Non-invasive monitoring of ovarian function in Asian elephants (Elephas maximus) by measurement of urinary 5 beta-pregnanetriol.

Abstract

The development of an enzymeimmunoassay for 5 beta-pregnanetriol and its use for non-invasive monitoring of reproductive cycles in Asian elephants is described. Gas chromatography-mass spectrometry (GCMS) and high performance liquid chromatography (HPLC) confirmed the presence of 5 beta-pregnane-3 alpha,17 alpha,20 alpha/beta-triols as the two most abundant urinary progesterone metabolites. The assay developed used the antiserum anti-5 beta-pregnane-17 alpha,20 alpha-diol-3 alpha-gamma l glucuronide but was designed to measure the free steroid in urine samples after hydrolysis and extraction. HPLC confirmed the presence of immunoreactive pregnanetriol in urine, but indicated that the measurement was nonspecific. Immunoreactive pregnanetriol concentrations were significantly correlated with the concentrations of both progesterone (r = 0.98, n = 269, P < 0.01) and 17 alpha-hydroxyprogesterone (r = 0.95, n = 205, P < 0.01), the metabolic precursor of pregnanetriol. The mean +/- SEM deviation of cycles as determined by measurements of plasma progesterone, 17 alpha-hydroxyprogesterone and urinary pregnanetriol, respectively, were 15.54 +/- 1.5 (n = 23, where n = number of cycles), 15.21 +/- 1.7 (n = 15) and 15.45 +/- 0.94 weeks (n = 20). These results demonstrate that it is possible to monitor ovarian function in Asian elephants by the measurement of urinary immunoreactive pregnanetriol concentrations.

Full text

Page 1

Non-invasivemonitoringof ovarian function in Asianelephants
(Elephasmaximus)bymeasurement ofurinary
C. A. Niemuller,
^Instituteof Zoology, Zoological Society ofLondon,Regent'sPark, London, NWl 4RY, UK,
and2Divisionof ReproductiveBiology,German Primate Centre, Kelinerweg 4, D-3400, Göttingen,
Germany
5\g=b\-pregnanetriol
H.J.Shaw andJ.K.Hodges
Thedevelopmentof anenzymeimmunoassayfor5\g=b\-pregnanetrioland its use for
invasive monitoring
chromatography\p=n-\massspectrometry(GCMS)andhigh performance liquid chromatography
(HPLC) confirmed thepresenceof5\g=b\-pregnane-3\g=a\,17\g=a\,20\g=a\/\g=b\-triolsas the two most
abundanturinary progesteronemetabolites. Theassay developedused the antiserum anti\x=req-\
5\g=b\-pregnane-17\g=a\,20\g=a\-diol-3\g=a\-\g=g\lglucuronidebut wasdesignedto measure the free steroid
in urinesamplesafterhydrolysisand extraction. HPLC confirmed thepresenceof immuno-
reactivepregnanetriolin urine, but indicated that the measurement
Immunoreactivepregnanetriolconcentrations weresignificantlycorrelated with the concen-
trations of bothprogesterone(r=0.98, n =269,P <0.01)and17\g=a\-hydroxyprogesterone
(r=0.95, n =205, P <0.01),the metabolicprecursorofpregnanetriol.The mean \m=+-\SEM
deviation ofcycles
hydroxyprogesteroneandurinary pregnanetriol, respectively,were 15.54 \m=+-\1.5 (n
where n= number ofcycles), 15.21 \m=+-\1.7 (n=15) and 15.45 \m=+-\0.94 weeks (n
These results demonstrate that it ispossibleto monitor ovarian function in Asianelephants
bythe measurement ofurinaryimmunoreactivepregnanetriolconcentrations.
non\x=req-\
Gas
ofreproductive cycles
in Asian elephants
is described.
wasnonspecific.
as determined by measurements ofplasma progesterone,
17\g=a\\x=req-\
=23,
=20).
Introduction
The twospeciesofelephant existing today (Elephasmaximus
and Loxodonta africana)are classified asendangered (CITES,
1992), with current estimated wildpopulationsof 50 000 and
500 000, respectively.For both species, habitat loss and
poachinghave beenmajor factorscontributingto the decline
andfragmentationof the wildpopulations.At the same time,
zoos andprivatereserves have becomeincreasinglyactive in
thedevelopmentofmanagementandbreeding programmes
for animals
conditions conducive to successful reproductionis often
difficult.
Duringthe last 10years,hormonal methods formonitoring
ovarian cycles in both Asian and Africanelephantshave been
described.Typically,the duration of an ovariancycleis 14—16
weeks and is characterizedbyan 8—10 week lutealphaseand a
4-6 week interluteal or follicularphase, accordingtopatternsof
circulatingconcentrations ofprogesterone(Hess et al, 1983;
Brannian etal, 1988;Plotkaetal, 1988; Brown etal, 1991;Taya
et al, 1991). Concentrations of oestradiol and of LH have also
been measured inplasma,but concentrations aregenerally very
low or results inconsistent between studies, and their secretion
has notproveduseful forindicatingovarian status (Hess et al,
1983; Brannian et al, 1988; Plotka et al, 1988; Mainka and
in captivity, although
the establishment of
Lothrop,1990; Brown etal, 1991;Tayaet al, 1991; Gross etal,
1991).
Measurement ofcirculatingconcentrations ofprogesterone
thereforeprovidesthemost reliable monitor ofovulatory cycles
inelephantsand as such are widely used inmany zoological
collections.Unfortunately,thedependenceonassessingovarian
function by plasmahormoneanalysishas been a hindrance for
those zoological collections in which collection of blood
samplesis notpossibleand cannot be used in studies monitor¬
ingthereproductive dynamics of wild animals. There is there¬
fore a need todevelopa non-invasiveapproachfor endocrine
reproductiveassessment inelephants.On the basis of our
limitedknowledgeofelephantovarianphysiology,this should
be most feasible bythe measurement of urinary progesterone
metabolites.
There are noreportsof theidentityofurinarymetabolites of
progesterone
ovarian functionby measuringunmetabolizedprogesteronein
urine from Asiaticelephantswas unsuccessful (Mainka and
Lothrop,1990).Althoughmeasurement ofurinary pregnanediol
glucuronideor20ct-dihydroprogesteroneimmunoreactivityhas
providedvaluable information on ovarian function in a variety
of mammalianspecies (see, forexample,Loskutoff et al, 1986;
Kirkpatricket al, 1990; Hindle et al, 1992; Hodges, 1992
for review), our ownunpublishedobservations (J. Hindle and
C. Niemuller)indicated that this did not hold true for Asian
elephants.The aims of thepresent study were therefore to
(i) identify the major urinary progesterone metabolite(s) in
in elephants.A recent attempt
to monitor
'Present address: RR 1Cambridge,Ontario, N1R 5S2, Canada.
Received IIFebruary1993.

