| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Breeding Centre for Endangered Arabian Wildlife,3 Sharjah, United Arab Emirates
Department of Veterinary Medicine,4 University of Cambridge, Cambridge CB3 0ES, United Kingdom
Center for Conservation and Research of Endangered Wildlife,5 Cincinnati Zoo & Botanical Garden, Cincinnati, Ohio 45220
Nelson Institute of Environmental Medicine,6 NYU School of Medicine, Tuxedo, New York 10987
ABSTRACT
Many nondomestic felids are highly endangered, and captive breeding programs have become essential components of holistic conservation efforts for these species. The ability to diagnose pregnancy early in gestation is fundamental to developing effective breeding programs. The purpose of this study was to develop a radioimmunoassay (RIA) for the detection of urinary relaxin in felids and assess its applicability for early, noninvasive pregnancy diagnosis in domestic cats (Felis silvestris catus) and leopards (Panthera pardus). Urine was collected from pregnant and nonpregnant domestic cats and leopards at mating, and then weekly thereafter for the duration of gestation. Paired serum samples were also collected from the domestic cats. A RIA for relaxin that uses an antiserum against synthetic canine relaxin was validated for felid urine and shown to detect relaxin immunoreactivity in pregnant cat urine subjected to acid-acetone extraction. In the cat, urinary relaxin was first detected between Days 21 and 28 of gestation; levels peaked at 42–49 days, and the concentrations then declined over 2 wk prior to parturition. The urinary relaxin profiles of the cat mirrored those in serum. In the leopard, urinary relaxin was first detected at Day 25–28 of gestation; levels peaked at Day 60–64 and declined in the last 3–4 wk of pregnancy. These results indicate that measurement of urinary relaxin in the cat and leopard provides a reliable method for pregnancy determination from as early as 3–4 wk of gestation. This method of pregnancy diagnosis and monitoring may prove useful in the breeding management of domestic cats and other felid and canid species, and provides a foundation for future studies on pregnancy in captive exotic carnivores.
conservation, felid, implantation, pregnancy, relaxin
INTRODUCTION
Of the 36 cat species in the world, most are facing threats to survival in the wild throughout all or part of their natural range [1–2]. Captive breeding programs for imperiled cat species represent an important component of holistic conservation efforts. The ability to diagnose pregnancy early in gestation is fundamental to developing an effective breeding program. However, methods for diagnosing pregnancy in wildlife are only practical if they are neither stressful to the animal nor disruptive of its daily routine.
In exotic felids, methods used for diagnosing pregnancy have included the observation of a nonreturn to estrus, the detection of visible changes in physical appearance (i.e., abdominal enlargement, mammary development, etc.), the evaluation of fecal steroid profiles [3], and the transabdominal ultrasonographic imaging of the uterus [4]. Of these methods, only ultrasonography is pregnancy-specific, but this technique usually requires handling and anesthesia of the female, potentially stressing both the dam and developing offspring. Other diagnostic methods are much less reliable. For example, estrous behavior is not always overt in captive felids, especially if the male has been separated in anticipation of impending parturition. Changes in physical appearance do not usually occur until the third trimester in large felids, and similar morphological changes may be observed in the absence of pregnancy. Finally, fecal progestagen concentrations are indicative of pregnancy only after the termination of the mating-induced luteal phase, which typically lasts 5–8 wk in felids [3, 5, 6]. Development of a urine-based pregnancy test, similar to the urinary hCG test widely used in women, would prove extremely useful for breeding management of wildlife species.
Relaxin is a 6 kDa polypeptide hormone that varies markedly in expression and structure between species [7]. Relaxin has many and diverse physiological roles [8], but is perhaps best characterized as a hormone of pregnancy. In 1976, O'Byrne and Steinetz [9] utilized a homologous porcine relaxin radioimmunoassay (RIA) to measure serum relaxin concentrations in the reproductive cycles of a range of mammals, and found markedly raised values during pregnancy. They noted large variations between the species in the amounts of immunoactive relaxin measured by the anti-porcine relaxin antiserum and suggested that this was because of the species-specific differences in relaxin secretion and/or the structure of the hormone. This same RIA was later used to measure the plasma relaxin levels in dogs, where it readily distinguished between true pregnancy and pseudopregnancy, because no relaxin was detected in the latter condition [10].
