1. family and parenting
2. motherhood
3. pregnancy

BIOLOGY OF REPRODUCTION 52, 814-823 (1995)
Daily Patterns of Pituitary Prolactin Secretion and Their Role in Regulating Maternal
Serum Progesterone Concentrations across Pregnancy in the Djungarian Hamster
(Phodopus campbelli)"
HEATHER E. EDWARDS, CATHARINE J. REBURN, and KATHERINE E. WYNNE-EDWARDS
2
Queen's University, Department of Biology, Kingston, Ontario, Canada K7L 3N6
ABSTRACT
Profiles of serum prolactin (PRL) and progesterone (P4) were determined in repeatedly (every 2 h) sampled female Djungarian
hamsters (Pbodopus campbelli) over a 24-h period on Days 3, 6, 9, 12, and 15 of the 18-day pregnancy. The first half of pregnancy
was characterized by significant surges of PRL within a 2-h period around dawn (0500 h) and dusk (1900 h), with some females
also showing a weak midday surge at 1300 h. By Day 9, dusk and midday surges were absent, but the dawn surge remained at
its initial amplitude. On Day 12, no PRL surges were seen. Resumption of both the dusk and dawn PRL surges occurred on Day
15 of gestation. Considerable interindividual variability in the amplitude of PRL surges, the timing of PRL surges, and the number
of surges per day was detected and would complicate any assessment of PRL levels based on single samples per female. Serum
P4 concentrations were 8-10 ng ml- ' before doubling on Day 15. A 3-day treatment of 50 and 300 g bromocryptine (CB 154;
on Days 13-15) effectively suppressed PRL during late pregnancy (Day 15) but did not alter serum P4 concentrations or interfere
with parturition. Therefore, surges of PRL are not an essential luteotropic stimulus during late gestation. Compared to oil-injected
controls, CB 154-treated females had a higher incidence of infanticide postpartum. Growth rates of the pups, mammary gland
development, and successful delivery of milk to pups, however, did not differ between groups. Further studies will be required
to determine the function of late-gestation PRL surges.
INTRODUCTION
pregnancy. The lack of measurable cross-reactivity with placental extracts or placental incubations and the recognition
of large quantities of PRL in pituitary incubations confirmed
that the PRL was of pituitary origin [1]. There are two hypotheses to explain this result. The first is that pituitary PRL
remains the primary luteotropin throughout pregnancy but
that the PRL surges shift in time such that the once-a-day
sampling paradigm of Edwards et al. [1] missed detecting
them during midgestation. Such temporal shifts in the timing of PRL surges occur in the golden hamster [7]. According to the second hypothesis, although placental luteotropins maintain the CL of midgestation, pituitary PRL surges,
of unknown function, are reestablished during late gestation. PRL is important for the initiation and/or maintenance
of lactation in many mammals (see [8] for review), and
suppression of PRL results in reduced milk yield and/or
reduction in offspring growth in several species [9-11].
Therefore it is possible that late-gestation PRL is mammotropic or lactogenic in function.
However, late-term PRL surges could also be luteotropic
and reflect pituitary rather than feto-placental regulation of
CL function. Preimplantation pregnancy block [12] occurs
at the level of the hypothalamic-pituitary axis through the
suppression of PRL release [13]. In some species, pregnancy
blocks can also occur during late pregnancy [14]. Edwards
et al. [1] postulated that temporary decreases in serum P4
on Days 6 and 12 of the Djungarian hamster pregnancy might
represent sensitive windows for the termination of maternal investment in the current reproductive attempt. Such
postimplantation pregnancy blocks would be adaptive for
this species, since short-lived rodents in extreme habitats
experience strong selection for rapid reproductive cycles
Maintenance of pregnancy in the Djungarian hamster,
Pbodopus campbelli, is critically dependent on the continued secretion of progesterone (P4) by ovarian CL, even during late gestation, since bilateral ovariectomy during that
time causes P4 levels to fall and pregnancies to fail [1]. As
in the rat [2] and mouse [3], extension of the CL life span
during early pregnancy in the Djungarian hamster is linked
to the establishment of a coitally induced pituitary prolactin
(PRL) secretion pattern that consists of surges occurring daily
at dusk and dawn [4].
