Response of Simian Immunodeficiency Virus to the Novel Nucleoside Reverse Transcriptase Inhibitor 4′-Ethynyl-2-Fluoro-2′-Deoxyadenosine In Vitro and In Vivo

Abstract

Nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) are essential components in first-line therapy for human immunodeficiency virus (HIV) infection. However, long-term treatment with existing NRTIs can be associated with significant toxic side effects and the emergence of drug-resistant strains. The identification of new NRTIs for the continued management of HIV-infected people therefore is paramount. In this report, we describe the response of a primary isolate of simian immunodeficiency virus (SIV) to 4′-ethynyl-2-fluoro-2′-deoxyadenosine (EFdA) both in vitro and in vivo. EFdA was 3 orders of magnitude better than tenofovir (TFV), zidovudine (AZT), and emtricitabine (FTC) in blocking replication of SIV in monkey peripheral blood mononuclear cells (PBMCs) in vitro, and in a preliminary study using two SIV-infected macaques with advanced AIDS, it was highly effective at treating SIV infection and AIDS symptoms in vivo. Both animals had 3- to 4-log decreases in plasma virus burden within 1 week of EFdA therapy (0.4 mg/kg of body weight, delivered subcutaneously twice a day) that eventually became undetectable. Clinical signs of disease (diarrhea, weight loss, and poor activity) also resolved within the first month of treatment. No detectable clinical or pathological signs of drug toxicity were observed within 6 months of continuous therapy. Virus suppression was sustained until drug treatment was discontinued, at which time virus levels rebounded. Although the rebound virus contained the M184V/I mutation in the viral reverse transcriptase, EFdA was fully effective in maintaining suppression of mutant virus throughout the drug treatment period. These results suggest that expanded studies with EFdA are warranted.

INTRODUCTION

Nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) have proven highly effective in both the treatment of chronic human immunodeficiency virus (HIV) infection (11) and, more recently, as a promising microbicidal prevention strategy for sexually transmitted virus (12). There are seven NRTIs approved for current clinical use, including the nucleoside emtricitabine (FTC) and the nucleotide tenofovir (TFV). Unfortunately, the frequent use of these drugs has resulted in the emergence of resistant virus strains (7, 8, 10, 18). New compounds with potent activity on a wide range of isolates, including NRTI-resistant strains, are critically needed.

We and others have previously described a group of 4′-substituted NRTIs that are more potent and have higher selectivity indices in vitro than existing NRTIs (9, 13, 14, 15, 16, 17). One of the most potent of these compounds, 4′-ethynyl-2-fluoro-2′deoxyadenosine (EFdA), inhibits HIV-1 replication in primary peripheral blood mononuclear cells (PBMC) with a 50% effective concentration (EC₅₀) of 50 pM, a potency 4 orders of magnitude greater than that of TFV and 400-fold greater than that of AZT (15). It is also nontoxic in vitro at concentrations as high as 10 μM, which results in an in vitro selectivity index greater than 200,000. Furthermore, EFdA retains significant potency against a broad range of clinically important drug-resistant HIV isolates (13, 17).

The striking anti-HIV properties of EFdA prompted us to examine its activity on simian immunodeficiency virus (SIV) replication in vitro and in vivo. High potency against SIV would enable detailed preclinical assessments of EFdA for potential therapeutic use, as well as for use in preexposure prophylaxis and topical microbicide applications, using the SIV-macaque model for AIDS. EFdA could also provide an important tool in experiments designed to increase our understanding of host-virus interactions during latent viral infection.

In the present report, we show that EFdA surpassed the potency of TFV, AZT, and FTC in blocking replication of a primary, virulent isolate of SIV in primary monkey PBMC in vitro, and in a preliminary study using two SIV-infected macaques with advanced AIDS, it was highly effective at treating SIV infection in vivo. Both animals showed substantial decreases in plasma virus burden within 1 week of EFdA therapy that, with the exception of a single time point in one animal, remained below 100 copies/ml plasma throughout treatment. The clinical signs of disease (diarrhea, weight loss, and poor activity) also resolved during therapy. Repeated analysis of blood chemistry values during therapy, and complete histopathological examination of tissues at necropsy, also failed to reveal clinical or histopathological signs of drug-induced toxicity. Together, these results demonstrate a potential role for this compound in the treatment of AIDS.

