Author for correspondence: Victor E. Buckwold, Veracity Biotechnology, LLC, 401 Rosemont Ave., Third Floor Rosenstock Hall, Frederick, MD 21701, U.S.A. (fax +1 301 644 3939, e-mail Vbuckwold@veracitybiotech.com).
Abstract: Human ω-interferon (IFN-ω) has been shown to be well-tolerated in man and to induce reductions of hepatitis C virus RNA levels in a series of human clinical trials. Here we provide an overview of our preclinical safety evaluation of the fully-glycosylated human IFN-ω produced from CHO-SS cells that is currently being evaluated clinically. IFN-ω was not associated with any biologically-relevant adverse effects in a series of 10 safety pharmacology experiments, in the Ames mutagenicity test, in the micronucleus test, or in intraarterial, intravenous, paravenous or subcutaneous local tolerance studies. Acute, subacute, subchronic and reproductive toxicity studies performed in cynomolgus monkeys and rats showed a toxicity profile similar to that of human α interferon (IFN-α). Except for the acute (single-dose) toxicology study, all of the other toxicity studies showed evidence for the formation of anti-IFN-ω antibodies over time in the animals. These antibodies were found to neutralize IFN-ω antiviral activity in vitro in a dose-dependent manner. The average pharmacokinetic parameters following a single subcutaneous dose of IFN-ω in rabbits, rats and monkeys were determined and found to be similar to that of human IFN-α. These findings demonstrate that IFN-ω has a safety profile consistent with that required for its use in man. IFN-ω might be beneficial for the treatment of patients infected with hepatitis C virus who fail to respond to IFN-α or as a first-line treatment option.
Infection with the hepatitis C virus is the most common cause of chronic hepatitis in the United States and is a major risk factor for the development of liver cirrhosis (Liang et al. 2000). Hepatitis C virus infection is currently treated using a combination therapy of pegylated IFN-α and ribavirin. This treatment leads to a sustained virologic response (SVR=absence of detectable hepatitis C virus RNA in patients six months after cessation of therapy) in approximately 55% of all patients (Pearlman 2004). Hepatitis C virus is classified into one of six different genotypes, each composed of a variety of related subtypes (Simmonds et al. 2005). The response of hepatitis C virus to IFN-α and ribavirin is known to be genotype-specific with an SVR=40–50% for patients infected with hepatitis C virus genotype 1 and an SVR=70–80% for those infected with genotypes 2 or 3 (Pawlotsky 2003). The current literature, although limited, suggests that hepatitis C virus genotype 4 responds poorly to this treatment, that genotype 5 responds favourably and that genotype 6 is intermediate in response (Nguyen & Keefe 2005). Other antiviral therapies to treat those patients in whom IFN-α and ribavirin treatment has failed are desperately needed.
The interferons (IFNs) were originally discovered as biological agents that “interfered” with viral replication (Stark et al. 1998). Human ω interferon (IFN-ω), like other IFNs, is secreted from cells in response to viral infection and displays antiviral, antiproliferative and immunomodulatory activities (Adolf 1995). This type I IFN has 62% amino acid identity with α-interferon (IFNα-2) and 33% amino acid identity with β-interferon (IFN-β) respectively, but it is unrelated to the type II γ-interferon (IFN-γ) (Adolf 1995). As a distinct IFN, IFN-ω, alone or in combination with ribavirin or other antiviral therapies, might therefore be beneficial for the treatment of patients who fail to respond to IFN-α or as an additional first-line treatment option. Recombinant human IFN-ω produced from Chinese hamster ovary cells adapted to serum free growth in suspension culture (CHO-SS) has been shown to be well-tolerated in man and induced a sustained virologic response in patients infected with hepatitis C virus genotypes 1, 2 and 3 in clinical trials (McHutchison et al. 2001; Jetschmann et al. 2002; Plauth et al. 2002; Gorbakov et al. 2005). Here we provide an overview of our preclinical safety evaluation of the CHO-SS cell-derived IFN-ω that we are evaluating for the treatment of hepatitis C virus infection.
Materials and Methods
IFN-ω. IFN-ω was from stably-transfected CHO-SS cells (Adolf et al. 1991). Vials of IFN-ω containing 5.0 mg human serum albumin (HS; Bayer Biological Products), 0.20 mg KCl, 0.20 mg KH2PO4, 8.0 mg NaCl, 2.89 mg Na2HPO4.12H2O were reconstituted in sterile water (H2O) before use and the concentration of IFN-ω was verified by enzyme linked immunosorbent assay (ELISA, see below). Vehicle control was the same formulation without IFN-ω.