Page 2

Asianelephants(ii)developa microtitreplate enzymeimmuno-
assayfor its determination and (iii)evaluate its use as a non-
invasive method formonitoring reproductionin thisspecies.
Materials and Methods
Animals andsamplecollection
Wherepossible,matched blood and urinesampleswere
collected once a week for 1—3yearsfrom 11 mature female
elephants aged12-25yearsfrom fourzoologicalcollections. A
5-10ml bloodsamplewas collected from either thesaphenous
vein or an ear vein intoheparinizedtubes, centrifugedat 2000g
for 10 min immediately after collection,and theplasmastored
at
morning, divided into 5 mlaliquots and stored frozen at
-20°C.
Oestrous behaviour in theeightfemales that had access to
males was recordedbytheelephant keeperson the basis of
increased interestbythe bull and eventualcopulationwith the
females. Three of the females monitored had no contact with
malesthroughoutthe duration of the study. Of theeight
females thathad access to males, three werekeptwith thebull(s)
dailybutseparatedatnight.Theremainingfive cows had
continuous access to a bullexcept duringthe winter months.
Duringthis time, access to the male was restricted todaylight
workinghours andonlyfor a cow in the latter halfofthe follicu¬
larphaseaspredicted by analysis ofplasma progesterone.As
soon as concentrations ofplasma progesterone beganto rise,
the cow had no further contactwith the bull.
—20°C. Mid-stream urinesampleswere collectedduringthe
Creatinine determination
All urinesampleswereanalysed for creatinine concentration
after the initialthawing by the method ofHodgesand Green
(1989) and aspreviouslyvalidated forelephants by Ramseyet
al (1981) and Poole et al (1984), to correct for variations in
glomerularfiltration rate of urine. Theassay sensitivity was
approximately0.1mgml-1 and the intra- andinterassaycoef¬
ficients of variation were 9.3 and 8.8%,respectively («=30).
Urinarycreatinine concentrationrangedfrom below detection
(samplesexcluded) toapproximately2mgml~\ with a mean of
0.6mgml^1.
Gaschromatographymassspectrometry
Ten urinesamples(each 20ml)from threepregnantcows and
from sevennonpregnantcows in the mid-luteal(n=4), and
follicular (n=3)phasewerepreparedforgas chromatography
massspectrometry(GCMS)analysisbythemethodofShackleton
et al(1976). In brief,sampleswere extractedusingaSep-pack
C18 column (Waters, Division ofMillipore,UK Ltd, Watford),
hydrolysed usingsodium acetate hydrolysis buffer(pH=5)
and25mg ß-glucuronidase-arylsulfatase(Sigma,Poole, Dorset,
sulfataseactivity: 12.5 Umg-1 solid, ß-glucuronidase activity:
1000 Umg-1 solid)and then re-extracted viaSep-packbefore
beingdried down on arotary evaporator. Sampleswere recon¬
stituted in
a
4:1 (v:v) cyclo-hexane:ethanol solution and
separatedinto three distinct fractions withSephadexLH-20
(Pharmacia,St Albans, Herts.). The fractions were dried down,
reconstituted in ethanol and derivatized withmethoxyamine
hydrochloride andtrimethylsilylimadazole. The finalsample
was reconstituted in 500pi cyclohexane of which 1pi was
injectedinto thegas chromatographymassspectrometer.
Massspectrometer(MS)profilesthus obtained were scanned
and identification of unknownpeaks attempted initially by
retention times andbycomputerMSlibrarysearch(Shackletonet
al, 1980). Peaks fromgas chromatography(GC) werecompared
withreferencetemplatesand re-runthroughthe GC withstraight
chain alkanes to determineméthylène units for further identifi¬
cation. This was followed whereverpossible by coinjection of
the unknownpeakswith steroid reference standards to deter¬
minechangeinpeak heightandprovidefurther confirmation of
theidentityof the steroid metabolite.
Preparation of[3H]pregnanetriol
As[3H]pregnanetriolswere not commercially available,
[ H]-labelledform of one of the isomers waspreparedin our
laboratory bytheenzymaticreduction of[3H]17a-hydroxy-5ß-
pregnane-3a,20a-diol (Amersham, Buckinghamshire)with 3a-
hydroxysteroid dehydrogenaseand NADH(method provided
by A. P. Scott; Ministry ofAgriculture,Fisheries and Food
(MAFF), Fisheries Laboratory, Lowestoft).Thereagentswere
incubated for 2h at roomtemperature,and theproductsether
extracted before thinlayerchromatography. Preliminary identi¬
fication of5ß-pregnane-3a,17a,20a-triol as thepredominant
labelledproductwas confirmedby recrystallizationof themajor
radioactivepeakandpurestandard to constantspecific activity
(SA).Thetracer thuspreparedhad an activity of 8.8pCiml-1
and wassubsequentlyused to indicate steroidretention time on
HPLC.
a
HPLC
Unconjugatedneutral steroids wereseparated usingHPLC.
Sample preparationwas doneaccording to the method of
Hindle et al (1992).Urinesamples(1 ml)wereadjusted topH5
withhydrolysisbuffer andhydrolysedovernightwith1000 FU
(50 pi)-1 ofhydrolysis enzyme (Sigma, sulfatase activity:
4500 Uml-1, ß-glucuronidase activity:100 000 Uml-1).Since
conjugated pregnanetriol
hydrolysis was monitored by the addition of[3H]pregnanediol
glucuronide ([3H]PdG) topooledurinesamplesmonitoredsep¬
arately.Sampleswereadjusted topH
and extracted with 5ml distilleddiethylether. Before extrac¬
tion[3H]steroids (progesterone, 20a-dihydroprogesterone,17a-
hydroxyprogesteroneandpregnanetriol)were added to the
sample
efficiency.The ether was dried undernitrogenandsamples
were reconstituted in 0.2ml ofdoubledistilled watenacetonitrile
(70:30 v:v).
TheprogesteronemetaboliteswereseparatedusingaNovapak
RP C-18 column(3.9 x 150 mm, 5pm particle size; Waters)
and an isocratic solventsystemof 30:70 acetonitrile:water at a
flow rate of 2ml min-1 ( . Junemann,unpublished).Fractions
(1 ml) were collected for 36 min, dried and reconstituted in
was not available, efficiency of
= 7 with 3 mol NaOH 1~J
as reference standards and to monitor extraction

