Abstract
Thalidomide produces numerous birth defects, the most notable being phocomelia. Mechanisms behind thalidomide-induced malformations have not been fully elucidated, although recent evidence suggests a role for reactive oxygen species. A thalidomide-resistant (rat) and -sensitive (rabbit) species were used to compare potential inherent differences related to oxidative stress that may provide a more definitive understanding of mechanisms of thalidomide embryopathy. Limb bud cells (LBCs) were removed from the rat and rabbit embryo, dissociated, and plated in culture for 24 h. A fluorescence (6-carboxy-2′,7′-dichlorofluorescin diacetate; DCF) assay for oxidative stress was used with varying concentrations of thalidomide (5–100 μM). Thalidomide (100 μM) showed a 6-fold greater production of oxidative stress in rabbit cultures than in rat. Lower concentrations (50 and 25 μM) also showed a significant increase in reactive oxygen species. Confocal microscopy revealed DCF fluorescence preferentially in rabbit LBC nuclei compared with the uniform distribution of DCF fluorescence in rat LBC. Localization of glutathione (GSH) was determined using 5-chloromethylfluorescein diacetate fluorescent confocal microscopy. In rat cultures, significant thalidomide-induced GSH depletion was detected in the cytosol but the nuclei maintained its GSH content, but rabbit LBC showed significant GSH depletion in both compartments. GSH depletion was confirmed by high-performance liquid chromatography analysis. These observations provide evidence that thalidomide preferentially produces oxidative stress in the thalidomide-sensitive species but not the thalidomide-resistant species. Nuclear GSH content in the rabbit LBC is selectively modified and indicates a shift in the nuclear redox environment. Redox shifts in the nucleus may result in the misregulation of transcription factor/DNA interactions and cause defective growth and development.
Thalidomide (α-phthalimidophthalimide), a drug originally marketed for its sedative-hypnotic effects, was removed from public use due to the discovery of its ability to cause human birth defects, mainly limb reduction malformations (McBride, 1961; Lenz, 1962, 1988). More recently, thalidomide was shown to have numerous therapeutic effects and was reintroduced for medicinal use in treating diseases such as erythema nodosum leprosum, human immunodeficiency virus-related wasting syndrome and esophageal ulcers, graft versus host disease, arthritis, and tuberculosis (Nightingale, 1998; Calabrese and Fleischer, 2000). Many hypotheses have been proposed to explain thalidomide teratogenesis, including DNA intercalation, acetylation of macromolecules, interference in glutamate metabolism, folic acid antagonism, and others (Faigle et al., 1962; Kemper, 1962; Jonsson, 1972; Stephens, 1988). Two recently presented hypotheses have suggested a thalidomide-induced disruption in angiogenesis and the decreased expression of adhesion receptors as possible teratogenic mechanisms (D'Amato et al., 1994; Neubert et al., 1996). Although many hypotheses have been proposed to explain the mechanism of thalidomide-induced teratogenesis, they do not sufficiently describe the mechanistic basis for thalidomide-induced limb reduction across different species.
Parman et al. (1999) showed that rabbits treated with thalidomide (400 mg/kg/day) produced fetuses with numerous defects, including limb reduction defects. Mice, which, like rats and certain other rodents, are thalidomide-resistant, were treated with 4-fold higher dosages of thalidomide (1600 mg/kg/day) and produced no significant increase in fetal malformations compared to controls. Furthermore, thalidomide-treated rabbit fetuses showed a significant increase in DNA oxidation, whereas thalidomide-treated mouse fetuses showed no such increase. Pretreatment with α-phenyl-N-t-butylnitrone (PBN), a free radical-trapping agent, lowered limb malformation rates in thalidomide-treated rabbit fetuses as well as decreasing the level of DNA oxidation. Investigators concluded that the increased production of reactive oxygen species (ROS) was involved in thalidomide embryopathy because malformations could be effectively blocked with the addition of a free radical trap (Parman et al., 1999).
Thalidomide has been shown to be a potent inhibitor of angiogenesis (D'Amato et al., 1994), although these mechanisms are also not fully understood. Sauer et al. (2000) used murine embryonic stem (ES) cells that differentiate into vascular structures and demonstrated that these cells have significantly increased levels of ROS with thalidomide treatment. After 5 days of culture, ES cells showed that angiogenesis could be effectively prevented. Coincidentally, cotreatment with the hydroxyl radical scavengers mannitol and 2-mercaptoethanol effectively prevented thalidomide-related inhibition of angiogenesis, and normal vessel differentiation of ES cells was restored. Investigators concluded that angiogenic processes were impeded due to thalidomide-generated ROS.
