**2. Mechanism of action**

Retroviruses such as HIV-1 carry their genomic information in the form of (+)strand RNA, but are distinguished from other RNA viruses by the fact that they replicate through a double-stranded DNA that is integrated into the host cell's genomic DNA (Temin & Mizutani, 1970; Baltimore, 1970; DeStefano et al., 1993). While the conversion of viral RNA into double-stranded DNA intermediate is a complex process, all chemical steps are catalyzed by the multi-functional viral enzyme reverse transcriptase (RT). HIV-1 RT exhibits two types of DNA polymerase activity, an RNA-dependent DNA polymerase activity that synthesizes a (-)strand DNA copy of the viral RNA, and a DNA-dependent DNA polymerase activity that generates the (+)strand DNA (Peliska & Benkovic, 1992; Cirno et al., 1995). RT also has ribonuclease H activity that degrades the RNA in the intermediate (+)RNA/(-)DNA duplex (Ghosh et al., 1997).

Once metabolized by host cell enzymes to their triphosphate forms (described in more detail below), nucleoside analogs inhibit HIV-1 reverse transcription. As such, they are typically referred to as nucleoside RT inhibitors (NRTI). NRTI-triphosphates (NRTI-TP) inhibit RTcatalyzed proviral DNA synthesis by two mechanisms (Goody et al., 1991). First, they are

Molecular Pharmacology of Nucleoside and Nucleotide HIV-1 Reverse Transcriptase Inhibitors 65

Like zidovudine, stavudine (2',3'-didehydro-3'-deoxythymidine, d4T) is a thymidine analog that undergoes metabolic activation by the sequential action of thymidine kinase and dTMP kinase (Figure 1). However, stavudine is inefficiently phosphorylated to its 5'-MP form by thymidine kinase (August et al., 1988; Zhu et al., 1990). As such, this first phosphorylation step is rate-limiting and most intracellular stavudine is not phophorylated (Balzarini et al., 1989). Maximal plasma concentrations of stavudine are achieved within 2 hours of oral administration and increase linearly as the dose increases, with an absolute bioavailability approaching 100 % (Rana & Dudley, 1997)). The drug distributes into total body water and appears to enter cells by non-facilitated diffusion (passive transport). Penetration into the central nervous system, however, is far less than zidovudine. Stavudine is cleared quickly with a terminal plasma half-life of 1-1.6 hours by both renal and nonrenal processes (Dudley

Initially, 2',3'-dideoxyadenosine (ddA) was evaluated as a clinical candidate but was ultimately discovered to cause nephrotoxicity. ddA is acid labile and oral administration leads to exposure to the acidic pH of the stomach and degradation to adenine (Masood et al., 1990). Adenine is further metabolized to 2,8-dihydroxyadenine which causes nephrotoxicity by crystallization in the kidney. Interestingly, ddA was shown to be metabolized to 2',3' dideoxyinosine (ddI, didanosine) by adenosine deaminase (Figure 2), and that much of the antiviral activity of ddA resides in didanosine (Cooney et al., 1987). Furthermore, the administration of didanosine avoids the production of adenine and the resulting nephrotoxicity. Didanosine is phosphorylated to didanosine-MP by cytosolic 5'-nucleotidase, which uses either inosine monophosphate (IMP) or guanosine monophosphate (GMP) as

**3.2 Stavudine** 

et al., 1992).