Statistical analysis. Unless otherwise indicated P<0.05 was considered significant for all the statistical analyses performed when comparing the control and treatment groups.
In vitro safety experiments. These experiments were conducted at Boehringer Ingelheim, Germany (BI) according to Good Laboratory Practice (GLP).
Haemolysis test. Samples were mixed 1:1 with citrated human blood (0.38% sodium citrate), incubated 45 min. at 37 °, centrifuged (1,000×g for 5 min.), the supernatant diluted 20 times with H2O and the absorbance at 578 nM read. The supernatant of 1:1 citrated blood:saline was used as the blank and 1:1 citrated blood:1% saponin in saline was used as the complete haemolysis standard. The blood of 3 different donors was used. No haemolysis was indicated when samples showed less than 2% of the absorbance of the complete haemolysis standard.
Micronucleus test. This study was conducted as previously (Schmid 1975). IFN-ω was given intravenously to SPF Chbb:NMRI mice. Negative (vehicle control, water) and positive control (20 mg/kg cyclophosphamide) groups were utilized for each experiment. Animals were euthanized 24 hr after drug administration (48 hr for the 2,000 μg/kg study). Statistical analysis was via the Fisher-Pitman permutation test (Leimer et al. 1991).
Safety experiments in mice. These studies were conducted at BI. Fasted SPF Chbb:NMRI mice were treated with intravenous IFN-ω and general behaviour was assessed 30 min. after drug administration (Irwin 1968). Open-field behaviour was assessed by monitoring catalepsy, loss of the righting reflex, stereotypy, as well as measuring when animals did not cross squares over time. Analgesic effects were assessed by measuring reaction time on a hot plate at 52 ° for 45 sec. Rectal temperature was taken. The significance of the difference between groups was evaluated using the one-way analysis of variance (ANOVA) including the Bonferroni t-test.
Experiments using rabbits (GLP). These studies were conducted at BI. Respiratory and cardiovascular function. New Zealand White (NZW) rabbits were anesthetized and the right femoral artery cannulated for the measurement of mean systemic arterial blood pressure (BP) using an external pressure transducer. A microtip Millar catheter was introduced into the left ventricle for the measurement of left ventricular pressure (LVP) as LVPmax and its first derivative (dP/dtmax), as well as respiratory rate and maximal inspiratory and expiratory flow. Tidal and esophageal volumes were recorded (ADAS system, Thomae Biberach) and an electrocardiogram (ECG) was used to calculate heart rate. A jugular transducer was cannulated for the intravenous administration of substances. A low pressure differential transducer connected to the trachea and oesophagus allowed for the measurement of oesophageal pressure. Respiratory flow was recorded through a Fleisch tube connected to a tracheal tube. Animals received increasing doses of IFN-ω with 30 min. between doses; control animals received vehicle control. Parameters were measured just before and then 1, 5, 10, 15 and 30 min. after drug administration. At each time point a between groups comparison was performed using ANOVA followed by a post hoc t-test when the ANOVA was significant.
Intraarterial tolerability. NZW rabbits were used. The hair on the ears was removed, the injection region disinfected and 0.5 ml of saline, vehicle control or IFN-ω was injected into the arteria auricularis; haemoperfusion was stopped for 20 min. using a soft clamp. The ear was inspected daily until day 7. Local reactions were scored numerically for the diverting part of the arteria. Statistical analyses were not utilized.
Intravenous tolerability. Rabbits were used as above except that test solutions were injected into the vena auricularis rostralis, the ear was inspected until day 8, and reactions were scored for the diverting part of the vein. Statistical analyses were not utilized.
Rabbit pharmacokinetic studies. NZW rabbits were used to examine IFN-ω subcutaneous administration. One animal in the 2 μg/kg group had detectable IFN-ω by ELISA in pre-dose samples and this data was not used. The injections were given as a bolus into a right ear vein of 12 hr fasted animals. Blood was sampled via an indwelling cannula in the left ear before and at 0, 0.5, 1, 2, 4, 6, 8, 10, 24, 32 and 48 hr after drug administration. Serum was frozen and analyzed for IFN-ω content by ELISA. Pharmacokinetic parameters were determined using the non-compartmental methods of TopFit (Tanswell & Koup 1993).
