Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The advancement of antisense oligonucleotides to human clinical trials as anticancer (Bayever et al., 1993; Stevenson et al., 1999; Chen et al., 2000), anti-inflammatory (Glover et al., 1997), or antiviral (Perry and Balfour, 1999; Séréni et al., 1999) therapeutics has been greatly enhanced by progress in the synthesis of nuclease-resistant oligonucleotides and an understanding of their pharmacological properties. Most antisense oligonucleotides in clinical trials are phosphorothioate (PS)¹ oligonucleotides, DNA molecules in which single nonbridging oxygen has been replaced by sulfur. This modification results in enhanced resistance to endonucleases, the enzymes primarily responsible for the catabolism of oligonucleotides. Studies with mixed backbone oligonucleotides (MBOs) where there is a substitution of the 2'-position of ribose with alkyl substituents appear to hold promise of increased safety and more desirable pharmacokinetic properties, including a greater resistance to endonucleases than PS oligonucleotides (Agrawal et al., 1997a; Chen et al., 2000).

Human cytomegalovirus (HCMV), a common opportunistic infection in immunocompromised persons, is a major cause of life-threatening disease. CMV-induced retinitis, if left untreated, will cause severe, irreversible damage to the retina, resulting in progressive loss of vision in the involved eye(s). Currently available therapies for CMV-induced retinitis include ganciclovir, foscamet, cidofovir, and the recently approved PS antisense oligonucleotide, fomivirsen (ISIS 2922).

GEM132 is a 20-mer mixed backbone oligonucleotide (MBO) with 2'-O-methylribonucleosides at the two 3'- and four 5'-terminal nucleotides. It is antisense, complementary to the intron-exon boundary of the UL35 and UL37 premRNA transcripts of HCMV, and has shown antiviral activity in vitro with an IC₉₀ of about 0.1 µM against standard laboratory strains of the virus (Pari et al., 1995). Viruses resistant to current therapies do not show cross-resistance to GEM132, which has shown activity against a battery of over 20 clinical HCMV isolates with an IC₉₀ of about 0.1 µM (Field et al., 1997). GEM132 has entered clinical trials as an antiviral for patients with HCMV-induced retinitis who can no longer benefit from or tolerate available therapies.

Although studies have been published on the pharmacokinetics of PS oligonucleotides after intravitreal dosing (Leeds et al., 1997, 1998), none have described the disposition and toxicity of an MBO after intravitreal administration. Because GEM132 had prolonged persistence in tissues after i.v. administration to rats (Hybridon internal report), this investigation was designed to characterize the acute and chronic toxicity and disposition of GEM132 in Dutch Belted rabbits for 6 months after a single intravitreal dose.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Synthesis and Purification of GEM132. GEM132 is a 20-base phosphorothioate MBO (5'-UGGGGCTTACCTTGCGAACA-3'), 6605-Da molecular mass (free acid), in which two deoxynucleosides at the 5' end and four deoxynucleosides at the 3' end (underlined) are substituted with 2'-O-methylribonucleosides. GEM132 was chemically synthesized using deoxynucleoside phosphoramidites (Milligen, Milford, MA) and 2'-O-methylribonucleoside phosphoramidites (Glen Research, Sterling, VA) on an automated synthesizer as previously described (Padmapriya et al., 1994; Zhang et al., 1995) and purified by preparative reversed phase HPLC. The purity was 91% as determined by capillary gel electrophoresis. Radiolabeled GEM132 (specific activity, 15 µCi/mg) was synthesized with [4,6-¹⁴C]thymidine phosphoramidite. The radiolabel was located in the first thymidine from the 5' end. Purity was 90% by capillary gel electrophoresis.

Experimental Design. Male Dutch Belted rabbits were given a single intravitreal injection (50 µl) of GEM132 in both eyes. Control animals received saline only. For the disposition study, [¹⁴C]GEM132 was administered; eyes and selected tissues were harvested at various times postdosing (n = 3 eyes/time point). The oligonucleotide was dissolved in 0.9% saline, and the doses administered were 3.7 µg/eye (0.07 µCi), 15.7 µg/eye (0.3 µCi), and 78.5 µg/eye (1.5 µCi). For the toxicity study, unlabeled GEM132 was administered at doses of 5, 20, or 100 µg/eye.