MATERIALS AND METHODS

Macaque studies.

Male Indian-origin rhesus macaques (Macaca mulatta), 4 to 6 years of age, were obtained after completion of another study examining the adjunctive effect of DNA immunization and antiretroviral therapy with (R)-9-(2-phosphonomethoxypropyl) adenine (PMPA; Gilead Sciences, Foster City, CA) and lopinavir-ritonavir (Kaletra; Abbot Laboratories, Abbott Park, IL) (6). They were infected by intravenous inoculation of a cryopreserved stock of SIV/DeltaB670 propagated in rhesus macaque PBMC. Beginning 41 days postinoculation (p.i.), they received daily subcutaneous (s.c.) injections of 20 mg/kg of body weight PMPA and twice-daily oral administration of 16 mg (approximately 13 mg/kg) lopinavir-ritonavir. Therapy was continued without interruption for 40 weeks, during which time each animal received 6 monthly immunizations with a DNA vaccine expressing SIV gag, pol, and env and E. coli enterotoxin. Both immunizations and antiviral therapy were discontinued after 40 weeks to monitor the effects of vaccination on viral rebound and progression to disease. Although both animals responded to PMPA plus lopinavir-ritonavir and treatment appeared to prolong survival, after therapy was discontinued it had no apparent effect on controlling virus rebound in either animal, and both progressed to AIDS.

Treatment with EFdA (0.4 mg/kg delivered s.c. twice daily) was initiated approximately 34 months into the infection (24 months after cessation of PMPA plus lopinavir-ritonavir) and continued for 4 and 6 months. Monkey R395 was sacrificed after 4 months of treatment while still on therapy. EFdA was discontinued after 6 months in monkey R393 due to depletion of the drug supply, and he was sacrificed 2 months later. Tissues from all organs from both monkeys, one during therapy (R395) and one 2 months after therapy was discontinued (R393), were subjected to detailed histopathological examination.

Monkeys were maintained in accordance with the NIH Guide to the Care and Use of Laboratory Animals under the approval of the University of Pittsburgh institutional animal care and use committee. The University of Pittsburgh is accredited by the American Association for the Accreditation of Laboratory Animal Care International.

Analysis of viral loads in plasma and tissues.

Virion-associated RNA in plasma was quantified by real-time PCR in a Prism 7700 (Applied Biosystems, Inc., Foster City, CA) using primers specific for the SIV long terminal repeat (LTR) as described previously (4, 5, 6, 21). This assay is linear over an 8-log range of template copy numbers and has a sensitivity threshold of 10 copies per reaction. Control amplifications of each sample omitting reverse transcriptase were also performed to control for contaminating DNA. RNA copy numbers from the unknown plasma samples were calculated from a similarly amplified external standard and expressed as viral RNA copies/ml plasma.

Full-length cell-associated SIV transcripts were quantified in TRIzol-extracted RNA from tissues or purified mononuclear cells as described previously (6). Briefly, reverse-transcribed cellular RNA was subjected to real-time quantitative reverse transcription-PCR (qRT-PCR) using TaqMan chemistry and forward and reverse primers with probe specific for the U5 region of the 5′ long terminal repeat (LTR) and the downstream primer binding site. These primers amplify a 92-bp fragment found either in the virion itself or in transcripts prepared for packaging into virions. When cell associated, they are a reliable indicator of virus production in the tissues. The assay has an amplification efficiency of 97% and a sensitivity of 10 copies/reaction.

Genotyping of plasma SIV.

cDNA obtained from purified RNA from high-speed plasma pellets was used for analysis as described previously (1). Amplicons spanning nucleotides 3353 to 3682 of the viral genome were obtained from a pool of 5 independent PCRs of each sample, cloned, and sequenced. Deduced sequences spanning amino acids 108 to 408 of the viral reverse transcriptase were aligned using SIVmac239 as the consensus sequence using software available in the Los Alamos database.

In vitro drug inhibition.