**3.3 Didanosine** 

Fig. 1. Metabolic pathways of zidovudine and stavudine

competitive inhibitors for binding and/or catalytic incorporation with respect to the analogous natural dNTP substrate. Second, they terminate further viral DNA synthesis due to the lack of a 3'-OH group. Chain termination is the principal mechanism of NRTI antiviral action (Goody et al., 1991). In theory, NRTI-TPs should be ideal antivirals. Each HIV virion carries only two copies of genomic RNA. There are about 20,000 nucleotide incorporation events catalyzed by RT during the synthesis of complete viral DNA, thus providing about 5000 chances for chain-termination by any given NRTI. Since HIV-1 RT lacks a formal proof-reading activity, a single NRTI incorporation event should effectively terminate reverse transcription. In reality, however, NRTIs are less potent than might be expected. The two primary reasons responsible for this are: (i) HIV-1 RT can effectively discriminate between the natural dNTP and NRTI-TP, and the extent of this discrimination is dramatically modulated by nucleic acid sequence (Isel et al., 2001); and (ii) HIV-1 RT can excise the chain-terminating NRTI-monophosphate (NRTI-MP) by using either pyrophosphate (pyrophophorolysis) or ATP as a substrate (Meyer et al., 1998; Goldschmidt & Marquet, 2004).

#### **3. NRTI approved for clinical use**

#### **3.1 Zidovudine**

Zidovudine was first synthesized in 1964 as a potential anticancer drug, but was not further developed for human use because of toxicity concerns. However, as described in the Introduction, it was found to have potent anti-HIV activity and, in 1987, was the first antiviral drug to be approved for clinical use. Zidovudine is a thymidine analog in which the 3'-OH group has been replaced with an azido (-N3) group (Figure 1). Zidovudine permeates the cell membrane by passive transport and not via a nucleoside carrier transporter (Zimmerman et al., 1987). It has good oral bioavailability and shows efficient penetration into the central nervous system. Zidovudine is efficiently metabolized to its 5'- MP form by cytosolic thymidine kinase (Ho & Hitchcock, 1989). The phosphorylation of zidovudine-MP to zidovudine-DP is catalyzed by thymidinylate monophosphate kinase (dTMP kinase; Furman et al., 1986). Interestingly, the apparent Michaelis constant (Km) of zidovudine-MP for dTMP kinase is almost equivalent to that of dTMP, however its maximum kinetic rate (Vmax) is only 0.3 % that of dTMP (Furman et al., 1986). Therefore, zidovudine-MP acts as a substrate inhibitor of dTMP kinase and limits its own conversion to the 5'-DP form. In this regard, there is a marked accumulation of zidovudine-MP and only low levels of the 5'-DP- and 5'-TP derivatives are detected in human T-lymphocytes (Balzarini et al., 1989). Cellular nucleoside diphosphate kinase (NDP kinase) is likely responsible for the further conversion of zidovudine-DP to zidovudine-TP. Zidovudine is metabolized to its 5'-O-glucuronide in the liver, kidney, and intestinal mucosa (Barbier et al., 2000). Because of the extensive glucuronidation of ZDV, other drugs that are also glucuronidated or that inhibit this process cause an increase in zidovudine plasma levels. Fourteen percent of the parent compound and 74% of the glucuronide have been recovered from the urine after oral administration in normal subjects (Ruane et al., 2004). Renal excretion of zidovudine is by both glomerular filtration and active tubular secretion. In some cells zidovudine can be metabolized to the highly toxic reduction product 3'-aminothymidine (Weidner & Sommadossi, 1990).

#### **3.2 Stavudine**

64 Pharmacology

competitive inhibitors for binding and/or catalytic incorporation with respect to the analogous natural dNTP substrate. Second, they terminate further viral DNA synthesis due to the lack of a 3'-OH group. Chain termination is the principal mechanism of NRTI antiviral action (Goody et al., 1991). In theory, NRTI-TPs should be ideal antivirals. Each HIV virion carries only two copies of genomic RNA. There are about 20,000 nucleotide incorporation events catalyzed by RT during the synthesis of complete viral DNA, thus providing about 5000 chances for chain-termination by any given NRTI. Since HIV-1 RT lacks a formal proof-reading activity, a single NRTI incorporation event should effectively terminate reverse transcription. In reality, however, NRTIs are less potent than might be expected. The two primary reasons responsible for this are: (i) HIV-1 RT can effectively discriminate between the natural dNTP and NRTI-TP, and the extent of this discrimination is dramatically modulated by nucleic acid sequence (Isel et al., 2001); and (ii) HIV-1 RT can excise the chain-terminating NRTI-monophosphate (NRTI-MP) by using either pyrophosphate (pyrophophorolysis) or ATP as a substrate (Meyer et al., 1998; Goldschmidt