Electroretinograms (ERGs). Electroretinograms were obtained from all animals in the study at regular intervals. Light stimuli included scotopic blue, red, and white single flash and 30 Hz white flicker. The animals were anesthetized with ketamine (50 mg/kg)/xylazine (5 mg/kg) and dark-adapted at least 30 min before recording of the ERG. A mydriatic agent was applied to the eyes to dilate the pupils before testing.

Sample Processing and Storage Conditions. At each time point, the left eye from the even-numbered animals and both eyes from the odd-numbered animals were dissected, and retina, vitreous humor, and remaining ocular tissue were removed and weighed. The weighed vitreous humor (0.724 ± 0.031 g, mean ± S.D.) and retina (0.0283 ± 0.0003 g) samples were homogenized in 2 ml of 0.9% saline in preweighed glass tubes with a Teflon probe. After homogenization, aliquots of retina were solubilized in 35% tetraethylammonium hydroxide. Ten milliliters of Hionic Fluor scintillation fluid (Canberra Packard, Ontario, Canada) was added to vitreous humor and solubilized retina, and radioactivity was measured by liquid scintillation spectroscopy. Remaining ocular tissue samples (1.47 ± 0.18 g) were minced and digested in toto in 35% tetraethylammonium hydroxide. Duplicate aliquots of the digest were added to 10 ml of scintillation fluid, and radioactivity was determined. Samples having a radioactivity level of less than or equal to double background were considered below the limit of quantitation and considered zero.

Autoradiography and Microautoradiography Conditions. At each time point, the right eye of each even-numbered animal was removed and fixed in a 2.5% gluteraldehyde solution, stored, refrigerated, and transferred within 24 h to 10% neutral buffered Formalin where they remained for at least 24 h before trimming. Eyes were then trimmed, dehydrated, infiltrated, and imbedded in paraffin within 8 days after necropsy. Six-micrometer-thick sections were prepared. The following procedures were conducted under reduced safety light conditions: sections were subjected to deparaffination, dipped in diluted and warmed Kodak NTB-2 nuclear track emulsion, followed by a 30-min rinse in hot distilled water and controlled drying. After drying, the emulsion-coated slides were stored in lightproof plastic boxes at 4-8°C. All slides were developed after 15 days of exposure using a solution of Kodak D-19 developer, washed, fixed in Kodak Rapid Fixer, and washed. Slides were stained with H&E, mounted, and evaluated for the localization and relative concentration of radioactivity (visualized as small, reduced silver grains lying on cellular surfaces). A four-grade (1, slight; 2, mild; 3, moderate; and 4, severe) grading scheme was used to semiquantify the amount of radioactivity. The amount of radioactivity present in underlying sclera was used to establish a threshold between a negative and a slight deposition of radioactivity in the retina.

Polyacrylamide Gel Electrophoresis (PAGE) and Phosphor Image Analysis. Retinal homogenates were subjected to protein digestion by incubation with proteinase K enzyme (0.25 ml of a 20 mg/ml solution) containing 20 mM Tris EDTA for 3 h at 60°C. Samples were extracted twice with Tris ETDA-saturated phenol-chloroform solution (1:1, v/v) and once with chloroform to remove proteinase K, digested protein, and lipids from nucleic acids. After extraction, an aliquot of the organic and aqueous phase from each sample was removed, and total radioactivity was determined by liquid scintillation spectroscopy. This allowed for the determination of extraction recovery. The average recoveries (±S.D.) were 79.7 ± 13.2% (quality control standards), 74.3 ± 19.1% (day 7), 75.7 ± 6.4% (day 14), 72.4 ± 10.3% (day 28), 74.1 ± 8.2% (day 56), 87.3 ± 3.6% (day 114), and 87.4 ± 12.0% (day 149). No radioactivity (2× background) was measurable in the organic phase, indicating that all the GEM132-associated radioactivities were retained in the aqueous phase following extraction. Loss of radioactivity was believed to be due to the difficulty in completely separating the two phases; a thin layer of the aqueous phase was always left behind. An 8-µl aliquot of the aqueous phase was loaded on a 20% polyacrylamide gel containing 7.5 µM urea and subjected to electrophoresis (70-290 V, 3-14 mA, 60-153 min). Quality control standards (0.11, 0.60, and 4.1 mg/ml, corresponding to 780, 4,500, and 21,000 dpm) were prepared by adding known amounts of radiolabeled GEM132 to retinal homogenates obtained from control rabbits. Duplicate standards were loaded on each of the two outer channels of the gel. After electrophoresis, Southern blots of the gels were performed on 0.45-µ nitrocellulose membranes over a period of about 12 h. The nitrocellulose membranes were then removed and allowed to dry at room temperature. Membranes were exposed to the phosphor screen for 25 to 288 h (Molecular Dynamics Storage Phosphor Screen, Sunnyvale, CA) at room temperature, after which the phosphor screen was scanned by a phosphor imager (Molecular Dynamics Phosphor Imager SI) to obtain radioactivity profiles. The lower limit of detection was determined to be an area of 15 phosphor image units. Concentrations of intact GEM132 were calculated by multiplying the measured total radioactivity in the retina (total MBO) by the fraction eluting with the same mobility as the standards, as determined by PAGE/phosphor image analysis and correcting for purity (90%).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Ophthalmic Examinations. Signs of ocular irritation attributed to the injection of GEM132 were infrequent, mild, and transient. Mild cellular infiltrate in the anterior vitreous occurred almost exclusively in the high-dose animals from 2 to 16 weeks postdose. Signs of iritis were also mild and infrequent. Retinal lesions were observed almost exclusively in the high-dose animals. Changes in intraocular pressures were mild and transient and considered only potentially related to treatment.

