of this protease inhibitor. Furthermore, HU at a concentration of 50 μM was found to inhibit viral growth by approximately 30%, or 0.3-fold; therefore, the fold decreases in drug IC₅₀s were not overly influenced by the inherent anti-HIV activity of HU.

The HU-induced decreases in drug IC₅₀s were greatest for the HIV_K65R, HIV_L74V, and HIV_{L74V, M184V} recombinant isolates (Table 1). Recent studies have shown that the specific activity is diminished for mutant RT enzymes containing ddI-resistant mutations including enzymes with K65R or L74V mutations (20, 31). The combination of reduced specific activity and HU-induced reduction in cellular nucleotide pools may cause the increased susceptibilities of these recombinants to inhibition by antiretroviral drugs in the presence of HU.

The IC₅₀s of HU remained relatively constant (approximately 83 μM) for wild-type and resistant viral strains. These observations suggest that the RT gene mutations of the viral strains in this study have little effect on the inherent anti-HIV activity of HU (Table 2).

TABLE 2.

Antiviral susceptibilities of a laboratory HIV-1 isolate and drug-resistant isolates to two-drug combinations

Isolate and drug	IC50 (μM)a
	ddI	PMEA	HU
HIV-1NL4-3
Single drugs	0.38	3.1	89
ddI-HU (1:8)	0.24		2.0
PMEA-HU (1:8)		1.3	10
HIV-1V75I, F77L, F116Y, Q151Mb
Single drugs	38	8.7	83
ddI-HU (1:8)	5.1		41
PMEA-HU (1:8)		3.0	24
HIV-1M41L, D67N, M184V, L210W, T215Y, K219N
Single drugs	3.5	5.1	76
ddI-HU (1:8)	1.2		9.7
PMEA-HU (1:8)		1.9	15

^aAll experiments were performed at least twice. The variations in the raw data were <20%. 

^bSubscripts are resistance mutations of the RT gene. 

At clinically achievable concentrations, the combinations ddI-HU (1:8) and PMEA-HU (1:8) synergistically inhibited the two drug-resistant viral strains tested, whereas the combination of ddI-HU synergistically inhibited the wild-type isolate at high concentrations of drug (Fig. 1). The mechanism for the synergistic interactions is unknown and may reflect the different modes of action of HU (decreasing cellular nucleotides) and the nucleoside and nucleotide analogs (RT inhibition) (16, 33).

FIG. 1.

Inhibition of HIV-1 isolates, a wild-type strain (HIV_NL4-3 [□]) and two drug-resistant strains (HIV_{V75I, F77L, F116Y, Q151M} [ ] and HIV_{M41L, D67N, M184V, L210W, T215Y, K219N} [⊕]), by the combinations ddI-HU (A) and PMEA-HU (B). The variations in the raw data were <20%.

Analysis of thymidine uptake assay results revealed a reduction in cellular DNA synthesis in the presence of HU (Table 3). Although the measurements of thymidine uptake may be affected by the HU-induced increase in intracellular dTTP pools and the extended half-life of these pools in the presence of HU, this observation suggests that HU has a cytostatic effect on cells (4, 22).

TABLE 3.

Effects of ddI, PMEA, and PMPA on the thymidine uptake of PHA-stimulated PMBCs in the presence or absence HU at designated concentrations

Drug and drug concn (μM)	Counts per minute (104)
	0 HU	25 μM HU	50 μM HU	100 μM HU	500 μM HU
ddI
0	1.2 ± 0.1	1.0 ± 0.08	0.86 ± 0.07	0.68 ± 0.03	0.04 ± 0.006
3.1	1.3 ± 0.07	0.9 ± 0.1	0.9 ± 0.03	0.5 ± 0.04	0.03 ± 0.006
12.5	1.1 ± 0.1	0.9 ± 0.08	0.9 ± 0.1	0.5 ± 0.1	0.03 ± 0.007
50	0.9 ± 0.1	0.8 ± 0.009	0.8 ± 0.09	0.5 ± 0.05	0.03 ± 0.006
PMEA
0
3.1	1.0 ± 0.1	1.0 ± 0.06	0.6 ± 0.06	0.4 ± 0.02	0.04 ± 0.006
12.5	0.9 ± 0.09	0.8 ± 0.03	0.6 ± 0.04	0.4 ± 0.04	0.03 ± 0.003
50	0.5 ± 0.03	0.5 ± 0.06	0.5 ± 0.04	0.2 ± 0.01	0.03 ± 0.005
PMPA
0
3.1	1.2 ± 0.2	1.0 ± 0.05	0.8 ± 0.02	0.5 ± 0.06	0.04 ± 0.003
12.5	1.2 ± 0.2	0.9 ± 0.06	0.8 ± 0.1	0.5 ± 0.1	0.04 ± 0.003
50	1.1 ± 0.1	0.9 ± 0.1	0.8 ± 0.02	0.5 ± 0.03	0.04 ± 0.005

In conclusion, the presence of HU at low, clinically tolerated concentrations enhances the anti-HIV activities of ddI, PMEA, and PMPA. This HU-induced increase in the activities of ddI, PMEA, and PMPA against HIV is also observed in clinical isolates that are resistant to one or more of these compounds. The two-drug combinations ddI-HU and PMEA-HU synergistically inhibited drug-resistant viral strains. This study provides evidence that supports the need for clinical trials with HU (i) in combination with ddI against ddI-resistant patient isolates and (ii) in combination with two recently developed adenosine analogs, PMEA and PMPA. Moreover, the strategy of combining highly specific HIV inhibitors with relatively nonspecific inhibitors is an approach to drug resistance that should be tested clinically.