Experiments using rats. These studies were conducted at BI. Statistical significance between the control and treatment groups was evaluated using the one-way ANOVA including the Bonferroni t-test, except for as indicated.
Exploratory motility study. Exploratory motility was assessed from 0–60 min. after drug administration using the Actiframe system (GERB electronics, Germany) in fasted male SPF Chbb: THOM strain rats.
Hexabarbitone-induced sleeping time. Was assessed in fasted SPF Chbb:THOM rats, 30 min. after drug administration (Winter 1948). Rats received 70 mg/kg sodium hexabarbitone intraperitoneally, were placed on a metal plate at 37 ° and the time from the disappearance of the righting reflex up to its recovery was measured. The change relative to the placebo group was determined.
Hepatic and renal function. Chbb:THOM strain rats received intravenous IFN-ω. Urine was collected from 8 animals (4 males, 4 females)/group and blood samples were taken from the other animals at 4, 8 and 24 hr after drug administration. Animals used for urine collection received a tap water loading of 2 ml/100 g when urine and blood were collected. Part of the urine was centrifuged (1500 r.p.m. for 5 min.), filtered through G-25 Sephadex (Pharmacia) and the columns washed twice with saline. Samples were stored frozen for blood chemistry and urinalysis. At each collection time, a comparison between groups was performed using ANOVA and a post hoc Dunnett test when the ANOVA was significant.
Paravenous tolerability (GLP). Chbb:THOM SPF rats were shaved in the ventral and lateral region of the neck and the right jugular region was disinfected. Vehicle control or IFN-ω (0.2 ml) was injected into the jugular vein into the subcutaneous tissue. At 6 and 24 hr after drug administration, animals were anaesthetized, exsanguinated and the injection sites subjected to a gross pathological examination. A numerical score based on both skin reaction and macroscopic findings was assigned followed by a cumulative score. Statistical analyses were not used.
Subcutaneous tolerability (GLP). Chbb:THOM SPF rats were shaved in the os sacrum area and the region disinfected. Test solutions (0.5 ml) of saline, vehicle control, positive control (40 g calcium chloride hexahydrate, 15 g magnesium chloride, 1 g caffeine monohydrate, 50 g invert sugar, 500 ml water) or IFN-ω were injected subcutaneously and the area was marked with a felt-tipped pen. The animals were examined as in the paravenous tolerability study.
Toxicology studies. All animals in the toxicology studies were monitored for signs of ill health and reaction to treatment, rectal temperature, blood pressure, electrocardiogram, body weight, and food consumption. At the end of the study all animals were euthanized and a full necropsy was performed. These data, as well as those specific additional measurements taken in the various toxicological studies, were analyzed in all of the control versus treatment group comparisons.
Subacute toxicity in rats (GLP). These studies were conducted at BI. Sprague-Dawley rats were dosed once daily by subcutaneous bolus injection for 28 days. Four injection sites were alternated. Blood samples for the determination of IFN-ω concentration were collected on: day 1: pre-dose (0), 2, 4, 8 and 24 hr after drug administration; days 7, 14, 21: pre-dose (0), 2 hr after drug administration; day 28: pre-dose (0), 2, 4, 8, 24 and 48 hr after drug administration. Blood samples for the detection of antibodies to IFN-ω were collected on: day 1 and 14: pre-dose (0) and >48 hr after drug administration day 28. Serum was frozen. Laboratory investigations of haematology, clinical chemistry and urinalysis were undertaken and each animal was subjected to full histological evaluation. Haematology, clinical chemistry and organ weight data were analyzed for homogeneity using the F-max test. If the group variances appeared homogeneous, a parametric ANOVA was used and pair-wise comparisons were made via Student's t-test. If the variances were heterogeneous, log or square root transformations were used. If the variances remained heterogeneous, then the non-parametric Kruskal-Wallis ANOVA was used.
Acute toxicity and toxicokinetic monkey study (GLP). This study was conducted at Inveresk Research International, Ltd. (Tranet, Scotland, UK) (Inveresk). Cynomolgus monkeys (Macaca fascicularis) received a single subcutaneous or intravenous injection. The animals first received the intravenous injection followed by a washout period of 2 weeks before subcutaneous drug administration. Intravenous injections were in the saphenous vein while subcutaneous injections were in the mid-dorsal lumbar region. Blood samples were taken at 15 and 1 days before dosing, just before dosing and at 0.5, 1, 4, 24, 48 and 72 hr after drug administration for the analysis of IFN-ω and anti-IFN-ω antibody levels. Statistical analyses were not utilized. Pharmacokinetic parameters were derived as above.