In 1985, Stewart and Stabenfeldt [11] identified relaxin as a pregnancy-specific hormone in the domestic cat (Felis silvestris catus). They showed that relaxin was first detected in plasma between Days 20 and 25 of pregnancy, levels rose to a plateau by Day 30–35 and declined again in the last 10–15 days of pregnancy. Relaxin was no longer detectable following the expulsion of the placentas and there was no detectable relaxin in the plasma during the follicular cycle or in the male cat. Subsequent research has shown that relaxin is produced specifically by the feline placenta [12] and is localized exclusively to the villous trophoblast cells of the lamellar placental labyrinth in both cats and dogs [13, 14]. These findings suggest that relaxin may be an important placental growth factor in these two species.
It is plausible that the small molecular weight and size of the relaxin molecule might allow it to pass through the glomeruli of the kidney and into the urine. Radiolabeled studies have shown a rapid uptake of radioactivity in the kidney of rodents following injection of 125I-labeled relaxin, which suggests a renal excretory route [15]. The study subsequently examined the urine of pregnant women and showed that low levels of relaxin immunoreactivity were present in extracted urine using an antibody against porcine relaxin. More recently, the detection of relaxin immunoactivity in human urine was reported using a sensitive homologous RIA for human relaxin-2 [16].
In research on the giant panda (Ailuropoda melanoleuca), Steinetz et al. [17] reported the first application of relaxin detection in urine for conservation purposes. They measured relaxin in paired serum and urine samples and found that, despite serum relaxin appearing to be pregnancy-specific, detection of relaxin in the urine was not a reliable marker of pregnancy. Relaxin was detected in multiple spikes around the time of implantation, and again in the last 10 days of pregnancy. Furthermore, a few relaxin spikes were also detected in mated pandas, but these failed to give birth and it was unknown whether they were pregnant or nonpregnant at the time of sampling. Nonetheless, the data did support the feasibility of measuring relaxin in urine specimens, potentially providing a noninvasive method for diagnosis of pregnancy in wild animals.
The objective of the present study was to determine whether the detection of urinary relaxin could be used for early, noninvasive pregnancy diagnosis in felids. As a preliminary assessment, the leopard was chosen as an exotic felid species to evaluate the possibility of a cross-species applicability of urinary relaxin detection. Our specific aims were to: 1) develop and validate a RIA for the detection of urinary relaxin in cats; 2) determine whether pregnant domestic cats excrete detectable levels of relaxin in urine; and 3) assess the application of the urinary relaxin RIA for pregnancy diagnosis and temporal monitoring in two felid species, the domestic cat and the leopard (Panthera pardus).
MATERIALS AND METHODS
Animals
Thirteen adult female domestic cats and four adult female Arabian leopards (Panthera pardus nimr) were used in the study. All leopards and eight of the cats were raised and maintained at the Breeding Centre for Endangered Arabian Wildlife in Sharjah, United Arab Emirates (UAE), which conforms to the husbandry guidelines set by the European Association of Zoos and Aquaria. The remaining five female cats were maintained at Cincinnati Zoo's Center for Conservation and Research of Endangered Wildlife and housed in compliance with U.S. Public Health Service guidelines. Male leopards at the UAE facility and male domestic cats housed at both sites were available for breeding purposes. The females had no access to the males for at least 3 mo prior to the study and were therefore known to be nonpregnant.
For breeding of domestic cats and leopards at the UAE facility, a male cat was introduced to a female when she first exhibited overt estrous behavior. The two were housed together continually during the mating period, which lasted 5–7 days. The male cat was removed at the end of estrus; however, the male leopard was left with the female for the following two months to ensure a nonreturn to estrus [18]. At the Cincinnati Zoo's cat research facility, group-housed female domestic cats were monitored daily for signs of behavioral estrus and estrual females were bred, beginning on the second to fifth day of estrus, by one of two proven breeder males. Females were mated to the same male three times each day, at 3-h intervals, for 2 consecutive days of natural estrus [19].
Blood and urine samples were collected from each female on the first day of mating and thereafter at weekly intervals throughout pregnancy and until after parturition. Cats that did not become pregnant were sampled weekly for 5–7 wk. Paired serum and urine samples were collected from all the domestic cats, whereas only urine was collected from the leopards.