In the absence of a viable feto-placental unit, this neuroendocrine response spontaneously terminates after Day
6 of gestation (as measured by the loss of dusk surges of
pituitary PRL in pseudopregnant females), resulting in a drop
in serum P4, regression of the CL, and termination of the
overall reproductive effort [5]. The loss of dusk PRL surges
after Day 6 also occurs in pregnant females; however, CL
persist and secrete large quantities of P4, maintaining high
levels in the serum [1]. As in the mouse and rat, it would
be expected that luteotropic hormones produced by the
placenta (i.e., placental lactogens) would maintain the functional integrity of the CL after pituitary PRL surges end in
midgestation [6].
However, in the Djungarian hamster, dusk surges of immunoreactive pituitary PRL unexpectedly resume during late
Accepted November 29, 1994.
Received July 14, 1994.
'This study was supported by an NSERC research grant to KE.W.-E. and a postgraduate NSERC scholarship to H.E.E.
2Correspondence. FAX: (613) 545-6617.
814
PROLACTIN SECRETION DURING GESTATION IN PHODOPUS
and early detection of failed reproductive attempts [15].
Evolution of the ability to block pregnancy during late gestation may require maternal (pituitary) rather than feto-placental control of CL function.
The present study tested the hypothesis that the pituitary
PRL secretion pattern induced by coitus is maintained
throughout pregnancy and 1) is necessary for the maintenance of ovarian CL and high serum P4 and 2) affects the
lactogenic potential of the dam. A technique for repeated,
small-volume sampling of individuals without induction of
handling responses was developed to minimize the number of animals used. This technique also permitted, for the
first time, an analysis of temporal variability between females in the PRL secretion pattern.
MATERIALS AND METHODS
Animals
The animals used were Phodopus campbelli raised in our
breeding colony, for which the details of origin have been
previously described [16]. The colony was maintained at 18
+ 1IC on a 14L:10D schedule, with 0000 h corresponding
to the middle of the dark phase. Dim red light provided
illumination during the dark portion of the light cycle. Animals were housed in 27 x 21 x 14-cm cages with wood
chip bedding and food and water given ad libitum.
All females used in this study (n = 68) were adult virgins
between 90 and 110 days of age (mean = 98 days). All males
(> 120 days) had proven their fertility by siring at least one
litter. Pairing occurred in the early afternoon, and pairs were
observed each day of the estrous cycle (4 days [17]) for
behavioral receptivity on proestrus. Mating occurred during the 2-h period surrounding lights-off (dusk) and was
confirmed by the direct observation of at least one behavioral ejaculation [16]. Day 1 of pregnancy began at midnight
(0000 h) following the evening on which mating occurred.
All pairs remained together after mating to minimize pregnancy-blocking responses [12]. After matings, behavioral
observations continued on a daily basis to screen for successful pregnancy initiation. Animals that remated on Days
4-5 or Days 9-11 had pregnancy blocked or were pseudopregnant [5] and were not included in analyses.
Sample Collection
After mating, females were scheduled for repeated sampling every 2 h for a 24-h period on one of Days 3, 6, 9,
12, or 15 of the 18-day pregnancy (n = 6-8 females per
day), with sampling beginning at 0100 h. The apparatus allowed four pairs to be sampled during any 24-h period. All
samples were collected during January through March of
1994. Twelve hours before sampling, pairs were removed
from their home cage, transferred to a 16 x 26 x 30-cm
experimental cage that had opaque siding and two vents at
its base, and allowed to acclimatize. Animals were main-
815
tained in a continuous air stream that was delivered through
one vent except during sampling, when it was turned off
via a remote switch. The experimental apparatus was kept
under a fume hood in the breeding colony, where routine
maintenance by humans was common.
