0022-3565/05/3123-1027–1033$20.00 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 312, No. 3 Copyright © 2005 by The American Society for Pharmacology and Experimental Therapeutics 76950/1191417 JPET 312:1027–1033, 2005 Printed in U.S.A.

Ciglitazone-Induced Lenticular Opacities in Rats: In Vivo and Whole Lens Explant Culture Evaluation

Michael D. Aleo, Colleen M. Doshna, and Kimberly A. Navetta Pfizer Global Research and Development, Worldwide Safety Sciences–Groton, Groton, Connecticut Received September 13, 2004; accepted October 27, 2004

ABSTRACT Downloaded from The cataractogenic potential of the thiazolidinedione ciglita- 28, and 42 Ϯ 8%, respectively. Lens wet weight increased 17 Ϯ zone (CIG) was investigated in vivo and in vitro. In the rat, CIG 4% with a concomitant decrement in lens clarity. Pretreating caused a dose-dependent (30–300 mg/kg/day) increase in in- lenses with the mitochondrial calcium uniport inhibitor ruthenium cidence and severity of nuclear cataract formation during a red (RR) partially or fully protected lenses from toxicity. In contrast, 3-month nonclinical safety assessment study. Potential mech- the antioxidant dithiothreitol, aldose reductase inhibitor sorbinil, anisms of toxicity were surveyed using whole rat lens explants and selective cell-permeable calpain inhibitors [calpain II inhibitor

exposed to CIG with or without various inhibitors of cataract and (2S,3S)-trans-epoxysuccinyl-L-leucylamido-3-methylbutane jpet.aspetjournals.org formation. In vitro, CIG caused a concentration- (0.375–30 ␮M) ethyl ester (E64d)] were ineffective in providing protection under and time-dependent (3–24 h) change in biochemical [ATP con- the present testing conditions. Early and selective changes in tent or mitochondrial reduction of the tetrazolium dye 3-(4,5- lenticular ATP content and the partial or full protective effect of RR dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT) and reduced suggest that alterations in lens bioenergetics may play an impor- glutathione (GSH) content] and morphometric (lens wet weight tant role in CIG-induced cataract formation. Lens explant cultures and clarity) markers of damage. Within3hofexposure, 7.5 ␮M were successfully used to select two thiazolidinediones that CIG decreased lens ATP content 37 Ϯ 7% (percentage of differ- lacked cataractogenic activity when evaluated in 3-month rat ence from control, p Ͻ 0.05). After 24 h of exposure, lens ATP safety assessment studies. at ASPET Journals on January 18, 2020 content, MTT reduction, and GSH content declined 57 Ϯ 5, 30 Ϯ

Thiazolidinediones represent a major new therapy in the relatively unknown safety profile of ciglitazone (CIG), an treatment of noninsulin-dependent diabetes. However, there early drug candidate in this chemical class. Upon oral ad- is new interest in exploring the use of thiazolidinediones in ministration, CIG was associated with the formation of len- the treatment of ophthalmic diseases (Murata et al., 2000; ticular opacities during nonclinical safety assessment studies Aoun et al., 2003). Although the safety profile of these agents conducted in rats by the Upjohn Company (Kalamazoo, MI). is well established based on oral exposure, such new indica- This adverse finding limited further development of CIG. tions and routes of exposure could potentially alter the safety Years later, based on anecdotal information, we indepen- profile of marketed drugs and would influence the selection of dently used CIG as a positive control to survey basic mech- new drug candidates. This may occur since intravitreal or anisms of cataract formation (Aleo et al., 2000). Although topical ocular administration may lead to higher intraocular these independently produced investigations were not re- drug concentrations and/or the potential for a longer half-life ported in the peer-reviewed literature at that time, we be- of the compound in ocular tissue compared with oral expo- lieve it is appropriate to document and disseminate the ocu- sure. Of even more relevance to the exploration and potential lar safety profile of CIG based on these earlier findings use of thiazolidinediones for ophthalmic indications is the because of the current interest in developing thiazolidinedio- nes for treating ophthalmic diseases. Lenticular opacities are caused by compounds that either In vitro portions of this work were reported in abstract form at the 36th penetrate the lens and directly disrupt normal cellular pro- Annual Meeting of The Society of Toxicology; Mar 9–11, 1997; Cincinnati, OH. Portions of this work have been reprinted from Aleo et al. (2000) with permis- cesses important in maintaining lens transparency or inter- sion from the New York Academy of Sciences. act indirectly with the lens surface through the generation of Article, publication date, and citation information can be found at http://jpet.aspetjournals.org. reactive oxygen species (Bhuyan et al., 1973; Bhuyan and doi:10.1124/jpet.104.076950. Bhuyan, 1979). CIG seems to interact directly with the lens

ABBREVIATIONS: CIG, ciglitazone (sodium salt); DTT, dithiothreitol; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium; RR, ruthenium red; E64d, (2S,3S)-trans-epoxysuccinyl-L-leucylamido-3-methylbutane ethyl ester; CI-II, calpain inhibitor II; SD, Sprague-Dawley; ANOVA, analysis of variance; LSD, least significant difference; GSH, reduced glutathione; SORB, sorbinil. 1027 1028 Aleo et al. since it has been shown to distribute to the rat eye after oral water ad libitum. Drug administration occurred once a day (15 administration (Torii et al., 1984). Cataractogenic com- animals/dose/gender) via the oral route using a suspension of cigli- pounds that penetrate the lens have an intrinsic ability to tazone (0.1% sorbic acid, 0.25% methylcellulose, 0.5% polysorbate 80, adversely perturb the lenticular environment, resulting in 2.0% Avicel RC-591, and purified water). opacity formation. In some cases, such as acetaminophen, the Three-Month Safety Assessment Studies of CIG in Canines. compound that penetrates the lens and causes the damage is Male and female purebred beagles were obtained from Laboratory a reactive metabolite that is formed via the cytochrome P450 Research Enterprises (Kalamazoo, MI) and White Eagle Laborato- system found in the ciliary body of the eye (Zhao and Shichi, ries, Inc. (Doylestown, PA), respectively. Animals ranged in age between 10 and 12 months for males and 18 and 20 months for 1995) or liver (Lubek et al., 1988). In other cases, such as females. Animals were housed individually and fed Purina certified S-(1,2-dichlorovinyl)-L-cysteine (Walsh Clang and Aleo, canine diet (#5007). Drug administration occurred once a day (4 1997), naphthalene (Lou et al., 1996), or glucose (Yeh et al., animals/dose/gender) via the oral route using gelatin capsules con- 1986, 1987; Yeh and Ashton, 1990), the compound seems to taining ciglitazone. require metabolic activation within the lens itself. Lens Explant Studies. Male SD rats (120–200 g) from Charles Lenticular opacities are caused by at least three broad and River (Raleigh, NC) were housed individually in wire rack cages with sometimes overlapping mechanisms: 1) oxidative stress free access to standard laboratory chow and water in an environ- ϩ (Spector, 2000) and/or Ca2 -dependent activation of calpains mentally controlled room on a 12-h light/dark cycle. The rats were

