Summary Basis of Decision for Isentress ™ *

3.2.1 Pharmacodynamics

Primary Pharmacodynamics

Raltegravir was found in vitro, to be a selective inhibitor of strand transfer activity of purified recombinant HIV-1 integrase, with no HIV-1 reverse transcriptase inhibitory activity. It did not appear to have any significant inhibition of human DNA polymerases α, β, and γ, at concentrations up to 50 μM, and no in vitro activity in enzymes or receptors up to 10 µM.

Raltegravir showed potency similar to indinavir when tested against HIV-2. Furthermore, it retained activity against M-tropic (CCR5 receptors) and T-tropic strains (CXCR4 receptors), PI- and NNRTI-resistant isolates, and SI and NSI HIV-1 strains.

The sponsor believes that raltegravir will not inhibit the major double-strand DNA break-repair pathway in human cells as it was shown to inhibit strand transfer activity in vitro as cells infected with HIV-1 accumulated 2-LTR circle products.

Resistance

Among integrase mutations selected in cell culture by other structurally diverse experimental integrase inhibitors, the single mutation that caused the greatest reduction of raltegravir activity was N155S, with an approximate 19-fold resistance. The N155S mutation was shown to confer limited cross-resistance and reduced infectivity, although it appears that the reduced infectivity was specific for integrase inhibitors. Other mutations that demonstrated some level of change in the inhibitory concentration (IC50) of raltegravir were F121Y/T/125K (~9-fold); T66I/L74M/V151 (~8-fold); F121Y, S153Y/N155S, and T125K/S153Y/N155S (~5-fold); T66I, T125K, V151I, S153Y, T66I/M154, and T66I/S153Y, T66I/V151I (~1-fold).

In cell culture, raltegravir showed the resistance mutation Q148K which increased resistance by approximately 46-fold. When Q148K was combined with E138A, resistance increased to approximately 90-fold, and when combined with G140A, it increased to approximately 508-fold. The E138A and G140A mutations alone conferred an approximate 2-fold and zero resistance, respectively. Q148K played a major role in raltegravir resistance as it was the first mutation to occur in cell culture and all viruses with this mutation showed reduced infectivity.

In an interim analysis of 16-week data from the Phase II treatment-experienced study, Protocol 005, fifteen subjects with virologic failure from the clinical trials had full mutation profiles created. Most viruses from patients failing a raltegravir regimen tended to have a mutation at positions 148 or 155. Mutations Q148H and G140S were observed in nine subjects; Q148R was observed in three subjects; both Q148R and G140S were observed in one subject; N155H was observed in two subjects, and E92Q was observed in one subject. Secondary mutations frequently observed with the above mutations were G140S or E138A/K.

Mutations Q148H and Q148R conferred approximately 24-fold and 27-fold resistance, respectively. G140S added to the resistance if it occurred with either of the above mutations; G140S/Q148R showed approximately 405-fold increased resistance, and G140S/Q148H showed approximately 521-fold increased resistance. G140S was only observed with mutations at position 148 and so appears to be only a secondary resistance mutation, though it does appear to increase virological fitness and infectivity. The E92Q mutation, when combined with the N155H mutation, confers a 64-fold resistance and improves the infectivity of the N155H mutated virus from approximately 82% to 59%.

Three clinical isolates from virologic failure subjects were tested to ensure that integrase mutations confer resistance in the context of endogenous integrase genes and not only when introduced into a wild-type background. At virologic failure, Subject 1 had the double mutation G140S/Q148H with 211-fold reduced susceptibility to raltegravir. Subject 2 had an N155H mutation, which together with D232N, did not conform to common polymorphisms. Subject 3 was unusual in that it contained no mutations at N155 or Q148, but instead had six other changes. Substitution positions included 74, 121, and 151, and all had low-level raltegravir resistance.