Page 3

assaybuffer. Mean recoveries forthe four tracers were between
71 and 78% («=14).The overallefficiencyof theprocedure,
including hydrolysisandextraction,was 68.8 + 4.2% (n=14)
based on therecoveryof [Hjpregnanediol.
Urinaryhormoneassays
Sample preparation.
extracted beforeassay.Analiquotof 250piof urine was incu¬
batedovernightat 37°C with 250FU(50pi)-1ofhydrolysis
enzyme (Sigma,sulfataseactivity:4500 Umr1,ß-glucuronidase
activity: 100 000 U ml-1) and 250pi of hydrolysis buffer.
Hydrolysisefficiencies were determined by addingintriplicate
tracer amounts of[3H]oestronesulfateandoestroneglucuronide
to urinepools (250pi)beforehydrolysis.Amounts of uncon-
jugatedsteroid recovered were 81.1 + 6.1% and 84.3 + 3.2%
(mean + SEM, =30),respectively.ThepHwasreadjusted to
7 with 3 mol NaOH 1_I andsampleswere extracted with
diethylether and reconstituted inassaybuffer. Procedural losses
duringextraction were monitored bythe addition of tracer
amounts of [3H]17a-hydroxyprogesterone to each sample.
Individual recoveries were used to correct for losses when
calculating
mean ± SEM extraction was 89.1 ±2.1%(«=200).
Urinesampleswere hydrolysedand
the final hormone concentration. The overall
Pregnanetriol
measuredusinga microtitreplateEIA method similar to that
describedbyHindle et al.(1992).Theassayused the antiserum
5ß-pregnane-17,20a-diol-3a-yl glucuronidewhich was raised
in rabbitagainsttheimmunogenic complexof5ß-pregnane-
17,20 diol-3a-yl glucuronide coupledto BSA(Samarajeewa
andKellie, 1985). The enzyme label,provided by E. Mosti
(Institut
für Biochemie,
Veterinärmedizinische Universität,
Vienna), wasprepared by conjugating 4-pregnene-17a,20a-
diol-3-one-carboxymethyloximeto horse radishperoxidasevia
the mixed acid anhydride procedureof Lieberman et al (1959).
The antiserum, which crossreacted 62.5% with the free steroid
as determined by radioimmunoassay (Samarajeewaand Kellie,
1985), was stored inaliquotsof 1:1000 in Tris buffer at
while theenzymelabel was stored inglycerol(1:1 v:v) also at
-20°C.
Microtitreplates(Nunc Immuno 1: Gibco Ltd, Paisley)were
coatedovernightwithsheepanti-rabbitIgG (1pi perwell).The
plates
(20 nmol 1
and stored at 4°C until needed. Just before use, theplateswere
emptied,rinsed with a 0.005% Tween solution and blotteddry.
Standard (5ß-pregnane 3a,17a,20a-triol, Sigma, Poole) was
doubledilutedin Tris buffer over arangeof 156pgml-I-10ng
ml-1 (7.8-500pg (50pi)"1)and 0.05 mlaliquots dispensed
in duplicate. Sampleswere diluted in Trisassaybuffer and
added induplicate(0.05 ml).Antiserum (1:160 000) and label
(1:30 000) were added toall wellsexceptthe blanks. Plates were
covered and incubatedovernightin the dark at 4°C.
After incubation,plateswereemptied by inversion, washed
with 0.005% Tween solutionandtappeddry. Enzymesubstrate
was prepared immediately before
hydrogen peroxidesolution (30%, 4 mol 1
O-phenylenediamine;(1,2, benzenediamine;Sigma)in 28 ml of
Immunoreactive
5ß-pregnanetriol
was
—20°C
were emptied by inversion, filled with Tris buffer
;0.25 mlper well) containing0.1% sodium azide
use by adding 0.025ml
, BDH) to 100mg
01-
0.010
0.100
1.000
10.000
Concentration (ngml^1)
100.000
Fig.1.Bindinginhibition curvesdemonstrating parallelisminserially
diluted Asianelephanturinepool samplesfrom the mid-luteal ( )
and follicularphase (O)as well as human urine from latepregnancy
(O);pregnanetriolstandard (·).
a citricacid substrate buffer (0.1mol citric acidmonohydrate1
pH
wereincubatedinthe darkat 4°Cfor 45 minand thereaction was
stopped bythe addition of 0.05 ml of sulfuric acid (3 mol 1_1).
Absorbance was measured at 492 nm (with a reference filter of
dualwavelengthat 620 nm) on an automaticplate reader
(DynatechMR 700,DynatechLaboratories Ltd,Billinghurst).
The amount of immunoreactivepregnanetriolin each well
was read from a standard curve constructed by plottingthe
percentageofsampleboundagainstthe amount of standard
added.
;
=5). A volume of 0.25ml was added to each well;plates
PregnanetriolEIA validation.
determined at 90%bindingwas 8.5pg perwell (170pg ml-1).
Serial dilutions ofelephanturine collectedduringthe mid-luteal
and follicularphaseand of human latepregnancyurine (a
known source ofhighconcentrationsofpregnanetriol) gavedis¬
placement
(Fig. 1).Theaccuracyof theassay,determinedby recoveryof
unlabelled 5ß-pregnane-3a,17a,20a-trioladded to elephant
urinepools containinglow concentrations ofendogenoushor¬
mone was 98.7% ± 1.1% (n=10). Theinterassay CV for
repeateddetermination ofelephanturinepool containing high
and low concentrations ofpregnanetriolwas 10.8 and 5.6%
(n=30), respectively, while the intra-assay CV was 7.8%
(n=30).
Steroids showing crossreactivity determined
50% inhibition ofbinding,were5a-pregnane-3ß,17,20a-triol
(109%),5ß-pregnane-3a,6a,17a,20a-tetrol(189%),5ß-pregnane-
3a,6a,17a,20ß-tetrol (113%), 5a-pregnane 3ß,17a,20ß-triol
(106%), 5ß-pregnane-3a,17ß,20ß-triol (2.9%), progesterone
(1.45%),5ß-pregnanediol(1.1%).All of the other C21, C19 and
C18 steroids tested crossreacted less than 1%.Specificityof
measurement ofpregnanetriolin Asianelephanturine was
determinedby co-chromatographyon HPLC. Thepresenceof
pregnanetriol
pregnanetrioltracer (Fig. 2) confirmed that theassaycould
detectpregnanetriolin urine from Asianelephants.However,
thepresenceoflargeamounts ofimmunoreactivity in fractions
precedingthepeakof[3H]pregnanetriolindicated that the
measurement ofpregnanetriolwas not specific.The values
Thesensitivityof theassayas
curvesparallel to the5ß-pregnanetriol standard
at
immunoreactivity co-eluting
withthe 5ß-