As handling stress is a confounding variable during repeated sampling, such effects were controlled for by anesthetizing pairs briefly (20 sec) with vaporized AErrane
Isoflurane before each sample was taken. The anesthetic
was delivered to the pair through the second vent and rapidly cleared from the cage when air flow through the first
vent resumed. Since Isoflurane is denser than air, it accumulated at the level of the hamsters, and 30 sec was required to induce anesthesia. The continuous air stream and
fume hood ensured that residual anesthetic did not remain
in the bedding. The fume hood was constructed of clear
acrylic, which allowed for normal physiological entrainment by the light-dark cycle. For each sample, 1.75 ml Isoflurane was delivered through tubing 1 m long (inner diameter = 1.5 cm) connected to an Ohmeda Isotec 3
Isoflurane vaporizer. Vaporization was performed with
medical grade oxygen as the carrier gas at a flow rate of
3.5 L-min'.
Isoflurane is a respiratory depressant and was chosen for
this experiment on the basis of 1) fast induction and recovery owing to its insolubility in the blood and its ability
to be almost completely eliminated in exhaled air [18], 2)
margin of safety, which is greatest among all the inhalant
anesthetics with respect to the cardiovascular system [18],
3) absence of known toxicities and lack of residual effect
in tissues [19], and 4) mode of administration, i.e., selfadministration such that the animal doses itself depending
on its respiratory rate [19].
Blood Collection
Blood samples from the orbital sinus of the right eye
were collected in 7 5-pLl heparinized microcapillary tubes
and immediately centrifuged, and serum was stored at -20°C
until assayed for PRL and P4 content. With 12 repeated samples, a total of 900 il of blood was removed per femalea total volume comparable to that of single samples obtained in previous studies, which had no adverse effect on
reproductive cycles or behavior [17]. After 24 h of sampling,
pairs were removed from the experimental cage and returned to their home cage. Only females that gave birth on
Day 18 were used for analyses. The number of pups for
each litter was recorded.
Experimental Controls
Sample collection and apparatus. Animals were prevented from associating the experimenter with sampling by
1) remote delivery of the anesthetic, 2) opaque siding of
the cage, and 3) being allowed to remain within.the general animal colony. However, the odor of Isoflurane was a
816
EDWARDS ET AL.
potential conditioning signal that could affect stress responses and serum PRL concentrations of the hamsters. It
was also possible that the odor was initially stressful but
that the animals became acclimatized with later sampling.
To test for these effects, two groups of females were sampled 12 h out of phase relative to the Day 6 experimental
females. Specifically, sampling began at 1300 h on Day 5
and ended at 1100 h on Day 6 of gestation in one group,
but began at 1300 h on Day 6 and ended at 1100 h on Day
7 of gestation in the second group. Serum PRL concentrations were compared between the groups to test for pattern
differences and potential "initial sample" effects.
To minimize the potential for artifactual changes in PRL
levels in a cohort of females sampled together (n = 4),
synchrony in gestation age for these 4 females was avoided.
In addition, sample sizes ranged between 6 and 8 females
per day.
Hematocrit. Another potential confounding variable was
changing blood hematocrit levels and hemodilution or
hemoconcentration associated with the repeated disruption
of the blood clot in the orbital sinus. Blood hematocrit was
measured in an additional 20 females, which were repeatedly sampled every 2 h over a 14-h period (sampling occurred between 0500 h and 1900 h). Immediately after the
eighth sample (from the right sinus), blood was also drawn
from the left orbital sinus for comparison between eyes. In
order to determine the full recovery period for the percentage hematocrit, single samples from the right eye were
taken at 48 h and 120 h after the last (eighth) sample at
1900 h.
