Significance of 2-Methoxypropionic Acid Formed from ß-Propylene Glycol Monomethyl Ether: Integration of Pharmacokinetic and Developmental Toxicity Assessments in Rabbits

E. W. Carney*,1, L. H. Pottenger{dagger}, K. A. Johnson*, A. B. Liberacki*, B. Tornesi*, M. D. Dryzga*, S. C. Hansen* and W. J. Breslin{ddagger}

* Toxicology & Environmental Research and Consulting, 1803 Building, The Dow Chemical Company, Midland, Michigan 48674; {dagger} Dow Europe, GmbH, Horgen, Switzerland; and {ddagger} Eli Lilly & Company, Greenfield, Indiana

Received July 23, 2002; accepted September 30, 2002


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Commercial grade propylene glycol monomethyl ether (PGME), which is composed of > 99.5% {alpha}-isomer and < 0.5% ß-isomer, has been shown in several studies to have a low potential for developmental toxicity. Nonetheless, questions have been raised about potential human developmental toxicity due to ß-PGME, because it can be metabolized to 2-methoxypropionic acid (MPA), a compound bearing structural similarity to the teratogen, methoxyacetic acid (MAA). Accordingly, a series of in vivo developmental toxicity, whole embryo culture, and in vivo pharmacokinetic experiments were conducted in New Zealand White rabbits (highly sensitive to these compounds) to better understand the developmental toxicity potential of MPA and the kinetics of its formation from ß-PGME. For the in vivo developmental toxicity studies, groups of 20 inseminated rabbits were gavaged with 0, 10, 26, or 78 mg/kg/day of MPA on gestation day (GD) 7–19, followed by fetal evaluation on GD 28. Results with MPA were compared with those of rabbits similarly dosed with 0, 2.5, 7.5, or 15 mg/kg/day of MAA. Developmental toxicity no-observable-effect levels (NOEL) were approximately 10-fold higher for MPA (26 mg/kg/day) than for MAA (2.5 mg/kg/day). Also, the severity of effects caused by MPA was less than that of MAA, and unlike MAA, MPA was not selectively toxic to the fetus. This differential toxicity was also seen in whole embryo cultures of GD 9 rabbit embryos, in which there were no adverse effects of MPA (1.0, 5.0 mM) or its parent compound, ß-PGME (0.5, 2.0 mM), but severe dysmorphogenesis in 100% of embryos cultured in 5.0 mM MAA. The pharmacokinetics study showed rapid and complete conversion of ß-PGME to MPA, with a relatively long elimination half-life (33–44 h) for MPA. However, peak and AUC concentrations of MPA in blood associated with the MPA LOEL dose of 78 mg/kg/day were 1.3 mM and 52.9 mM-h/l, respectively, suggesting a relatively high threshold based on internal dosimetry. Taken together, these data indicate a negligible risk of developmental toxicity due to MPA formation from the small amounts of ß-isomer present in commercial PGME.

Key Words: glycol ethers; whole embryo culture; maternal pharmacokinetics.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The glycol ethers comprise a large family of compounds distinguished by the presence of both ether and alcohol moieties. As a result, they are miscible with both aqueous and organic solvents, making them ideally suited for a wide variety of industrial applications. Glycol ethers are used extensively in the manufacture of lacquers, paints, dyes, inks, hydraulic brake fluids, and cleaning agents, among others.

While it is tempting to view the glycol ethers as a single class of chemicals with common properties, the toxicity profiles across different glycol ether subclasses vary significantly. In the case of developmental toxicity, some of the ethylene glycol ethers have received considerable attention due to their ability to induce fetal malformations (reviewed by Boatman and Knaak, 2001Go; Hardin, 1983Go; Hevelin, 1989Go). The prototype ethylene glycol ether is ethylene glycol monomethyl ether (EGME; also known as 2-methoxyethanol), which causes digit, skeletal, and cardiovascular defects, as well as fetal resorptions in mice, rats, and rabbits (Hanley et al., 1984bGo; Horton et al., 1985Go; Nelson et al., 1984Go; Toraason et al., 1985Go). EGME is converted by alcohol/aldehyde dehydrogenase (ADH/ALDH) to an alkoxy acid metabolite (Fig. 1Go), methoxyacetic acid (MAA), which is considered the proximate developmental toxicant (Brown et al., 1984Go; Sleet et al., 1988Go; Yonemoto et al., 1984Go). Rabbits appear to be the most sensitive laboratory animal species to EGME, with teratogenic effects being induced in 63% of fetuses at 50 ppm (Hanley et al., 1984bGo).



View larger version (23K):
[in this window]
[in a new window]
 
FIG. 1. Metabolism of EGME and PGME isomers. ADH, alcohol dehydrogenase; ALDH, aldehyde dehyrogenase; MFO, mixed function oxidase.

 
In contrast, propylene glycol monomethyl ether (PGME) has a very low potential for developmental toxicity, based on studies in rats, mice and rabbits exposed by the oral, subcutaneous, and inhalation routes (Hanley et al., 1984aGo; Stenger et al., 1972Go). In the inhalation developmental toxicity studies, the fetuses of rats and rabbits exposed to 0, 500, 1500, or 3000 ppm PGME exhibited little to no developmental effects, even though the highest exposure concentration induced maternal sedation and decreased body weight gains in both species (Hanley et al., 1984aGo). The most likely explanation for the lower toxicity of PGME relative to that of EGME is isomer-specific metabolism (Fig. 1Go). Commercial PGME is composed mainly of the {alpha}-isomer (1-methoxy-2-propanol), which is metabolized via the microsomal mixed function oxidase (MFO) system through O-dealkylation to propylene glycol (Miller et al., 1983Go). In turn, propylene glycol is either excreted or further oxidized to CO2, thus forming products with minimal toxicity potential. On the other hand, the ß-isomer of PGME (2-methoxy-1-propanol) is oxidized to the alkoxy acid, 2-methoxypropionic acid (MPA; Miller et al., 1986Go), a molecule bearing structural similarity to the known teratogen, MAA. The ß-isomer content of the PGME sample tested in the aforementioned developmental toxicity studies (Hanley et al., 1984aGo) was 1.3%, but current standards limit the ß-isomer content of commercial PGME to less than 0.5%.