Page 4

Fractions
Fig.2.Co-chromatography profilesfrom HPLC of immunoreactive
progesteronemetabolites in Asianelephanturine. The elutionprofile
(·) ofpregnanetriol immunoreactivityinhydrolysedand extracted
urine iscomparedwith that of[3H]-labelled pregnanetriol(P3), 20a-
dihydroprogesterone(20A),pregnanediol(PDIOL) andprogesterone
(P4)(0).
reported
immunoreactivity.
here
are
therefore expressed
as pregnanetriol
20a-DihydroprogesteroneEIA.
Hindle etal(1992). Immunoreactive20a-dihydroprogesterone
was measured in fractions collected from the HPLC. Theassay
using
antiserum andenzyme conjugateof horse radishperoxidase
conjugatedto20a-dihydroprogesterone (providedbyE.Mosti)
showed crossreactivities with5ß-pregnane-20a-ol-3-one (11.9%),
5ß-pregnene 3a-ol-20-one (0.92%), 5ß-pregnane-3a-20adiol
(0.57%),progesterone(0.52%),pregnenolone(<0.1%)and5ß-
pregnane-3a,17a,20a-triol(<0.01%). Theassay sensitivity as
determined at 90%bindingwas 150pgml"1. Theintra-assay
CV was 8.4%(n=30),whereastheinterassayCV was 12.1 and
16.1% (n=15) forhighand low valuequalitycontrolpools,
respectively.
Theassaywas as describedby
ananti-4-pregnene-20a-ol-3-one-carboxymethyloxime
Assays ofplasmahormone
Progesterone.
assayasinitiallydescribedbyShaw et al(1989)with modifi¬
cations. Theassayused1,2,6,7-[3H]progesterone (Amersham)
and sheep anti-progesterone lla-hemisuccinate—ovalbumin
(from J. Foulkes, MAFF, Shinfield,Reading)as antiserum which
showed thefollowingcrossreactivities: 1la-hydroxyprogesterone
(29.8%), 11ß-hydroxyprogesterone (16.5%),5ß-pregnanedione
(16.1%),5a-pregnanedione(2.63%)and less than 0.1% with all
other steroids tested.Samples (500 pi) were extracted with
petroleumether(recovery83.6 + 2.3%, =200)and reconsti¬
tuted inassaybuffer. Thesensitivityoftheassayat 90%binding
was 20pgml-;intra- andinterassayCVs were 8.9% and 10.7%,
respectively.Serial dilutions of extractedelephant plasma gave
displacementcurvesparallelto theprogesteronestandard. The
accuracyof theassaydeterminedbythe mean + SEMrecovery
of unlabelledprogesteroneadded to extractedplasma poolsof
lowendogenoushormone was 99.4 + 2.5%.
Progesteronewas measuredby radioimmuno¬
17a-Hydroxyprogesterone.
progesterone in plasma and HPLCsamples followed the
Themeasurement of17a-hydroxy-
sameprotocol
assay. The assay used1,2,6,7-[3H]17a-hydroxyprogesterone
(Amersham) with
carboxymethyloxime
antiserum crossreacted withprogesterone (2.2%) and 5a-
pregnanedione (1.1%) as determined by Steranti. Other C21
and Cl9 steroids tested crossreacted less than 1%.Samples
(500pi)
87.6 + 3.1%, =200) and reconstituted inassaybuffer. The
sensitivityoftheassaybased on 90%bindingwas 20pg ml-1,
whileintra- andinterassayCVs were 6.3 and9.7%,respectively.
Serial dilutions ofsamplesextracted fromplasmaofpregnant
and mid-lutealphase elephants gave displacementcurvesparal¬
lel to the17a-hydroxyprogesteronestandard. Theaccuracyof
theassaydeterminedbythe mean + SEMrecoveryofunlabelled
17a-hydroxyprogesterone(10-160pgml-)added to extracted
plasma poolsof lowendogenoushormone was 101.7 + 1.4%.
as used for theprogesterone radioimmuno¬
an anti-sheep-17-hydroxyprogesterone-3-
antiserum (Steranti,
St Albans).
The
were extracted withpetroleum
ether (recovery
Analysis ofdata
Stages ofthe ovariancycle.
follicularstagesof theelephant reproductive cyclesbased on
analysisofplasmahormones for thisstudywas modelled after
the definitionoriginallydescribed by Plotka el al. (1988). For
bothprogesteroneand17a-hydroxyprogesterone,the onset of
the lutealphaseof thecyclewas defined as the firstpoint (after
therespective plasmaconcentrations fell below 200pg ml-1)
that increased by 50pg and remainedhighfor at least two
weeks; the end ofthe lutealphasewas defined as the first of two
consecutive values within 50pgthat were less than 200pgand
at least 50pglower than theprecedingvalue. Thecomposite
datapresentedwere thereforealignedto the onset of the luteal
phase (week 1)accordingtothis definition.
Similarly, accordingtourinary pregnanetriol concentrations,
the onset ofthe lutealphasewas defined as the firstpointafter a
fall in values below 200ng mg-1creatinine which increasedby
50ng mg~
weeks.Similarly,the end of the lutealphasewas taken as the
first of two consecutive values within 50ng mg_I creatinine,
which were less than 200ng mg_Icreatinine and at least
50ng mg-1creatinine lower than theprecedingvalue.
The duration of the ovarian cycle was determined as the
interval between the onset of two successive lutealphasesas
definedseparatelyfor each hormone measured.
The definition for the luteal and
creatinine and remainedhighfor at least two
Statisticalanalysis
Statistical analysis was carried outby standard analyses of
mean and SEM and Student's I test asappropriate.
Results
According
derivatives of5ß-pregnanetriol (as described by Quilliam and
Westmore, 1980) were the most abundant of theprogesterone
metabolitessuccessfullyidentified.5ß-Pregnane-3a,17a20 /ß-
triols
to the
mass spectra obtained, the silyl ether
werepresent in every urine sample analysed from