At the UAE facility, domestic cats were sampled under sedation induced by intramuscular medetomidine (Domitor, 0.02 mg/kg; Pfizer) and butorphanol (Torbugesic, 3 µg/kg; Fort Dodge). Blood was collected by jugular venipuncture and urine was collected by manual expression of the bladder. The uterus of each cat was imaged to detect the number of conceptuses using ultrasonography (5-MHz convex array; Aloka 900) on Day 14 after mating and weekly thereafter throughout gestation. Sedation was reversed using intramuscular atipamezole (Antisedan, 0.05 mg/kg; Pfizer). At the Cincinnati Zoo's cat research facility, jugular blood samples were collected from nonsedated females using manual restraint and urine was collected the same day via natural voiding into a litter tray filled with nonabsorbent plastic beads. Pregnancy and the number of conceptuses were determined by ultrasonographic exam (7.5MHz linear array; Aloka 500) on Days 20–24 of gestation. Serum samples from nonpregnant females on Day 21 postbreeding were assessed for progesterone using a semiquantitative assay (Target Rapid Feline Progesterone Kit; Biometalics) to determine the presence or absence of functional corpora lutea. At both facilities, aliquots of serum and urine were frozen at –20°C to –80°C until assayed. Additionally, a large volume of tomcat urine was obtained for use as control urine in the relaxin assay.
Leopard urine samples were collected by restricting each female to a clean 4-m2 cage within her enclosure until urination had occurred, typically within 3 h. Cage floors were coated with industrial floor paint to prevent urine absorption into the concrete base. The leopards avoided their own urine instinctively, which meant that the urine sample was always untouched by the time the researcher had gained access to it. Urine puddles were aspirated from the floor and centrifuged, and the supernatant was frozen in aliquots at –20°C.
Relaxin Extraction
Relaxin was extracted from domestic cat urine according to Doczi [20] as modified by O'Byrne et al. [15]. A 10-ml sample of pregnant domestic cat urine was mixed with water, absolute acetone, and concentrated HCl to give concentrations of 0.15 M HCl in 70% acetone; unlike most polypeptides, relaxin is soluble in 70% acid-acetone. The solution was stirred in a cold box overnight, and the acetone layer was recovered and added to additional acetone to a concentration of 87% to precipitate the relaxin. The supernatant was aspirated and discarded, leaving a white precipitate that was air-dried. The precipitate was then dissolved in 10% aqueous acetic acid and filtered through Whatman #541 filter paper. Without further purification, the crude extract was dried under a stream of nitrogen and then dissolved in 1 ml of PBS containing 1% BSA prior to assay.
Relaxin RIA
Relaxin concentrations were measured using the double-antibody canine relaxin RIA as described by Steinetz et al. [21]. The standard used was synthetic canine relaxin (cRlx) dissolved in PBS-1% BSA to make a standard curve that ranged from 0.06 to 32 ng in an aliquot size of 0.1 ml. When using the assay on urine samples, 25 µl of tomcat urine was added to each standard in order to compensate for any change in pH or solute. The primary antibody used (no. 79888) was raised in a rabbit against synthetic canine relaxin. It showed no significant cross-reactivity with LH, FSH, insulin, hCG, or prolactin, and variable cross-reactivity with purified relaxins from other species [21].
A 50-µl aliquot of each serum sample or a 25-µl aliquot of urine was assayed in triplicate in borosilicate tubes, and total binding, nonspecific binding, and zero tubes were included in each assay. The primary antibody mixture (100 µl of a 1:10,000 dilution of anti-cRlx 79888 in 0.05 M EDTA-PBS, pH 7.0, containing 6% normal rabbit serum) was added to each tube, followed by 100 µl iodinated cRlx (125I-cRlx in PBS-1% BSA, 30 000 cpm/0.1ml). The tubes were vortexed and incubated overnight at 4°C. Second antibody mixture (500 µl of 4% ice-cold goat anti-rabbit IgG and 5% polyethylene glycol in PBS-1% BSA) was added to each tube to bind and precipitate the labeled antigen-antibody complex. The complexes were pelleted by centrifugation at 1000 x g for 30 min at 4°C. The supernatant was decanted and the radioactivity present in the remaining precipitate was measured in a gamma counter (Mini-gamma; LKB Wallac).