Zidovudine was first synthesized in 1964 as a potential anticancer drug, but was not further developed for human use because of toxicity concerns. However, as described in the Introduction, it was found to have potent anti-HIV activity and, in 1987, was the first antiviral drug to be approved for clinical use. Zidovudine is a thymidine analog in which the 3'-OH group has been replaced with an azido (-N3) group (Figure 1). Zidovudine permeates the cell membrane by passive transport and not via a nucleoside carrier transporter (Zimmerman et al., 1987). It has good oral bioavailability and shows efficient penetration into the central nervous system. Zidovudine is efficiently metabolized to its 5'- MP form by cytosolic thymidine kinase (Ho & Hitchcock, 1989). The phosphorylation of zidovudine-MP to zidovudine-DP is catalyzed by thymidinylate monophosphate kinase (dTMP kinase; Furman et al., 1986). Interestingly, the apparent Michaelis constant (Km) of zidovudine-MP for dTMP kinase is almost equivalent to that of dTMP, however its maximum kinetic rate (Vmax) is only 0.3 % that of dTMP (Furman et al., 1986). Therefore, zidovudine-MP acts as a substrate inhibitor of dTMP kinase and limits its own conversion to the 5'-DP form. In this regard, there is a marked accumulation of zidovudine-MP and only low levels of the 5'-DP- and 5'-TP derivatives are detected in human T-lymphocytes (Balzarini et al., 1989). Cellular nucleoside diphosphate kinase (NDP kinase) is likely responsible for the further conversion of zidovudine-DP to zidovudine-TP. Zidovudine is metabolized to its 5'-O-glucuronide in the liver, kidney, and intestinal mucosa (Barbier et al., 2000). Because of the extensive glucuronidation of ZDV, other drugs that are also glucuronidated or that inhibit this process cause an increase in zidovudine plasma levels. Fourteen percent of the parent compound and 74% of the glucuronide have been recovered from the urine after oral administration in normal subjects (Ruane et al., 2004). Renal excretion of zidovudine is by both glomerular filtration and active tubular secretion. In some cells zidovudine can be metabolized to the highly toxic reduction product 3'-amino-

& Marquet, 2004).

**3.1 Zidovudine**

**3. NRTI approved for clinical use** 

thymidine (Weidner & Sommadossi, 1990).

Like zidovudine, stavudine (2',3'-didehydro-3'-deoxythymidine, d4T) is a thymidine analog that undergoes metabolic activation by the sequential action of thymidine kinase and dTMP kinase (Figure 1). However, stavudine is inefficiently phosphorylated to its 5'-MP form by thymidine kinase (August et al., 1988; Zhu et al., 1990). As such, this first phosphorylation step is rate-limiting and most intracellular stavudine is not phophorylated (Balzarini et al., 1989). Maximal plasma concentrations of stavudine are achieved within 2 hours of oral administration and increase linearly as the dose increases, with an absolute bioavailability approaching 100 % (Rana & Dudley, 1997)). The drug distributes into total body water and appears to enter cells by non-facilitated diffusion (passive transport). Penetration into the central nervous system, however, is far less than zidovudine. Stavudine is cleared quickly with a terminal plasma half-life of 1-1.6 hours by both renal and nonrenal processes (Dudley et al., 1992).