Merkle et al.(1987)Go were the first to investigate the developmental toxicity of ß-PGME. In an initial study, they tested the acetate of ß-PGME and observed teratogenic effects in Himalayan rabbits and Wistar rats, with rabbits being considerably more sensitive than rats (Merkle et al., 1987Go). In a follow-up dose-response study of purified ß-PGME (Hellwig et al., 1994Go), rabbits were exposed via inhalation (6 h/day) to 0, 145, 225, 350, or 545 ppm on GD 6–18, followed by maternal necropsy and fetal examination on GD 29. At 350 and 545 ppm, the percentage of resorbed implantations was increased, as was the incidence of fetal malformations and variations. The predominant malformations observed were enlarged sternum and enlarged rib cartilage (described as a bony plate), missing gall bladder, fused ribs, and digit defects. Maternal toxicity was quite apparent at these dose levels, as evidenced by a 72 g weight loss during the dosing period at 545 ppm. At 350 ppm, rabbits gained only 15% of the control body weight gain during the dosing period. Marginal effects occurred at 225 ppm in the form of a slight, nonstatistically significant increase in bony plate/enlarged cartilage in the presence of slight (12%), nonstatistically significant decreases in maternal body weight gain during the dosing period. The lowest exposure concentration of 145 ppm was devoid of treatment-related maternal or developmental effects.

The striking differences in developmental toxicity profile between commercial ({alpha})-PGME and ß-PGME parallel the differences in metabolism of these two isomers, and suggest that MPA is a potential developmentally toxic metabolite of ß-PGME. Therefore, in vitro and in vivo developmental toxicity studies were conducted to thoroughly characterize the developmental toxicity potential of MPA in a highly sensitive test species, the rabbit. Parallel studies with MAA were run as a benchmark compound and positive control. In conjunction with this work, blood time-course studies to describe the formation of MPA from ß-PGME, and the pharmacokinetics of directly administered MPA in rabbits were also conducted. By establishing a critical internal dose of the alleged proximate toxicant, MPA, the risk of human exposure to ß-PGME present as a minor impurity of commercial PGME can then be assessed with greater accuracy and confidence.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Test materials and dosing solutions.
The sample of MPA (CAS 107-98-2) used in this study was custom synthesized by Schweizerhall, Inc. (Piscataway, NJ) and was found to be 98.6% pure by gas chromatography (GC)/mass spectrometry. MAA was obtained from Aldrich Chemical Company (Milwaukee, WI) and was 98.9% pure by GC. Structural confirmation of the MAA and MPA samples was obtained via nuclear magnetic resonance imaging. The sample of ß-PGME was provided by The Dow Chemical Company (Plaquemine, LA) and analyzed at 98.2% ß-PGME, 1.6% {alpha}-PGME by GC coupled with flame ionization detection and infrared spectroscopy. All dosing solutions were prepared in distilled water, and were corrected for purity. The dosing solutions assayed within 97–110% of target and were stable throughout the period of use. The MPA and ß-PGME used in the whole embryo culture studies were provided by The Dow Chemical Company and were determined by GC to be 97.9 and 99.8% pure, respectively.

Test animals.
Adult, virgin female New Zealand White (NZW) rabbits ranging in age from 5.0–6.5 months were obtained from Covance (Denver, PA) and acclimated to the laboratory (AAALAC International accredited) for at least two weeks. The rabbits were housed individually, maintained on 4 (nonpregnant rabbits) or 8 oz/day (bred rabbits) of Purina Certified Laboratory Rabbit Chow No. 5322 and water ad libitum. The animal rooms of the facility were designed to maintain humidity at approximately 40–70%, temperature at approximately 22°C and photoperiod at 12 h light:12 h dark. All treated rabbits were identified using uniquely coded alphanumeric metal ear tags.

In Vivo Developmental Toxicity Studies
Study design.
Groups of 20 artificially inseminated NZW rabbits (Gibson et al., 1966Go) were randomly assigned to treatment groups in either the MAA or MPA developmental toxicity studies (OECD guideline 414). Dose levels for the MAA study were 0, 2.5, 7.5, or 15.0 mg/kg/day, while 0, 10, 26, or 78 mg/kg/day were used for the MPA study. Dose levels were chosen on the basis of previously conducted probe studies. Dose solutions were administered once daily on GD 7–19 via gavage. Body weights were recorded on GD 0, 7–20, and 28 (day of insemination was considered GD 0). The dose volume for all test materials was 2 ml/kg body weight and was adjusted daily based on body weight. Feed consumption was determined daily throughout the study.

Maternal and fetal examinations.
On GD 28, all surviving animals were euthanized by iv injection of Beuthanasia-D (Schering, Kenilworth, NJ) via the marginal ear vein and necropsied. Any gross pathologic changes in the maternal rabbits were recorded, as were maternal liver, kidney, and gravid uterus weights. The uterine horns were exteriorized and the following data were collected: (1) the number and position of fetuses in utero, (2) the number of live and dead fetuses, (3) the number and position of resorptions, (4) the number of corpora lutea, (5) the body weight of each fetus, and (6) any gross external alterations. The uteri of animals that appeared nonpregnant were stained with a 10% aqueous solution of sodium sulfide (Kopf et al., 1964Go) and examined for evidence of early resorptions. All fetuses were immediately examined by dissection under a low power stereomicroscope for evidence of visceral alterations, determination of sex, and changes in brain structure as assessed by serial sectioning (Staples, 1974Go). All fetuses were then preserved in alcohol, eviscerated, stained with alizarin red-S (Dawson, 1926Go), and examined for skeletal alterations. In judging the toxicological significance of fetal alterations, the following criteria were considered: (1) statistical significance, and/or (2) historical control values, (3) dose-dependence, (4) occurrence in more than one fetus, and (5) consistency with previously reported developmental effects of EGME (Hanley et al., 1984bGo).