The accuracy of the assay was tested by recovery and parallelism. A range of known quantities of synthetic cRlx (0.06, 0.13, 0.25, 0.5, 1.0, 2.0, 4.0, and 8.0 ng) was added to 25-µl aliquots of tomcat urine as well as to nonpregnant leopard urine, and the raw and spiked samples were assayed. Because of the high interference of 100 µl of urine in the assay, only 25-µl aliquots of urine were utilized. To further test the accuracy of the assay, serial dilutions of pregnant cat serum and urine (from 1:2 to 1:512) were run in parallel with the cRlx standards, with 25 µl of tomcat urine added to the standards in the urine assay. The 4-parameter curves obtained were compared to the calibration curve by analysis of covariance (ANCOVA).
Urine Creatinine Assay
Creatinine was measured in all the urine samples using a commercial colorimetric kit (Sigma Diagnostics Creatinine), based on the Jaffe reaction, to enable the relaxin concentration to be expressed relative to the concentration of creatinine.
Data Analyses
Means (± SEM) were calculated for relaxin concentrations measured weekly in the pregnant and nonpregnant felid groups. Differences in weekly mean value between pregnant and nonpregnant groups were analyzed using the Mann-Whitney test, and relaxin concentrations were considered to be significantly increased when P < 0.05. The relationship between the relaxin concentrations measured in serum versus those measured in urine was expressed as a correlation coefficient. The area under the curve was calculated for individual urinary relaxin profiles and the relationship of these with litter size was analyzed using Spearman's rank correlation.
RESULTS
Validation of the Urinary Relaxin RIA
Relaxin was extracted successfully from pregnant cat urine and the extract exhibited relaxin immunoreactivity in the heterologous RIA; this confirmed that relaxin is excreted in urine during pregnancy and is detectable using an antibody raised against canine relaxin. An average of 3.2 ng immunoreactive relaxin per ml could be extracted from pregnant cat urine, whereas the same unextracted sample showed a value of 34 ng immunoreactive relaxin per ml. Thus, extraction losses were approximately 10-fold, and for this reason all urine samples were assayed unextracted.
The average recoveries of immunoreactive relaxin from spiked aliquots of tomcat urine and nonpregnant leopard urine were 103.0 ± 8% and 94.2 ± 9%, respectively. Higher quantities of urine (e.g., 100 µl) interfered with antibody binding and yielded unusable results. Based on the results of the recovery experiment, a 25-µl aliquot of tomcat urine was added to each standard in the urinary RIA to compensate for any pH or solute differences incurred by the use of unextracted urine samples. Assessment of the standard curves obtained in the assay, with and without tomcat urine, for parallelism showed no significant difference between the two (Fig. 1; ANCOVA, F = 0.013, P > 0.05), thereby indicating that the modification of the assay to measure relaxin in urine samples had not influenced binding. The sensitivity of the serum and urine assays at binding level of 90% was 0.1 ng/tube and 0.085 ng/tube, respectively. Analysis of serial dilutions of both pregnant cat serum and pregnant cat urine against the respective standard curves both showed parallelism (Fig. 2a, ANCOVA serum, F = 1.33, P > 0.05; Fig. 2b, ANCOVA urine, F = 2.73, P > 0.05).
|
|
Profile of Relaxin in Felids
Among the domestic cats, a total of 18 mating periods yielded nine pregnancies. Litter size varied from one to eight kittens (median five kittens) and only one cat showed a reduction in the number of conceptuses from two to one by Day 24 after mating. Gestation lasted 61–68 days (median 64 days) and all the pregnancies produced viable kittens. All the nonpregnant cats returned to estrus at the expected time.
Figure 3, a and b, shows the temporal changes of serum and urinary relaxin concentrations respectively throughout gestation (n = 9). Relaxin was first detected in the serum at Day 14 postbreeding in four of the pregnant cats, and all nine showed raised serum relaxin concentrations by Day 21. In urine, however, relaxin was not detected until Day 21 in seven pregnant cats, and all nine showed raised urinary relaxin concentrations by Day 28. Serum and urinary relaxin concentrations both peaked around 42–49 days postbreeding and declined again during the last 2 wk of gestation. The mean (± SEM ) peak serum and urinary relaxin concentrations were 7.0 ± 3.3 ng/ml and 6.5 ± 1.7 ng/mg creatinine, respectively. There was no discernible lag between the serum and urinary detection from Day 28 onward, and the relaxin concentrations measured in paired serum and urine samples were positively correlated (r = 0.76, P < 0.01). Although serum relaxin profiles differed between individual cats, there was no correlation (P > 0.05) between the areas under the curve and the increase in litter sizes.
|
Among the leopards, a total of nine mating periods resulted in three pregnancies. One leopard (#3) unfortunately aborted at Day 48 of gestation, but the other two showed an 88- and a 90-day gestation period, producing two and three viable cubs respectively. All the nonpregnant leopards returned to estrus at the expected time.