The production of ROS can result in many deleterious outcomes, but ROS concentrations can be effectively controlled by antioxidants, such as α-tocopherol, ascorbic acid, and glutathione (GSH). Glutathione is the most abundant intracellular thiol and accounts for approximately 90% of the entire intracellular pool of reducing equivalents (Cotgreave and Gerdes, 1998). Damage to lipid membranes, DNA, and proteins can result from excessive ROS exposure (Fantel, 1996), but lower and more transient levels of ROS may also have detrimental effects by altering redox status, the ratio of reducing equivalents to oxidized equivalents (Wells et al., 1997). Changes in the intracellular redox status have been associated with alterations in apoptosis, differentiation, and proliferation (Allen and Venkatraj, 1992; Hutter et al., 1997; Zhao et al., 1997; Cotgreave and Gerdes, 1998; Kirlin et al., 1999).
Herein, we provide evidence that thalidomide causes oxidative stress in rabbit limb bud cells (LBCs) compared with rat LBCs in culture. Furthermore, we also provide evidence that depletion of GSH in the rat LBCs occurs mostly in the cytosol, whereas oxidation of GSH in the rabbit LBCs involves the cytosol, organelles, and nucleus. Understanding the different responses to thalidomide between the rat and rabbit may provide critical information related to mechanisms involved in thalidomide teratogenesis. The effect of ROS formation and GSH oxidation, and subsequently modified redox status, may provide valuable insights into misregulation of specific developmental pathways connected to thalidomide-induced embryopathy.
Materials and Methods
Chemicals.
Glutathione, cysteine, methanesulfonic acid (MSA), diethylenetriaminepentaacetic acid, andN-[2-hydroxyethyl]-piperazine-N′-[3-propane-sulfonic acid] were obtained from Sigma Chemical (St. Louis, MO). Thalidomide (99.5% pure) was purchased from Andrulis Pharmaceuticals (Beltsville, MD). Monobromobimane (thiolyte) was obtained from Calbiochem (La Jolla, CA). MitoTracker Green FM, MitoTracker Red, Redox sensor red CC-1, 6-carboxy-2′,7′-dichlorofluorescin diacetate (DCF), and 5-chloromethylfluorescein diacetate (CMFDA) were purchased from Molecular Probes (Eugene, OR). Hanks' balanced salt solution (HBSS) and CMRL-1066 media were acquired from Invitrogen (Carlsbad, CA). Sodium methanesulfonate was purchased from Aldrich Chemical (Milwaukee, WI). All other reagents were obtained from common commercial sources.
Animals.
Primagravida Sprague-Dawley rats were obtained from the Toxicology Program Small Animal Facility, University of Michigan (Ann Arbor, MI) on gestational days (GD) 6 to 9. Day 0 was determined by a sperm-positive vaginal smear on the morning after copulation. Pregnant rats were maintained on a 12-h light/dark cycle until explantation on GD 13. Food and water were given ad libitum.
Primagravida New Zealand White rabbits (5.5–7 months old) were purchased from Covance Research Products, Inc. (Kalamazoo, MI). Females were naturally mated with males of the same strain with the day of copulation assigned as GD 0 at the suppliers facility, shipped on GD 1, and then acclimated to the new environment for 7 to 11 days before explantation (GD 12). Pregnant rabbits were kept on a photoperiod of 12-h light/dark. Food and water were also given ad libitum.
Micromass Culture.
The micromass limb bud cell culture technique used was slightly modified from the original rat method as described by Flint and Orton (1984), Kistler and Howard (1987), andKistler (1987). On GD 13 in rats and GD 12 in rabbits (similar developmental periods), animals were anesthetized with ether, and embryos were gathered and placed in warmed HBSS, pH 7.4. With a dissection microscope, both the forelimbs and hindlimbs were removed from each embryo, and each respective species' limbs were pooled. Limbs were washed twice in HBSS and placed in a sterile microcentrifuge tube containing calcium/magnesium-free phosphate-buffered saline, pH 7.4, supplemented with 0.1% trypsin (w/v) and 0.1% EDTA (w/v). The tube was placed in a water bath and heated to 37°C for 18 to 20 min or until limb buds were sufficiently dissociated. The tube was inverted two to three times during dissociation and was gently pipetted in and out at the end of the trypsinization to promote a single cell suspension. Nu-serum (BD Biosciences, Franklin Lakes, NJ) was added to the tube to stop the dissociation, and the cells were centrifuged for 5 min at 12,000g at 5°C after which the supernatant was removed and discarded. Cells were resuspended in CMRL-1066 media that had been supplemented with 10% Nu-serum (v/v) and antibiotics (penicillin and streptomycin, 100 U/ml and 100 μg/ml, respectively). The suspension was filtered using a 40-μm nylon filtered cell strainer to remove any undissociated cells. Single cells were then resuspended to a final concentration of 2.0 × 107 cells/ml. Aliquots consisting of 2.0 × 105 cells (10 μl) were placed in the center of separate wells on a 24-well plate. Special care was taken to not allow the droplet to touch the wall of each well. Wells where the aliquot did merge with the well wall were excluded from any further analysis. The plates were then placed in a humidified incubator for 1.5 h at 37°C in an atmosphere of 95% O2 per 5% CO2 (v/v), allowing the cells to attach to the plate in a concentrated area. After the initial incubation for attachment, each well was flooded with 1.0 ml of CMRL-1066 media containing 10% Nu-serum and antibiotics. The addition of thalidomide or other test chemicals did not constitute more than 1.0% of the total media in each well. Micromass cultures were maintained in a humidified incubator at 95% O2 per 5% CO2 (v/v) at 37°C.