Role of PRL during Late Gestation
PRL as a luteotropin. A group of 5 females repeatedly
sampled on Day 15 were treated beginning 3 days earlier
(Days 13-15 at 1600 h) with daily 0.6-ml injections (s.c.)
containing 50 ig of CB 154 (bromocryptine mesylate, a dopamine agonist [Sigma Chemical Co., St. Louis, MO; B2134])
to suppress endogenous PRL levels. A single 50-gig injection
in male mice (-30 g BW) suppresses PRL levels for 24 h
[20], and both a 3-day treatment of 1 mg CB 154 per 300g rat (equivalent to 100 ig per 30-g animal) and a single
300-gig injection reduce surge levels of PRL in early-pregnant rats and cause abortion of entire litters [21, 22]. Thus
it was expected that a 3-day treatment of 50 gg would similarly reduce PRL levels in the Djungarian hamster. CB 154
was first dissolved in 95% ethanol and then suspended in
sesame oil [20]. To subsequently evaporate ethanol, the
mixture was stirred for 10 h. The suspension was stirred
for 1 h before injection to ensure homogeneity of the injection sample. Before the first injection, pregnancy was
confirmed in all females by external palpation of the abdominal wall. Females that were not pregnant were eliminated from the study. After injection, females were monitored several times daily for signs of pregnancy failure
including vaginal bleeding and/or decreased activity levels
[1].
PRL as a mammotropin and/or lactogen. The hypothesis that late-gestation PRL surges at dusk are necessary
to initiate lactation was tested by using three additional
groups of females, which were treated daily at 1600 h on
Days 13-15 with either 1) oil injections, 2) 50-g injections
of CB 154, or 3) 300-gig injections of CB 154 prior to a
single 75-1l blood sample taken at 1900 h (dusk) on Day
15 of pregnancy. The 300-11g dose was used for comparison
with the 50-g dose of CB 154 to ensure that PRL levels
were sufficiently lowered to detect a response. The presence (or absence) and degree of teat development (i.e.,
amount of teat swelling and visible protrusion through the
ventral pelage relative to that of a midpregnant [Day 10]
female) on Days 13 and 15 were recorded for all females.
For females that successfully gave birth on Day 18, the following were recorded: 1) litter size at birth and on each
consecutive day of lactation, 2) individual pup weight on
postnatal Days 0-6, and 3) the presence or absence of milk
visible in the stomach of the pups [23]. Milk was quantified
into three categories: 1) low amounts, 2) moderate amounts,
or 3) plenty of milk visible through the stomach wall. The
weight gained over this time was used as an index of the
dam's lactational ability [9]. Both parents remained with the
pups during lactation.
RIA
The RIA procedure for P4 (progesterone antibody no. 337,
supplied by G.D. Niswender, Colorado State University, Fort
Collins, CO) has been validated previously in this species
[17, 24]. All samples were assayed in duplicate at 5 gl1. The
conservative limits of sensitivity were defined as 85% binding at 13.2 pg/tube and 20% binding at 134.3 pg/tube, giving a range of 2.4-32.7 ng ml - '. Inter- and intraassay coefficients of variation were 9.4 and 11.8%, calculated from
duplicate determinations of a serum pool in 12 assays.
The RIA procedure for PRL measurements in the Djungarian hamster was modified from those of Dr. A.F. Parlow
(Pituitary Hormones and Antisera Center, Harbor-UCLA
Medical Center, Torrance, CA) and Dr. F. Talamantes [7].
Validation for use in this species is available in Edwards et
al. [1] and from Dr. Parlow. The protocol used anti-hamster
PRL (rat; #AFP-7472988) as the primary antibody, hamster
PRL (#AFP-10302-E) as the reference preparation, and goat
anti-rat gamma globulin (titer P-3, lot #9TA05Y; Antibodies
Inc., Davis, CA) as the second antibody. For purposes of the
present study, values are reported as ggml - 1 (relative potency of 1.00), although the heterologous golden hamster
antibody and reference preparation may not provide absolute determination of PRL concentrations in Djungarian
hamsters.