Figure 4 shows the urinary relaxin immunoreactivity in three pregnant leopards. Relaxin was first detected at Day 25–28; levels remained low until Day 40, rose thereafter to peak at Day 60–64 and then declined again in the last 3–4 wk. The peak concentrations measured in the two pregnancies that went to term were 3.6 and 4.6 ng/mg creatinine respectively. The leopard that aborted at Day 48 showed a decline in relaxin concentration from Day 32 to Day 53.
|
Relaxin could not be detected in the serum (Fig. 3a) or urine (Fig. 3b) of nonpregnant cats (n = 9) or leopards (n = 6; Fig. 4). Of the nonpregnant cats, serum progesterone assessment determined that five of the nine were luteal at Day 21 postbreeding, indicating induction of ovulation in these cats without apparent conception. No attempt was made to determine if nonpregnant leopards had ovulated in response to mating.
DISCUSSION
This study accurately detected relaxin in the urine of pregnant cats and leopards using a RIA for canine relaxin. This is the first report of the detection of relaxin in the urine of felids and has strong implications for reproductive management of both domestic and endangered nondomestic cats. The assay was sufficiently sensitive and accurate to be used to diagnose pregnancy in both domestic cats and leopards between 21 and 28 days of gestation. Urinary relaxin profiles were obtained from each species throughout gestation, allowing continual monitoring of placental function and possible detection of postimplantation pregnancy loss. Importantly, the assay also appears to be very specific for pregnancy because no relaxin was detected in any mated, nonpregnant felid.
All the pregnant cats studied showed serum relaxin profiles that were similar to those reported in pregnant cats by Stewart and Stabenfeldt [11]. These authors measured average plasma relaxin concentrations of 
7.0 ng/ml between Days 35 and 55 of gestation in their porcine relaxin RIA using a porcine standard, after which the concentrations of plasma relaxin declined. When they used a different anti-porcine relaxin antibody, the average concentrations of plasma relaxin were only 1.5 ng/ml. The lack of a homologous immunoassay for feline relaxin was also a limitation in the present study, because the results obtained were relative rather than quantitative and relied on the cross-reactivity of the canine relaxin antibody with the feline relaxin molecule. However, canine relaxin could be accurately recovered from spiked urine in the assay, and serially diluted pregnant cat urine demonstrated good parallelism to the canine relaxin standards. The positive results in these validation experiments supported the decision to use the canine relaxin RIA to measure relaxin concentrations in felids.
The question of whether or not relaxin is excreted in the urine unchanged was not addressed in this study, and it must remain speculative whether some losses will occur through reabsorption or degradation in the kidney. The concentration of a freely-filtered molecule tends to be much higher in the urine than in the serum, yet, in the present study, the relaxin concentrations measured in paired serum and urine samples from pregnant cats were remarkably similar and are suggestive of a loss of relaxin immunoreactivity within the urinary system.
Results of this study suggest that urinary relaxin in cats is pregnancy-specific because only pregnant felids excreted detectable levels of relaxin in their urine, in contrast to observations in mated, nonpregnant felids. The urinary profiles of relaxin concentration mirrored that of serum, which further justified the use of urine for noninvasive pregnancy testing. Urine is easily collected from most small captive, nondomestic felids or from domestic cats using nonabsorbent beads in a litter tray. For larger felids, urine may be recovered off any clean surface, such as the cage floor, as with leopards in the current study. It is likely that canids will also excrete measurable relaxin levels in urine, a prospect that can be tested using the same canine relaxin RIA.
Relaxin was first detectable in the urine of the domestic cat by Day 21 of gestation, and that of the Arabian leopard by Day 28 of gestation. This early detection period is well before the end of the nonpregnant luteal phase in both species and precedes the time of expected return to estrus for mated, nonpregnant individuals. In the leopard, urinary relaxin levels, although significantly elevated above nonpregnant values, remained low until Day 42 and thereafter assumed a profile similar to that seen in the domestic cat, but at relatively lower concentrations. Accordingly, assays for early pregnancy detection in the leopard require a higher sensitivity than those for the domestic cat. The difference in the urinary relaxin profile between the leopard and cat could relate to differences in placental growth and structure, or in renal physiology, between the two species. Relaxin concentrations during pregnancy have been shown to vary considerably between species [8] and, in equids, between breeds. For example, Stewart et al. [22] reported differing concentrations and profiles of relaxin in serum during pregnancy between Thoroughbreds, Standardbreds, and ponies. The differences were unrelated to placental size or fetal sex, and the authors questioned whether they were of any functional significance.