Page 5

Table 1.Summaryof results ofgas chromatography-mass spectrometryof urine
from Asianelephants
Reproductive
state
Identified steroid
metabolite
Incidence
Follicular
(n=3)
Mid-luteal
(«=4)
Pregnancy
( =3)
11ß-hydroxyandrosterone/aeticholanolone
No detectablepregnanetriols
5ß-pregnane-3a,l7a,20a-triol
5ß-pregnane-3a,l7a,20ß-triol
5ß-pregnane-3a,6a,20a-triol
5a-pregnane-3a,llß,20a-triol
5a-pregnane-3ß,17,20a-triol
3a,20a dihydroxy-5ß-pregnan-11-one
11ß-hydroxyandrosterone/aeticholanolone
5ß-pregnane-3a,l7a,20a-triol
5ß-pregnane-3a,l7a,20ß-triol
5a-pregnane-3ß,17,20a-triol
5a-pregnane-3a,11ß,20a-triol
3 ,20 dihydroxy-5ß-pregnan-11-one
pregnantAsianelephantsand those at the mid-lutealphase
(Table 1). Quantitatively,these two steroid metabolites were
also the most abundant forms ofpregnanetriol present(datanot
shown). It was notpossibleto detectany pregnanetriolin urine
from the follicularphase,andprogesterone, pregnanedioland
20a-dihydroprogesteronewere not detected inany samples.
The results of HPLCanalysisof asampleofhydrolysedand
extracted urine from Asianelephantsat the mid-lutealphase
are shown (Fig. 2). Immunoreactivity co-eluting with the
[3H]pregnanetriolmarker
(n=6), thus indicatingthepresence of immunoassayable
pregnanetriolin urine of Asianelephantsat the lutealphase.As
indicated earlier, relatively largeamounts of additional immu¬
noreactivitynotco-elutingwith the[3H]pregnanetriollabel was
alsopresentin fractions 4-9 in allsamples analysed (n=6).
In contrast,progesterone, 20a-dihydroprogesteroneand 17a-
hydroxyprogesteronewere not detectable in HPLC fractionsby
theirspecific assaysat the volumes and dilutions used.
The profileof urinary pregnanetriol immunoreactivity in
relation to that ofcirculating progesteroneconcentrations dur¬
ingthree consecutive ovariancyclesin oneelephantis shown
(Fig. 3). The urinarymeasurements revealed a cyclic patternof
excretion in which the luteal and interlutealperiods,as deter¬
minedbyintervals ofhighand lowcirculating progesterone
concentrations, could beclearly distinguished.In this animal,
maximumpregnanetriolconcentrations at the lutealphasewere
between 300and 750ng mg-1creatinine,approximatelyfive¬
foldgreaterthan those of thecorrespondinginterlutealperiod.
There is a closetemporal relationshipbetween the defined rise
inplasma progesteroneand urinary pregnanetriol following
predictedovulation in each of the threecyclesillustrated.
Composite profilesof the mean and SEM ofplasmaconcen¬
trations of progesterone (n=11 animals, =23 cycles)
and urinary concentrations ofpregnanetriol (n=8 animals,
=20 cycles) throughoutthe ovariancyclein Asianelephants
are shown (Fig. 4). Correspondingdata forplasma 17a-
hydroxyprogesterone,themajor circulating precursorforpreg-
was found in all samplestested
o>—
2E
. )
E
2500
2000
1500
1000
500
0
MO
li
MMO
O
\ A
10
20
30
40r\
;
50
o
M
c
tV
it
fir
'
6070
80
90
100
1500

cd
c
S
-
co1000
CDu
Q.7
> CD
a E
.E en
5-s
500
(b)
Ul
10
20
30
loo
40
Time (weeks)
50
60
Fig.3.Representative reproductive profileof weekly (a)plasma pro¬
gesteroneand (b) urinary pregnanetriolconcentrations for one female
Asianelephant throughfour consecutivereproductive cyclesand into
earlypregnancy.(O:predictedovulation; M:mating;C: conception).
nanetriol (Axelrodand Goldzieher, 1960), are also shown from
15 ofthesecycles(n=8 animals).All data werealignedto the
rise inplasma progesterone reflectingthe start of the luteal
phase.The overall duration of theoestrous cycle (mean + SEM)
calculated from the interval between successive lutealphases
was 15.54 + 1.5(n=23), 15.21 ± 1.7 (n=15) and 15.45 ±
0.94 (n=20) weeks, based
progesterone, 17a-hydroxyprogesteroneand urinary preg¬
nanetriol,respectively.Immunoreactivepregnanetriolconcen¬
trations weresignificantlycorrelated with the concentrations
of bothprogesterone (r=0.98, =269 cycles, < 0.01)
on measurements ofplasma

Page 6

700,(a)
en
2E
. coil
E
—
600
500
400
300
200
100
y
500
(c)
£;c 400]
CD'c
?î 3001
ÇD

>- O)
||100J
U
200
;
Kl
1/
510
15
Time (weeks)
20
Fig. 4. Mean (+SEM)profilesofweeklyconcentrations of (a)pro¬
gesterone,(b) 17a-hydroxyprogesteroneand (c)pregnanetriol samples
for Asianelephants(n=11)representingonereproductive cycle). All
data werealignedto the elevation inplasma
concentrations(week 1) until thefollowingweek I.
orurinary steroid
and17a-hydroxyprogesterone (r=0.95,
< 0.01)throughoutthe ovariancycle.Plasma concentrations
of17a-hydroxyprogesteronewere alsosignificantlycorrelated
withcirculatingconcentrations ofprogesterone during the
ovariancycle (r=0.93, =205cycles, < 0.01).
The datashowclearcyclicpatternsofpregnanetriolexcretion,
with valuesincreasingfrom an overall mean of 45ng mg
creatinine(range5-172ng mg^1 creatinine) duringthe inter¬
lutealperiodto a mean of 237ng mg-1creatinine in the luteal
phase (range75-802ng mg-1creatinine).Although variability
was seen inpregnanetriolvalues, both between cycles and
between animals, the extent of this was notmarkedly different
from that seen in measurements ofplasmahormones(as evi¬
dencedbythe size of the SE bars inFig. 3). Mostimportantly
thetrend of high urinary pregnanetriol
throughoutthe interlutealperiodwas maintainedinallprofiles
and of the 20cycles withmatching progesteroneandpregna¬
netriol data, defined rises in both hormones occurred on the
same date 18 times (90%). In theremainingtwo cycles, the
rises wereseparated by one week(for example,onesample
apart).
=205 cycles,
J
concentrations
Table 2.Relationshipbetweenpredictedoestrus andmating
events in six Asianelephantsfrom one zoo from January
1989-November 1991
Nomatingwith
predictedoestrus
Mating
matching
predicted
oestrus
(%)
Total
number of
cycles
Wrong
time
(%)
No
mating
(%)
Resulting
pregnancy
84 (50)
4 (36.4)
3 (50)
3 (37.5)
2 (25)
1 (25)
17 (37.8)
1 (12.5)
4 (36.4)
1 (16.7)
1 (12.5)
3 (37.5)
0 (0)
10 (22.2)
3 (37.5)
3 (27.3)
2 (33.3)
4 (50)
3 (37.5)
3 (75)
18 (40)
1
0
0
0
2
0
3
11
6
8
8
4*
45
Overallconceptionrate: 3/17 x 100=17.6%
"This cowcompleteda full termpregnancyinJulyof 1990.
In thosecycles in whichmatingdid occur(Fig. 3), only dur¬
ingthe last(conception) cycle didmating correspondto the
time ofpredictedovulationresultinginpregnancy.In all of the
other cases,matingeither did not occur at all, or it occurred at
least a week before the time ofexpectedovulation and did not
result inconception. Furthermore, theapparent asynchrony
between behavioural and endocrine(physiological)events was
not confined to this individual,but was seen in all six adult
females in thegroupto which shebelonged (Table 2).The data
were collected for 34 months and indicated that, out of a total
of45 ovarian(presumed ovulatory) cyclesfrom six adult cows,
matingcoincided with the time ofpredictedovulation in only
17 (37.8%) cases. In theremaining62.2% of the cycles, mating
either did not occur (40%) or occurred at a timeseparated
frompresumedovulation (22.2%). Of the 17cycles in which
mating appearedto coincide with ovulation, three resulted in
conception (17.6%).Nopregnancyoccurred fromthe mis-timed
matings.
Discussion
This studyhas shown that5ß-pregnanetriolis a major urinary
progesteronemetabolite in Asianelephants.The establishment
of anenzymeimmunoassayfor the measurement of immuno¬
reactivepregnanetriolhas enabled the firstdescriptionof the
pattern of excretion of urinary progesterone metabolites
throughoutthe ovariancycle.The resultssuggestthatmeasure¬
ment ofpregnanetriolshouldprovide
method formonitoring reproductivestatus in thisspecies.
The initial GCMSfindingsofrelatively highconcentrations
of5ß-pregnanetriolin urine of female Asianelephants during
variousreproductivestates was unusual,butmay explainthe
previouslack of success inattemptingto monitorreproductive
cycleswith morecommonlymeasuredprogesteronemetabolites
suchas20a-dihydroprogesteroneandpregnanediolglucuronide.
Measurement ofurinarypregnanediolor20a-dihydroprogester-
one haveprovedto be invaluable formonitoring reproductive
a useful non-invasive