Fig. 1. Metabolic pathways of zidovudine and stavudine

#### **3.3 Didanosine**

Initially, 2',3'-dideoxyadenosine (ddA) was evaluated as a clinical candidate but was ultimately discovered to cause nephrotoxicity. ddA is acid labile and oral administration leads to exposure to the acidic pH of the stomach and degradation to adenine (Masood et al., 1990). Adenine is further metabolized to 2,8-dihydroxyadenine which causes nephrotoxicity by crystallization in the kidney. Interestingly, ddA was shown to be metabolized to 2',3' dideoxyinosine (ddI, didanosine) by adenosine deaminase (Figure 2), and that much of the antiviral activity of ddA resides in didanosine (Cooney et al., 1987). Furthermore, the administration of didanosine avoids the production of adenine and the resulting nephrotoxicity. Didanosine is phosphorylated to didanosine-MP by cytosolic 5'-nucleotidase, which uses either inosine monophosphate (IMP) or guanosine monophosphate (GMP) as

Molecular Pharmacology of Nucleoside and Nucleotide HIV-1 Reverse Transcriptase Inhibitors 67

the GI tract with peek plasma levels of 85-93% achieved within 2 hours post oral administration. Lamivudine has a plasma half-life of 5-7 hours and is eliminated unmetabolized by active organic cationic excretion (Johnson et al., 1999). Emtricitabine persists in plasma with a half-life of 10 hours and is eliminated primarily in urine by glomerular filtration and active tubular secretion but approximately 14% is eliminated in feces. Oxidation of the 3'-thiol by unidentified enzymes yields 3'-sulfoxide diasteriomers

Abacavir (1S,4R)-4-[2-amino-6-(cyclopropylamino)-9H-purin-9-yl]cyclopent-2-en-1-yl]methanol) is a prodrug of carbovir (2-Amino-1,9-dihydro-9-[(1R,4S)-4-(hydroxymethyl)-2-cyclopenten-1-yl]-6H-purin-6-one), a deoxyguanosine analog (Figure 4; Daluge et al., 1997). Abacavir permeates T lymophoblastoid cell lines by passive diffusion. Abacavir is phosphorylated to abacavir-MP by adenosine phosphotransferase (Faletto et al., 1997). A yet unidentified cytosolic deaminase then converts abacavir-MP to carbovir-MP. Phosphorylation to the diphosphate derivative occurs via guanidinylate monophosphate kinase. The final phosphorylation step can be catalyzed by a number of cellular enzymes including 5' nucleotide diphosphate kinase, pyruvate kinase, and creatine kinase (Faletto et al., 1997). A linear dose relationship with carbovir-mono-, di-, and tri- phosphate derivatives over a 1000-fold dose range in vitro suggests there are no rate limiting steps in abacavir anabolism. The active metabolite carbovir-TP has been shown to persist with an elimination half-life of greater than 20 hours (McDowell et al., 2000). Abacavir bioavailability is ~83 % and is rapidly absorbed after oral dosing reaching peak plasma levels within 1 hour (Chittick et al., 1999). However, abacavir is extensively catabolized in the liver and only 1.2% is excreted as unchanged abacavir in urine. Abacavir oxidation by alcohol dehyrogenases to form the 5' carboxylic acid derivative represents 36% of metabolites recovered from urine, while the 5'- O-glucuronide corresponds to 30% of metabolites from urine (Chittick et al., 1999). Fecal excretion also accounts for approximately 16 % of the given dose. Abacavir is not

metabolized by cytochrome P450 enzymes and does not inhibit these enzymes.

and 2'-O-glucuronidation also occurs.

**3.5 Abacavir** 

Fig. 3. Metabolic pathways of lamivudine and emtricitabine

phosphate donors (Johnson & Fridland, 1989). Didanosine-MP is then converted to ddAMP by adenylosuccinate synthetase and 5' adenosine monophosphate-activated protein (AMP) kinase (Ahluwalia et al., 1987). The enzymes involved in phosphorylation of ddAMP to ddADP and ddATP have not been identified, although AMP kinase and NDP kinase have been proposed to play a role. ddATP is the active metabolite that is recognized by HIV-1 RT and incorporated into the nascent viral DNA chain causing chain-termination. No evidence has been provided for the formation of didanosine-DP or didanosine-TP. Didanosine is hydrolyzed to hypoxanthine by purine nucleoside phosphorylase (PNP) and further anabolized by hypoxanthine-guanine phosphoribosyl transferase to IMP (Ahluwalia et al., 1987). ATP and GTP are formed from IMP through the classical purine nucleotide biosynthetic pathways.