Subacute toxicity and toxicokinetic study. This study was conducted at Inveresk. Cynomolgus monkeys were dosed once daily subcutaneously in the mid-dorsal region using four alternating injection sites. Ocular examinations and laboratory investigations of haematology, clinical chemistry and urinalysis were undertaken. Blood samples were collected for the determination of IFN-ω levels on days 1 and 27 (pre-dose and 2, 4, 8, 12 and 24 hr after drug administration), 7, 14 and 21 (pre-dose and 2 hr after drug administration). Samples for the determination of anti-IFN-ω antibody levels were taken on days 1, 14 and 28. The animals were subjected to a full histological evaluation. Statistical analyses were not utilized. Pharmacokinetic parameters were derived as above.
Subchronic toxicity and toxicokinetic study. This study was conducted at Inveresk in cynomolgus monkeys treated once daily with saline, vehicle control or IFN-ω subcutaneously in the mid-dorsal region for 13 weeks. Four injection sites were alternated. The high dose group contained an additional 2 female and 2 male animals that were retained for a 4 week recovery period. Laboratory investigations of haematology, clinical chemistry and urinalysis were undertaken. Blood samples were collected for the quantification of IFN-ω and anti-IFN-ω antibody levels on day 1 and the last day of weeks 1–13. Another blood sample was taken at the end of the recovery period. Statistical analyses were not utilized.
Reproductive toxicity study. This study was performed at Covance Laboratories GmbH (Münster, Germany). The effects of IFN-ω on embryos, foetotoxicity and teratogenicity was assessed using sexually mature female cynomolgus monkeys. Pregnancy was identified via ultrasound on day 16 past-coitus. IFN-ω was administered daily into the lateral abdominal region from day 16–80 of gestation. Pregnancies were terminated on day 100±1 past-coitus by cesarean section. Vaginal smears, serum progesterone and 17-β-oestradiol levels were monitored. Serum samples were analyzed for IFN-ω and IFN-ω antibody content from days 16–71 of gestation. The necropsy included an analysis of the fetus for skeletal defects (Wilson & Warkany 1965) and organ histopathology. Data were analyzed using the Bartlett's test for homogeneity of variance followed by a rank transformation. In the case of heterogeneity only the Bartlett's test was employed. For homogenous data, the one-way ANOVA was performed with the Dunnett's two-tailed t-test used when the ANOVA was significant. In the case of heterogeneity, the Kruskal-Wallis test was used together with the Wilcoxon-rank-sum test.
Antiviral activity evaluation and antibody neutralization studies. These studies were performed as previously described (Adolf 1990).
Quantitative IFN-ω ELISA (GLP). Briefly, 100 μl of 3 μg/ml anti-IFN-ω antibody OMG 5 (Adolf 1990) in 50 mM sodium carbonate buffer, pH 9.6 was pipetted into the wells of 96-well plates and incubated for 1 hr at 37 ° or overnight at 4 °. The plates were washed with H2O in a plate washer (Skatron Multiwash II). 200 μl assay buffer (2.5 g bovine serum albumin, 0.05% Tween:phosphate buffered saline (PBS) pH 7.4)/well was added and the plate incubated for 1–2 hr at room temperature. The plates were washed and samples, IFN-ω standards (7.8–1,000 pg/ml in sample diluent (10% thimerasol in normal HS pool (PAA Labor)) or sample diluent were added to the appropriate wells in duplicate (100 μl/well). Next, 50 μl anti-IFN-ω antibody OMG 7 (Adolf 1990) conjugated with horseradish peroxidase was added/well, the plates covered and shaken for 30 min. at room temperature. Plates were washed in 3 cycles and blotted dry. Two hundred μl/well substrate solution (10 mg tetramethyl-benzidine dihydrochloride/3 ml dimethyl sulfoxide to 50 ml 100 mM citrate buffer, pH 5.0 followed by the addition of 50 ml 10% NaBO2.H2O2.3H2O:H2O) was added, the plates were covered and shaken for 30 min. at room temperature. Reactions were stopped with 50 μl 2 M sulphuric acid and the plates read at 450 nM. Absorbance was proportional to the concentration of IFN-ω. Calculations were performed using 4-parameter curve fitting, log-logit transformations or by plotting log absorbance versus log standard concentration followed by interpolation based on the best-fit curve generated.