The sponsor noted several important points with respect to resistance to raltegravir. It appeared that in both cell culture and clinical use, substitution of Q148 with any of the amino acids H, K, or R, conferred significant resistance and reduced infectivity of the virus. Q148 mutations appeared to be frequently linked to G140S mutations and this substantially increased resistance and improved virus infectivity. N155H alone conferred significant resistance and reduced virus infectivity but did not appear to be linked to other mutations, although its resistance was augmented and viral infectivity was increased in the presence of the E92Q mutation. It is important to watch for secondary mutations in the presence of N155H and Q148H/K/R mutations as the secondary mutations in general appeared to increase infectivity.

None of the previously identified HIV-1 integrase mutations G140S, V165I, F185K, V249I, C280S, and C280Y, were identified as conferring resistance to raltegravir. None of the mutations identified as conferring resistance to raltegravir are significant in cross-resistance to other antiretrovirals.

Safety Pharmacology

The appropriate battery of non-clinical safety pharmacology tests were conducted as per the ICH guidelines. With respect to cardiovascular effects, no meaningful effects on PR<QRS, QT, or heart-rate corrected QTc intervals were noted at any doses of raltegravir providing exposures similar or higher than those in humans. Furthermore, no significant drug-related increases in heart rate, mean arterial blood pressure, or blood flow were noted at doses of 1, 3, or 5 mg/kg in dogs.

After oral administration of 5 mg/kg raltegravir in dogs, no effect on urine output, urinary sodium, potassium excretion, or plasma electrolyte concentrations were noted. Glomerular filtration rate or filtration fraction did not appear to be affected and no emesis or changes in demeanour were noted in the 3-hour post-dose interval.

A 5 mg/kg intravenous (IV) dose of raltegravir in dogs showed no notable effect on respiratory mechanics, ventilation, blood pressure, heart rate, hemostasis, or platelet function.

Intestinal transit in the mouse was significantly increased after administration of 30 mg/kg of raltegravir compared to the vehicle alone (although the sponsor notes the vehicle group had below average transit). In a second study, gastrointestinal (GI) mobility was not significantly altered at a dose of 10 mg/kg. No adverse effects (behavioural, diarrhea) were noted.

A statistically significant but small increase in body temperature was noted in raltegravir-dosed mice during a 120-minute assessment phase. No effect on central nervous system behaviour including spontaneous activity, neural reflexes, or thermoregulation was noted after a 100 mg/kg dose.

Pharmacodynamic drug interactions

Antiviral activity was evaluated by measuring p24 antigen production in a multiple-cycle replication assay using the laboratory HIV-1 isolate H9IIIB and the human T-cell line MT4. Raltegravir did not show any antagonism with any antiretrovirals at any

concentration tested (from below-effective levels to levels at >99% inhibition). Depending on the concentrations used, interactions were all additive or synergistic. Synergism occurred at concentrations approaching IC95 and higher.

3.2.2 Pharmacokinetics

Absorption

Absorption was relatively rapid in both rats and dogs. Exposure in both species was good, with dose-proportional exposures achieved over the 40 to 120 mg/kg dose range in rats and the 5 to 45 mg/kg dose range in dogs. Absolute bioavailability in both dogs and rats was good at approximately 62% for rats and 70% for dogs. In rats it was found that the milled free phenol form of raltegravir was poorly bioavailable while the potassium salt form of the drug had reasonable and dose-proportional exposure when administered in 0.5% methylcellulose twice a day. A PEG400/water formulation yielded good exposures over the 40 to 120 mg/kg dose range without needing to be dosed twice a day, so it was selected for the sub-chronic safety studies in rats. In dogs, the potassium salt single daily dose in 0.5% methylcellulose showed good dose-proportional exposure over the dose range of 5 to 45 mg/kg. Exposure in dogs was higher than in rats.

Distribution

Single oral or IV doses of radiolabelled raltegravir were administered to rats and dogs. In rats, after a 3 mg/kg IV dose, plasma clearance was 45.5 mL/min/kg, steady-state volume of distribution was 2.2 L/kg, and half-life was 0.2 hours. After oral dosing at 6 mg/kg, Cmax was 1.7 µg at 0.4 hours post-dose. Clearance after oral dosing was 7.4 mL/min/kg in dogs.