For the in vitro experiments, viably frozen rhesus macaque PBMC (rhPBMC) were thawed, washed thoroughly, and stimulated for 72 h in growth media (1× RPMI 1640 with 15% fetal bovine serum, 1% streptomycin-penicillin, and 1% l-glutamine) containing 2 μg/ml phytohemagglutinin. After thorough washing, rhPBMC were resuspended in growth media containing 10 U/ml interleukin-2 (IL-2), infected with 16.7 μl of SIV/DeltaB670 (66 50% tissue culture infectious doses [TCID₅₀]) per 1.0 × 10⁶ cells, and plated in 96-well plates at 4.5 × 10⁴ cells per well. Twofold serial dilutions of the assayed compounds were produced in IL-2 growth media and added to infected rhPBMC, generating final drug concentration ranges of 0 to 12.5 nM EFdA, 0 to 1,000 nM AZT, 0 to 10,000 nM TFV, and 0 to 10,000 nM FTC. Infected, treated rhPBMCs were cultured at 37°C in a 5% CO₂ incubator for 10 days. Half of the cell supernatant was harvested at days 4, 7, and 10 postinfection; after each sampling, medium was replenished with fresh IL-2 growth medium containing half the concentration of drug added at day 0. The supernatants harvested on day 10 postinfection were tested for virus replication using an SIV p27 enzyme-linked immunosorbent assay (ELISA) kit according to the manufacturer's protocols (Advanced Bioscience Laboratories, Rockville, MD). The percent inhibition of SIV replication was determined by measuring the reduction of p27 in the culture supernatant at each escalating drug concentration versus untreated virus. Results are the averages from three replicates for each dilution of drug. The EC₅₀s and EC₉₀s for each drug were determined by four-parameter logistic nonlinear regression analysis.

RESULTS

Susceptibility of SIV to EFdA in vitro.

To determine whether EFdA was as effective in controlling SIV replication as it had been for HIV, we evaluated its ability to inhibit in vitro replication of the primary SIV isolate SIV/DeltaB670 in rhesus PBMC (rhPBMC). The inhibitory effect of two other well-studied NRTIs, TFV and FTC, were also evaluated for comparison (Fig. 1). SIV/DeltaB670 is highly virulent in macaques (19), has been propagated in vitro only in rhPBMC, and is comprised of a well-characterized genetic swarm (1, 2, 21, 22). A comparison of the percent inhibition and the 50 and 90% effective concentrations (EC₅₀ and EC₉₀) for each drug are shown in Fig. 1A and B, respectively. The inhibitory activity of EFdA against SIV (EC₅₀ of 50 pM) was identical to that observed for inhibition of HIV-1 replication in human PBMCs (15) and was 3 orders of magnitude more potent than that of either TFV or FTC. No cytotoxicity was noted upon exposure of monkey PBMCs for 10 days to EFdA concentrations up to 10 μM, providing an in vitro selectivity index of greater than 200,000 (data not shown).

Fig 1.

In vitro response of SIV/DeltaB670 to escalating doses of EFdA, TFV, and FTC in rhesus macaque PBMC. Sixty-six TCID₅₀ of virus was incubated with serial dilutions of each drug, and infected cultures were monitored for p27 levels in the culture supernatant over time postinfection. (A) The percent inhibition relative to untreated control cultures for each dilution. (B) The 50 and 90% effective concentrations for each drug, determined using four-parameter logistic nonlinear regression. Values represent the means ± standard errors from one experiment with three or two replicates per dilution, respectively.

Susceptibility of SIV to EFdA in vivo.

The in vitro potency of EFdA against the same virulent primary stock used for our in vivo experiments encouraged us to treat two macaques chronically infected with SIV/DeltaB670 despite their advanced disease. The virological history of these animals is shown in Fig. 2. Both animals were originally used in a study to evaluate the immunotherapeutic potential of a DNA vaccine administered during treatment of chronic infection with TFV and lopinavir-ritonavir (6). Animals had been inoculated intravenously with the same SIV stock as that used for the in vitro experiments described above. Plasma virus burden was monitored by qRT-PCR (4, 5, 21). Forty-one days postinoculation (after the viral set point had been reached but while they were clinically asymptomatic), daily treatment with 20 mg/kg TFV and twice-daily administration of 16 mg/kg lopinavir-ritonavir was initiated; therapy was continued for 279 days and then discontinued. Both macaques were additionally immunized at monthly intervals during therapy with the SIV DNA vaccine.