Quantitative anti-IFN-ω antibody ELISA (GLP). Plates were coated with 100 μl/well streptavidin solution (2 mg/ml streptavidin in 0.01% thimerasol:PBS, diluted to 0.3 μg/ml in 50 mM sodium carbonate buffer, pH 9.6), incubated 1 hr at 37 ° or overnight at 4 °, washed with H2O, coated with 100 μl biotinylated IFN-ω (1:10,000 in assay buffer) and incubated 1 hr at 37 ° or overnight at 4 °. The plates were washed, samples of 100 μl anti-IFN-ω reference antibody IFN-ω-AK-ELISA (0.78–100 ng/ml in antibody ELISA sample diluent (10% normal horseradish peroxidase pool in 0.01% thimerasol:PBS), unknowns or antibody ELISA sample diluent were added to the appropriate wells in duplicate. Next 50 μl/well horseradish peroxidase-conjugated human IFN-ω (1:10,000 in assay buffer) was added and shaken for 2 hr at room temperature. Plates were washed for 3 cycles and blotted dry. Next 200 μl/well of substrate solution was added, shaken 30 min at room temperature, reactions were stopped, absorbance read and the data analyzed as above. Absorbance was proportional to the concentration of anti-IFN-ω antibodies.
The concordance between the anti-IFN-ω antibody ELISA and the neutralization assay was examined using 4 samples from 3 monkeys and 18 samples from 9 rats from the toxicokinetic studies. All samples with an ELISA antibody concentration of >1,000 ng/ml also had a measurable antibody titer in the neutralization assay, while samples with <100 ng/ml antibodies by ELISA showed no neutralizing activity (data not shown). The minimum concentration of IFN-ω-binding antibodies by ELISA that was necessary to show a titer in the antibody neutralization assay was 100–1,000 ng/ml. When the results were plotted as the log antibody titer by ELISA versus the log antibody titer using the neutralization assay, a correlation coefficient=0.815 was determined (statistically significant, F-test P<0.01).
There were no important pharmacological effects of IFN-ω treatment in a wide variety of safety pharmacology experiments that were performed (table 1). Some minor drug-related renal effects were observed in rats (table 2). Specifically, serum Mg2+ was reduced at 8 hr after drug administration, but not at 4 or 24 hr after drug administration. Serum blood urea nitrogen (BUN) was reduced only at the 4 hr after drug administration time-point at the highest dose tested. Unfiltered urine pH was reduced, while unfiltered urine Cl− was increased, but only at the 4 hr time-point at the highest dose tested. These changes appear to have little biological relevance. IFN-ω at up to 500 μg/kg caused no major acute renal effects in rats.
Table 1. Safety pharmacology experiments performed with IFN-ω.
There was no evidence of a toxic or genotoxic effect of IFN-ω in the Ames test with or without metabolic activation or in the mouse micronucleus test (table 3). Intra-arterial, intravenous, paravenous and subcutaneous local tolerance tests conducted in rabbits and rats demonstrated that all doses and routes of administration of IFN-ω were well tolerated (table 3).
Table 3. Mutagenicity and local tolerance studies performed with IFN-ω.
A summary of the toxicology experiments undertaken is shown in table 4. A single intravenous or subcutaneous administration of 2,000 μg/kg IFN-ω into cynomolgus monkeys was not associated with any signs of acute toxicity. Also, the single administration of IFN-ω did not result in the induction of anti-IFN-ω antibodies during the study period.
Increased incidence of abortions at 10 and 50 μg/kg/day Neutralizing antibodies produced
Daily administration of various doses of IFN-ω subcutaneously for 28 days was undertaken in a subacute toxicity study in cynomolgus monkeys. Summaries of the drug-related changes observed are shown in tables 2 and 4. A slight decrease in red blood cell indices in the male animal at the intermediate and high dose group was observed, but no significant differences in these parameters were noted for the female animals or for any animal treated at 1 μg/kg/day. A decrease in lymphocytes and platelet levels was noted for the male animal at the highest dose tested. A slight decrease in platelets was also observed in the female animal at the highest dose tested. No effects on lymphocyte levels in the other groups of females were observed. No other toxic effects were noted. Anti-IFN-ω antibodies as detected by ELISA were present in one animal each (50%) in the 10 and 100 μg/kg/day groups only at day 28. All other samples did not contain detectable antibodies.