(Azuma et al., 2003); 2) osmotic stress (Chung et al., 2003) euthanized by CO2 asphyxiation. After enucleation, lenses were ex- caused by the accumulation of active osmolytes such as glu- tracted from the globes using a posterior approach and placed in a Downloaded from cose metabolites (polyols) within the lens; or 3) disregulation 20-ml glass scintillation vial with a 2-mm hole in the cap. The vial of normal bioenergetic processes (Walsh Clang and Aleo, contained 4 ml of warmed bicarbonate-based Media 199 (without 1997; Martynkina et al., 2002; Belusko et al., 2003) or thiol phenol red) supplemented with L-glutamine (2 mM), penicillin status (Lou, 2003). (50,000 U/l), and streptomycin (50,000 mg/l). Final osmolality of the Here we report the cataractogenic effects of CIG in rats. base medium was adjusted to 295 to 300 mOsM/kg. The medium was bubbled for 10 min with 95:5% air, CO2, filter sterilized, and stored

Using rat lens explants in organ culture as a short-term, jpet.aspetjournals.org at 4°C. The lenses were incubated on an orbital shaker in a humid- mechanistically based ex vivo model, we examined the rela- ified 37°C incubator in a 95:5% air, CO2 environment. Lenses were tive contributions of either perturbation of mitochondrial harvested and cultured as described previously (Walsh Clang and function, osmotic stress, oxidative stress, and/or calpain ac- Aleo, 1997) and after 24 h were inspected visually for damage ac- tivation in the mechanism of CIG toxicity to the rat lens. The cording to Shearer et al. (2000). Only lenses without procedure- effects of CIG on rat lens explants in culture were assessed by induced damage (i.e., partial opacification of the lens or rupture of analysis of selective biochemical markers (lenticular ATP the capsule) were retained for further experimentation. Lenses were