Fig. 2. Metabolic pathways of ddA and didanosine

#### **3.4 Lamivudine and emtricitabine**

The structurally related cytidine analogs lamivudine ((-)-3'-thia-2',3'-dideoxycytidine; 3TC) and ematricitabine ((-)-3'-thia-5-flouro-2',3'-dideoxycytidine; FTC) both contain the unnatural L-enantiomer ribose with a sulfur atom replacing the C3' position (Figure 3). Emtricitabine has an additional 5-flouro moiety on the cytosine ring. Lamivudine and emtricitabine are both metabolized to their respective 5'-mono- and di- and triphosphate derivatives by deoxycytidine kinase, deoxycytidine monophosphate kinase, and 5' nucleoside diphosphate kinase, respectively (Chang et al., 1992; Cammack et al., 1992; Stein & Moore 2001; Darque et al., 1999; Bang & Scott, 2003). There is no evidence that lamivudine or emtricitabine are deaminated to their uridine analogs by cellular cytidine or deoxycytidine deaminases (Starnes & Cheng, 1987). Formation of the free base by cellular pyrimidine phosphorylases has also not been observed. Lamivudine-DP and emtricitabine-TP accumulate to higher levels in peripheral blood mononuclear cells than their monophosphate forms. It has been suggested that conversion of lamivudine-DP to lamivudine-TP is rate limiting. Lamivudine and emtricitabine are rapidly absorbed through the GI tract with peek plasma levels of 85-93% achieved within 2 hours post oral administration. Lamivudine has a plasma half-life of 5-7 hours and is eliminated unmetabolized by active organic cationic excretion (Johnson et al., 1999). Emtricitabine persists in plasma with a half-life of 10 hours and is eliminated primarily in urine by glomerular filtration and active tubular secretion but approximately 14% is eliminated in feces. Oxidation of the 3'-thiol by unidentified enzymes yields 3'-sulfoxide diasteriomers and 2'-O-glucuronidation also occurs.

Fig. 3. Metabolic pathways of lamivudine and emtricitabine

#### **3.5 Abacavir**

66 Pharmacology

phosphate donors (Johnson & Fridland, 1989). Didanosine-MP is then converted to ddAMP by adenylosuccinate synthetase and 5' adenosine monophosphate-activated protein (AMP) kinase (Ahluwalia et al., 1987). The enzymes involved in phosphorylation of ddAMP to ddADP and ddATP have not been identified, although AMP kinase and NDP kinase have been proposed to play a role. ddATP is the active metabolite that is recognized by HIV-1 RT and incorporated into the nascent viral DNA chain causing chain-termination. No evidence has been provided for the formation of didanosine-DP or didanosine-TP. Didanosine is hydrolyzed to hypoxanthine by purine nucleoside phosphorylase (PNP) and further anabolized by hypoxanthine-guanine phosphoribosyl transferase to IMP (Ahluwalia et al., 1987). ATP and GTP are formed from IMP through the classical purine nucleotide biosynthetic pathways.