This laboratory has recently changed the method of PRL
iodination relative to that used in previous studies [1, 4, 5].
Specific alterations in the iodination procedure are the use
PROLACTIN SECRETION DURING GESTATION IN PHODOPUS
of 1) a much smaller reaction chamber in the paraffin well
than that described by Butt [25], 2) a larger reaction volume
(50 i1) into which the 3 pIl (300 pCi) of NaI' 25 was diluted,
and 3) a 0.05 M sodium phosphate (pH = 7.6, Na 2PO 4 +
NaH2 PO4 ) diluent into which choramine T (Sigma Chemical
Co.; C-9887) was dissolved prior to the gaseous phase reaction with the NaCI-saturated filter paper disk. With the
use of this gentler procedure, the iodinated PRL peak showed
markedly improved binding to the primary antibody. This
in turn increased the range of the assay (85-20% binding)
from 15-390 pg/tube [1, 4, 5] to 33-1880 pg/tube with the
LOGIT slope unchanged. The primary result of this change
in assay sensitivity was in the amplitude of PRL peaks resolved. As sample volumes were small (20-25 p1I serum),
capillary tubes were difficult to reseal, and freeze-thaw cycles
were rigorously avoided; all PRL samples were assayed in
triplicate (at 1 pI), but it was not possible to reassay samples that were outside the sensitivity range. Therefore, values outside assay limits were entered into the calculations
as the limiting values, and statistical analyses were chosen
to reflect this necessity. The inter- and intraassay coefficients of variation were 13.6% and 9.6%, calculated from
triplicate determinations of a blood pool in 12 assays.
StatisticalAnalyses
Mean values are expressed + SEM. Serum P4 concentrations were normally distributed and were analyzed by parametric techniques. A repeated measures one-way ANOVA
was used for comparisons between times on a given day
of gestation. A global mean serum P4 concentration per female over the 12 samples was calculated and used for comparisons between 1) different days, 2) Day 6 females and
out-of-phase control females, and 3) Day 15 pregnant and
CB 154-treated females. The first two comparisons used a
factorial one-way ANOVA, while the latter used an unpaired
t-test. All parametric tests were two-tailed with a critical significance level of 0.05.
Unlike serum P4, serum PRL concentrations were not
normally distributed and, as an intrinsic feature of the RIA,
had variances that increased at both the high and low ends
of the range. Rounding of out-of-range values to the calculated assay limit artificially reduces the variance associated with precisely those samples that intrinsically have the
highest variances. Parametric statistics are sensitive to this
artifact and the truncation of any normal distribution in
transformed data. Therefore, more conservative nonparametric techniques that employ ranking and do not assign
increased weight to extreme values were used for analyses.
As there is no nonparametric equivalent to the repeated
measures ANOVA, pairwise Wilcoxon matched-pairs, signedrank tests were used to retain individual female identity and
to permit comparisons between time points within a single
day of sampling. Since such analyses involved 66 comparisons per day, the typical significance level of 0.05 would
have resulted in an unacceptable level of type III statistical
817
errors. Instead, a critical significance level of 0.0125 was
used. For the averaged values, any sample significantly different from the nadir sample for those animals was considered to represent a PRL surge.
For comparisons of PRL concentrations between 1) different days, 2) Day 6 and out-of-phase control females, and
3) Day 15 pregnant and CB 154-treated females, independence between values was retained by calculating a mean
concentration for an individual female (the average of her
12 samples) and using that value in analyses. The first two
comparisons used a one-way factorial ANOVA, while the latter used an unpaired t-test. Provided that the overall ANOVA
was significant, post hoc analyses used the Fisher PLSD
(Probable Least Squares Difference) test. All of these tests
had a critical significance level of 0.05.