After oral dosing, levels of radioactivity in plasma, blood, and blood cells reached maximum mean concentrations at 0.5 hours post-dose, with values of 601, 341, and 157 ng equivalents/g, respectively. Maximum mean concentrations were observed in most tissues at 0.5 hours post-dose. The highest mean concentrations were found in the GI tract and organs of excretion, particularly the stomach (30900 ng equivalents/g), small intestine (8970), liver (3910), kidney (1720), and urinary bladder (437). Maximum mean concentrations in other tissues ranged from 9.71 ng equivalents/g (brain) to 221 ng equivalents/g (lungs) at 0.5 hours post-dose. At 2 hours post-dose, maximum mean concentrations were observed in the small intestine and mesenteric lymph nodes, and by 6 hours post-dose, in the cecum and large intestine. Levels in most tissues were below the limit of quantitation by 24 hours post-dose. Aside from the GI tract, the highest concentrations of radioactivity were found in the liver, stomach at 0.5 hours, and small intestine at 2 hours. A single IV dose of raltegravir was administered to P-gp competent and deficient mice and concentrations in plasma and brain were determined. Raltegravir penetrated the brain poorly.

Metabolism

Raltegravir was mainly eliminated by metabolism in both rats and dogs, with changed drug in urine and bile accounting for 10 to 30% of the administered dose. The major metabolite detected was the phenolic glucuronide derivative of the parent compound (M2) which accounted for 62% of dose in rats and 31% in dogs. Other major metabolites included the glucose conjugate of the parent compound (M1) and the acetyl hydrazine derivative (M3).

No significant metabolism was observed in the liver microsomes fortified with NADPH from all three species tested. The major metabolite in rat and dog hepatocytes was the phenolic glucuronide conjugate of the parent compound (M2), and the same conjugate also was a major metabolite in hepatocytes from humans. Small quantities of the glucose derivative of the parent (M1) and an acetyl hydrazine derivative (M3) were also detected. The metabolic pathways were qualitatively and quantitatively similar in hepatocytes from all three species. The formation of the glucuronide metabolite M2 in human liver microsomes was characterized and UGT1A1 was found to play a major role in the glucuronidation of raltegravir. Raltegravir was not a potent reversible inhibitor (IC50 >100 μM) of CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4, or CYP2B6. Similarly, raltegravir did not potently inhibit (IC50 > 50 μM) β-estradiol (3-OH) and AZT glucuronidation mediated by UGT1A1 and UGT2B7, respectively. Also, raltegravir (up to 10 mM) did not induce CYP3A4 RNA expression or CYP3A4-dependent testosterone 6β-hydroxylase activity. Raltegravir was modestly bound to plasma proteins from rat, dog, and human. Raltegravir was not an inhibitor of human P-gp. The major metabolite in mice after a 20 mg/kg oral dose was the phenolic glucuronide derivative of the parent compound (M2, 96.3% of the radioactivity recovered in bile). M2 was also a major circulating metabolite along with the unchanged parent compound.

Excretion

In rats, the total recovered radioactivity after IV administration was 70.2% (19.9% in urine and 50.3% in feces) and after oral administration, 87.7% (23.7% in urine and 64.0% in feces). In dogs, the recoveries were 88.4% from IV dosing (30.8% from urine and 57.6% from feces) and 87.1% from oral dosing (13.2% in urine and 73.9% in feces).

3.2.3 Toxicology

Acute Toxicity

Acute oral toxicity studies were conducted in mice and rats to determine the approximate lethal dose (LD50), and in dogs for a toxicokinetic (TK) analysis. Raltegravir administered to female and male mice as a single oral dose of up to 2000 mg/kg caused treatment-related mortality in a third of the male mice at 2000 mg/kg. The LD50 of raltegravir was ≥2000 mg/kg in mice. Raltegravir administered to three female rats by oral gavage produced no treatment-related mortality, physical signs, or effects on body weight at 2000 mg/kg. The LD50 for raltegravir in female rats was >2000 mg/kg. Administration of single-rising doses of up to 1000 mg/kg of raltegravir to dogs produced treatment-related emesis at ≥500 mg/kg and no mortality.