Electroretinograms. Representative data for ERGs are shown in Table 1. The scotopic white flash measures the response of both rods and cones and is representative of overall retinal function. The mean values and standard deviations for B-wave amplitude and response latency both demonstrate significant dose-related decreases. No changes were observed in the low-dose animals for any ERG parameter, when compared with controls. Statistically significant changes that were observed in the mid-dose animals included a decrease in amplitude in the scotopic blue and red parameters during weeks 4 to 20, decreased latency in the scotopic red parameter during week 10, and decreased amplitude in the scotopic white parameter, weeks 10 to 20. Statistically significant changes occurred in both amplitude and latency for all parameters (scotopic blue, red, white and white flicker) at all time points measured in the high-dose animals. Due to the severely decreased ERG potential observed at 16 weeks in the high-dose group, we chose to terminate this group at this time. The decreased ERG potential that persisted for up to 20 weeks in the mid-dose group was the reason we chose to terminate the mid- and low-dose groups at week 21. 

Histopathology. Histopathological findings attributed to the administration of GEM132 were seen in the retina, lens, and optic nerve of a few animals in the mid-dose group at 8 and 21 weeks and the high-dose group at 4, 8, and 16 weeks postdose. Slight retinal degeneration characterized by partial disorganization of the inner and outer plexiform layer and the ganglion cell layer was observed in one high-dose animal at week 4 and in three high-dose animals at week 8. This finding was also observed in one mid-dose animal at weeks 8 and 21 and in one control and two low-dose animals at week 21. Retinal detachment was observed in one high-dose animal at week 8. A basophilic granular material was observed in the retina of one high-dose animal at week 4, in three high-dose animals at week 8, and in two high-dose animals at week 16. One mid-dose animal at week 8 also presented with this finding. A slight to mild mononuclear infiltration was seen in the optic nerves of some high-dose animals at weeks 4 and 6 and in one high-dose animal at week 16. This was also observed in some mid-dose animals at week 8.

Systemic Exposure following Intravitreal Administration. Plasma concentrations of total radioactivity were lower by 3 to 4 orders of magnitude compared with that in ocular tissues. GEM132-derived radioactivity in tissues such as kidneys, liver, and spleen was also low compared with that in ocular tissues. The highest concentration of radioactivity in nonocular tissues was seen in the liver 14 days postdose and equaled 2.3% of the intraocular dose.

Disposition Kinetics of GEM132-Derived Radioactivity in Vitreous Humor. AUC values in vitreous humor increased somewhat less than proportional to dose (Table 2). GEM132-derived radioactivity was cleared rapidly from the vitreous humor (Fig. 1). At 7 days postdose, 93% of the dose had cleared the vitreous humor, which increased to 99% of the dose by 28 days postdose. Radioactivity distributed rapidly throughout the tissues of the eye (Table 3).