Page 7

status in a wide variety ofspecies includingblack and white
rhinos(Hindle et al, 1992),gorillas (HodgesandGreen, 1989),
giant pandas (Hodgesetal, 1984), white-tailed deer (Knox et al,
1992),killer whales(Walkerel al, 1988)andmacaques (Monfort
et al, 1986). In contrast, however, theimmunoreactivityassoci¬
ated with these metabolites in Asianelephanturine is low
and variable and theirpatternsof excretion do not correlate
with ovarian events (J.Hindle and C.Niemuller, unpublished).
Furthermore, in allsamples analysed,neitherprogesteronenor
its metabolites, pregnanedioland20a-dihydroprogesterone,
were detected bythe GCMSproceduresused. In anindepen¬
dent study,Mainka andLothrop(1990)attemptedbut failed to
measure urinary progesterone cyclesin two Asianelephants,a
findingconsistent with the above results.
Pregnanetriolis aunique major urinarymetabolite of 17a-
hydroxyprogesterone produced by reductions inringA, C-3
and C-20 (GowerandHonour, 1984). Thus, the formation of
pregnanetriolfollows aseparate pathwayfrom that used in
theproductionof20a-dihydroprogesteroneandpregnanediol
(GowerandHonour, 1984),forwhich17a-hydroxyprogesterone
is not anobligatoryintermediate (Axelrod and Goldzieher,
1960).
The measurement ofpregnanetriol, originallydiscovered in
the urine of two women withcongenitaladrenalhyperplasia
(ButlerandMarrian, 1937), is normally used as an indicator of
adrenalhyperplasiarather thanformonitoringovarian function.
Human ovaries, however, are known to be a source of itspre¬
cursor, 17a-hydroxyprogesterone (Zander, 1958, Short and
London, 1961) and measurement ofpregnanetriolfor deter¬
miningovarian function in women has beenreported (for
examplePickett et al, 1959;Fotherby, 1960; Pickett and Kellie,
1962).
The resultsreportedhere are ofnonspecificmeasurements of
unconjugated pregnanetriol immunoreactivity, as HPLC data
clearlyindicated thepresenceof other, morepolarimmunoreac¬
tive substances. These unidentified substancesmayeven be
quantitativelymoreimportantthanpregnanetriol;however,
their nature remains unknown.Althoughtheantibody cross-
reacts extensively with otherpregnanetriolisomers and with
pregnanetetrols,all of which can beexpectedto elute from
HPLC atpositionssimilar to that of the unknown, their contri¬
bution to the immunoreactive profilesdescribed cannot be
determined.
Thenature of therelationshipbetweenurinarypregnanetriol
andplasma 17a-hydroxyprogesterone promptedthe initial
measurements of 17a-hydroxyprogesterone throughoutthe
ovarian cycle of Asianelephantsin this study.Thesignificant
correlation
of circulating plasma
hydroxyprogesteronewith progesterone suggeststhat the
elephant ovaryis amajorsource of this steroid and thaturinary
pregnanetriolconcentrations are related to ovarianprogester¬
one secretion. In other mammals, lutealphase 17a-hydroxy-
progesteroneconcentrations are much lower thanprogesterone
concentrations. Furthermore, thesignificantcorrelation between
pregnanetrioland bothprogesteroneand17a-hydroxyproges-
teronethroughoutthereproductive cycle providesadditional
evidence thatpregnanetriolis a major urinary progesterone
metabolite inelephantsand that itspatternof excretion reflects
ovarian function. The results presentedheresuggestthat
elephants may have significantly higher 17a-hydroxylase
concentration of
17a-
activityinboth theovaryand theliver, whichmay explainwhy
pregnanetriolis themajor urinary progestagenmetabolite.
Thegeneralfeatures of thereproductive cycles described
in this study were baseduponreference measurements of
circulatingconcentrations ofplasma progesterone. Thus, the
cyclic changesin thepatternofplasma progesteroneconcen¬
trations indicated that the duration of thereproductive cycleof
the 11 Asianelephantsin thisstudy rangedbetween 14 and 18
weeks, which matched with earlierdescriptions (Hess et al,
1983; Plotka et al, 1988; Brown et al, 1991; Taya et al,
1991). Datapresentedfrom both the individual as well as the
composite profilesrevealed a cyclic patternofpregnanetriol
excretionhighlycorrelated withcirculating plasma progester¬
one and 17a-hydroxyprogesterone concentrations. Despite
higher variability in urinary values ascomparedwithplasma
steroid concentrations, there was
lutealphasevaluescomparedwith interlutealphasevalues.
Furthermore, 90% of the cycles analysed demonstrated
increase inpregnanetriolconcentrations concurrent with the
defined rise inplasma progesterone. Although plasma pro¬
gesteronehas to date been the most useful measurement in
determining reproductive cycles,these results demonstrate that
the measurement ofpregnanetriol closelyreflectsprogesterone
secretion andcorpusluteum function.
Althoughit isgenerallyassumed thatcyclicfluctuationsinthe
concentrations ofcirculating progesterone(or its metabolites)
reflectovulatory cycles,there is no information on theelephant
from which a cycle can be defined asendocrinologicallyor
behaviourallynormal. The resultspresentedhere indicate that
in individuals from oneparticular group, only25-50% of
apparently 'normal' ovulatory cycles
matingbehaviourduringtheperiovulatory period, despite
access to bulls ofproven fertility throughoutthis time. Further¬
more,of thecycles in whichmatingdid occur at the time of
expected ovulation, less than 20% resulted inconception.In
domesticspeciessuch ascows, horses,goatsandsheep, concep¬
tion rateper ovulatory event ishigher, varyingbetween 40
and 75%(Bristol, 1986; McDonald, 1986; Smith, M. C, 1986;
Smith, R. D., 1986). Data collected by researchers on wild
Africanelephantsin Amboseli National Park, Kenya, have
noted aconceptionrate ofapproximately75%(Moss, 1983).At
this time, the reason for thehigh percentageof non-fertile
cycles in the captive elephant population is unclear. Our
data demonstrate that in only a smallproportionofcycles
doesmatingoccur at the time ofpresumedovulation. This
behavioural/physiologicalimbalancemaybe related to the
highlyvariablepatternofoestrogensecretionduringthecycle
and theinabilityin most studies todescribe a clearpreovulatory
increase inoestrogens(Hess et al, 1983; Mainka andLothrop,
1990; Brown et al, 1991;Tayaet al, 1991). Furthermore, with
theexceptionof the studies of Hess et al. (1983)and Mainka
andLothrop (1990), the above studies lacked thematching
behavioural data for theirendocrinological descriptionsof the
elephant reproductive cycle.These two studies indicated that
theprevalenceof non-fertilecycles in thepresent study group
of cows was not an isolated event. Mainka andLothrop(1990)
described flehmen but notmatingevents of a bull towards a
cow (whodid not conceive)throughfivecomplete cycles and
comment on threematingtimes ofFebruary, July and October
in a second cow beforeconceptionoccurred. Likewise, Hess
a clear fivefold increase in
an
were associated with