Repeat-Dose Toxicity

Repeat-dose toxicity studies with raltegravir were conducted in mice, rats, and dogs.

Raltegravir, administered by oral gavage once daily (OD) to female mice at doses up to 2500 mg/kg/day and male mice at doses up to 1000 mg/kg/day for approximately 5 weeks, caused mortality at ≥500 mg/kg/day. At the minimum lethal dose (MLD) of 500 mg/kg/day, the systemic exposure to raltegravir (45.2 μM•hr, or 13.56 μM•hr unbound drug for mice) was approximately 1.48-fold greater than the exposure (AUC) (54 μM•hr, or 9.18 μM•hr unbound drug) at the 400 mg, twice daily maximum recommended dose (BID MRD). The maximum-tolerated dose (MTD) in mice was exceeded at 500 mg/kg/day. Raltegravir, administered to mice at doses up to 5000 mg/kg/day for 14 weeks, produced treatment-related mortality at doses ≥500 mg/kg/day, as well as treatment-related physical signs and body weight loss, and/or decreases in body weight gain (-30%). Histomorphologic observations were consistent with the clinical signs of gastrointestinal bloating and were suggestive of an irritant effect from the compound. The no-effect level (NOEL) for mice was 50 mg/kg/day.

Repeat-dose oral toxicity studies of 5, 14, and 27 weeks duration were conducted in rats. Raltegravir doses up to 120 mg/kg/day caused post-dosing salivation. At doses up to 600 mg/kg/day for 5 weeks, minor increases in alanine aminotransferase (ALT) and evidence of irritation to the stomach mucosa were evident. Twice-daily dosing of raltegravir at 600 mg/kg/day did not provide a significant increase in systemic exposure to rats. Raltegravir caused mortality at 600 mg/kg/day in the 27-week study that was associated with body weight loss, urine staining, decreased food consumption prior to death, as well as decreases in mean body weight gain. Degeneration of the glandular mucosa of the stomach was seen at ≥120 mg/kg/day. Additional observations in rats included stomach inflammation in the 5-week study, and post-dosing salivation, audible respiratory sounds, and inflammation in the nose and nasopharynx in the 27-week study. Therefore, the NOEL in rats following 27 weeks of treatment was 30 mg/kg/day. The no-adverse-effect-level (NOAEL) was 120 mg/kg/day, which provided an exposure margin of 1.6-fold above the expected exposure in patients with 400 mg/BID.

Raltegravir was tested in oral toxicity studies of 5, 14, 27, and 53 week's duration in dogs. Doses up to 45 mg/kg/day were administered for 14 weeks, then subsequently doses were administered at 500 mg/kg/day for 5 weeks or 360 mg/kg/day for 53 weeks. There were no treatment-related deaths or post-mortem changes. Treatment-related changes included transient and/or intermittent emesis with or without body weight loss. Therefore, the NOEL in dogs, with the exception of emesis, following 53 weeks of treatment, was ≥360 mg/kg/day. The exposure margin was 9-fold above the expected human exposure at the 400 mg/BID MRD.

Genotoxicity

Raltegravir was not mutagenic when tested in vitro in bacteria with and without metabolic activation. It did not induce a significant increase in DNA single strand breaks in the in vitro rat hepatocyte DNA alkaline elution assay. Raltegravir treatment groups did not show statistically significant increases in the frequency of chromosome aberrations in the Chinese hamster ovary cell chromosome aberration assay with and without metabolic activation. In an in vivo mouse micronucleus assay at doses up to 1500 mg/kg, there was no evidence of an increase in micronuclei or on the proportion of polychromatophilic erythrocytes at any of the doses tested.