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Mean concentration of GEM132 in retina and total radioactivity in vitreous after a single intravitreal injection of GEM132 to rabbits.

Disposition Kinetics and Metabolism of GEM132-Derived Radioactivity in Retina. The AUC values in retina increased approximately in proportion to the dose (Table 2). The profiles of radioactivity by PAGE/phosphor image analysis of retina homogenates are presented in Table 4. Figure 2 shows representative electropherograms from different gels. Differences in the mobility of the calibration standard from gel to gel were corrected by running calibration standards with each gel. There was no apparent difference in the recoveries at each sampling time between dose groups or, overall, between samples of retina homogenates spiked with radiolabeled GEM132. At day 7 postdose, all extracted radioactivities migrated with the same mobility as GEM132. From day 14 to day 149 postdose, the mean radioactivity in extracted samples with a mobility similar to GEM132 declined to 45.1% on day 149. In the low- and mid-dose groups, radioactivity associated with shorter chain length oligonucleotides was first detectable on day 28 (9.3-12.5%) and increased to about 45% (range 43.3-48.9%) by day 149 (Table 4). There were no apparent differences in the radioactivity compositions related to the different doses administered.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Representative PAGE/phosphor image scans of radioactivity in rabbit retina after a single intravitreal injection of 14C GEM132 to rabbits. Top, calibration standard; middle, high-dose, day 29; bottom, high-dose, day 114. Region A, radioactivity with the same mobility as GEM132 and n-1 + n-2 impurities/metabolites. Region B, radioactivity of impurities/metabolites of chain length n  3. Region C, radioactivity of molecular species with higher molecular mass than GEM132.

View larger version (75K): [in this window] [in a new window]   Fig. 3.   Microautoradiography of the ventral retina from rabbits 56 days after a single intravitreal injection of [l4C]GEM132. H&E. Bar, 30 µm. A, low-dose animal. Slight radioactivity in most retinal layers; one radioactive mononuclear cell in the vitreous humor (arrowhead). B, mid-dose animal. Radioactivity is more intense in outer nuclear layer (ONL), outer plexiform layer (OPL), and inner nuclear layer (INL) than in other retinal layers. Superficial retinal atrophy is noted (between arrowheads). C, high-dose animal. Superficial retinal atrophy is marked with disruption of inner nuclear layer (arrows). Note the strong and patchy distribution of the radioactivity in the outer nuclear layer (ONL) and more inner layers of the retina.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This report is the first to describe the combined disposition profile and toxicity for a mixed backbone oligonucleotide after intravitreal administration. Although the disposition and toxicity arms of these studies were, in fact, run in different groups of animals and with slightly different dose levels, there were numerous remarkable consistencies in the patterns of responses in the eyes and the patterns of radiolabeled drug distribution.

After intravitreal administration, GEM132 distributed rapidly out of the vitreous into the retina and remaining ocular tissue. Although concentrations in the retina were greater than those in the remaining ocular tissue, the large difference in weight between these two tissue masses (remaining ocular tissues/retina ~50:1) most likely explains why the fraction of the dose present in the remaining ocular tissues is greater than that in the retina. The reasons for the apparent increased uptake (on a per gram of tissue basis) of GEM132 by retinal tissues compared with the remaining ocular tissue may reflect unequal distribution of GEM132-associated radioactivity within the remaining ocular tissue. Autoradiography of frozen sagittal sections of rabbit eyes indicates that 5 days postdose, radioactivity in the remaining ocular tissue was associated primarily with the iris/ciliary body (data not shown).

The clearance of GEM132 from the vitreous humor followed apparent first-order kinetics. The estimated half-life for GEM132 in the vitreous humor was about 48 h. This is similar to that reported for other compounds of similar molecular weight (Sirossian et al., 1995; Leeds et al., 1997). Given the enhanced stability of MBOs (Agrawal et al., 1997a; Chen et al., 2000), the data reported here suggest that the clearance of GEM132 from the vitreous humor is primarily due to distribution to ocular tissues and via outflowing aqueous humor. The radioactivity in the vitreous humor at time points beyond 14 days postdose were below the sensitivity of the methodology, and therefore data are not available regarding the profile of the radioactivity in this ocular compartment.