content and/or mitochondrial MTT reduction and GSH con- incubated and appearance assessed at least daily in treated medium at ASPET Journals on January 18, 2020 tent) and morphometric markers (lens weight and daily ob- containing a combination of vehicle, CIG, and/or other test com- servations of changes in lens clarity) of damage to the crys- pound(s) for 3, 6, 24, 48, or 96 h. Ethanol was used as a vehicle at a talline lens. The results of experiments using whole lens final concentration of 0.2%. Lens appearance was assessed and explants in culture were used to propose a mechanism for graded at least daily by visual inspection. Lenses were graded in the cataract formation by CIG and to screen other drug candi- following manner: stage 0, clear; stage 1, hazy cortical region; and dates for cataractogenic potential in vitro. stage 2, haziness characterized by a visible demarcation between nuclear and cortical regions. Grading system is based on a modifi- cation from Dickerson et al. (1997). For multiday exposure experi- Materials and Methods ments, treated medium was exchanged every 24 h for 4 days as indicated. At the end of the incubation period, lenses were rinsed CIG was the generous gift of Dr. David Clark (Pfizer Global with ice-cold saline, blotted, and weighed before processing for bio- Research and Development, Groton, CT). Dithiothreitol (DTT), chemical analysis. Lenticular ATP content, mitochondrial MTT re- 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT), mono- duction, and GSH content were determined as described previously chloroacetic acid, sodium octyl sulfate, and ethylenediamine-tet- (Walsh Clang and Aleo, 1997). raacetic acid were purchased from Sigma-Aldrich (St. Louis, MO). Three-Month Safety Assessment Studies of Englitazone in Ruthenium red (RR) was acquired from ICN (Cleveland, OH). Media Rats. Male and female Long-Evans [Crl(LE)BR] rats were obtained 199, glutamine, penicillin, and streptomycin were obtained from from Charles River Laboratories, Inc. (Wilmington, MA). Animals Invitrogen (Carlsbad, CA). E64d and calpain inhibitor-II (CI-II) were (120–200 g) were housed individually and fed Agway Prolab RMH obtained from Calbiochem (San Diego, CA). All other reagents were 3200 diet (Agway Country Foods Inc., Syracuse, NY) and water ad at least analytical or high-pressure liquid chromatography grade. libitum. Drug administration occurred once a day (15 animals/dose/ Animal care and use were conducted in accordance with all appli- cable state and federal regulations and guidelines and were in full gender) via the oral route (15, 45, and 100 mg of free base/kg/day) compliance with national legislation [Animal Welfare Act Regula- using a suspension of englitazone in its sodium salt form (0.1% tion, Title 9 Code of Federal Regulations parts 1 to 3 and the Asso- methylcellulose). ciation for Assessment and Accreditation of Laboratory Animal Care, Three-Month Safety Assessment Studies of Darglitazone in International Standards as set forth by the Guide for the Care and Rats. Male and female Long-Evans [Crl(LE)BR] rats were obtained, Use of Laboratory Animals (1996 National Academy Press, Wash- housed, fed, and dosed as described previously (Aleo et al., 2003). ington, DC)] and approved by the Animal Care and Use Committee Statistical Analysis. Comparisons were made using a one-way of this research site. analysis of variance (ANOVA). When necessary, the data were trans- Three-Month Safety Assessment Studies of CIG in Rodents. formed to natural logarithms to meet better the assumptions of the Male and female Sprague-Dawley (SD) Rats [Upj:TUC(SD)spf] were ANOVA. Multiple means were compared using Fisher’s protected obtained from the Upjohn Company. Animals ranged in age between least significant difference (LSD) test. Statements of significance 6 and 7 weeks for males and 7 and 9 weeks for females. Animals were were based on a p Ͻ 0.05. All results are expressed as mean Ϯ S.E.M. housed individually and fed Purina certified rat chow (#5002) and unless noted otherwise. Cataractogenic Effects of Ciglitazone in Vivo and in Vitro 1029 Results In Vivo Safety Assessment of CIG in Rodents and Canines. CIG caused a dose-related increase in the inci- dence of opacity formation during the conduct of a 3-month nonclinical safety assessment study in SD rats (15 animals/ sex/group) at daily oral doses of 30, 100, and 300 mg/kg/day (Table 1). Opacities were first noted in some animals on day 34 of study. An ophthalmic exam conducted on day 49 re- vealed four animals (male and female) in the 100 and 14 animals in the 300 mg/kg/day group with varying degrees of opacity formation (bilateral changes ranging from stress lines to stress lines with early cataracts). After 86 days of daily exposure, 4, 9, and 16 animals receiving 30, 100, and 300 mg/kg/day were diagnosed with varying degrees of opac- ity formation (bilateral changes ranging from posterior cor- tical capsular cataracts with or without nuclear cataracts to mature complete cataracts). There was no evidence of opacity Downloaded from formation identified in a 3-month canine study at daily doses Fig. 1. Concentration-dependent changes in the transparency of ex- planted Sprague-Dawley rat lenses treated with CIG for 24 h. Gradation up to 300 mg/kg/day (data not shown). Drug exposure data of shading represents percentage of lenses within each treatment group were not obtained during the in-life portion of either the rat at the various stages of lens clarity indicated. Total number of treated or the canine 3-month safety assessment study. lenses (n) is listed in parentheses; stage 0, clear; stage 1, hazy cortical Preliminary Exposure and Lens Penetration Stud- region; stage 2, haziness characterized by a visible demarcation between nuclear and cortical regions. Grading system is based on a modification ies. In separate animal experiments lens drug levels after a from Dickerson et al. (1997). Figure is based on the collective results of jpet.aspetjournals.org single dose of CIG (25 mg/kg p.o.) approached 0.15 and 0.26 several independent experiments. ␮g/g lens weight after 2 and6hofexposure, respectively. This dose resulted in plasma levels that ranged between 8.8 1) and 27% were hazy with a visible demarcation between the and 12.6 ␮g/ml (26–38 ␮M). After a single dose of 300 mg/kg, nuclear and cortical regions (stage 2). Progression to stage 2 lens levels approached 0.18 and 0.70 ␮g/g lens weight after 2 graded opacities was also concentration-dependent with 27, and6hofexposure, respectively. Plasma levels of drug at 68, and 100% of all the lenses treated with 7.5, 15, and 30 ␮M this dose ranged between 22 and 30 ␮g/ml (65–90 ␮M). Based CIG, respectively, being affected within 24 h of exposure. on protein binding information (Torii et al., 1984), the free There was a correlation between adverse changes in the at ASPET Journals on January 18, 2020 circulating concentration of CIG would range between 1.85 clarity of lenses exposed to CIG and several biochemical and 2.7 ␮M for these two doses. markers of toxicity. Lenses exposed to CIG for 24 h showed In Vitro Assessment of CIG in Rodent Whole Lens concentration-dependent alterations in lens ATP content, mi- Explant Culture. CIG adversely affected the transparent tochondrial MTT reduction, GSH content, and wet weight quality of explanted lenses obtained from SD rats within 24 h (Fig. 2). Although lenticular GSH content increased after of exposure in treated culture medium. The incidence and exposure to 3.75 ␮M CIG, progressive decrements in lentic- severity of these adverse changes in lens clarity were concen- ular ATP content, MTT reduction, and GSH content occurred tration-dependent (Fig. 1). Although exposure to 0.375 and after exposure to concentrations Ն7.5 ␮M for 24 h. Lenticular 3.75 ␮M CIG for 24 h caused no discernible adverse effect on glutathione disulfide content was below the lower limit of lens clarity, concentrations Ն7.5 ␮M caused considerable quantitation (10 nmol/lens) in all samples tested (data not damage. Lenses exposed to 7.5 ␮M CIG for 24 h showed a shown). Lens wet weight progressively increased as a func- graded response. Of the lenses treated, 23% showed no effect tion of concentration at exposure levels Ն7.5 ␮M. Changes in (stage 0), whereas 50% were hazy in the cortical area (stage these representative biochemical markers of toxicity and lens

TABLE 1 Ocular (lens) effects of ciglitazone during 3-month safety assessment study in Sprague-Dawley rats

Day 49 Day 82 Treatment Incidence Description Range Incidence Description Range mg/kg/day p.o. Control (male) 0/14 Normal 0/14 Normal background Control (female) 0/15 Normal 0/15 Normal background 30 (male) 0/14 Normal 1/14 Posterior cortical capsular cataract with nuclear opacities 30 (female) 0/15 Normal 3/15 Posterior cortical capsular cataract with nuclear opacities 100 (male) 3/15 Stress lines (anterior ϩ posterior) 5/15 Posterior cortical capsular cataract 100 (female) 1/15 Total cataract 4/15 Posterior cortical capsular cataract with nuclear cataract 300 (male) 6/14 Stress lines (anterior ϩ posterior) with early 8/14 Posterior cortical capsular cataract, early cataract, total cataract mature complete cataract 300 (female) 8/13 Stress lines (anterior ϩ posterior) with early 8/13 Posterior cortical capsular cataract, early cataract, total cataract mature complete cataract 1030 Aleo et al.