Rabbit whole embryo culture studies.
In order to more fully characterize the intrinsic teratogenicity potential of MPA and its parent compound, ß-PGME, rabbit whole embryo culture studies were performed using early somite stage embryos cultured for 48 h according to methods described by Ninomiya et al.(1993)Go and Pitt and Carney (1999a)Go. In brief, untreated, pregnant rabbits were euthanized on the morning of GD 9 via carbon dioxide inhalation, exsanguinated, and the uterine wall was cut open to expose the conceptuses. Healthy, normal appearing conceptuses (approximately 12–14 somites) with a visible peripheral visceral yolk sac (VYS) blood vessel, and fused cardiac tubes were considered acceptable for culture. A cut was made around the entire conceptus, immediately parallel and distal to the peripheral VYS vessel in order to isolate the embryo, placenta, and VYS apart from the maternal implantation site. Underlying placental tissue was then trimmed so that only a thin layer remained attached to the VYS and embryo. Each conceptus was then transferred to a culture bottle containing 5 ml of culture medium, and embryos were balanced by litter across treatment groups. The culture medium was composed of 75% (v/v) heat-inactivated rabbit serum, and 25% (v/v) phosphate-buffered saline (PBS), supplemented with penicillin-streptomycin solution (Life Technologies, Gaithersburg, MD). Test materials were dissolved in the PBS portion of the media prior to mixing with serum. Test material concentrations were 0, 1.0, or 5.0 mM of MPA or 5.0 mM MAA in the first embryo culture experiment, and 0, 0.5, or 2.0 mM of ß-PGME in the second experiment. The highest concentration of each compound was based on results of the pharmacokinetic study (see below), and slightly exceeded the peak concentration of each compound following dosing with a high dose (270 mg/kg) of ß-PGME. All media were equilibrated to culture conditions for at least 30 min prior to the start of culture. The culture bottles were placed in a rotating culture bottle unit (BTC Engineering, Cambridge, UK) under continuous gassing with 5% CO2 in air for the first 24 h, and 5% CO2/95% O2 for the remaining 24 h. The gas mixture flow rate (approx. 20 cc/min) and temperature (37°C) remained constant throughout the culture.

After approximately 48 h in culture, embryos were evaluated for viability (based on heart beat), presence of blood flowing through the yolk sac, head length, somite number, morphology score, and incidence of malformations. The morphology score (Pitt and Carney, 1999aGo,bGo) is an index of development in which individual embryonic structures are assigned numerical scores based on their appearance in relation to defined expectations for specific developmental ages. For example, an embryo with two branchial bars would be given a score of 10 for that structure, based on the fact that normal rabbit embryos are expected to have two branchial bars on GD 10. Individual scores for the various embryonic structures were averaged to derive an overall morphological score for each embryo. This approach allowed one to quantify generalized delays in embryo development, as well as helping to identify specific delays in certain organs or tissues. Embryo malformations were defined as specific deviations from the normal pattern of development, clearly distinguishable from generalized delays in development.

Pharmacokinetics study design.
Four groups of three virgin, nonpregnant female rabbits were fitted with auricular artery catheters (Heim, 1989Go) to allow for serial blood sampling. Feed was removed from the rabbits in the afternoon on the day before dosing, and was returned 4 h postdosing. Target dose levels were 26 mg (0.25 mmol)/kg and 78 mg (0.75 mmol)/kg of MPA, and 67.5 mg (0.75 mmol)/kg and 270 mg (3.0 mmol)/kg of ß-PGME. All doses were administered as aqueous solutions via gavage. Blood samples were collected at 0 (immediately prior to dosing), 0.25, 0.5, 0.75, 1, 1.5, 2, 2.5, 3, 4, 6, 8, 12, 24, 36, 48, 72, 96, 120, 144, and 168 h postdosing. Blood samples ({approx} 0.5 ml) were drawn at each time point, and approximately 0.5 ml of heparinized saline was flushed through the catheter to clear the blood from the catheter and provide a "heparin lock." In cases of catheter failure, samples were collected from the marginal ear vein. Following collection of the final blood sample, the animals were euthanized by iv injection of Beuthanasia-D (Schering, Kenilworth, NJ).

Sample analysis.
All blood samples were analyzed for both MPA and ß-PGME using a double derivatization followed by GC/negative chemical ionization/mass spectroscopy (NCI/MS). Weighed aliquots of rabbit blood were either immediately derivatized or frozen until derivatized. For same-day derivatizations, samples of blood were transferred to an HPLC vial containing a weighed amount of ion-pairing agent (IPA) and mixed well. The IPA was a 0.3 N NaOH solution containing 0.1 M tetrabutylammonium hydrogen sulfate and D3-labeled internal standards of MPA and ß-PGME (0.3–0.7 µg/ml). The samples were derivatized to the pentafluorobenzoyl and pentafluorobenzyl esters by the addition of 4 µl of pentafluorobenzoyl chloride and 60 µl of pentafluorobenzyl bromide, respectively, with heated (46°C) vortexing for 20 min. The esters were then extracted with 1 ml of toluene during another 20-min heated vortex cycle. After centrifuging the samples at 3000 rpm for 10 min, the toluene extracts were transferred to GC vials and analyzed by GC/NCI/MS. Some samples of blood were transferred to HPLC vials containing 10 µl of heparin, mixed well and frozen at –80°C. Prior to derivatization, these samples were removed from the freezer, thawed, and mixed well with the IPA solution. The samples were derivatized using the same procedure described above. Blood samples were shown to be stable after 7 days storage at –80°C (data not shown). Detection limits were 50 ng/g blood for both MPA and ß-PGME.

Pharmacokinetic parameters were calculated using a noncompartmental analysis conducted with the aid of an Excel®-based computer software package (PK Solutions®, 1997, V2.0.2; Summit Research Services, Ashland, OH). Area under the curve (AUC) for MPA and ß-PGME blood concentration-time curves was estimated by PK Solutions® based on the trapezoidal rule as described in Gibaldi and Perrier (1982)Go:


1

where t2 and t1 are consecutive time points and Cp1 and Cp2 are blood concentrations of analyte at the respective time points. AUCs over all the time points were summed to give a total AUC for a particular blood time course.