After administration of the lowest dose (3.7 µg), retinal concentrations of GEM132 at 7 days postdose reached a value equal to about 10 times the IC₉₀ (1.2 µM, 8 µg/g, Fig. 1). Retinal concentrations of intact GEM132 reached a maximal value 56 days postdose (4.1 µM, 26.9 µg/g) and persisted at levels ~10 times the IC₉₀ for months (Fig. 1). The observation that retinal levels of metabolites increased with time, approaching 50% of the total radioactivity in the retina, suggests that clearance of GEM132 metabolites from the retina was slower than their rate of formation.

The single, bilateral administration of GEM132 at a dose level of 5 µg/eye did not result in any adverse test article related effects. Although clinical ophthalmological evidence of retinal toxicity was seen in a few animals receiving 20 µg/eye, decreased potential by ERG suggests a higher incidence of functional retinal changes persisting for up to 20 weeks postdose. At a dose of 100 µg/eye, there was clear evidence of retinal toxicity in the form of severely decreased ERG electrical potential and retinal degeneration observed during the ophthalmology exams and microscopic pathology present for up to 16 weeks postdose. Due to the severely decreased ERG potential observed at 16 weeks in the high-dose group (Table 1), we chose to terminate this group at this time. The decreased ERG potential that persisted for up to 20 weeks in the mid-dose group (Table 1) was the reason we chose to terminate the mid- and low-dose groups at week 21.

Figure 3 shows the distribution profile in the retina for [¹⁴C]GEM132 in each dose group evaluated 8 weeks after dosing. In the low-dose group, there was only slight radioactivity in most retinal layers (GEM132 retinal concentration, 24 µg/g). Correspondingly, there was no significant pathology or functional toxicity evident in the eyes of animals from that low-dose group. In the mid-dose group, radioactivity was more intense in both the inner and outer nuclear layers and the outer plexiform layer (GEM132 retinal concentration, 85 µg/g). In the toxicology study, slight to mild retinal degeneration, characterized by a partial disorganization at the outer and inner plexiform layer and the ganglion cell layer, was evident in both the mid- and high-dose animals by 4 weeks postdose (high-dose; GEM132 retinal concentration, 350 µg/g) or 8 weeks postdose (mid-dose). In addition, by 16 weeks postdose, both eyes in high-dose animals showed slight to mild retinal degeneration at the inner nuclear and plexiform layers, ganglion cell layer, and the optic nerve fiber layer. This is also consistent with the radiolabel distribution seen in Fig. 3, where superficial retinal atrophy was evident in the high-dose, along with more intense and patchy distribution of radioactivity in the outer nuclear layer and the more inner layers of the retina.

ERGs were performed at regular intervals during the course of the toxicology study. Effects on B-wave amplitude and increased white flicker latency were indicative of a loss of function of the rods and cones, and, while the effect tended to remain stable throughout the observation period, there was no evidence of recovery to levels comparable with control values. This functional observation is consistent with the microscopic evaluations that revealed a basophilic granular material in the retina of both the mid- and high-dose animals. This material was considered to be potentially an accumulation of the test article. The accumulation was also not reversed, as expected from the ERG effects and the slow apparent clearance of radiolabeled compound from the retina.

A comparison of the ocular disposition of the MBO GEM132 after administration of 3.7 µg/eye and the PS oligonucleotide ISIS 2922 at 66 µg/eye (Leeds et al., 1997) indicate that both are cleared rapidly from the vitreous with similar half-lives. Retinal concentrations of GEM132 and ISIS 2922 at C_max were about the same (3.5 and 4.1 µM, respectively), although the time to achieve C_max was later for GEM132 than for ISIS 2922. Despite an 18-fold difference in dose, retinal concentrations of intact GEM132 were greater than that of intact ISIS 2922 at the comparable time points (days 7, 14, 20, and 28). The clearance of GEM132 from the retina was significantly slower than that of ISIS 2922, most likely due, in part, to the enhanced stability of the MBO compared with the PS oligonucleotide. Other factors, such as the presence of four contiguous guanosines (4G motif) may also impact the tissue clearance of the compound (Agrawal et al., 1997b). The combination of the higher oligonucleotide concentration in the retina and slower retinal elimination of GEM132 suggest that less frequent dosing with this MBO compared with the PS oligonucleotide might be required to achieve and maintain therapeutic concentrations above those associated with in vitro antiviral activity.