␮ Fig. 2. Concentration-dependent alterations in ATP content, GSH con- Fig. 3. Protective effects of 100 M RR on ATP content, GSH content, and ␮ tent, MTT reduction, and wet weight of explanted Sprague-Dawley rat wet weight of explanted Sprague-Dawley rat lenses treated with 15 M Downloaded from ϩ lenses treated with CIG for 24 h. n ϭ 3 to 32 lenses per treatment, CIG for 24 h. Lens clarity was significantly improved in RR CIG- mean Ϯ S.E.M. Values with different superscripts within a given param- treated lenses compared with CIG treatment alone (cloudy versus cloudy ϭ eter are significantly different from each other using one-way ANOVA with nuclear cataract formation, respectively). n 3 to 4 lenses per Ϯ followed by Fisher’s LSD test, p Ͻ 0.05. Figure is based on the collective treatment, mean S.E.M. Values with different superscripts are signif- results of several independent experiments. icantly different from each other using one-way ANOVA followed by Fisher’s LSD test, p Ͻ 0.05. ␮ weight indicate that 7.5 M CIG was the threshold concen- jpet.aspetjournals.org tration at which both biochemical and morphological evi- itous declines in lenticular ATP and GSH content (up to dence of opacity formation became evident after 24 h of 100% compared with control values) were partially (ATP) or exposure. completely (GSH) attenuated by pretreating lenses with RR. In separate experiments, lenticular ATP content was the Pretreatment with RR also attenuated alterations in lens most sensitive biochemical marker of toxicity used in the weight (Fig. 3) and lens clarity compared with CIG treatment present investigation. A decrease of 37 Ϯ 7% in lenticular alone (stage 1 versus stage 2 grade opacities, respectively). In Ϯ Ϯ

ATP content (control 250 27 versus treated 157 39 nmol separate experiments using a lower concentration of CIG (7.5 at ASPET Journals on January 18, 2020 ATP/lens; mean Ϯ S.D.; p Ͻ 0.05) occurred within 3-h expo- ␮M) and RR (30 ␮M), RR completely prevented the loss of sure to 7.5 ␮M CIG. Decrements in lenticular ATP content ATP caused by CIG (control 76 Ϯ 11a versus 30 ␮MRR occurred before any other potential adverse changes in len- treatment 75 Ϯ 4a versus 7.5 ␮M CIG treatment 41 Ϯ 6b ticular GSH content (control 77 Ϯ 12 versus treated 79 Ϯ 9 versus 30 ␮MRRϩ 7.5 ␮M CIG treatment 73 Ϯ 8a nmol nmol GSH/lens; p Ͼ 0.05), wet weight (control 31 Ϯ 2 versus ATP/lens; mean Ϯ S.D.; n ϭ 5 lenses/treatment; p Ͻ 0.05). In treated 31 Ϯ 1 mg/lens; mean Ϯ S.D., n ϭ 5 lenses/treatment; this experiment, lens wet weight and transparency were not p Ͼ 0.05), and lens clarity were noted (data not shown). adversely affected by CIG relative to control (data not The effects of lower CIG concentrations during longer ex- shown). posure periods were also explored. Low concentrations of CIG Because the general protective effects of RR suggest that (3.75 ␮M) decreased lenticular ATP content 20 Ϯ 6, 57 Ϯ 6, CIG may perturb intracellular free calcium regulation, we and 57 Ϯ 4% after 24, 48, and 96 h, respectively. Lenticular also explored whether cataract formation by CIG would be GSH content showed a biphasic response, increasing 33 Ϯ 8% blocked with inhibitors of calcium-dependent calpain pro- after 24 h and then decreasing 39 Ϯ 6% at 48 h before teases. Pretreating lenses with the membrane-permeant cal- returning to control values after 96 h. Lens weight did not pain inhibitors E64d (100 ␮M) and CI-II (1 mM) did not increase relative to controls during the 96-h exposure period. protect lenses from CIG toxicity after a 24-h exposure period Except for an isolated reduction in lens GSH content (48 h) (Table 2). Aside from a statistically significant improvement and lens weight (96 h), there were no significant changes in on lens weight by CI-II, both calpain inhibitors did not pro- any other measured biochemical markers of toxicity in lenses tect lenses from decrements in lenticular ATP content (CI-II exposed to 0.375 ␮M CIG. These minimal and isolated effects treatment tested) or mitochondrial MTT reduction (E64d at 0.375 ␮M CIG suggest it was a relatively nontoxic concen- treatment tested) and GSH content caused by exposure to 15 tration in this short-term assay. There were no discernable ␮M CIG for 24 h. More importantly, these calpain inhibitors changes in lens transparency compared with controls at any had no apparent protective effect on the adverse changes in of the concentrations (0.375 and 3.75 ␮M) or exposure periods lens transparency caused by CIG exposure (i.e., treated (24, 48, and 96 h) tested (data not shown). lenses were still graded as stage 2 opacity). The stage 1 Since a decline in lenticular ATP content was noted before opacity observed with CI-II treatment alone is a known effect any other biochemical changes, the role of changes in lentic- at concentrations Ն250 ␮M (Lampi et al., 1992). ular mitochondrial function in the cataractogenic potential of To determine whether CIG caused toxicity by oxidative CIG was further explored. Lenses pretreated with 100 ␮M stress, we investigated whether antioxidants protected ex- RR, a mitochondrial calcium uniport inhibitor, 15 min before planted lenses from the cataractogenic potential of CIG. addition of CIG were partially protected from the adverse Lenses were exposed to 15 ␮M CIG for 24 h with or without effects of 15 ␮M CIG after 24 h of exposure (Fig. 3). Precip- the antioxidant and sulfhydryl reducing reagent DTT (2 Cataractogenic Effects of Ciglitazone in Vivo and in Vitro 1031

TABLE 2 Lack of protective effects of DTT, CI-II, E64d, and SORB on the toxicity of ciglitazone to explanted Sprague-Dawley rat lenses Results for ATP content are expressed as percentage of difference from control, mean Ϯ S.E.M., of three to five lenses per treatment. Lenses were incubated in the presence of CIG for 24 h using 0.2% ethanol as vehicle. CI-II and E64d were added 15 min or 2 h before CIG, respectively. DTT and SORB were added to the treated medium at the same time as CIG. Stage 0, clear; stage 1, hazy cortical region; stage 2, haziness characterized by a visible demarcation between nuclear and cortical regions. Grading system is based on a modification from Dickerson et al. (1997).