Redox Status in LBC Cultures.
Histochemical methods used to evaluate intracellular LBC redox status are as originally described byChen et al. (1998) and the supplier (Molecular Probes). After a preincubation period allowing cells to attach to the plate surface, 50 μM hydrogen peroxide and 100 μM thalidomide were added directly to cell culture media. After a 60-min treatment, the dyes MitoTracker Green FM (0.5 μM) and Redox sensor red CC-1 (3 μM), diluted in media, were added, and cultures were placed back into the incubator for 15 min. Media were then removed from each well, and cells were fixed with 4% paraformaldehyde for 20 min. Special care was taken to keep cells from exposure to sunlight to prevent degradation of the dyes. LBCs were washed with PBS twice and then mounted with a coverslip by using Gelmount (Fisher Scientific, Chicago, IL) and visualized with a fluorescence microscope (100×). MitoTracker Green FM was visualized at an excitation of 490 nm and an emission of 516 nm. Redox sensor red CC-1 was visualized at an excitation of 560 nm and an emission of 600 nm.
During the 15-min incubation, MitoTracker Green FM is preferentially translocated only to the mitochondria. Redox sensor red CC-1, if introduced into an oxidative environment, is oxidized and translocated to lysosomes and mitochondria, but if introduced into a normal or reducing environment, it is translocated only to the lysosome (Chen et al., 1998). Visualization of the colocalization of the MitoTracker Green FM and Redox sensor red CC-1 in the mitochondria is indicative of an oxidative intracellular environment. Absence of dye colocalization provides evidence of a less oxidative (more reducing) environment.
DCF Fluorescence Detection of Oxidative Stress.
The method for fluorescence detection of oxidative stress is slightly modified from the original as described by Wang and Joseph (1999). Micromass cultures were grown at least 24 h before assay. Media were removed and replaced with loading media (CMRL-1066 with 1% fetal bovine serum) containing 100 μM DCF. Cultures were placed back into the incubator for 45 min followed by washing with PBS. Cultures were flooded with media containing 100, 50, 25, 10, or 5 μM thalidomide, 25 μM hydrogen peroxide, or 100 μM thalidomide and 200 μM PBN and placed in a multiwell fluorescence microplate reader (Gemini XS; Molecular Devices, Sunnyvale, CA) maintained at 37°C. Culture fluorescence was followed for 120 min at an excitation 488 nm and an emission at 530 nm. Data points were taken every 2.5 min. After the final measurement at 120 min, the percentage of increase was determined using the formula [(FT120 −FT0)/FT0∗ 100], where FT0 = fluorescence at 0 min and FT120 = fluorescence at 120 min.
DCF Fluorescence.
DCF techniques used for the localization of oxidative stress in rat and rabbit LBCs were performed as described by the supplier (Molecular Probes). Rat and rabbit limb bud cells were plated on a glass coverslip in CMRL-1066 media as indicated above for 24 h before treatment and preparation for microscopy. Incubation conditions were as described above for micromass cultures. Media were removed and replaced with loading media (CMRL-1066 with 1% fetal bovine serum) containing 100 μM DCF for 30 min. Cells were treated with 100 μM thalidomide in fresh media for 120 min. Coverslips containing the cells were removed and placed in a coverslip holder set in a heated plate and viewed using fluorescent confocal microscopy with a FluoView confocal microscope (Olympus, Melville, NY) at an excitation of 488 nm and an emission of 530 nm for DCF.
CMFDA Microscopy.