Daily administration of various doses of IFN-ω subcutaneously for 28 days was also undertaken in a subacute toxicity study in rats. A summary of the drug-related effects observed is shown in tables 2 and 4. IFN-ω was associated with a reduction in body weight gain for all animals in the high dose group and also for female animals in the intermediate dose group. Food intake was slightly reduced in high dose male animals only. There were no other treatment-related changes observed. Dose-proportional plasma concentrations of IFN-ω were measured on day one, however there did not seem to be any significant accumulation of IFN-ω following day 1 after drug administration. From day 7 onwards no measurable plasma concentrations of IFN-ω could be detected in most animals 2 hr after drug administration. Anti-IFN-ω antibodies were assayed from day 1, 14 and 28 samples by ELISA. No antibodies were detected from any animals on day 1. At day 14, 50%, 33% and 50% of the 1, 10 and 100 μg/kg/day-treated animals respectively showed low levels of anti-IFN-ω antibodies (12–221 ng/ml). By day 30, high levels of anti-IFN-ω antibodies were detected in all animals with mean levels of 1,140, 7,430 and 25,000 ng/ml in the 1, 10 and 100 μg/kg/day groups, respectively. These antibodies were shown to be neutralizing in the EMCV neutralization assay (data not shown).
Subchronic toxicity after daily subcutaneous injection was assessed in cynomolgus monkeys in a 13 week study (table 4). One animal dosed at 10 μg/kg/day IFN-ω died for unknown reasons. There was no evidence to indicate an effect of IFN-ω and the death of this animal was attributed to chance. No substance-related effects were found on clinical signs, body weight development, food consumption, clinical pathology, ECG, blood pressure, ophthalmology, body temperature, organ weights, or gross or histological pathology in any animal. There were no toxicological effects noted at any dose range tested and thus the no toxic effect dose was found to be 50 μg/kg/day. A clear inverse relationship between the serum concentration of IFN-ω and the concentration of anti-IFN-ω antibodies was observed (fig. 1). The serum concentration of IFN-ω was proportional to the dose administered; however the concentration of IFN-ω began to wane from week two or three on. All animals showed the presence of antibodies by week three. The titers of these antibodies tended to increase in all animals throughout the study, except in one animal from the 1 μg/kg/day treatment group in which the anti-IFN-ω antibody levels were reduced to low levels after week three. This animal showed high levels of IFN-ω throughout the study.
The potential reproductive toxicity of IFN-ω was assessed using four groups of female cynomolgus monkeys (tables 2 and 4). In each group, 12 pregnant animals received IFN-ω at doses of 0, 1, 10 and 50 μg/kg/day subcutaneously from day 16–80 of gestation. Table 2 lists the specific effects observed. No foetal defects or foetotoxicity were observed. The incidence of abortions was not statistically significantly different from the control group, but was increased in the intermediate and high dose groups, indicating a treatment-related abortofacient drug effect. All other findings were unremarkable with no other treatment related maternal or foetal effects observed. The administration of 1 μg/kg/day did not elicit statistically significant maternal toxicity, embryotoxicity or teratogenicity and is therefore considered to be the no toxic effect dose. IFN-ω was detected in serum from day 16 of gestation, after the first dose was given, in dose-dependent concentrations. The serum concentration increased in dose-dependent concentrations until day 28 of gestation; from day 44 of gestation onward the serum concentrations of IFN-ω were drastically reduced and were close to or at the lower limit of detection in the majority of samples. Anti-IFN-ω antibodies were detected as early as 10 days following treatment (day 26 of gestation) in 0, 33, 42 and 67% of the control, 1, 10 and 50 μg/kg/day treatment groups, respectively. Antibody concentrations and the number of animals displaying antibodies in serum increased with time. At the end of the study all samples from the IFN-ω treated animals showed moderate to high levels of anti-IFN-ω antibodies (mean=8,600 ng/ml). The decline of IFN-ω serum concentrations as determined using both the ELISA and EMCV bioassay correlated directly with the serum anti-IFN-ω antibody concentration (data not shown).