Treatment ATP Content MTT Reduction GSH Content Lens Weight Lens Appearance Calpain inhibitors Control 0 Ϯ 10a ND 0 Ϯ 5a 0 Ϯ 3a Stage 0 CI-II (1 mM) 7 Ϯ 8a ND Ϫ14 Ϯ 3b Ϫ2 Ϯ 1a Stage 1 CIG (15 ␮M) Ϫ74 Ϯ 3b ND Ϫ53 Ϯ 4c 31 Ϯ 5c Stage 2 CI-II ϩ CIG Ϫ70 Ϯ 6b ND Ϫ51 Ϯ 6c 15 Ϯ 2b Stage 2 Control ND 0 Ϯ 3a ND 0 Ϯ 5a Stage 0 E64d (100 ␮M) ND 25 Ϯ 7b ND Ϫ2 Ϯ 3a Stage 0 CIG (15 ␮M) ND Ϫ53 Ϯ 4c ND 43 Ϯ 9b Stage 2 E64d ϩ CIG ND Ϫ79 Ϯ 1d ND 27 Ϯ 3b Stage 2 Antioxidants Control 0 Ϯ 12a ND 0 Ϯ 10a 0 Ϯ 1a Stage 0 DTT (2 mM) Ϫ6 Ϯ 5a ND 5 Ϯ 4a Ϫ12 Ϯ 2b Stage 0 CIG (15 ␮M) Ϫ70 Ϯ 2b ND Ϫ67 Ϯ 2b 30 Ϯ 8c Stage 2 DTT ϩ CIG Ϫ77 Ϯ 2b ND Ϫ75 Ϯ 7b 29 Ϯ 2c Stage 2 Control 0 Ϯ 5a ND 0 Ϯ 4a 0 Ϯ 2a Stage 0 Downloaded from DTT (2 mM) Ϫ6 Ϯ 5a,b ND 5 Ϯ 4a Ϫ12 Ϯ 2b Stage 0 ␮ Ϫ Ϯ b Ϫ Ϯ b Ϯ c H2O2 (300 M) 12 3 ND 29 3 10 2 Stage 1 ϩ Ϫ Ϯ a Ϯ a Ϫ Ϯ a DTT H2O2 0 4 ND 9 8 2 2 Stage 0 Control ND 0 Ϯ 3a ND 0 Ϯ 5a Stage 0 DTT (2 mM) ND Ϫ2 Ϯ 3a ND Ϫ8 Ϯ 1a Stage 0 ␮ Ϫ Ϯ b Ϯ b H2O2 (400 M) ND 73 2 ND 12 3 Stage 1 ϩ Ϯ a Ϫ Ϯ a DTT H2O2 ND 4 3 ND 5 3 Stage 0 Aldose reductase inhibitors jpet.aspetjournals.org Control ND 0 Ϯ 3a ND 0 Ϯ 6a Stage 0 SORB (50 ␮M) ND 7 Ϯ 6a ND Ϫ4 Ϯ 5a Stage 0 CIG (15 ␮M) ND Ϫ55 Ϯ 1b ND 40 Ϯ 7b Stage 2 SORB ϩ CIG ND Ϫ53 Ϯ 7b ND 42 Ϯ 1b Stage 2 ND, not determined. a–d Values within a given experiment significantly different from each other, p Ͻ 0.05. at ASPET Journals on January 18, 2020 mM). As shown in Table 2, DTT had no protective effect on mg/kg i.v.) fed a sucrose diet for 2 to 5 days (data on file Pfizer any of the biochemical (ATP content, GSH content, and lens Global Research and Development). wet weight) markers of toxicity. DTT also did not protect In Vitro Ranking of Thiazolidinediones for Catarac- against adverse changes in the transparent quality of the togenic Potential. To demonstrate the utility of the explant lens caused by CIG. In contrast, DTT completely protected culture model system to distinguish cataractogenic com- lenses from the oxidative damage caused by 300 and 400 ␮M pounds, we rank ordered the thiazolidinediones: ciglitazone, ␮ H2O2, concentrations associated with either slight (300 M) darglitazone, and englitazone in vitro. Explanted lenses, ob- or substantial (400 ␮M) reductions in several biochemical tained from SD rats, exposed to equivalent concentrations markers of toxicity, equivalent increases in lens wet weight, (15 ␮M) of all three thiazolidinediones for 48 h showed and similar adverse changes in lens transparency. marked differences in their clarity (Fig. 4). CIG-treated Finally, because the hypoglycemic activity of CIG is related lenses were graded as stage 2, whereas darglitazone- and to its ability to facilitate translocation of glucose transporters englitazone-treated lenses were clear or very slightly hazy from intracellular storage pools to the plasma membrane and (barely gradable as stage 1), respectively. Based on this sin- isoforms of the transporter are differentially expressed in gle concentration comparison the rank order of cataracto- normal adult rat lenses (Merriman-Smith et al., 2003), an genic potential based on visual assessment alone was cigli- increase in lens weight may indicate osmotic stress caused by tazone ϾϾ englitazone Ͼ darglitazone. Visual assessment the abnormal accumulation of polyols (sugar metabolites). We examined whether the cataractogenic potential of CIG was associated with its pharmacological mechanism of ac- tion. Under the culture conditions described, sorbinil (SORB), a specific inhibitor of aldose reductase, the enzyme responsible for converting glucose to its active osmolyte sor- bitol, had no discernible protective effect on the biochemical or morphological indicators of CIG-mediated toxicity after a 24-h exposure period (Table 2). Furthermore, independent work by Chang and colleagues at the Upjohn Company showed that CIG (50 or 100 mg/kg/day for 2 to 4 days, respectively) inhibits aldose reductase activity in an uncom- Fig. 4. Visual appearance of rat lenses treated with the thiazolidinedio- ␮ petitive manner and lowers sorbitol accumulation in the lens nes ciglitazone englitazone, and 15 M darglitazone after 48 h of expo- sure (representative examples). Note presence of stage 2 opacity in cigli- (18–55% compared with control) and sciatic nerve (14–33% tazone-treated lens. Reprinted from Aleo et al. (2000) with permission compared with control) of streptozotocin-treated rats (70 from the New York Academy of Sciences (New York, NY). 1032 Aleo et al.