CMFDA techniques used for the detection and analysis of GSH content in rat and rabbit LBCs were performed as described by the supplier (Molecular Probes). Rat and rabbit LBCs were plated on a glass coverslip in CMRL-1066 media as indicated above for 24 h before treatment and preparation for microscopy. Incubation conditions were as described above for micromass cultures. Cells were treated with 100 μM thalidomide for 120 min. The media were removed, and cultures were subsequently washed with PBS. Fresh serum-free media containing 100 nM CMFDA and 200 nM MitoTracker Red, to determine nuclear area, were added and cultures were placed back into the incubator for 30 min. Media were then removed and replaced with normal serum-containing media for another 30 min. Cultures were fixed with 4% paraformaldehyde for 15 min, placed on a coverslip, and viewed on a FluoView confocal microscope (Olympus) at an excitation of 488 nm and an emission of 530 nm for CMFDA and an excitation of 578 nm and an emission of 600 nm for MitoTracker Red. Computer analyses of pixel intensity were performed with the Olympus FluoView imaging software. Relative intensities were downloaded into spreadsheets and were analyzed for statistically significant differences in the nucleus and the entire cell for CMFDA fluorescent differences.
HPLC Analysis.
After the plating procedure as described above, thalidomide dissolved in dimethyl sulfoxide (DMSO) was given in various concentrations to each well (100, 50, 25, 10, and 5 μM) and incubated for 5 days. Treated cultures did not receive more than 0.5% DMSO per total media volume. Control cultures received 0.5% DMSO. On day 7 of culture, micromass cultures were prepared for GSH and cysteine analysis by washing twice with HBSS to remove excess media. Two hundred microliters of 200 mM MSA was added to each well, and then the cells were removed from the plate surface manually using a cell scraper. Samples in MSA were pipetted from the 24-well plate, placed in separate microcentrifuge tubes, snap-frozen in liquid nitrogen, and stored at −70°C until prepared for HPLC analysis. GSH and cysteine concentrations were measured using the HPLC method as outlined by Fahey and Newton (1987) and as modified by Harris (1993). Frozen micromass samples were thawed and homogenized by ultrasonic disruption. Sodium methanesulfonate (4 M, 200 μl) was added to each tube to precipitate protein in the sample and was followed by centrifugation at 13,900 rpm for 5 min at 5°C. The resulting supernatants were removed and separately placed into new microcentrifuge tubes containing 160 μl of 5 mM diethylenetriaminepentaacetic acid in 1 MN-[2-hydroxyethyl]-piperazine-N′-[3-propane-sulfonic acid], pH 8.5. Derivatization was carried out by the addition of 20 μl of 0.2 mM monobromobimane (thiolyte) to each tube followed by gentle mixing. Samples were allowed to incubate for 20 min in the dark at room temperature. Derivatization was terminated by the addition of 380 μl of 400 mM MSA after which samples were snap-frozen in liquid nitrogen and stored at −70°C.
HPLC analyses were carried out by using a Waters NovaPak C 4-μm Radial-Pak cartridge fitted with a NovaPak Guard-Pak guard precolumn. Samples were eluted using an isocratic mobile phase consisting of 14.2% methanol (v/v) and 2.5% glacial acetic acid (v/v) at a flow rate of 1.0 ml/min. After elution of both the cysteine and GSH peaks, the column was washed with 90% methanol (v/v) and 2.5% glacial acetic acid (v/v) for 15 min. Detection and quantitation of GSH-bimane and cysteine-bimane was accomplished using a Waters model 470 scanning fluorescence detector (λ excitation 360 nm; λ emission 455 nm) followed by analysis and quantitation by a Waters model 746 data module. Authentic standards were prepared and used to identify each peak of interest. These same standards were also used to quantify each sample. This method is sensitive enough to accurately detect GSH and cysteine levels as low as 10 pmol/200-μl injection.
Statistical Analysis.
Significant differences between species for control values (cysteine and glutathione) were determined by the Student's t test. Differences within the treatments of similar species were determined by a one-way ANOVA followed by a Tukey's studentized range (honestly significant difference).P values that were at or below 0.05 were considered a significant difference between treatments.
Results
Redox-Sensitive Dye.
Both rat and rabbit LBCs showed near identical patterns of staining after both hydrogen peroxide and thalidomide treatments. Thalidomide (100 μM)-treated rabbit LBCs showed that MitoTracker Green FM colocalized with Redox sensor red CC-1 in the mitochondria, a sign of an oxidative intracellular environment (under Materials and Methods; Fig.1, A and B). Similarly, hydrogen peroxide treatment, used as a positive control to promote an oxidative environment, showed almost identical staining as seen in thalidomide LBCs (Fig. 1, C and D). Both thalidomide and hydrogen peroxide treatments showed staining that differed from controls in that the Redox sensor red CC-1 colocalized with the MitoTracker Green FM dye in the mitochondria, whereas controls showed very little dye colocalization (Fig. 1, E and F).