In women who undergo superovulation treatment and gestate multiple conceptuses, plasma relaxin concentrations at 10–12 wk of gestation correlate well with the number of fetuses in utero [23]. The cats in the present study carried between 1 and 8 fetuses, and there was no apparent change in the relaxin profile relating to the increase in litter size. One leopard carrying triplets did show a greater total relaxin excretion (i.e., based on the area under the curve) than the other leopard, which carried twins, but the small sample size precludes any assessment of correlation.
In mammals, relaxin is produced mainly by two tissues; the corpora lutea, which are the major source of relaxin during pregnancy in women [24], sows [25], cows [26], and rats [27], and the placenta, which is the primary source in mares [28], rhesus monkeys [29], rabbits [30], hamsters [31], dogs [32], and cats [12]. Implantation commences at Day 13–14 of gestation in the cat [33]; soon afterwards, by Day 18–22, relaxin activity can be detected in plasma, and the subsequent rise in concentrations mirror those in placental tissue [12].
The time of implantation in the leopard is not known. However, because relaxin is first detected in the urine around Day 28, it seems reasonable to speculate that functional placental tissue also becomes established at about this time. However, relaxin concentrations remain low relative to those in the pregnant cat, and rise markedly only after Day 42. Although undoubtedly valuable for pregnancy diagnosis in cats, measurement of urine relaxin also is potentially useful as a monitoring tool for fetal and placental health. For example, in the present study, one leopard experienced a midterm abortion at Day 48 of gestation. Urinary relaxin concentrations began to decline 10–14 days prior to abortion, indicating possible fetal death or placental dysfunction. Similarly, Stewart et al. [22] reported that a Standardbred mare that suffered a midterm abortion also showed declining serum relaxin concentrations for 10 days prior to fetal expulsion; the horse, like the felids, secretes most of its relaxin from the placenta during pregnancy. A similar drop in relaxin levels was also noted in relation to fetal resorption in taurine-deficient cats [34], which supports the suggestion that relaxin can be a measure of placental sufficiency in the cat [13].
To date, noninvasive methods of pregnancy diagnosis in exotic felids have been based on changes in physical appearance or fecal steroid analyses, neither of which provides early or pregnancy-specific diagnoses. As was shown in this study, measurement of urinary relaxin concentration represents a sensitive and specific method for pregnancy diagnosis in both domestic cats and leopards. The work presented in this study therefore provides a foundation for future studies on pregnancy in captive carnivores. The detection of pregnancy at an earlier stage of gestation in nondomestic cats would necessitate the development of a homologous feline relaxin immunoassay with a higher sensitivity. Further research is also necessary to assess whether urinary relaxin could be used to diagnose pregnancy in other felids and canids. Although only three leopard pregnancies were monitored for urinary relaxin in the present study, urine sample collection from additional felid species is ongoing to further substantiate these findings. If the detection of urinary relaxin proves to have broad cross-species applicability, this method for pregnancy diagnosis should greatly facilitate conservation efforts for carnivores in captivity and possibly in the wild.
ACKNOWLEDGMENTS
We are sincerely grateful to His Highness Dr. Sheikh Sultan bin Mohammed al Qassimi and the Environment and Protected Areas Authority in Sharjah for their continued support for the conservation of Arabian wildlife. We also thank the staff at the Breeding Centre for Endangered Arabian Wildlife for their help in sample collection, Dr. T. Klonisch and Dr. P.G.G. Jackson for their helpful discussions, and Professor W.R. Allen and Dr. J.A. Skidmore for their kind assistance with the preparation of the manuscript.
FOOTNOTES
1 Supported in part by NIEHS Center grant P30 ES00260 and NCI Center grant P30 CA016087. 
2 Correspondence: Bernard G. Steinetz, Nelson Institute of Environmental Medicine, NYU School of Medicine, 57 Old Forge Road, Tuxedo, NY 10987. FAX: 845 351 4510; steinetz{at}env.med.nyu.edu
Received: 27 December 2005.
First decision: 19 January 2006.
Accepted: 15 February 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