Carcinogenicity

The carcinogenicity studies in mice and rats are ongoing. The in-life phase of these studies is to be completed in 4Q 2007 with final study reports planned to be completed in 3Q-4Q 2008. In mice, doses up to 400 mg/kg/day (females) or 250 mg/kg/day (males) were used. In rats, doses up to 300 mg/kg/day (males) or 600 mg/kg/day (females) were selected. Preliminary results up to 76 weeks showed increasing mortality of up to 56% in female mice and 46% in female rats. Effects of drug irritation to the nose/nasopharynx including inflammation, epithelial hyperplasia, and squamous metaplasia were seen in rats. Six neoplasms of the nose/nasopharynx were detected in rats. Similar effects of drug irritation to the nose, nasopharynx, and trachea including inflammation, epithelial hyperplasia, and squamous metaplasia were seen in mice; however, no malignancies were detected. In a 27-week TK study with mice, systemic exposure at the high dose was approximately 1.8-fold (females), or equal to (males), the human AUC at the 400 mg/BID MRD. A 26-week parallel TK study in rats indicated that at 300-600 mg/kg/day (female), or 150-300 mg/kg/day (male), systemic exposure was approximately 10-fold (female), or 1.7-fold (male), of the human AUC at the 400 mg/BID MRD.

Reproductive and Developmental Toxicity

Administration of raltegravir orally OD to rats [females: for 14 days prior to and during co-habitation, and through gestation day (GD) 7; males: from prior to co-habitation for approximately 8 weeks] with doses up to 600 mg/kg/day did not demonstrate any treatment-related effects on mating performance, fertility parameters, F1 embryonic/fetal survival, sperm count, or sperm motility. No treatment-related gross changes in the thoracic or abdominal cavities, testicular weights, or gross or microscopic changes in the testes or epididymides were observed. The NOEL for effects of raltegravir on fertility in F0 female and male rats was ≥600 mg/kg/day.

Raltegravir, administered orally OD to pregnant rats at doses up to 600 mg/kg/day from GD 6 through lactation day (LD) 20, did not cause treatment-related changes. Following raltegravir administration to female rats, doses up to 600 mg/kg/day from GD 6 to 20 or from GD 6 to LD 20 showed no treatment-related findings in the F0 and F2 generation. Treatment-related increases in incidences of supernumerary ribs were observed in the F1 generation of the 600 mg/kg/day group. The NOEL for maternal toxicity was ≥600 mg/kg/day. The NOEL for developmental toxicity was 300 mg/kg/day (3.4-fold the exposure in patients receiving the 400 mg/BID MRD).

There were no indications of maternal or developmental toxicity following oral administration OD to rabbits at doses up to 1000 mg/kg/day from GD 7 to 20. The NOEL of raltegravir for maternal and developmental toxicity in rabbits was ≥1000 mg/kg/day.

Doses up to 600 mg/kg/day were orally administered to juvenile rats from post-natal day (PND) 5 to postnatal week (PNW) 8. No treatment-related changes were observed except for treatment-related histomorphologic findings consisting of vacuolation of the nonglandular mucosa at the limiting ridge and inflammation at ≥200 mg/kg/day. The NOEL for juvenile rats was 50 mg/kg/day. The findings in juvenile rats were consistent with the stomach irritation effects seen in adult rats. There were no additional toxicities noted in juvenile rats indicating that, overall, raltegravir effects were similar between juvenile and adult rats.

Local Tolerance

Raltegravir potassium was non-irritating based on 104.2% cell viability in an in vitro human epidermal skin culture system (EpiDerm™).

Based upon assessment of the induction of opacity and changes in permeability to fluorescein (irritation score) with raltegravir tested as a 20% solution, raltegravir free base was classified as a mild irritant and raltegravir potassium salt was classified as a severe irritant in the bovine corneal opacity (BCOP) in vitro assay.

Groups of mice were treated topically with raltegravir, later injected with tritiated thymidine or BrdU, sacrificed and incorporation of radioactivity in the auricular lymph node measured as an index of sensitization potential. No significant increase in cell proliferation in the lymph node was detected, and as such, raltegravir, as either free base or potassium salt, was not considered a dermal sensitizer using the local lymph node assay.