DCF Fluorescent Assay.
As a positive control, hydrogen peroxide substantially increased DCF fluorescence in both rat and rabbit LBC cultures, measuring 210 ± 12.1 and 203 ± 8.5%, respectively (Fig. 2).
Rat LBC cultures did not show any significant increase in oxidative stress by using 25, 10, or 5 μM thalidomide, measured as an increase in fluorescence from control of 13.1 ± 3.1, 10.2 ± 2.1, and 2.0 ± 10.4%, respectively. Significant differences were detected at the two highest concentrations, 50 and 100 μM thalidomide, and showed an increase in fluorescence from control of 20.6 ± 4.3 and 29.2 ± 5.6%, respectively.
Rabbit LBC cultures treated with 10 and 5 μM thalidomide did not produce results significantly different from control values, measuring 14.9 ± 5.4 and 3.0 ± 13.7% increases from control fluorescence (Fig. 2). However, 25, 50, and 100 μM thalidomide concentrations all produced significant levels of oxidative stress and differences in fluorescence. Cultures treated with 25 μM thalidomide increased from control fluorescence by 26.7 ± 6.0%. Both treatments of 50 and 100 μM thalidomide resulted in increased fluorescence of 86.4 ± 16.4 and 183.8 ± 19.4% in rabbit LBC cultures. Comparisons between thalidomide treatments in rat and rabbit LBC cultures showed that significant differences in fluorescence were shown at the two highest concentrations, 50 and 100 μM.
Addition of PBN caused a decrease in ROS-induced DCF fluorescence in both rat and rabbit LBC cultures treated with thalidomide. Neither rat nor rabbit LBC cultures showed DCF fluorescence comparable with control levels of fluorescence. Cotreatment with PBN decreased fluorescence near to that of the controls (Fig. 2). Differences between controls and thalidomide/PBN-treated rabbit LBC cultures were not statistically different, but significant differences were evident between rabbit LBC cultures treated with thalidomide and those receiving both thalidomide and PBN.
DCF Dye Fluorescence Microscopy.
Using DCF to determine areas within the cellular environment that are prone to produce oxidative stress revealed that in the rat, LBC produced oxidative stress but it was primarily localized to the cytosol (Fig.3A). Rabbit LBCs also produced evidence of oxidative stress in the cytosol (Fig. 3C), but the area where oxidative stress seemed most concentrated was in the nucleus.
CMFDA Dye Fluorescence Microscopy.
Rat control LBCs showed a fairly even distribution of GSH throughout the cell with indications of slightly higher concentrations in the perinuclear cytosol as demonstrated with CMFDA. Rat LBCs treated with 100 μM thalidomide (Fig. 4C) showed that there was a significant depletion of GSH in the cytosol compared with LBC controls (Fig. 4A). Using MitoTracker Red, we were able to demonstrate that some of the intracytosolic fluorescence foci are mitochondria as evidence by colocalization of the two dyes (data not shown). However, some cytosolic areas of dye accumulation did not overlap mitochondria and are presumed to be other organelles, such as the endoplasmic reticulum and Golgi apparati. Nuclear pools of GSH did not seem to be affected by thalidomide treatment.
Rabbit LBCs differed from rat LBCs in that general overall staining was lower. This may be due to lower GSH content as we have demonstrated below. Interestingly, visualization of the nucleus was much easier in rabbit LBCs and generally contained more GSH than the cytosol (Fig.4E). Nuclei were confirmed by absence of MitoTracker Red staining. Treatment with 100 μM thalidomide produced GSH depletion throughout the cytosol as was seen in rat LBCs (Fig. 4G). Nuclear GSH was significantly depleted to the point where the nucleus and the cytosol could not be distinguished from one other as in the controls without the use of MitoTracker Red.
Intensity of CMFDA fluorescence could be effectively quantified using pixel intensity measurements and correlates directly with available reduced GSH. CMFDA intensity found in the rat limb bud cell averaged 2447 ± 49.2 relative fluorescence units (RFU/cell; Fig.5). The rabbit limb bud cell measured 1400 ± 30.9 RFU/cell, 57% less than the rat. Treatment with 100 μM thalidomide resulted in a decrease in fluorescence in the rat limb bud cell to 1910 ± 45.0 RFU/cell, a decrease of 22% from control fluorescence. In rabbit limb bud cells, fluorescence was decreased to a greater extent, 42% from control fluorescence, measuring 817 ± 17.6 RFU/cell.