Statistics.
Maternal body weights, body weight gains, organ weights, maternal reproductive parameters, mean fetal body weight/litter, and continuous measures of in vitro embryo development were first evaluated by Bartlett’s test (p = 0.01) for equality of variances (Winer, 1971Go). Based upon the outcome of Bartlett’s test, either a parametric (Steel and Torrie, 1960Go) or nonparametric (Hollander and Wolfe, 1973Go) ANOVA was performed (p = 0.10). If the ANOVA was significant, a two-sided Dunnett’s test or the Wilcoxon Rank-Sum test with Bonferroni’s correction (Miller, 1966Go) was performed, respectively (p = 0.05). The number of corpora lutea, implantations, litter size, and embryo morphology score were evaluated using the nonparametric ANOVA followed by the Wilcoxon Rank-Sum test with Bonferroni’s correction. Pregnancy rates and percentage data for whole embryo culture were analyzed by the Fisher exact probability test (p = 0.05) and Bonferroni’s correction was used for multiple testing of groups in comparison to a single control (Siegel, 1956Go). Evaluation of the fetal sex ratio was performed by the binomial distribution test (Steel and Torrie, 1960Go). The frequency of preimplantation loss, resorptions, and the incidence of fetal alterations were analyzed by the Wilcoxon test (p = 0.05) as modified by Haseman and Hoel (1974)Go. Nonpregnant females, females pregnant following staining, or females having totally resorbed litters were excluded from the appropriate analyses. Statistical outliers were identified using a sequential method (Grubbs, 1969Go), but only values for feed consumption were routinely excluded unless justified by sound scientific reasons unrelated to treatment.

Good laboratory practices.
All studies and resulting data were audited and found to be compliant with appropriate Good Laboratory Practice (GLP) standards.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
In Vivo Developmental Toxicity Studies
MPA.
In the high-dose group (78 mg/kg/day), maternal toxicity was evidenced by clinical observations of reduced fecal output, corresponding with decreases in feed consumption (47–73% of control consumption; data not shown), decreased body weights, and body weight gains during the dosing period (Table 1Go). Absolute and relative kidney weights in this group were significantly increased compared to controls. Also, three high-dose rabbits aborted their litters prior to the scheduled necropsy. Marked inanition, deceased fecal output, and body weight losses were observed for all of these rabbits in the preceding days. At gross examination, two of these rabbits had indications of low-grade pneumonia, while the other rabbit exhibited nonspecific signs. An additional high-dose rabbit died on GD 15 as a result of gavage error. In the middle-dose group (26 mg/kg/day), maternal toxicity was also apparent, albeit of lesser severity. Signs of maternal toxicity in the middle-dose group consisted of decreased fecal output, slightly decreased feed consumption (84–95% of controls; data not shown), and a slight, but statistically significant decrease in body weight gain after the first few days of dosing (GD 7–10). There was no evidence of maternal toxicity in the low-dose group.


View this table:
[in this window]
[in a new window]
 
TABLE 1 MPA: Maternal Toxicity and Litter Data
 
An increase in resorption rate was statistically identified in the high-dose group (Table 1Go), but no such effect occurred at 10 or 26 mg/kg/day. A decrease in fetal body weights (approximately 8%) was seen at the high-dose level only, and while not statistically significant, was considered to be treatment-related. There were no effects at any dose levels on pregnancy rate, number of corpora lutea, implantations or viable fetuses per litter, percent preimplantation loss, fetal sex ratio, or gravid uterine weights.

Table 2Go lists all of the fetal malformations observed in the MPA study, as well as those variations that exhibited a treatment-related increase. Overall, there were 0, 3, 1, and 6 malformed fetuses in the 0, 10, 26, and 78 mg/kg/day MPA groups, respectively, with the incidence of total malformed fetuses in the high-dose group being statistically increased relative to controls. The only specific malformation observed in more than one high-dose fetus was fused ribs, which was seen in three fetuses from three different litters. While not statistically significant, the incidence of fused ribs (4.8% of fetuses, 27.3% of litters) was outside the range of historical control data (maximum of 2.4% of fetuses, 15.4% of litters). Other malformations among the high-dose group were limited to single occurrences and included short tail, spina bifida, ectopic gall bladder, hemivertebrae, scoliosis, and oligodactyly (missing digit I). There were also increased incidences of several fetal variations noted in the high-dose MPA group, including retrocaval ureter, irregular pattern of sternebrae ossification, and delayed ossification of the hyoid, talus, and forelimb phalanges. The incidence of paraovarian cysts (6.3% of fetuses, 27.3% of litters) also was higher than that of contemporary controls, and while not statistically identified, was outside the range of historical controls (maximum of 1.2% of fetuses, 10% of litters) and thus was considered an effect of treatment. There were no treatment-related changes in the incidence of fetal malformations or variations in the middle- and low-dose groups. The few malformations observed showed no dose-response and all incidences were within historical control ranges.


View this table:
[in this window]
[in a new window]
 
TABLE 2 MPA: Fetal Malformations/Variations
 
MAA.
In the MAA study, treatment-related maternal toxicity was restricted to the high-dose group (15 mg/kg/day), and consisted of decreases in fecal output, soft feces, decreased feed consumption, decreased body weight gain (with net body weight losses over the course of treatment, followed by recovery posttreatment) and increased relative liver weight (Table 3Go). There were no maternal deaths in the study, but two high-dose dams aborted their litters prior to the scheduled necropsy. One of these high-dose rabbits, as well as one control rabbit aborting on GD 24, had only one implantation site and no gross pathological signs, suggesting that the number of implantations may have been insufficient to maintain pregnancy (Feussner et al., 1992Go). The other high-dose group abortion was preceded by marked inanition and weight loss, but again, no lesions were grossly visible. In the middle-dose group (7.5 mg/kg/day), the only maternal findings were a few incidences of decreased and/or soft feces and slight decreases in feed consumption (data not shown), but these were not accompanied by any effects on body weight or body weight gain and thus were not considered toxicologically significant. There was no evidence of maternal toxicity at the low dose of 2.5 mg/kg/day.


View this table:
[in this window]
[in a new window]
 
TABLE 3 MAA: Maternal Toxicity and Litter Data
 
A statistically identified increase in resorption rate was shown for the high-dose MAA group, which was also reflected in a decreased number of viable fetuses per litter (Table 3Go). Accompanying these high-dose group effects were statistically identified decreases in fetal body weight, and decreased gravid uterus weights. Fetal body weights were also significantly decreased in the middle-dose group, but there were no effects on resorption rate, litter size, or gravid uterus weight. There were no significant changes in any of the other reproductive parameters in the middle- and high-dose groups, and there were no reproductive effects whatsoever in the low-dose MAA group.