The pharmacokinetic profile of IFN-ω following a single subcutaneous administration was investigated in rabbits and cynomolgus monkeys. Some toxicokinetic data were obtained by examining the serum-time concentration of IFN-ω from samples taken from the subacute toxicity studies performed in rats and cynomolgus monkeys given a single daily s.c. injection of various doses of IFN-ω over 28 days. However, due to the appearance of anti-IFN-ω antibodies in the samples only the day 1 data were analyzed. The results of these studies are summarized in table 5. In each study IFN-ω displayed a characteristic biphasic elimination profile (data not shown). The terminal elimination half-life (t1/2) of IFN-ω ranged from ∼3 hr in rabbits to 3–6 hr in cynomolgus monkeys. Serum concentrations of IFN-ω increased proportional to the administered dose with no significant deviation from dose proportionality observed for Cmax and AUC other than in the 2,000 μg/kg IFN-ω administration to cynomolgus monkeys. In this study a lower than dose-proportional increase in Cmax and an increased t1/2 was observed relative to that which was observed at lower doses indicating that both the absorption and elimination processes may be dose-dependent.
Table 5. Average preclinical pharmakokinetic and toxicokinetic parameters for animals given a single subcutaneous dose of IFN-ω.
IFN-ω produced from CHO-SS cells was not associated with any biologically-relevant adverse effects in a series of 10 safety pharmacology experiments. This included an in vitro haemolysis test and a series of nine commonly-employed indices of general safety pharmacology conducted in mice, rabbits and rats. No mutagenic or genotoxic effects were observed in the Ames test or in the micronucleus test. No effects were observed in intraarterial, intravenous, paravenous or subcutaneous local tolerance studies conducted in rabbits or rats.
No effects were observed in acute toxicity experiments performed in cynomolgus monkeys when 2,000 μg/ml IFN-ω was administered by intravenous or subcutaneous routes. In subacute toxicity studies conducted in cynomolgus monkeys using 1, 10 or 100 μg/kg/day IFN-ω administered daily subcutaneous for 28 days we observed a slight decrease in haemoglobin and haematocrit levels in the intermediate and high dose male animals as well as a decrease in lymphocyte levels for the high dose male animals. Both the high dose male and female animals showed a decrease in platelet levels. The reduction in red blood cell indices, lymphocytes and platelets that was induced by IFN-ω is comparable to the same effects induced by IFN-α when it is used clinically (Dusheiko 1997) and this result was not unexpected. The results of this study allowed us to define a maximum no toxic effect concentration as 1 μg/kg/day. At the end of the study half the animals in the intermediate and high dose groups showed the presence of neutralizing anti-IFN-ω antibodies.
An identical subacute toxicity study was also undertaken using rats. Treatment with IFN-ω was associated with a reduced body weight gain for all animals in the high dose group (100 μg/kg/day) and for female animals in the intermediate dose group (10 μg/kg/day). The male animals in the high dose group also showed a reduced food intake. Flu-like symptoms, including poor appetite, are seen as a side effect of IFN-α treatment in the majority of treated patients (Dushieko 1997) and was therefore not unexpected. No other effects were noted. Based on this result we would again define the maximum no toxic effect concentration as 1 μg/kg/day. However, by day seven of the study no IFN-ω could be detected in most of the animals at 2 hr after drug administration and by the end of the study all the animals treated with IFN-ω contained high levels of neutralizing anti-IFN-ω antibodies.
A subchronic toxicity study was undertaken in cynomolgus monkeys using 1, 10 or 50 μg/kg/day IFN-ω administered daily subcutaneously for 13 weeks. In this study there was no evidence for any substance-related toxicological effects at any dose range. All the animals showed the presence of neutralizing antibodies directed against IFN-ω by week 3. The anti-IFN-ω antibody titers rose to high levels in all but one animal during the course of the study. In this experiment there was a clear inverse relationship between the serum concentration of IFN-ω and the presence of neutralizing anti-IFN-ω antibodies.