Inspection of nuclear CMFDA fluorescence intensity showed that the rat LBCs fluoresced 54% more than rabbit LBCs, measuring 4095 ± 16.2 and 2662 ± 23.0 RFU/nucleus, respectively (Fig. 5). Thalidomide did not affect CMFDA fluorescence in rat LBC nuclei (3980 ± 14.9 RFU/nucleus). However, thalidomide treatment in rabbit LBCs significantly decreased nuclear CMFDA fluorescence by 47%, 1400 ± 18.9 RFU/nucleus.
Glutathione and Cysteine in LBCs in Micromass Cultures.
Control rat LBCs contained 1830 ± 193.1 pmol of GSH per micromass (Fig. 6). Upon treatment with low concentrations of thalidomide (5 μM), GSH levels increased substantially to 2515 ± 171.1 pmol of GSH per micromass, a significant increase of 37% from control GSH concentrations. Thalidomide concentrations of 10 and 25 μM produced no difference in rat LBCs from control LBC values, measuring 1875 ± 128.2 and 1577 ± 99.1 pmol of GSH per micromass, respectively. Significant differences were evident in the subsequently higher thalidomide concentrations of 50 and 100 μM, measuring 1356 ± 86.5 and 1242 ± 80.0 pmol of GSH per micromass, 26 and 32% decreases, respectively.
Control rabbit LBCs were shown to contain 1211 ± 133.0 pmol of GSH per micromass (Fig. 6). Upon treatment with 5 μM thalidomide, rabbit LBC GSH significantly decreased by 39%, measuring 741.7 ± 35.2 pmol of GSH per micromass. Thalidomide concentrations that did not show any significant GSH depletion in rat LBC cultures did demonstrate significant GSH depletion in the rabbit LBC cultures, measuring 607.0 ± 40.4 pmol of GSH per micromass at 10 μM thalidomide and 640.5 ± 58.1 pmol of GSH per micromass at 25 μM thalidomide, decreases of 50 and 47%, respectively, from control GSH concentrations. Thalidomide concentrations of 50 and 100 μM caused a decrease in GSH concentrations, measuring 496 ± 48.5 and 392 ± 43.1 pmol of GSH per micromass, a decrease of 59 and 68%, respectively, from control GSH concentrations. Differences between control rat and rabbit LBCs in micromass were statistically significant. Rat LBCs were shown to contain nearly 50% more GSH per micromass than rabbit LBCs.
Rat LBC micromass cultures contained 670 ± 34.5 pmol of cysteine per micromass (Fig. 7). Thalidomide treatment significantly decreased cysteine concentrations to 485 ± 27.9, 468 ± 48.6, 392 ± 20.4, 330 ± 18.7, and 301 ± 14.5 pmol of cysteine per micromass for the 5, 10, 25, 50, and 100 μM concentrations, respectively.
Control rabbit LBC micromass cultures contained 150.8 ± 18.9 pmol of cysteine per micromass (Fig. 7). Thalidomide treatment significantly decreased cysteine concentrations to 13.6 ± 2.8, 21.5 ± 1.5, 27.0 ± 3.4, 24.1 ± 5.2, and 14.5 ± 6.5 pmol of cysteine per micromass for the 5, 10, 25, 50, and 100 μM concentrations, respectively.
As seen with GSH content, a major difference was also detected upon comparison between rat and rabbit control LBC micromass cultures. Rabbit LBC cultures contained significantly less cysteine than rat LBC cultures, a difference of 346% more in the rat LBC cultures than the rabbit LBC cultures.
Discussion
Hypotheses that oxidative stress induces thalidomide-induced limb malformations have received significant experimental support (Hansen et al., 1999; Parman et al., 1999; Sauer et al., 2000). Sauer et al. (2000) showed that thalidomide caused an increase in hydroxyl radical production in ES cells, Parman et al. (1999) confirmed that thalidomide-induced oxidative stress in utero leads to phocomelia, andHansen et al. (1999) demonstrated GSH depletion in rat and rabbit embryos treated in culture, but these studies do not specifically address effects in the most commonly affected tissue, the limb. After reacting with ROS, DCF is cleaved and converted to a form that fluoresces, making localization by microscopy and computer-assisted quantitation of ROS possible. Our findings support previous studies by showing that thalidomide does cause oxidative stress (Hansen et al., 1999; Parman et al., 1999; Sauer et al., 2000). However, unlike previous studies, the present study examines specific effects on the target susceptible to thalidomide-induced teratogenesis, the limb.