The structurally related cytidine analogs lamivudine ((-)-3'-thia-2',3'-dideoxycytidine; 3TC) and ematricitabine ((-)-3'-thia-5-flouro-2',3'-dideoxycytidine; FTC) both contain the unnatural L-enantiomer ribose with a sulfur atom replacing the C3' position (Figure 3). Emtricitabine has an additional 5-flouro moiety on the cytosine ring. Lamivudine and emtricitabine are both metabolized to their respective 5'-mono- and di- and triphosphate derivatives by deoxycytidine kinase, deoxycytidine monophosphate kinase, and 5' nucleoside diphosphate kinase, respectively (Chang et al., 1992; Cammack et al., 1992; Stein & Moore 2001; Darque et al., 1999; Bang & Scott, 2003). There is no evidence that lamivudine or emtricitabine are deaminated to their uridine analogs by cellular cytidine or deoxycytidine deaminases (Starnes & Cheng, 1987). Formation of the free base by cellular pyrimidine phosphorylases has also not been observed. Lamivudine-DP and emtricitabine-TP accumulate to higher levels in peripheral blood mononuclear cells than their monophosphate forms. It has been suggested that conversion of lamivudine-DP to lamivudine-TP is rate limiting. Lamivudine and emtricitabine are rapidly absorbed through

Fig. 2. Metabolic pathways of ddA and didanosine

**3.4 Lamivudine and emtricitabine** 

Abacavir (1S,4R)-4-[2-amino-6-(cyclopropylamino)-9H-purin-9-yl]cyclopent-2-en-1-yl]methanol) is a prodrug of carbovir (2-Amino-1,9-dihydro-9-[(1R,4S)-4-(hydroxymethyl)-2-cyclopenten-1-yl]-6H-purin-6-one), a deoxyguanosine analog (Figure 4; Daluge et al., 1997). Abacavir permeates T lymophoblastoid cell lines by passive diffusion. Abacavir is phosphorylated to abacavir-MP by adenosine phosphotransferase (Faletto et al., 1997). A yet unidentified cytosolic deaminase then converts abacavir-MP to carbovir-MP. Phosphorylation to the diphosphate derivative occurs via guanidinylate monophosphate kinase. The final phosphorylation step can be catalyzed by a number of cellular enzymes including 5' nucleotide diphosphate kinase, pyruvate kinase, and creatine kinase (Faletto et al., 1997). A linear dose relationship with carbovir-mono-, di-, and tri- phosphate derivatives over a 1000-fold dose range in vitro suggests there are no rate limiting steps in abacavir anabolism. The active metabolite carbovir-TP has been shown to persist with an elimination half-life of greater than 20 hours (McDowell et al., 2000). Abacavir bioavailability is ~83 % and is rapidly absorbed after oral dosing reaching peak plasma levels within 1 hour (Chittick et al., 1999). However, abacavir is extensively catabolized in the liver and only 1.2% is excreted as unchanged abacavir in urine. Abacavir oxidation by alcohol dehyrogenases to form the 5' carboxylic acid derivative represents 36% of metabolites recovered from urine, while the 5'- O-glucuronide corresponds to 30% of metabolites from urine (Chittick et al., 1999). Fecal excretion also accounts for approximately 16 % of the given dose. Abacavir is not metabolized by cytochrome P450 enzymes and does not inhibit these enzymes.

Molecular Pharmacology of Nucleoside and Nucleotide HIV-1 Reverse Transcriptase Inhibitors 69

However, the mono- and di-phosphate forms both inhibit purine nucleoside phosphorylase

Despite the widespread clinical success of NRTI-containing therapy, the currently FDA approved NRTIs display important limitations including the selection of drug resistance mutations that display cross-resistance to other NRTI, toxicity-related adverse events, and drug-drug interactions (for review see Cihlar & Ray, 2010). Thus, there is a need for novel NRTI that overcome these limitations. Here we will discuss the pharmacology of several