In contrast to the MPA study, there was a greater incidence and broader spectrum of fetal malformations and variations with MAA, despite the lower dose levels (Table 4Go). There were 1, 2, 9, and 32 malformed pups in the 0, 2.5, 7.5, and 15.0 mg/kg/day MAA groups, respectively, with the overall incidence of malformed fetuses statistically increased at 7.5 and 15 mg/kg/day. In particular, statistically significant increases in the incidence of three external malformations affecting the distal limb, namely forelimb flexure, oligodactyly, and anonychia (missing toenails), occurred in the high-dose MAA fetuses. In addition, a number of other external malformations (brachydactyly, limb rotation, hindlimb flexure, digit flexure) and variations (kinky tail) appeared to occur at increased incidence. Although these incidences were not statistically different from controls, they were interpreted to be treatment-related based on consideration of historical control ranges, dose-dependency, and/or consistency with the effects of EGME (Hanley et al., 1984bGo). In the middle-dose group, slightly increased incidences of forelimb flexure and kinky tail were also observed. There were no treatment-related external alterations among low-dose fetuses, with the incidence of externally malformed fetuses in this group (2/106) similar to that of control fetuses (1/119).


View this table:
[in this window]
[in a new window]
 
TABLE 4 MAA: Fetal Malformations/Variations
 
The only visceral malformation that was considered to be treatment-related was diaphragmatic hernia, the incidence of which was increased in the high-dose MAA group. However, several visceral variations occurred at increased incidence in both the middle- and high-dose groups. Among high-dose fetuses, increased incidences of retrocaval ureter, paraovarian cysts, and paratesticular cysts were statistically identified, while pale spleen and small spleen were considered to be effects of treatment based on the aforementioned criteria. In the middle-dose group, retrocaval ureter was the only statistically identified effect, with paraovarian and paratesticular cysts being considered elevated in a treatment-related manner as well. No visceral alterations attributable to treatment were found among the low-dose fetuses.

Increased incidences of several skeletal alterations were evident in the 7.5 and 15 mg/kg/day groups. In the 15 mg/kg/day dose group, an increased incidence of extra ribs (not statistically identified) was the only skeletal malformation considered to be treatment-related. Other skeletal malformations in this group were observed only in single fetuses, thus their relationship to treatment was difficult to ascertain. Treatment-related skeletal variations were numerous in this group and included delayed ossifications of the pubis, forelimb phalanges, hindlimb phalanges, talus, metacarpals, sternebrae, hyoid and centra, irregular pattern of sternebrae ossification, and fused sternebrae. In the middle-dose group, there were no treatment-related skeletal malformations, but several treatment-related variations, such as delayed ossification of the forelimb phalanges, talus, sternebrae, hyoid and pubis, along with fused sternebrae and irregular pattern of sternebrae ossification. No skeletal alterations attributable to treatment were found among the low-dose MAA fetuses.

Rabbit Whole Embryo Culture
MPA.
All measures of embryo development in both MPA dose groups were comparable to control values. This included the percentage of viable embryos, percentage of embryos with an active yolk sac circulation, head length, somite number, morphological score, and percentage of malformed embryos (Table 5Go). One embryo in the control group had an abnormally small head. In the 1 mM MPA group, one embryo had a misshapen head and also an enlarged pericardium, while another embryo from the same group had a blister over the hindbrain. Low incidences of such malformations in rabbit whole embryo culture are not unusual findings. In particular, small to moderately sized blisters over the head and neural tube have been seen in up to 25% of control rabbit embryos in culture (Pitt and Carney, 1999bGo), and thus, are unlikely to be due to treatment. There were no malformed embryos in the 5 mM MPA group.


View this table:
[in this window]
[in a new window]
 
TABLE 5 Comparison of MPA and MAA in Rabbit Whole Embryo Culture
 
MAA.
In contrast to the lack of effects with MPA, the positive control, MAA, induced severe effects on all measures of embryo development in culture. Only 73.3% of the embryos were viable and only one of them had an active yolk sac circulation, suggesting that even the survivors were in poor condition. Growth retardation was evidenced by a significant decrease in the number of somites. All embryos in the MAA group exhibited malformations, consistent with previous reports. Among the most common of these included malformations of the head and brain (swollen heads, lack of brain/facial differentiation), and misshapen optic and otic vesicles.

ß-PGME.
There were no differences between the ß-PGME groups and controls for any parameter of embryo development (Table 6Go). One embryo in the control group had an open neural tube in the region of somites 15–18, and also had a slightly irregular neural tube in the pontine flexure region of the hindbrain. Another control embryo exhibited a blister over the pontine flexure. In the 0.5 mM ß-PGME group, one embryo had an abnormally small head. In the 2.0 mM ß-PGME group, one embryo exhibited an irregular, zig-zag neural tube, while another had a blister over the telecephalon. A third embryo in this group was grossly malformed, with an abnormally shaped, small head, open neural tube, and missing pericardium.


View this table:
[in this window]
[in a new window]
 
TABLE 6 Evaluation of ß-PGME in Rabbit Whole Embryo Culture
 
Pharmacokinetics Study
Blood concentration-time courses (Fig. 2AGo) and mean pharmacokinetic values (Table 7Go) indicated a very rapid absorption following oral administration of ß-PGME (Tmax < 1 h), as well as rapid metabolic conversion of absorbed ß-PGME to MPA (Fig. 2BGo). ß-PGME levels in blood rapidly peaked and then declined to undetectable levels (LOD = 50 ng/g blood) by 4–8 h postdosing. Blood MPA levels following ß-PGME dosing peaked at approximately 2.5 h postdosing, but then decreased slowly and were still detectable in blood 168 h postdosing.



View larger version (12K):
[in this window]
[in a new window]
 
FIG. 2. (A) Blood ß-PGME concentration time-courses in nonpregnant female rabbits following oral gavage administration of either 67.5 or 270 mg/kg of ß-PGME. (B) Blood MPA concentration time-courses in nonpregnant female rabbits following oral gavage administration of 67.5 or 270 mg/kg of ß-PGME, or 26 or 78 mg/kg/day of MPA; n = 3 rabbits/treatment group.