The dermal irritation potential of raltegravir as a free base or potassium salt was determined following application of the drug to the intact skin of New Zealand White (NZW) rabbits. No evidence of dermal irritation was found in treated rabbits. Raltegravir, as either a free base or potassium salt, was not considered a dermal irritant based on the results of this study.

Other Toxicity Studies

The phototoxicity potential of raltegravir was determined after oral administration to female mice at doses up to 2000 mg/kg as a single dose then exposing mice to UVB, UVA, and visible light. Treatment-related mortality, body weight effects, or phototoxic effects were not noted. Treatment-related decreased activity was observed at 2000 mg/kg. Raltegravir was considered non-phototoxic at doses up to 2000 mg/kg.

Raltegravir was negative for in vitro hemolysis of human, dog, or rat erythrocytes prior to initiation of the intravenous (IV) toxicity studies. Raltegravir administered as a single IV injection to female rats at doses up to 1600 mg/kg caused treatment-related mortality associated with physical signs at ≥200 mg/kg. Evidence of drug irritation at the injection site was observed at >100 mg/kg. In a range-finding study, dogs received IV rising doses of 40 and 100 mg/kg/day for three consecutive days each. One dog died on Day 1 during IV administration of the 40 mg/kg dose. The death was attributed to cardiac arrhythmia secondary to the amount of administered potassium. Increases in alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP) at 100 mg/kg/day were observed. Raltegravir IV administered to dogs OD at doses of 30 or 100 mg/kg/day for seven days produced primary changes at both doses including transient post-dosing emesis/retching and local changes at the injection sites, as well as body weight loss, minimal increases in serum urea nitrogen, increases in ALT (up to 1967%), ALP (up to 374%), and cholesterol, and very slight, multifocal tubular dilatation in the cortex of the kidneys at 100 mg/kg/day. At the NOEL, the safety margin was 6.5-fold that of the expected exposure of the 400 mg/BID MRD in humans.

Raltegravir was administered to rats at 30, 120, 600 mg/kg/day by oral gavage OD for approximately 5 weeks. Treatment-related salivation at the mid- and high-dose levels was observed. Raltegravir caused very slight irritation to the limiting ridge of the non-glandular stomach accompanied by a very slight increase in inflammation at 600 mg/kg/day. The NOEL for the local effect on the stomach was 120 mg/kg/day. Administration of raltegravir containing 0.03% of the acid impurity/degradate to rats for one month did not cause any unique toxicity relative to raltegravir alone, nor was the NOEL for the gastric irritation significantly altered. Results from this study support a Drug Product Specification limit for the acid impurity/degradate up to 0.3% (on mg/kg basis).

No toxicity studies were conducted to assess antigenicity, immunotoxicity, or dependence. No studies were conducted to evaluate specific metabolites.

3.2.4 Summary and Conclusion

Raltegravir was found, in vitro, to be a selective inhibitor of strand transfer activity of purified recombinant HIV-1 integrase with no HIV-1 reverse transcriptase inhibitory activity. The non-clinical pharmacology and pharmacokinetic studies were considered to be thorough and acceptable. No major areas of concern were identified.

The toxicity profile of raltegravir after repeated oral dosing was characterized in rats, mice, and dogs and demonstrated that raltegravir was well tolerated in animals. Doses used in toxicity studies provided adequate systemic exposure and margins of safety were determined for each of the toxicities identified. Raltegravir was not genotoxic in a battery of in vitro assays in bacteria and mammalian cells or in vivo in mice designed to detect mutagenicity, direct DNA damage, or elastogenicity. An assessment of the carcinogenic potential of raltegravir is ongoing, preliminary results up to Week 76 indicated increased mortality, evidence of drug irritation, and neoplasms. The relevance of these findings to human safety will be assessed in the clinical studies. Raltegravir was shown not to pose a significant hazard to reproduction or to the developing fetuses based on studies in rats and rabbits.