Page 8

et al. (1983) described 15reproductive cyclesdivided between
six cows, from whichonly twoconceptionsoccurred. The
reason for theapparent infertility is not understood.
The incidence of asynchronous mating with respect to
presumedovulation observed in thisstudydoes notappearto
have been described before. The lack ofmatingormistiming
ofmating mayhave been overestimated as animals were not
constantlyobserved over 24 h, thusmatings occurringatnight
would not have been observed. The mostimportantexternal
cue receivedbythe male from the femalesignifyingonset
of oestrusappearsto be smell as determinedbythe flehmen
responses (Jainudeen et al, 1971; Hess et al, 1983). Miscueing
of thesepheromonal signalson thepartof the femalemay
be another reason for mis-timedmatingevents. Theremaybe
a link between thetiming of ovulation and miscueing of
pheromonal signals,whichmight explaintheasynchronous
matingevents.
Insummary,thispaperhas identified 5ß-pregnanetriolas a
major urinaryprogesteronemetabolite in Asianelephantsand
described its measurementthroughoutthe ovariancycleof this
species.Its measurement in urine should not only enable
routine, non-invasivemonitoringofreproductive cycles incap¬
tivity,but alsoprovide potential applicationtowardsassessing
reproductivestatus in wild Asianelephants.
The authors
Biochemistry, UCL for the donation of thepregnanetriol-glucuronide
antibody,
Universität, Vienna, for the enzymelabels and S. Scott (MAFF)for
[3H]17-hydroxy-5ß-pregnane-3a,20a-diol. They also owe thanks to
]. Honour and M.Schneider, MiddlesexHospitalfor instruction and
use of the GCMS and are indebted to M.Smith,B. Bliss and thekeep¬
ingstaff at Port Lympne Zoo, B. Harman and theelephantstaff at
London Zoo, C.Grayand staff at the African Lion Safari and D.Moore
and C. Doyle and staff of Burnet Park Zoo for their invaluable assist¬
ance insamplecollection.Finally,thanksgoto J. Hindle and D. Green
for advice and technicalsupport.Thisprojectwassupported by
private grantfrom Howlett's and PortLympne Estates Ltd and from
theUniversityFederation of Animal Welfare.
aregrateful
to
P. Samarajeewa, Department of
E. Mosti, Institut für Biochemie, Veterinärmedizinische
a
References
Axelrod
hydroxyprogesteroneand its relation tocongenitaladrenalhyperplasia
journal ofClinicalEndocrinology20 238-252
Brannian JD, GriffinF, PapkoffH and Terranova PF (1988)Short andlong
phasesofprogesteronesecretionduringthe oestrouscycleof the African
elephant(Loxoáonta africana) Journal of ReproductionandFertility84 357—3*65
Bristol F (1986) Estrus synchronizationin
Theriogenology2:Diagnosis, Treatment and Prevention of ReproductiveDiseases
in Small andLarge Animals, pp661-664 Ed. DA Morrow. W Saunders
Company, Philadephia
BrownJL,CitinoSC,Bush M, LehnhardtJandPhillipsLG(1991) Cyclic patterns
ofluteinizinghormone,follicle-stimulating hormone, inhibin,andprogester¬
one secretion in the Asianelephant (Elephas maximus) Journal of Zoo and
Wildlife Management22 49-57
Butler GC and Marrian GF(1937)The isolation ofpregnane-3,17,20-triolfrom
theurine of womenshowingtheadrenogenital syndromeJournal ofBiological
Chemistry119 565-572
Fotherby (1960)Excretion ofpregnanetriol duringthe normal menstrualcycle
British MedicalJournal1 1545-1546
Gower DB and Honour JH (1984) Part I. Steroid catabolism andurinary
excretion. Part 2.Biliaryexcretion andenterohepaticcirculation. In Steroid
Biochemistry, pp349-407 Ed.HJLMakin. Blackwell ScientificPublications,
Oxford
LR
andGoldzieher
JW
(I960)
The
metabolismof
17a-
mares. In Current Therapy in
GrossTS,Patton M,ArmstrongDL and Simmons LG(1991)Estrus detection and
synchronizationin Africanelephants (Loxodonta africana): vaginal cytology
and serum endocrineprofiles Biology of Reproduction44(Supplement 1)
Abstract 31
Hess DL, Schmidt AM and SchmidtMJ (1983) Reproductivecycleof the Asian
elephant {Elephas maximus)incaptivity Biology ofReproduction28 767-773
HindleJE,Mosti E andHodges JK (1992) Measurement ofurinary oestrogens
and20a-dihydroprogesterone duringovariancyclesof black(Diceros bícornís)
and white (Ceratotherium simum)rhinoceroses Journal of Reproductionand
Fertility94237-249
Hodges JK (1992) Detection of oestrous cycles andtimingof ovulation
Symposium oftheZoological Society ofLondon 64 73—88
Hodges JKand Green DI(1989)Asimplified enzymeimmunoassayforurinary
pregnanediol-3-glucuronide: applicationtoreproductiveassessment inexotic
speciesJournal of Zoology219 89-99
Hodges JK,BevanDJ,CelmaM,HearnJP, Jones DM, KleimanDG,Knight JA
and Moore HD(1984) Aspectsofreproductive endocrinologyof the female
giant panda (Ailuropodamelanoleuca)incaptivitywithspecialreference to the
detection of ovulation andpregnancyJournal ofZoology203 253-267
Jainudeen MR,Eisenberg JFand Tilakeratne (1971) Oestrouscycleof the
Asianelephant, Elephas maximus,incaptivity Journal of Reproductionand
Fertility27 321-328
Kirkpatrick JF, Lasley BL and Schideler SE(1990) Urinarysteroid evaluation to
monitor ovarian function in exoticungulates.VII. Urinary progesterone
metabolites in theEquidaeassessedby immunoassayZooBiology9 341—348
Knox WM, MillerKV,CollinsDC,Bush PB, Kiser TE and Marchinton RL(1992)
Serum andurinarylevels ofreproductivehormones associated with the
estrous cycle in white-tailed deer (Odocoileus virginianus) Zoo Biology 11
121-131
LiebermannS, Erlanger BF, Beiser SM andAgate FJ, )r (1959) Steroid—protein
conjugates:theirchemistry,immunochemistryandendocrinological properties
RecentProgress in Hormone Research 15 165—200
LoskutoffNM,WalkerL, Ott-Joslin JE, RaphaelBL andLasleyBL(1986) Urinary
steroid evaluations to monitor ovarian function in exoticungulates:II.
Comparison between thegiraffe (Giraffa camelopardalis)and theokapi
(Okapia johnstoni)ZooBiology5 355—361
McDonald MF(1986)Estroussynchronizationand control of the estrouscycle.
In CurrentTherapyinTheriogenology2:Diagnosis, Treatment and Preventionof
ReproductiveDiseases in Small andLargeAnimalspp887—889 Ed. DA Morrow.
WB SaundersCompany, Philadelphia
Mainka SA andLothropCD,Jr (1990) Reproductiveand hormonalchangesdur¬
ingthe estrouscycleandpregnancyinAsianelephants (Elephas maximus)Zoo
Biology9 411-419
Monfort SL, Jayaraman S, Schideler SE, Lasley BL and Hendrickx AG (1986)
Monitoring ovulation andimplantation
(Macaca fascicularis) throughevaluations ofurinaryestroneconjugates
andprogesteronemetabolites: atechniquefor the routine evaluation of
reproductive parametersJournal ofMedicalPrimatology15 17-26
Moss CJ (1983) Oestrous behaviour and female choice in the Africanelephant
Behaviour 86 167-196
Pickett MT and Kellie AE(1962) Urinary excretion ofpregnanetriolin the men¬
strualcycle.I. Additional evidenceofidentityActaEndocrinologica41 129—134
Pickett MT, Kyriakides EC, Stern MI and Sommerville IF (1959) Urinary
pregnanetriol throughoutthe menstrual cycle: observations on ahealthy
subject and an adrenalectomised woman The Lancet 2 829—830
Plotka ED, Seal US, ZarembkaFR,SimmonsLG,TeareA,PhillipsLG,Hinshaw
KC and Wood DG(1988)Ovarian function in theelephant: luteinizinghor¬
mone andprogesterone cycles in African and Asianelephants Biology of
Reproduction38309-314
PooleJH, Kasman LH, RamsayEC and Lasley BL (1984)Musth andurinary
testosterone concentrations in the Africanelephant (Loxodonta africana)
Journal ofReproductionandFertility70255-260
Quilliam MA and Westmore JA (1980) Massspectralstudies ofpartialand
mixedsilyl derivatives of 5ß pregnane3a,17a,20a-triol European Journal of
MassSpectrometry, Biochemistry,Medical and Environmental Research 1 53—56
Ramsey EC, LasleyBL and Stabenfeldt GH(1981) Monitoringthe estrouscycle
of the Asianelephant (Elephas maximus) using urinary estrogensAmerican
Journal of VeterinaryResearch 42256-260
Samarajeewa and Kellie A (1985)The radioimmunoassay ofpregnanetriol
3a-glucuronideJournal of Steroid Biochemistry22 843-849
ShackletonCHL, HonourJW,Dillon M and Milla (1976) Multicomponent gas
Chromatographie analysisofurinarysteroids excretedbyan infant with a
defect in aldosteronebiosynthesisActaEndocrinologica81 762-769
in thecynomologus macaque