 

View this table:
[in this window]
[in a new window]
 
TABLE 7 Pharmacokinetic Parameters of Blood ß-PGME and MPA for Female Rabbits following Oral Administration
 
Dosing directly with 78 mg/kg of MPA (0.75 mM/kg) resulted in an MPA blood concentration-time course that was essentially identical to that observed for 67.5 mg/kg of ß-PGME (also 0.75 mM/kg; Fig. 2BGo). Peak blood levels of MPA were reached within 2–4 h after dosing with MPA, indicating rapid absorption following oral administration. Again, blood levels of MPA declined slowly, and were still detectable 144 h postdosing for both MPA dose groups. Comparison of peak and AUC values across dose levels demonstrated that the pharmacokinetic values were proportionate to administered dose.


    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Although commercial PGME was negative for developmental toxicity when tested in three different species and by three different routes of exposure (Hanley et al., 1984aGo; Stenger et al., 1972Go), there has been a tendency to regard all glycol ethers, including PGME, as toxicants of heightened concern. Apparently this phenomenon is based on the toxicity of certain ethylene glycol ethers, such as EGME, that are well known for their developmental and reproductive effects. However, structure-toxicity relationships, metabolism studies, and safety evaluations all clearly demonstrate that individual glycol ethers do not necessarily share similar toxicity profiles. For developmental toxicity, the key determinant is the presence of a primary alcohol moiety that is subsequently metabolized to an alkoxy acid. As such, PGME has been the subject of additional scrutiny because commercial preparations contain some ß-isomer, a primary alcohol that is metabolized to the alkoxy acid, MPA. With a view toward increasing certainty for human risk assessment, experiments were conducted to (1) characterize the intrinsic developmental toxicity potential of MPA, (2) estimate the amounts of MPA that might be formed from commercial PGME, and (3) establish as benchmarks the MPA blood levels associated with developmentally toxic and nontoxic exposures to MPA.

The present in vivo developmental toxicity assessments conducted in NZW rabbits, the species most sensitive to developmental effects of ethylene glycol ethers, showed that MPA caused developmental effects only at maternally toxic dose levels. In fact, the developmental toxicity NOEL of 26 mg/kg/day was higher than that for maternal toxicity (NOEL of 10 mg/kg/day). In contrast, MAA was selectively toxic to the fetus, with a developmental NOEL of 2.5 mg/kg/day, but a maternal toxicity NOEL 7.5 mg/kg/day. These data suggest that, even in the most sensitive species, MPA is about 10-fold less potent than MAA for causing developmental toxicity.

The spectrum of fetal anomalies induced by the high-dose of MPA was more restricted than that of MAA. While both compounds induced fetal resorptions and caused increases in the incidence of retrocaval ureter, paraovarian cysts, irregular pattern of sternebrae ossification, and delayed ossification of the distal limb bones and hyoid (considered fetal variations), MAA induced a host of additional fetal effects that were not increased by MPA (Table 4Go). These included limb flexures and rotations, digit malformations, diaphragmatic hernia, pale/small spleen, extra ribs, paratesticular cyst, delayed ossification of the skull, sternebrae, pubis and centra, and fused sternebrae. Thus, MPA is not only a weaker developmental toxicant with respect to dose, but is also less potent with respect to the severity and range of effects it induces.

The in vivo studies mentioned above were complemented by rabbit whole embryo culture experiments with MPA, ß-PGME, and MAA in order to compare the intrinsic potency of these materials. Rabbit whole embryo cultures are conducted at a stage of very active development (GD 9–11), and one that is thought to be highly sensitive to teratogenic insult. Also, the system allows one to precisely control the level of test chemical to which embryos are exposed without the confounding factor of maternal metabolism, elimination, etc. For these reasons, rabbit embryo culture was considered an ideal approach for comparing the intrinsic teratogenic potential of these compounds.

The effects seen in rabbit whole embryo culture were largely consistent with the in vivo studies, in the sense that MAA was strongly teratogenic in vitro, whereas equivalent concentrations of MPA were without effect. This clearly indicated that the intrinsic developmental toxicity potential of MPA is much lower than that of MAA. Rabbit whole embryo culture evaluation of ß-PGME also indicated a lack of treatment-related effects. This finding is consistent with other studies suggesting that developmental toxicity of ethylene glycol ethers is due to their alkoxy acid metabolites, rather than a direct action of the parent compound (Rawlings et al., 1985Go; Yonemoto et al., 1984Go).

In conjunction with in vivo developmental toxicity and rabbit whole embryo culture studies, we also investigated the kinetics of MPA formed from ß-PGME, as well as the kinetics of MPA administered directly, in female (nonpregnant) rabbits. Both compounds were readily absorbed by the oral route (Figs. 2AGo and 2BGo), with peak blood levels reached within 1–4 h. Levels of MPA formed from ß-PGME peaked at 2.5 h, suggesting rapid conversion of ß-PGME to MPA. Increases in blood ß-PGME and MPA were proportional to dose administered, suggesting linear kinetics, at least in the dose-range examined. Comparison of MPA blood levels following the 67.5 mg/kg ß-PGME dose and 78 mg/kg MPA dose (equimolar doses) demonstrated their pharmacokinetic equivalence, as evidenced by the fact that the semilogarithmic plots of the blood concentration-time courses were virtually superimposable (Fig. 2BGo). The latter observation suggests that essentially 100% of the ß-PGME found in blood is converted to MPA.

Whereas parent ß-PGME was cleared rapidly from blood, MPA blood levels declined very slowly, and were still readily detectable 144 h postdosing. This long apparent t1/2 of elimination may indicate plasma binding of MPA and/or resorption of MPA by the kidneys resulting in recycling of MPA, both of which would result in slow elimination from blood. Although a long plasma t1/2 could also be caused by saturation of renal elimination of MPA, one would expect that to occur only at the early part of the time-course, and to change the slope of the high-dose elimination curve compared with the low-dose curve. The overall slow elimination from plasma could also be due to a combination of these different mechanisms. The long elimination half-life for MPA in rabbits is in direct contrast to the imputed value for rats, based on rapid elimination of radioactivity following administration of 14C-ß-PGME, where 90% of the administered dose was eliminated within 48 h (Miller et al., 1986Go). These pharmacokinetic data suggest that overall systemic exposure to MPA is much greater in rabbits than in rats exposed to equivalent doses of ß-PGME. Thus, pharmacokinetic differences might explain the greater sensitivity of rabbits relative to rats following inhalation exposure to ß-PGME (Hellwig et al., 1994Go). Interestingly, MAA appears to be eliminated slowly in humans (t1/2 = 77 h; Groeseneken et al., 1989Go), supporting the use of the rabbit as a preferred animal model for this compound.