alone was sufficient to rank compounds for cataractogenic ular ATP content 56% after extended exposure periods of 48 potential in vitro. and 96 h. Based on exposure data and tissue dosimetry In Vivo Safety Assessment of Englitazone and Dar- information in the eye after exposure in vivo and the toxic glitazone for Cataractogenic Potential. Since both engli- effects of this compound at similar in vitro concentrations, tazone and darglitazone were less toxic to the lens compared these data strongly suggest that investigations of various with CIG, both compounds were selected for evaluation in biochemical mechanisms of cataract formation in the present subacute safety assessment studies in animals. Unlike CIG, study are germane to the in vivo setting. neither darglitazone nor englitazone caused lenticular opac- Changes in lenticular ATP content by CIG seem to be ities during the in-life portion of 3-month safety assessment caused in part by mitochondrial dysfunction. RR, an inhibitor studies in Long-Evans rats at consecutive daily oral doses of of the mitochondrial Ca2ϩ uniport, effectively blocked toxic- up to 50 mg/kg (darglitazone) and 100 mg/kg (englitazone), ity and opacity formation at concentrations of CIG that did respectively. These doses were associated with serum con- not result in changes in lens weight (7.5 ␮M) and provided ␮ centrations at Cmax that ranged from 24 to 60 M (darglita- partial protection from the adverse effects of higher concen- zone) and from 170 to 340 ␮M (englitazone) in male and trations of CIG (15 ␮M). A protective effect of RR was also female rats, respectively. With protein binding greater than seen in this model system by us with another cataractogenic 99% for both compounds, the effective free concentrations agent S-(1,2-dichlorovinyl)-L-cysteine (Walsh Clang and were at the most 0.6 and 3.4 ␮M, respectively. Aleo, 1997), a compound known to cause its primary toxic

effects via disturbances in mitochondrial bioenergetics and Downloaded from Discussion mitochondrial calcium regulation. However, since certain portions of the lens obtain the majority of its ATP through Rat and bovine lens explant cultures have long been used anaerobic respiration, it is possible that the dramatic de- to effectively investigate mechanisms of cataract formation crease in lens ATP content also reflects inhibition of this in animals induced by several xenobiotics such as tamoxifen, pathway as well. In either case, periodic or sustained reduc-

naphthalene, lovastatin, simvastatin, and S-(1,2-dichlorovi- tions in lenticular ATP content over a 3-month period caused jpet.aspetjournals.org nyl)-L-cysteine (Mosley et al., 1989; Kalinowski et al., 1991; by low concentrations of CIG could easily disrupt or impair Xu et al., 1992; Zhang et al., 1994; Walsh Clang and Aleo, crystallin formation, cholesterol biosynthesis or proper ion 1997). The present study is the first combined report of homeostasis within the lens. The functional integrity of all of lenticular opacity formation by CIG in rodents in vivo and in these processes is necessary to maintain normal lens trans- vitro using rat lens explant culture. CIG was a potent cata- parency. ractogenic agent in vitro, producing a concentration-depen- Other mechanisms of cataract formation seem to be less

dent loss of lens transparency and formation of opacities likely given the lack of a protective effect of pharmacologi- at ASPET Journals on January 18, 2020 (graded stage 2) that were evident within 24 h of exposure at cally active agents. The sulfhydryl reducing agent DTT and concentrations Ն7.5 ␮M. Several biochemical markers (ATP inhibitors of calpain activation were ineffective in preventing and GSH content) were sensitive indicators of toxicity with CIG toxicity to the lens despite their effectiveness in prevent- changes in lenticular ATP content preceding any visually ing ionophore- and sugar-induced cataract formation in ex- perceptive changes in lens transparency. For example, before planted rodent lenses (Azuma et al., 1992, 1995). Future discernible effects on lens transparency were noted, lenticu- investigations should focus on direct determinations of per- lar ATP content declined 37% within 3 h of exposure to 7.5 tubations in these pathways using GSH/glutathione disulfide ␮M CIG. Additionally, lenses exposed to lower concentra- ratio and lens protein sulfhydryl status as markers of oxida- tions of CIG (3.75 ␮M) for extended treatment periods (Յ96 tive stress, tempol-H, or other agents as a pure antioxidant h) had pertubations in both lenticular ATP and GSH content (Zigler et al., 2003), and/or activation of calpains using casein before any adverse effect on lens transparency was observed. zymography or lens protein proteolysis to understand the Increased lens wet weight was not a predictive or sensitive relative role of each of these mechanisms compared with marker of toxicity since it occurred only after changes in lens early and significant reductions in lens ATP content. clarity were evident by visual inspection and was not consis- More importantly for this class of compounds, the catarac- tently reproducible between some experiments. togenic potential of CIG does not seem to be related to its Although tissue distribution of CIG to the eye is low com- pharmacological mechanism of action. Although an increase pared with other body tissues (Torii et al., 1984), concentra- of lens weight could indicate osmotic stress caused by polyol tions used in vitro to cause lens opacification were consistent accumulation within lens tissue, the cataractogenic effects of with in vivo exposure data. Peak plasma concentrations of CIG do not seem to be involved using the present approach. CIG 2 h after a 30-mg/kg or 6 h after a 300-mg/kg oral dose Despite the documented ability of the aldose reductase inhib- approach 37 (Torii et al., 1984) and 90 ␮M in the plasma, itor SORB to effectively inhibit sugar-induced cataract for- respectively. This range of oral doses caused a dose-related mation in vivo (Kador et al., 1986; Yeh et al., 1987; Yeh and degree and incidence of bilateral cataract formation in a Ashton, 1990) and in vitro (Kador et al., 1986; Yeh et al., 3-month subacute safety assessment study in SD rats. Since 1986), both in terms of adverse changes in lens weight and CIG is approximately 95 to 97% bound to rat plasma proteins clarity and lenticular ATP and GSH content, SORB did not at these concentrations (Torii et al., 1984), the free circulat- attenuate the toxicity of CIG to rat lenses either in terms of ing concentration of CIG at peak would range between 1.85 lens clarity or on the same biochemical markers of toxicity and 2.7 ␮M for these two doses. Thus, the concentrations under the present culture conditions (Table 2). The concen- used in the in vitro studies, especially results obtained at tration of SORB used in this study was 5 times higher than 3.75 ␮M are extremely relevant to discerning mechanisms of the concentration that has been shown to effectively inhibit toxicity. In vitro exposure to 3.75 ␮M CIG decreased lentic- aldose reductase (Ͼ90%) and sorbitol accumulation (Ͼ80%) Cataractogenic Effects of Ciglitazone in Vivo and in Vitro 1033