Fig 2.

Virological history of SIV/DeltaB670-infected macaques treated with TFV plus lopinavir-ritonavir followed by EFdA. Monkeys were infected by intravenous inoculation and longitudinally monitored for virus burden, clinical signs of disease, and changes in CD4⁺ T cell numbers in the blood over time during infection and treatment. Daily subcutaneous injections with TFV (20 mg/kg; once daily) and twice-daily oral administration of 13 mg/kg lopinavir-ritonavir was initiated on day 41 and discontinued on day 320 p.i. Twenty-four months later, twice-daily subcutaneous injections of 0.4 mg/kg EFdA were initiated on day 1025 and terminated on day 1224 (at sacrifice) in monkey R395 and day 1276 in monkey R393. Monkey R393 was sacrificed for tissue collection on day 1336. The black bars are indicative of the intervals of therapy. The asterisks indicate the timing of sacrifice of each animal. Plasma virus load measurements were performed over the course of the experiments as described in the text. Values fewer than 5,000 copies of SIV RNA/ml plasma is indicative of asymptomatic infection. Values of less than 100 copies were repeated to ensure accuracy. Values from monkey R395 are indicated by triangles connected by the dashed line. The circles connected by the solid line reflect values from monkey R393.

Initial response to TFV and lopinavir-ritonavir.

Both animals responded to the cocktail of TFV and lopinavir-ritonavir with a 5-log drop in plasma virus loads. R395 continued to respond well throughout therapy, whereas plasma virus loads in monkey R393 soon rebounded and remained at or above the threshold associated with disease progression (>10⁴/ml plasma) (21) throughout the remainder of therapy. Treatment was discontinued at 320 days p.i.; virus loads rebounded in both animals shortly thereafter. Subsequent to viral rebound, both animals developed progressive disease characterized clinically by chronic diarrhea that was unresponsive to treatment and persistent weight loss. Thirty-four months into the infection (24 months after cessation of PMPA plus lopinavir-ritonavir), both animals had declines in CD4⁺ T lymphocytes to less than 45% of their preinfection values, plasma virus loads greater than 10⁴ copies/ml plasma, and clinical signs of AIDS (diarrhea and weight loss greater than 20% of body weight; end-stage AIDS).

Response to EFdA treatment.

EFdA treatment (0.4 mg/kg delivered s.c. twice daily [BID]) was initiated during this time. Despite the poor clinical conditions and high virus loads at the onset of therapy, the response to treatment with EFdA exceeded that observed with TFV plus lopinavir-ritonavir initiated during the asymptomatic stage of the disease (Fig. 2). Within 1 week, a 3- to 4-log reduction in virus burden was observed in both animals that, except for barely detectable blips in monkey R393, further declined and remained below the threshold of detection until EFdA was discontinued (Fig. 3). Four months into the therapy, monkey R395 was sacrificed and the tissues collected for analysis. EFdA therapy was continued for an additional 2 months in monkey R393 (6 months total) and then discontinued. After discontinuation of the drug, a rebound in virus was observed in R393 that continued until his sacrifice 2 months later.

Fig 3.

In vivo response of SIV/DeltaB670 to EFdA treatment. Virus loads during EFdA therapy, excerpted from Fig. 2, are shown for better resolution. Therapy consisted of twice-daily subcutaneous injections of 0.4 mg/kg. Values from monkey R395 are indicated by triangles connected by the dashed line. The circles connected by the solid line reflect the values from monkey R393.

Viral resistance.

Although there was no obvious sign of emerging resistance to EFdA during treatment, the sporadic blips of viremia observed in monkey R393 suggested that, despite treatment, virus replication had continued in some unknown reservoir. This suspicion was further substantiated by the rebound of virus after EFdA was discontinued. We therefore obtained a partial sequence of the viral RT (nucleotides 3353 to 3682) from plasma virions at two time points during the study, 2 weeks prior to initiation of EFdA treatment, and during the postdrug rebound. We had previously determined that the rebounded virus after TFV plus lopinavir-ritonavir therapy did not contain the K65R mutation (24 and unpublished data). As expected, of 12 clones obtained 2 weeks before EFdA treatment, all contained wild-type virus. Sixty days after the discontinuation of EFdA therapy, however, the rebound virus contained 2 significant changes, conversion of the methionine at position 184 to either valine or isoleucine (M184V and M184I, respectively; 6 of 12 clones). Although these mutations are well known for conferring resistance to other NRTIs (23, 25), they did not result in a significant change in virus burden during EFdA therapy. Whether treatment with a higher dose of EFdA or a more prolonged therapy would have had a greater impact on infection and disease will require further study with a larger cohort of animals.