The combined weight of evidence from the developmental toxicity and pharmacokinetic studies reported herein offer valuable insight regarding the potential risks of developmental effects posed by the ß-isomer contained in commercial PGME. Whether one makes estimates based on delivered doses or internal dose metrics, it is clear that extremely high exposures to commercial PGME would be required to adversely affect development. For example, knowing that the LOEL for MPA of 78 mg/kg/day is pharmacokinetically equivalent to 67.5 mg/kg/day ß-PGME, and assuming that commercial PGME contains 0.5% ß-isomer, then developmental toxicity would require a delivered dose of 13,500 mg/kg/day (67.5/0.005). Making a similar analysis based on the pharmacokinetic values obtained in this study, it appears that developmental toxicity would require MPA blood levels of at least 0.5–1.3 mM (Cmax) or 19.9–52.9 mM-h/l (AUC), levels which generally are considered quite high. Based on human exposure data (Anundi et al., 2000Go; Dentan et al., 2000Go; Devanthery et al., 2000Go; Hubner et al., 1992Go; Jones et al., 1997Go; Laitinen, 1997Go) and occupational limits set for PGME (ACGIH threshold limit value =100 ppm, STEL = 150 ppm), it is highly unlikely that human exposure to ß-PGME (or MPA) would ever approach such developmentally toxic levels.

In summary, our data on the developmental toxicity and pharmacokinetics of MPA provide a mechanistic foundation to address concerns about the presence of ß-isomer in commercial PGME. The combined weight of evidence consists of negative developmental toxicity studies with the commercial product in several species and routes of exposure, in vivo and in vitro developmental studies showing that MPA’s intrinsic toxicity is much lower than that of MAA, and pharmacokinetic data indicating a high threshold based on internal MPA dose. These data collectively increase confidence in the conclusion that MPA formed from small amounts of ß-isomer in commercial PGME represents a negligible risk to the developing fetus.


    ACKNOWLEDGMENTS
 
The contributions of U. Vedula, C. Zablotny, E. Decker, F. Tiede, G. Zielke, A. Liberacki, M. Vennix, J. Hammond, A. Mendrala, J. Albe, S. Lick-Day, and M. Bartels to the conduct of this study are gratefully acknowledged. We also thank P. Spencer, S. Marty, J. Wilmer, and W. C. Hayes for critical review of the manuscript and J. Bus for conceptual contributions to the project.


    NOTES
 
1 To whom correspondence should be addressed. Fax: (517) 638-9863. E-mail: ecarney{at}dow.com. Reprint requests: Chemical Hazard Evaluation & Communication, 1803 Building, The Dow Chemical Company, Midland, MI 48674. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Anundi, H., Langworth, S., Johanson, G., Lind, M.-L., Akesson, B., Friis, L., Itkes, N., Soderman, E., Johsson, B. A. G., and Edling, C. (2000). Air and biological monitoring of solvent exposure during graffiti removal. Int. Arch. Occup. Environ. Health 73, 561–569.[ISI][Medline]

Boatman, R. J., and Knaak, J. B. (2001). Ethers of ethylene glycol and derivatives. In Patty’s Toxicology, 5th ed. Vol. 7, pp. 73–270. John Wiley and Sons, New York.

Brown, N. A., Holt, D., and Webb, M. (1984). The teratogenicity of methoxyacetic acid in the rat. Toxicol. Lett. 22, 93–100.[ISI][Medline]

Dawson, A. B. (1926). A note on the staining of the skeletons of cleared specimens with alizarin red-S. Stain Technol. 1, 123–124.

Dentan, A., Devanthery, A., de Peyer, J. E., and Droz, P. O. (2000). Propylene glycol monomethyl ether (PGME) occupational exposure. 1. Biomonitoring by analysis of PGME in urine. Int. Arch. Occup. Environ. Health 73, 311–315.[ISI][Medline]

Devanthery A., Dentan, A., Berode, M., and Droz, P. O. (2000). Propylene glycol monomethyl ether (PGME) exposure. 2. Identification of products containing PGME, their importance and their use in Switzerland. Int. Arch. Occup. Environ. Health 73, 349–351.[ISI][Medline]

Feussner, E. L., Lightkep, G. E., Hennesy, R. A., Hoberman, A. M. and Christian, M. S. (1992). A decade of rabbit fertility data: Study of historical control animals. Teratology 46, 349–365.[ISI][Medline]

Groeseneken, D., Veulemans, H., Masschelein, R., and Van Vlem, E. (1989). Experimental human exposure to ethylene glycol monomethyl ether. Int. Arch. Occup. Environ. Health 61, 243–247.[ISI][Medline]

Gibaldi, M., and Perrier, D. (1982). Pharmacokinetics. Marcel Dekker, New York.

Gibson, J. P., Staples, R. E., and Newberne, J. W. (1966). Use of the rabbit in teratogenicity studies. Toxicol. Appl. Pharmacol. 9, 398–408.[ISI]

Grubbs, F. E. (1969). Procedures for detecting outlying observations in samples. Technometrics 11, 1–21.[ISI]

Hanley, T. R., Calhoun, L. L., Yano, B. L., and Rao, K. S. (1984a). Teratogenic evaluation of inhaled propylene glycol monomethyl ether in rats and rabbits. Fundam. Appl. Toxicol. 4, 784–794.[ISI][Medline]

Hanley, T. R., Yano, B. L., Nitschke, K. D., and John, J. A. (1984b). Comparison of the teratogenic potential of inhaled ethylene glycol monomethyl ether in rats, mice, and rabbits. Toxicol. Appl. Pharmacol. 75, 409–422.[ISI][Medline]

Hardin, B. D. (1983). Reproductive toxicity of the glycol ethers. Toxicology 27, 91–102.[ISI][Medline]

Haseman, J. K., and Hoel, D. G. (1974). Tables of Gehan’s generalized Wilcoxon test with fixed point censoring. J. Statist. Comput. Simul. 3, 117–135.