A reproductive toxicity study was performed in cynomolgus monkeys using daily subcutaneous administration of IFN-ω at 1, 10 and 50 μg/kg/day. There was an increase in the incidence of abortions observed at the intermediate and high dose groups. Based on this finding we conclude that the maximum no toxic effect dose would be 1 μg/kg/day. However, since there were possible treatment-related abortofacient effects observed, the use of IFN-ω during pregnancy would be contraindicated. No well-controlled studies of the effects of IFN-α treatment in pregnant women have been conducted, however high doses of either IFNα-2a or IFNα-2b in rhesus monkeys were previously associated with an increased incidence of abortions (see product inserts for PEGASYS, Hoffman-La Roche, Inc. and PEG-Intron, Schering Corporation). Thus, our findings were not unexpected. Also, by day 44 of our study the serum concentrations of IFN-ω were drastically reduced in most animals and at the end of the study all animals showed the presence of moderate to high levels of neutralizing anti-IFN-ω antibodies in their serum.
Antibody formation due to the immunogenicity of recombinant human proteins in animal studies (Toon 1996) and against human recombinant IFNs used for a variety of indications (Antonelli & Dianzani 1999; Bertolotto et al. 2004), including the use of recombinant human IFN-α during the treatment of hepatitis C virus infection (Milella et al. 1993 & 1995; Giannelli et al. 1994; Antonelli 1995 & 1996; Haria & Benfield 1995; Bonino et al. 1997; Hoffman et al. 1999) is well known, so our observations were not unexpected. Currently both the extent of and the implications of such antibody formation during IFN-α treatment in patients are not well understood. Since a variety of methodologies have been employed in the determination of the presence of antibodies against human IFN-α formed during hepatitis C virus therapy, the reported incidence of these antibodies has varied enormously from 7–61% (Schellekens et al. 1997). Also it is as yet unclear to what extent the presence of neutralizing antibodies against IFN-α during hepatitis C virus treatment correlates with breakthrough viraemia during treatment and/or treatment failure (Milella et al. 1993 & 1995; Giannelli et al. 1994; Antonelli 1995 & 1996; Haria & Benfield 1995; Bonino et al. 1997; Hoffman et al. 1999). We have little insight into understanding why the combination treatment of IFN-α and ribavirin is not effective in 45% of patients (Pearlman 2004). It is possible that the formation of anti-IFN-α antibodies during therapy might explain some of these cases of treatment failure and that other recombinant human IFNs such as human IFN-ω produced from CHO-SS cells might be a useful second therapy to employ when IFN-α treatment has failed. This supposition is strengthened by our observation that half of 24 patients who previously did not respond to at least 12 weeks of therapy using IFN-α or pegylated IFN-α with or without ribavirin showed statistically-significant reductions in HCV RNA levels following monotherapy with IFN-ω in a phase 1b clinical trial (McHutchison et al. 2001).
The aggregate analysis of the five animal toxicology studies performed indicates that the no toxic effect toxic dose of IFN-ω is 1 μg/kg/day. However, due to the confounding effects of the neutralizing anti-IFN-ω antibodies that were produced in the animals over time, this finding should be viewed cautiously. Pharmacokinetic studies indicated that IFN-ω is rapidly absorbed after subcutaneous administration, with dose-proportional blood levels observed and a half-life similar to that observed for IFN-α (5 hr; Osborn et al. 2002).
In summary, human recombinant IFN-ω produced from CHO-SS cells was not associated with any biologically-relevant adverse effects in a series of safety pharmacology experiments or in local tolerance studies. Acute, subacute, subchronic and reproductive toxicity studies performed in cynomolgus monkeys and rats showed a toxicity profile similar to that of human IFN-α with use during pregnancy being a contraindication. These toxicity studies indicated a maximum no toxic dose of 1 μg/kg/day, however this dose should be viewed cautiously since neutralizing anti-IFN-ω antibodies were produced during these studies. The average pharmacokinetic parameters for IFN-ω determined following a single dose of IFN- ω in rabbits, rats and monkeys were similar to those of IFN-α. IFN-ω has a safety profile consistent with that required for use in man. IFN-ω, alone or in combination with other antiviral therapies, might be beneficial for the treatment of hepatitis C virus-infected patients who fail to respond to IFN-α or as an additional first-line treatment option.
This work was supported by Intarcia Therapeutics, Inc. We gratefully acknowledge the scientific contributions of Drs. M. Baumeister, I. Dean, A. Geischel, G. Heinzel, D. Innes, L. Jacob, B. Kruss, H. R. Lamche, A.B.M. Mauz, R.O. Oshodi, I. Osterburg, M. Pairet, J. Stangier, K. Thomae, D.C. Thompson and E. Weller.