Effect of EFdA treatment on disease.

Physical examination of both animals was carried out at biweekly intervals throughout the study. This examination included palpation of lymph nodes, assessment for potential opportunistic infections, weight measurements, and analysis of blood chemistry and complete blood counts. Chronic diarrhea in both animals completely resolved within the first month of treatment and both animals gained weight, with weights returning to preinfection levels within 4 months (Fig. 4). The 4-month therapeutic regimen, however, failed to completely resolve a preexisting Mycobacterium sp. infection in M395, and he was humanely sacrificed. Therapy was continued in the remaining monkey (animal R393) for an additional 2 months and then discontinued. Despite virus rebound after treatment was discontinued, R393 remained clinically asymptomatic until sacrifice.

Fig 4.

Effect of EFdA therapy on diarrhea and weight loss. Monkeys were observed daily in their cages for normal stool and weighed at weekly or biweekly intervals. Stool consistency was graded +1 (fully formed), +2 (loose),+3 (watery with no solids), or +4 (bloody). The black line depicts the interval of +3 grade stool. Weight is measured in kilograms. The triangles connected by the dashed line present weights of monkey R395. Circles connected by the solid line represent weights of monkey R393.

Toxicity.

Serum chemistries and complete blood counts were also analyzed at biweekly intervals during the study, and a complete histopathological examination of the tissues was performed at sacrifice to identify drug-induced toxicity. The mean chemistry values over the period of treatment are shown in Table 1. Other than the low platelet values in monkey R395, who was thrombocytopenic prior to the onset of therapy, no change in the normal levels of blood components that signal bone, liver, heart, and kidney disease were observed in either animal. Complete blood counts were also within normal range, confirming the lack of drug-induced bone marrow toxicity (data not shown).

Table 1.

Effects of EFdA therapy on multiorgan function

Assay	Normal range	Mean values for monkeya:
		R393	R395
Alanine aminotransferase (U/liter)	31–50	23	28
Aspartate transaminase (U/liter)	19–38	38	35
Alkaline phosphatase (U/liter)	55–237	155	175
Bilirubin (mg/dl)	0.1–0.3	0.3	0.03
Blood urea nitrogen (mg/dl)	16–30	19	22
Creatinine (mg/dl)	0.7–1.4	1.2	1.1
Blood urea nitrogen /creatinine (ratio)	10–20	17	20
Albumin (g/dl)	3.2–4.1	4.3	2.9
Platelets (#/dl)	260–361	371	40
Total protein (g/dl)	6.4–7.0	7.3	7.6
Calcium (mg/dl)	8.7–10.9	9.1	8.3
Potassium (mg/dl)	3.3–3.7	4	4.2

^aData are from ten tests total.

Histopathological examination of the organs at necropsy also revealed no evidence of drug toxicity. In particular, the histopathology of the heart was unremarkable, with the cardiovascular tissues in both animals lacking any sign of mononuclear cell infiltrates typical of those found in animals treated with other NRTIs (3) (data not shown). Together, these findings indicate that EFdA at this dose and duration of therapy had no effect on liver, kidney, heart, and bone marrow function.

Virus levels in tissue reservoirs.

Even though virus was undetectable in the plasma, virus production within tissues collected from both animals by either biopsy during treatment or at necropsy during (R395) or after (R393) treatment was analyzed to determine the impact of EFdA on SIV replication in tissues using primers that amplify full-length genomes as a measure of virus production in susceptible target cells. Not surprisingly, viral RNA was readily detected in most tissues obtained from the monkey (animal R393) that was sacrificed after he rebounded (Table 2), with the highest levels of expression observed in the peripheral lymph nodes and spleen. Interestingly, neither the seminal vesicles nor the prostate of the male reproductive tract contained detectable virus, despite the fact that this animal was sacrificed during the fall breeding season.