Ϫ in rat lenses exposed to 35 mM glucose (Yeh et al., 1986). The protection the ocular lens from oxidative damage by endogenous O2 ,H2O2 and/or OH, in Biochemical and Clinical Aspects of Oxygen, pp 797–809, Academic Press, early work by A. Y. Chang (unpublished results) further New York. indicates that CIG does not cause abnormal accumulation of Bhuyan DK, Bhuyan KC, and Katzin HM (1973) Amizol-induced cataract and inhibition of lens catalase in rabbit. Ophthal Res 5:236–247. lens polyols. Chung SS, Ho EC, Lam KS, and Chung SK (2003) Contribution of polyol pathway to To our knowledge, the cataractogenic potential of CIG diabetes-induced oxidative stress. J Am Soc Nephrol 14(Suppl 3):S233–S236. seems to be unique among other thiazolidinediones. There Dickerson JE, Dotzel E, and Clark AF (1997) Steroid-induced cataract: new perspec- tive from in vitro and lens culture studies. Exp Eye Res 65:507–516. have been no reports of cataract formation in rats with dar- Herman JR, Dethloff LA, McGuire EJ, Parker RF, Walsh KM, Gough AW, Masuda glitazone up to 3 months of treatment at 50 mg/kg/day (Aleo H, and de la Iglesia FA (2002) Rodent carcinogenicity with the thiazolidinedione antidiabetic agent troglitazone. Toxicol Sci 68:226–236. et al., 2003) or troglitazone during subacute (13 weeks) and Herman JR, Metz AL, McGuire EJ, de la Iglesia FA and Masuda H (1997) Sub- chronic studies (104 weeks) in rats at daily doses of up to 800 chronic toxicity of the antidiabetic troglitazone in Wistar rats (Abstract). Fundam Appl Toxicol 36 (Suppl):273. mg/kg (Herman et al., 1997, 2002). Troglitazone also did not Kador PF, Akagi Y, and Kinoshita JH (1986) The effect of aldose reductase and its cause lenticular opacities after 52 weeks of treatment in inhibition on sugar cataract formation. Metabolism 35 (Suppl 1):15–19. Kalinowski SS, Tanaka RD, and Mosley ST (1991) Effects of long-term administra- cynomolgus monkeys (Rothwell et al., 1997) at oral doses of tion of HMG-CoA reductase inhibitors on cholesterol synthesis in lens. Exp Eye Res up to 1200 mg/kg/day. Furthermore, troglitazone (20 ␮M) 53:179–186. actually has a protective effect on polyol-induced cataract Lampi KJ, Kadoya K, Azuma M, David LL, and Shearer TR (1992) Comparison of cell-permeable calpain inhibitors and E64 in reduction of cataract in cultured rat formation in rat lenses exposed to galactose in vitro lenses. Toxicol Appl Pharmacol 117:53–57. (Yokoyama et al., 1999). To our knowledge, there have been Lou MF (2003) Redox regulation in the lens. Prog Retin Eye Res 22:657–682. Lou MF, Xu GT, Zigler S Jr, and York B Jr (1996) Inhibition of naphthalene cataract no reports of lenticular opacities with currently marketed in rats by aldose reductase inhibitors. Curr Eye Res 15:423–432. Downloaded from thiazolidinediones. Lubek BM, Basu PK, and Wells PG (1988) Metabolic evidence for the involvement of enzymatic bioactivation in the cataractogenicity of acetaminophen in genetically This study establishes the cataractogenic potential of CIG susceptible (C57BL/6) and resistant (DBA/2) murine strains. Toxicol Appl Phar- in rats, shows the utility of explanted lenses for investigating macol 94:487–495. Martynkina LP, Qian W, and Shichi H (2002) Naphthoquinone cataract in mice: lenticular opacity formation, and indicates the ability to rank mitochondrial change and protection by superoxide dismutase. J Ocul Pharmacol order compounds in vitro. Preliminary investigation into Ther 18:231–239. Merriman-Smith BR, Krushinsky A, Kistler J, and Donaldson PJ (2003) Expression mechanisms of CIG toxicity in the rat lens suggests that patterns for glucose transporters GLUT1 and GLUT3 in the normal rat lens and jpet.aspetjournals.org mitochondrial dysfunction may play a significant role but in models of diabetic cataract. Investig Ophthalmol Vis Sci 44:3458–3466. more extensive investigative studies are warranted. The Mosley ST, Kalinowski SS, Schafer BL, and Tanaka RD (1989) Tissue-selective acute effects of inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A reductase on cho- foundation of biochemical information presented here should lesterol biosynthesis in lens. J Lipid Res 30:1411–1420. be useful in establishing the basis for future investigations Murata T, He S, Hangai M, Ishibashi T, Xi XP, Kim S, Hsueh WA, Ryan Sj, and Law RE (2000) Peroxisome proliferator-activated receptor-gamma ligands inhibit cho- into mechanisms of cataract formation by CIG. Because of roidal neovascularization. Investig Ophthalmol Vis Sci 41:2309–2317. the