Experiments used thalidomide concentrations of 0 to 100 μM to elicit oxidative stress. Previous studies have shown that the highest rate of malformation in the rabbit is manifested at oral doses of 400 mg/kg/day (Schardein, 1993). Tsenova et al. (1998) demonstrated that oral doses of 200 mg/day (∼100 mg/kg/day) resulted in a maternal blood concentration between 20 and 25 μg/ml (∼100 μM) and support the rationale for 0 to 100 μM thalidomide in vitro. The degree to which ROS are manifested differs significantly between limb bud cells from different species. Although specific reasons for increased ROS production in rabbits are unknown, thalidomide pharmacokinetics between sensitive and insensitive species demonstrates differences in embryonic regions of accumulation. Schmahl et al. (1996) showed that in utero exposure to radioactively labeled EM12, a highly teratogenic thalidomide derivative, to thalidomide-sensitive marmosets resulted in the preferential accumulation in the limb of, compared with other embryonic regions. Moreover, preferential accumulation of EM12 in the limb was not evident in thalidomide-resistant species, the rat and mouse (Schmahl et al., 1996). Specific pharmacokinetics of thalidomide in rat and rabbit embryos have not been identified as of yet, but may account for the increased susceptibility to thalidomide-induced oxidative stress in the rabbit LBC. Greater GSH depletion may be evident in rabbit LBCs due to selective, enhanced thalidomide transport or binding, resulting in greater intracellular accumulation.
ROS shifts redox potential by decreasing reducing equivalents and increasing oxidizing equivalents. To verify that thalidomide-induced ROS production was causing redox potential shifts, we used a redox-sensitive dye in thalidomide-sensitive rabbit LBCs. Untreated LBCs showed evidence of a natural reducing environment, but treatment with thalidomide altered intracellular redox status and produced staining typical of a pro-oxidizing environment. Although ROS production does not necessarily translate to changes in redox potential, our observations suggest that thalidomide modifies the intracellular environment to a more pro-oxidizing condition, and any subsequent processes that are redox-sensitive may be altered, attenuated, or inhibited.
Rat LBCs showed relatively uniform DCF staining between the cytosol and nucleus, whereas rabbit LBCs showed the nucleus to be the primary site of ROS formation. These data support previous studies where DNA oxidation occurred in thalidomide-sensitive rabbit embryos, whereas thalidomide-resistant mouse embryonic DNA was relatively unaffected by thalidomide exposure (Parman et al., 1999). Nuclear GSH concentrations are critical to combat the production of ROS and impede damage to the DNA. Bellomo et al. (1992) reported that rat hepatocytes contained glutathione ratios between the nucleus and cytosol were 3:1, indicating a large amount of GSH is concentrated in the nucleus, suggesting that the nucleus maintains a highly reductive environment, possibly to regulate redox-sensitive signals and/or protect DNA. Preferential GSH compartmentalization was evident in rabbit LBCs by using fluorescent microscopy, where the nucleus clearly contained higher concentrations of GSH than the cytosol. Interestingly, rat LBC visualization of nuclear GSH with CMFDA fluorescence did not show a clear demarcation between the cytosol (perinuclear) and nucleus because both contain similar GSH concentrations. Rat LBCs showed a ratio of nuclear to cytosolic GSH concentrations of 0.86, indicating a relatively even distribution of GSH between cellular compartments, whereas rabbit LBCs showed a ratio of 3.21, indicating that the bulk of GSH was localized in the nucleus (Table 1). Glutathione localization experiments after thalidomide treatment confirmed GSH depletion in both the rat and rabbit LBC. Thalidomide concentrations known to result in oxidative stress (100 μM) caused depletion in both rat and rabbit LBCs, but treated rat LBC nuclei still contained a generous amount of reduced GSH. Using the same thalidomide concentrations, rabbit LBC nuclei were significantly affected, and nuclear GSH was depleted, indicating that the nucleus may be the sensitive region for misregulation. The alteration of many cell processes (calcium sequestration, transcription factor activation, kinase activity) can result from changing the nuclear redox balance from a reducing to a pro-oxidizing environment. Glutathione concentrations affect many nuclear processes, including the regulation of the nuclear matrix organization (Dijkwel and Wenink, 1986); maintenance of cysteine residues on Zinc-finger DNA binding motifs in a reduced and functional state (Klug and Rhodes, 1987); chromosome consolidation (De Capoa et al., 1982); DNA synthesis (Suthanthiran et al., 1990); DNA protection from oxidative stress (Sandstrom and Marklund, 1990); and protection of DNA-binding proteins (Sen and Packer, 1996), indicating the importance for nuclear GSH.