Table 2.

Effect of EFdA therapy on SIV expression in tissues

Tissuea	Drug effects on monkey:
	R395		R393
	Pre-EFdA (2 months before therapy)	At necropsy (after 4 months therapy)	Day of EFdA stop (after 6 months therapy)	At necropsy (5 months after stopping therapy)
Plasma	3.1 × 104	Undetectable	Undetectable	2.6 × 104
PBMC	5.4 × 104	Undetectable	Undetectable	2.2 × 103
Axillary lymph node	2.7 × 103	8 × 101	NAb	1.9 × 105
Pancreas		Undetectable		Undetectable
Spleen		Undetectable		4 × 104
Thymus		Undetectable		Undetectable
Ileum		Undetectable		Undetectable
Tonsil		Undetectable		1.6 × 104
Lung		Undetectable		4.2 × 102
Seminal vesicles		Undetectable		Undetectable
Prostate		Undetectable		Undetectable
Jejunum		Undetectable		4.1 × 101
Mesenteric lymph node		Undetectable		3.5 × 103
Kidney		Undetectable		2.6 × 101
Duodenum	2.7 × 101	Undetectable	Undetectable	1.5 × 101
Colon		Undetectable		4.1 × 101
Liver				1.4 × 101
Cecum				7 × 100

^aCounts in plasma is measured in copies/ml plasma, and those in the other tissues are measured in copies/250 ng total RNA.

^bNA, not available.

In contrast, virus could not be detected in most tissues obtained by biopsy from R393 or at sacrifice of animal R395, both of which were obtained while the animals were undergoing therapy. Analysis of necropsied tissues from R395 showed that SIV RNA could be identified in only one organ, an axillary lymph node, and then only at a level barely exceeding the threshold for detection. Similar results were obtained from PBMC and a duodenal biopsy specimen taken from monkey R393 just prior to stopping treatment. These findings suggest that EFdA is readily distributed among the major organs, including those of the reproductive tract, where it can access susceptible target cells and maintain a persistent block to virus replication.

DISCUSSION

Nucleoside/nucleotide reverse transcriptase inhibitors have proven to be the most effective anti-HIV therapeutics to date (11). Indeed, first-line treatment (highly active antiretroviral therapy; HAART) usually consists of two NRTIs plus another drug from a different class. Not surprisingly, however, despite the potency of these compounds, HIV continues to be refractory from clearance, with strains of HIV resistant to more than one of these drugs now resident in the treatment-naïve population (7, 8). Effective therapy typically also requires high doses (200 to 600 mg daily) that must be taken long term to effectively treat chronic infection. Interest in increasing the repertoire of available drugs thus remains high.

EFdA offers several advantages (9, 13, 14, 15, 16, 17). In vitro, it is the most potent anti-HIV compound described to date (15), with an EC₅₀ that allows an effective dose that is significantly lower than that of other NRTIs. In this report, we show that EFdA inhibits SIV equally well and, like that observed for HIV-1, is more potent in vitro than TFV and FTC. The equivalent potency against both HIV-1 and SIV suggests that EFdA will retain excellent activity against all HIV-1 subtypes, a factor that further highlights its potential importance for global health.

Although the in vivo results reported here are confined to a 6-month therapeutic regimen in two macaques taken from another study, the rapid and profound reduction in virus burden observed in both plasma and tissues and the continued efficacy during 4 to 6 months of therapy without measurable toxicity in both animals are encouraging.

One potential concern is that EFdA treatment resulted in selection of RT mutations associated with high-level resistance to the widely used NRTIs lamivudine (3TC) and FTC (23, 25). This finding might discourage the use of EFdA as a first-line therapy, because it might preclude 3TC and FTC from subsequent use in these patients. However, M184I/V mutations are already prevalent in treatment-experienced patients due to the widespread use of both 3TC and FTC in HAART (20, 25). Indeed, despite the appearance of M184V/I-containing variants in the rebounded virus after drug was removed, EFdA successfully controlled virus burden to undetectable levels in both blood and tissues throughout the interval of therapy. Together, these findings provide support for further studies designed to address issues of drug resistance, toxicity, and the presence of viral reservoirs with this compound.