Apricitabine (ATC) is the (-)-enantiomer of 2'-deoxy-3'-oxa-4'-thiocytidine, a deoxycytidine analog that is currently in phase II/III clinical trials (Figure 6). Both the (+) and (-) enantiomers of apricitabine demonstrate potent inhibition of HIV-1 replication, however the (+)-enantiomer demonstrated significant mitochondrial and cellular toxicity in pre-clinical studies that was not observed with the (-) enantiomer (de Muys et al., 1999; Taylor et al., 2000). Racemic conversion of (-)-apricitabine to (+)-apricitabine is not observed in vivo (Holdich et al., 2006). Orally administered ATC is absorbed quickly, reaching maximal plasma levels within 2 hours with a plasma half-life of 3 hours. Maximal peripheral blood mononuclear cell (PBMC) intracellular concentrations of apricitabine -TP are achieved 3.5 – 4 hours after oral administration in healthy and HIV-infected patients. The intracellular halflife is 6 – 7 hours (Sawyer & Struthers-Semple, 2006; Cahn et al., 2008; Holdich et al., 2007). Apricitabine is not metabolized by hepatocytes *in vitro*, however a deaminated metabolite was observed likely due to gastrointestinal metabolism (Nakatani-Freshwater et al., 2006). This metabolite is excreted renally and does not demonstrate antiviral or pharmacologic effects. Apricitabine had no effect on cytochrome P450 or glucouronidase but was a weak inhibitor of P-glycoprotein (Sawyer & Cox, 2006). The first phosphorylation of apricitabine is

which is responsible for base removal of didanosine to form hypoxanthine.

Fig. 5. Metabolic pathways of tenofovir and TDF

**4. NRTI in the pipeline** 

novel drug candidates.

**4.1 Apricitabine** 

Fig. 4. Metabolic pathways for abacavir

#### **3.6 Tenofovir and tenofovir disoproxil fumerate**

The acyclic nucleoside phosphonate tenofovir (*R-*9-(2-phosphonylmethoxypropyl)-adenine) has no sugar ring structure but contains an acyclic methoxypropyl linker between the base N9 atom and a non-hydrolyzable C-P phosphonate bond. Thus tenofovir represents the only currently approved *nucleotide* HIV inhibitor. Tenofovir is poorly absorbed by the oral route and is therefore administered as a lipophilic orally bioavailable prodrug tenofovir disoproxil fumerate (TDF), a fumaric acid salt of the bis-isopropoxycarbonyloxymethyl ester of tenofovir (Figure 5). TDF is readily absorbed by the gastrointestinal epithelial cells with an oral bioavailability of 25% (Barditch-Crovo et al., 2001). Administration with a high fat meal increases absorption to 40%. Degradation of TDF to its monoester and subsequently to tenofovir occurs readily in the intestinal mucosa by the action of carboxylesterases and phosphodiesterases, respectively. The mono- or bis-ester forms of tenofovir are not observed in plasma suggesting efficient release of tenofovir following oral administration of TDF (Naesens et al., 1998). Following oral administration tenofovir has a long terminal half-life of 17 hours. The phosphonic acid linkage is chemically and metabolically stable and phosphorolysis back to the nucleoside does not occur (Naesens et al., 1998). Tenofovir is rapidly converted intracellularly to tenofovir-monophosphate and the active tenofovirdiphosphate forms by adenylate monophosphate kinase and 5'-nucleoside diphosphate kinase, respectively (Robbins et al., 1998). Tenofovir is not subject to intracellular deamination or deglycosylation. This stability results in a very long intracellular half-life for tenofovir-diphosphate of 15 hours in activated lymphocytes and 50 hours in resting lymphocytes (Robbins et al., 1998). Tenofovir is eliminated by glomerular filtration and active tubular secretion by organic anion transporter mediated uptake and MRP4 mediated efflux (Ray et al., 2006). At 72 hours post oral administration 70 - 80 % is recovered from urine as unchanged tenofovir. Tenofovir does not inhibit cytochrome P450 enzymes. However, the mono- and di-phosphate forms both inhibit purine nucleoside phosphorylase which is responsible for base removal of didanosine to form hypoxanthine.

Fig. 5. Metabolic pathways of tenofovir and TDF