potential for CIG to induce cataract formation in rats it is Rothwell CE, Bleavins MR, McGuire EJ, de la Iglesia FA, and Masuda H (1997) 52-week oral toxicity study of troglitazone in cynomolgus monkeys (Abstract). advisable to avoid the use of this compound to investigate Fundam Appl Toxicol 36 (Suppl):273. at ASPET Journals on January 18, 2020 ophthalmic indications for thiazolidinediones, especially Shearer TR, Ma H, Shih M, Fukiage C, and Azuma M (2000) Calpains in the lens and cataractogenesis. Methods Mol Biol 144:277–285. since other drugs in this chemical class do not have this Spector A (2000) Review: oxidative stress and disease. J Ocul Pharmacol Ther safety liability in animals. 16:193–201. Torii H, Yoshida K, Tsukamoto T, and Tanayama S (1984) Disposition in rats and dogs of ciglitazone, a new antidiabetic agent. Xenobiotica 14:259–268. Acknowledgments Walsh Clang CM and Aleo MD (1997) Mechanistic analysis of S-(1,2-dichlorovinyl)- We Kathy Bonnema for retrieving study information related to L-cysteine (DCVC)-induced cataractogenesis in vitro. Toxicol Appl Pharmacol 146:144–155. ciglitazone testing in rats and dogs generated by the late Dr. Henry Xu G-T, Zigler JS, and Lou MF (1992) The possible mechanism of naphthalene D. Webster and colleagues from the former Upjohn Company. cataract in rat and its prevention by an aldose reductase inhibitor (AL01576). Exp Eye Res 54:63–72. References Yeh L-A and Ashton MA (1990) The increase in lipid peroxidation in diabetic rat lens can be reversed by oral sorbinil. Metabolism 39:619–622. Aleo MD, Avery MJ, Beierschmitt WP, Drupa CA, Fortner JH, Kaplan AH, Navetta Yeh L-A, Rafford CE, Beyer TA, and Hutson NJ (1986) Effects of the aldose reductase KA, Shepard RM, and Walsh CM (2000) The use of explant lens culture to assess inhibitor sorbinil on the isolated cultured rat lenses. Metabolism 35 (Suppl cataractogenic potential. Ann NY Acad Sci 919:171–187. 1):4–9. Aleo MD, Lundeen GR, Blackwell DK, Smith WM, Coleman GL, Stadnicki SW, and Yeh L-A, Rafford CE, Goddu KJ, Ashton MA, Beyer TA, and Hutson NJ (1987) Kluwe WM (2003) Mechanism and implications of brown adipose tissue prolifer- ϩ ϩ ation in rats and monkeys treated with the thiazolidinedione darglitazone, a Na -K -ATPase pumping activity is not directly linked to myo-inositol levels after potent PPAR-␥ agonist. J Pharmacol Exp Ther 305:1173–1182. sorbinil treatment in lenses of diabetic rats. Diabetes 36:1414–1419. Aoun P, Simpkins JW, and Agarwal N (2003) Role of PPAR-gamma ligands in Yokoyama T, Yoshida Y, Inoue T, and Horikoshi H (1999) Inhibition of galactose- neuroprotection against glutamate-induced cytotoxicity in retinal ganglion cells. induced cataractogenesis by troglitazone, a new antidiabetic drug with an antiox- Investig Ophthalmol Vis Sci 44:2999–3004. idant property, in rat lens culture. J Ocul Pharmacol Ther 15:73–83. Azuma M, David LL, and Shearer TR (1992) Superior prevention of calcium iono- Zhang JJ, Jacob TJC, Valverde MA, Hardy SP, Mintenig GM, Sepulveda FV, Gill phore cataract by E64d. Biochim Biophys Acta 1180:215–220. DR, Hyde SC, Trezise AEO, and Higgins CF (1994) Tamoxifen blocks chloride Azuma M, Inoue E, Oka T, and Shearer TR (1995) Proteolysis by calpain is an channels: a possible mechanism for cataract formation. J Clin Investig 94:1690– underlying mechanism for formation of sugar cataract in rat lens. Curr Eye Res 1697. 14:27–34. Zhao C and Shichi H (1995) Histocytological study on the possible mechanism of Azuma M, Tamada Y, Kanaami S, Nakajima E, Nakamura Y, Fukiage C, Forsberg acetaminophen cataractogenesis in mouse eye. Exp Mol Pathol 63:118–128. NE, Duncan MK, and Shearer TR (2003) Differential influence of proteolysis by Zigler JS, Qin C, Kamiya T, Krishna MC, Cheng Q, Tumminia S, and Russell P calpain 2 and Lp82 on in vitro precipitation of mouse lens crystallins. Biochem (2003) Tempol-H inhibits opacification of lenses in organ culture. Free Radical Biol Biophys Res Commun 307:558–563. Med 35:1194–1202. Belusko PB, Nakajima T, Azuma M, and Shearer TR (2003) Expression changes in mRNAs and mitochondrial damage in lens epithelial cells with selenite. Biochim Address correspondence to: Dr. Michael D. Aleo, Pfizer Global Research Biophys Acta 1623:135–142. and Development, Groton Laboratories, Safety Sciences, MS 8274-1229, East- Bhuyan DK and Bhuyan KC (1979) Mechanism of cataractogenesis induced by ern Point Rd., Groton, CT 06340. E-mail: [email protected] 3-amino-1H-1,2,4-triazole II: superoxide dismutase of the eye and its role in