Mean PRL concentrations were associated with large variances. These could have been the result of 1) similar patterns but different absolute levels of hormone in different
individuals, 2) similar absolute hormone levels across females but variation in the timing of peak and nadir values,
3) fundamental differences between females in the occurrence of peak values, or 4) a combination of the above. The
repeated sampling of individual females allowed direct assessment of interindividual variability in PRL concentrations
and secretion patterns. As the amplitude of PRL peaks differed across females but every female had at least one sample at the lower limit of the assay, peaks were defined within
each female's profile as 70% of the highest level detected
(i.e., the highest PRL sample = 100%). Thus every female
had, by definition, at least one peak (except those for which
the value never rose above the lower limit of the assay).
The total number of peaks was counted for each female
and used in analyses.
For each litter, mean pup weight per day was calculated
and used in analyses. Pup weights were normally distributed. Pup weights from CB 154-treated females were compared against those for controls for each lactation day via
a repeated one-way ANOVA with a critical significance of
0.05.
RESULTS
Of the 68 females used in this study, pregnancy-blocking
responses occurred in 10.3%. All remaining females successfully gave birth on Day 18, with no significant difference found in litter size between females or experimental
groups (mean = 6.02
0.02 pups; range = 3-8). With
repeated sampling, hematocrit decreased from 63.4
0.7
in the first sample to 44.3
0.8 in the seventh sample and
beyond. No laterality between eyes in hematocrit was detected. Five days were required for hematocrit to recover
to initial values after the 16-h sampling.
Serum Levels of PRL and P4
Figure 1 shows the 24-h patterns found in serum levels
of PRL across pregnancy. On Days 3 and 6, significant peaks
818
EDWARDS ET AL.
Day 3
.I
r.P
Day 6
ii::?..
.?. -
::
t
E
c
r-Ir-
U
Day 9
,LO
2
.9to
. . . . .· .
bo
IC-,
8
0
2
4]
E
. .
. .
.- . .- .-
-5> :.i..::-?:~:,:~:.:.:-:.:&5~O
330
0D
12
Day
.@Ba
a
..
2
_
10
~~
~~~~~~~~~~~~~~~~
i.~~~~~~~~~~~~~~~~~~~~~~~~~~~
.?B
..........
BB
~~
?: 4?4.:t~~~~~~~~~
B:
aBX~~~~~~~~~~~g,
1
n.@0?a
, .....
Ba
55
.....,,,Bl;:,a,,
,,:',Bii:
Bi
a~~~~~~~~~~~~~~~~~~~~~i-
?:
iil.ffi0Ei,
A
o!
BBB
·
i
i
·
-i
,
,
,.
,
igii15
·.
·
Day
i.
ii ·
ie
I
.
.·
3
3,
,--
220
iii :
il???
:<S;
+/
w
1
O'
???
.
ia~~~~~~ia
'--@B''
-5i:itiiii::
~ @000?ili
ii40|ii
i~~~~~~~~::::::::--:-:-iliiii
a
,i',444ii~~~~j
i,ii titititt:·:
a . .?B i~~~~~~~~~~ililli
Pi 5 @W:
:!:g gggggg
?g
i~~i Bililil? | CaaB
°
a
iji ,i~iiiititii
0;it
: s?
'itiiiiii
;<@
i~~~~~li~~i
i~~~~~lX
E~~:::
5.Li.
i??
4 .EHi
O~~~~~~~~~~~~~~~~~~~~~~~~~
0
4
8
12
16
20
24
Time (hours)
FIG. 1. Mean SEM PRL concentration in the serum over a 24-h period on Days 3, 6, 9, 12, and 15 of the 18-day pregnancy. The number of
repeatedly sampled females for each day is shown in parentheses. Asterisks denote significant peaks of PRL relative to nadir values indicated with
(t) (Wilcoxon matched-pairs signed ranks, p < 0.0125). Shaded areas depict
the dark phase of the 14L:10D photoperiod.