Mechanisms behind GSH nuclear localization are not fully understood, but some compelling evidence has been reported concerning the role of Bcl-2 in the localization process. Bcl-2 is a pore-forming protein in both the mitochondrial membrane and the nuclear envelope (Voehringer et al., 1998; Voehringer, 1999). Overexpression of Bcl-2 causes the preferential localization of GSH to the nucleus, whereas underexpression of Bcl-2 causes the preferential localization of GSH in the cytosol (Voehringer et al., 1998). The ability for rats to maintain their nuclear GSH concentration may be due to inherently higher levels of Bcl-2 expression in the nuclear envelope. Limb bud cell oxidative stress may cause GSH to be transported into the nucleus via Bcl-2-mediated mechanisms. Movement of GSH into the nucleus would maintain a normal nuclear redox potential, preserving processes such as DNA/protein interactions and subsequent gene expression. Because rabbit cytosolic GSH concentrations are inherently low, facilitated transport of GSH may not be protective during oxidative stress, allowing for more severe effects on nuclear processes and redox state. Conversely, the rat LBC, containing abundant GSH cytosolic stores, would protect DNA from oxidative insult and continue to maintain a normal intranuclear redox status.
Another observation that suggests the rabbit LBC is more susceptible to redox modulation is that rat and rabbit LBCs contain inherently different quantities of GSH, as determined in control micromass cultures, where rabbit LBC GSH concentrations were considerably lower. In terms of GSH as an antioxidant, rabbit LBCs have a lowered capacity to deal with ROS, and in terms of GSH as a regulator of redox status, it suggests that less ROS can modify nuclear and cytosolic redox states in the rabbit LBCs more easily than the rat's, making the rabbit LBCs more susceptible to misregulation of redox-related transactivation pathways.
Redox misregulation may be further compounded by lower cysteine concentrations in rabbit LBCs, a fact that is supported in previous rat and rabbit embryonic comparisons (Hansen et al., 1999). Cysteine concentrations were appreciably lower in rabbit LBCs compared with rat LBCs. Cysteine, secondary to GSH, is one of the more abundant reduced thiols in the cell and contributes to redox status as a reducing equivalent. Lowered concentrations of cysteine would decrease the pool of available reducing equivalents. Furthermore, lowered cysteine concentrations would compromise the ability to produce GSH, synthesize proteins, and act in other important cellular functions. Thalidomide treatment significantly depleted cysteine concentrations in both rats and rabbit LBC cultures. In fact, treated rabbit LBC cultures were almost completely devoid of cysteine. Rat LBC cultures still contain enough cysteine to rejuvenate lost GSH via de novo synthesis, whereas rabbit LBC cultures cannot due to the absence of available cysteine. Species-specific cysteine depletion may explain the more dramatic GSH depletion and lack of the rebound effect in rabbit LBC cultures. Cysteine is the rate-limiting precursor for GSH synthesis de novo. Lower cysteine concentrations during a period where GSH is being actively lost would hinder the cells ability to replace large quantities of GSH via de novo synthesis. Again, these results suggest that the rabbit LBC redox balance is more easily perturbed during oxidative stress.
This study demonstrates that thalidomide depletes GSH and alters intracellular redox status. Moreover, rabbit LBCs are more sensitive to thalidomide-induced GSH depletion than rat LBCs. Because rabbits are thalidomide-sensitive and rats are thalidomide-resistant, excessive ROS production and nuclear redox potential shifts represent critical elements of a plausible mechanism for thalidomide embryopathy. Clearly, further study is needed to determine the effects of thalidomide on altered redox potentials and excessive ROS production in both rat and rabbit LBCs.
Acknowledgments
We thank Terry Miller for expertise and technical assistance with confocal microscopy and Surekha Akella and Melissa Beck for help with animal explantation procedures.
Footnotes
-
This study was supported by National Institutes of Health Grants ES05235 and ES07062, the Dow Chemical Company, Office of the Vice President of Research of the University of Michigan, and the School of Public Health of the University of Michigan.
- Abbreviations:
- PBN
- N-t-butyl-α-phenylnitrone
- ROS
- reactive oxygen species
- ES
- embryonic stem cell
- GSH
- glutathione
- LBC
- limb bud cell
- MSA
- methane sulfonic acid
- DCF
- 6-carboxy-2′,7′-dichlorofluorescin diacetate
- DCF
- 6-carboxy-2′,7′-dichlorofluorescin diacetate
- CMFDA
- 5-chloromethylfluorescein diacetate
- HBSS
- Hanks' balanced salt solution
- GD
- gestation day
- PBS
- phosphate-buffered saline
- HPLC
- high-pressure liquid chromatography
- DMSO
- dimethyl sulfoxide
- RFU
- relative fluorescence units
- Received September 17, 2001.
- Accepted November 14, 2001.
- The American Society for Pharmacology and Experimental Therapeutics