Heim, R. A. (1989). A practical technique for obtaining multiple blood samples from rabbits. Lab. Anim. Jan./Feb., 32–33.

Hellwig, J., Klimisch, H.-J., and Jackh, R. (1994). Prenatal toxicity of inhalation exposure to 2-methoxypropanol-1 in rabbits. Fundam. Appl. Toxicol. 23, 608–613.[ISI][Medline]

Hevelin, W. (1989). Reproductive and Hematopoietic Toxicity of the Glycol Ethers: An Update. Hazard Evaluation System and Information Service, Department of Health Services/Department of Industrial Relations. State of California.

Hollander, M., and Wolfe, D.A. (1973). Nonparametric Statistical Methods. John Wiley and Sons, New York.

Horton, V. L., Sleet, R. B., John-Greene, J. A., and Welsch, F. (1985). Developmental phase-specific and dose-related teratogenic effects of ethylene glycol monomethyl ether in CD-1 mice. Toxicol. Appl. Pharmacol. 80, 108–118.[ISI][Medline]

Hubner, B., Lehnert, G., Schaller, K. H., Welte, D., and Angerer, J. (1992). Chronic occupational exposure to organic solvents. XV. Glycol ether exposure during the manufacture of brakehoses. Int. Arch. Occup. Environ. Health 64, 261–264.[ISI][Medline]

Jones, K., Dyne, D., Cocker, J. and Wilson, H. K (1997). A biological monitoring study of 1-methoxy-2-propanol: Analytical method development and a human volunteer study. Sci. Total Environ. 199, 23–30.[ISI][Medline]

Kopf, R., Lorenz, D., and Salewski, E. (1964). The effect of thalidomide on the fertility of rats: In an examination of two generations. Naunyn-Schmiedebergs Arch. Exp. Pathol. Pharmakol. 247, 121–135.[ISI]

Laitinen, J. (1997). Biomonitoring of technical grade 1-alkoxy-2-propanol acetates by analysing urinary 2-alkoxypropionic acids. Sci. Total Environ. 199, 31–39.[ISI][Medline]

Merkle, J., Klimisch, H.-J., and Jackh, R. (1987). Prenatal toxicity of 2-methoxypropylacetate-1 in rats and rabbits. Fundam. Appl. Toxicol. 8, 71–79.[ISI][Medline]

Miller, R. G., Jr., (1966). Simultaneous Statistical Inference. McGraw-Hill, New York.

Miller, R. R., Hermann, E. A., Langvardt, P. W., McKenna, M. J. and Schwetz, B. A. (1983). Comparative metabolism and disposition of ethylene glycol monomethyl ether and propylene glycol monomethyl ether in male rats. Toxicol. Appl. Pharmacol. 67, 229–237.[ISI][Medline]

Miller, R. R., Langvardt, P. W., Calhoun, L. L. and Yahrmarkt, M. A. (1986). Metabolism and disposition of propylene glycol monomethyl ether (PGME) ß-isomer in male rats. Toxicol. Appl. Pharmacol. 83, 170–177.[ISI][Medline]

Nelson, B. K., Setzer J. V., Brightwell, W. S., Mathinos, P. R., Kuczuk, M. H., Weaver, T. E., and Goad, P. T. (1984). Comparative inhalation teratogenicity of four glycol ether solvents and an amino derivative in rats. Environ. Health Perspect. 57, 261–271.[ISI][Medline]

Ninomiya, H., Kishida, K., Ohno, Y., Tsurumi, K., and Eto, K. (1993). Effects of trypan blue on rat and rabbit embryos cultured in vitro. Toxicol. in Vitro 7, 707–717.[ISI]

Pitt, J. A., and Carney, E. W. (1999a). Development of a gestational age-based morphology evaluation system for in vivo and in vitro rabbit embryos. Teratology 59, 88–101.[ISI][Medline]

Pitt, J. A., and Carney, E. W. (1999b). Evaluation of various toxicants in rabbit whole-embryo culture using a new morphologically-based evaluation system. Teratology 59, 102–109.[ISI][Medline]

Rawlings, S. J., Shuker, D. E., Webb, M., and Brown, N. A. (1985). The teratogenic potential of alkoxy acids in post-implantation rat embryo culture: Structure-activity relationships. Toxicol. Lett. 28, 49–58.[ISI][Medline]

Siegel, S. (1956). Non-Parametric Statistics for the Behavioral Sciences. McGraw-Hill, New York.

Sleet, R. B., John-Greene, J. A., and Welsch, F. (1988). The relationship ofembryotoxicity to disposition of 2-methoxyethanol. Toxicol. Appl. Pharmacol. 93, 195–207.[ISI][Medline]

Staples, R. E. (1974). Detection of visceral alterations in mammalian fetuses. Teratology 9, 37 (Abstract).

Steel, R. G. D., and Torrie, J. H. (1960). Principles and Procedures of Statistics. McGraw-Hill, New York, pp. 101–112.

Stenger, E. G., Aeppli, L., Maachemer, L., Muller, D., and Trokan, J. (1972). The toxicology of propylene glycol monomethyl ether. Arzneim.-Forsch. 22, 569–574.[Medline]

Toraason, M., Stringer, B., Stober, P., and Hardin, B. D. (1985). Electrocardiographic study of rat fetuses exposed to ethylene glycol monomethyl ether (EGME). Teratology 32, 33–39.[ISI][Medline]

Winer, B. J. (1971). Statistical Principles in Experimental Design, 2nd ed. McGraw-Hill, New York.

Yonemoto, J., Brown, N. A., and Webb, M. (1984). Effects of dimethoxyethyl phthalate, monomethoxyethyl phthalate, 2-methoxyethanol, and methoxyacetic acid post implantation on rat embryos in culture. Toxicol. Lett. 21, 97–102.[ISI][Medline]





This Article
Abstract
FREE Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (4)
Disclaimer
Request Permissions
Google Scholar
Articles by Carney, E. W.
Articles by Breslin, W. J.
PubMed
PubMed Citation
Articles by Carney, E. W.
Articles by Breslin, W. J.