Page 9

Shackleton CHL, Taylor NF and Honour JW (1980)
Chromatographie Profiles ofNeutral UrinarySteroids in Health and Disease.
Packard-Becker,Delft
ShawHJ,Hillier GS andHodgesJK (1989) Developmental changesinluteinising
hormone/human chorionicgonadotrophin. Steroidogenic responsivenessin
marmosetgranulosacells: effects of folliclestimulatinghormone and
androgens Endocrinology124 1669-1677
Short RV and London DR(1961)Defectivebiosynthesisof ovarian steroids in
the Stein-LeventhalsyndromeBritish MedicalJournal1 1724-1727
Smith MC(1986) Synchronizationof estrus and the use ofimplantsandvaginal
sponges.In CurrentTherapy in Theriogenology2:Diagnosis, Treatment and
Prevention ofReproductiveDiseases in Small andLargeAnimalspp582 Ed. DA
Morrow. WB SaundersCompany, Philadelphia
An Atlas of GasSmith RD (1986) Estrus detection. In CurrentTherapy in Theriogenology 2:
Diagnosis, Treatment and Preventionof ReproductiveDiseases in Small andLarge
Animalspp153-157 Ed. DA Morrow. WB SaundersCompany, Philadelphia
Taya K,KomuraH,KondohM,OgawaY, NakadaK,WatanabeG,SasamotoS,
Tanabe ,SaitoK, Tajima and Narushima E(1991)Concentrations ofpro¬
gesterone, testosterone andestradiol-17ß in the serumduringthe estrous
cycleof Asianelephants (Elephasmaximus) Zoo Biology10299-307
WalkerLA,Cornell L, DahlKD,CzekalaNM,DargenCM,JosephB, HsuehAJW
and Lasley BL(1988) Urinary concentrations of ovarian steroid hormone
metabolites and bioactivefollicle-stimulatinghormone in killer whales
(Orcinas orchus)Biology of Reproduction39 I0I3-1020
ZanderJ (1958)Steroids in the humanovary Journal ofBiological Chemistry232
II7-122