occurred within 2 h of lights-on (dawn) and 2 h of lightsoff (dusk). A diurnal elevation in mean serum PRL was found
at 1300 h on both days, but was not statistically significant
because of the large variance between females (range: 0.033
Lgml-'-1.744 ,ag-ml-'). Peak levels of PRL were also ob-
0
4
8
12
16
20
24
Time (hours)
FIG. 2. Mean + SEM P4 concentration in the serum over a 24-h period
on Days 3, 6, 9, 12, and 15 of pregnancy. Sample sizes are shown in Figure
1. No significant differences between times within a day were detected.
Shaded areas depict the dark phase of the 14L:10D photoperiod.
served near dawn on Day 9, although the dusk surge was
absent. There were no PRL peaks on Day 12 of pregnancy.
On Day 15, both dawn and dusk surges were present (z >
2.20, p < 0.0125). On average, surge levels that were statistically significant were 0.90 + 0.07 Ig ml - ' (n = 69).
Average levels of PRL per day changed significantly across
gestation (F4 ,34 = 4.95,p < 0.01), with the value for Day 12
819
PROLACTIN SECRETION DURING GESTATION IN PHODOPUS
- 2- a,
lav 5
[av 6
7
flav
---
.:
Day3
I-.
2
_
I
'7
T
T
I
E
T\
tJ
0
t
11
.1
-;1
i1
1'
.
-
1 I
:
I
I
1
.
1
1
1
:J
I.
1
Time (hours)
FIG. 3. Mean SEM PRL concentration in the serum over a 24-h period from females sampled 12 h out of phase on Days 5/6 (n = 6) and 6/
7 (n = 7). Asterisks denote significant surge peaks of PRL relative to all
nadir values indicated with (t) (Wilcoxon matched-pairs signed ranks, p <
0.0125). The vertical line within the 10-h dark phase is provided for ease
of comparison with the Day 6 data in Figure 1.
.. .
being significantly lower than those for all other days in
post hoc analyses (Fisher PLSD, p < 0.05).
Figure 2 shows the 24-h patterns found in serum levels
of P4 across pregnancy. No significant changes were detected within any day (p > 0.26 in all cases). Correspondingly, there was no significant correlation between serum
concentrations of PRL and P4 within a given female (R =
0.08). Mean P4 concentrations varied significantly across
gestation (F4, 30 = 27.34, p < 0.01), with the value for Day
15 being significantly higher at 24.2
0.9 ng-ml-l than that
of any of the other days (average 8.9 + 0.2 ng ml-'). Pooled
data across Days 3-12 of gestation showed significant variation between those females in mean serum P4 concentrations (F28, 347 = 8.42, p < 0.01), with individual females exhibiting values consistently higher or lower than the mean
in successive samples. Conversely, late-pregnant females (Day
15) had higher absolute concentrations of P4 and higher
standard errors on the mean values, but interindividual
variability did not contribute significantly to this result. Individual females showed a range of P4 levels during the day.
Experimental Controls
PRL levels from midday on Day 5 through midday on
Day 7 are shown in Figure 3. All initial values at 0100 h in
Figure 1 were low, leading to a potential concern that the
sampling paradigm was affecting PRL levels. The out-of-phase
females not only replicated the low value at 0100 h on Day
6 and the overall pattern across Day 6 (correlation, R =
0.77), but began with an initial value that was a significant
peak at 1300 h on Day 5 (p = 0.009) and showed a second
initial value over 0.75 RIg-ml-' at 1300 h on Day 6. Matching
the dusk and dawn PRL peaks on Days 3 and 6 in Figure
1, significant peaks were detected at each 2100 h and 0300
h on Days 5, 6, and 7 (p = 0.009).
Interindividual Variability
Figure 4 shows individual PRL serum profiles over 24 h
for each sampling day. Each female had nadir values at or
a
To
0
i
-
14
Day15
I
i
(...I
2
4 .1 10
14
1.
Time (hours)
2
Day 9
......:.,
.........
2i
ii ? i
2
Z
10
14
1
22
Time (hours)
FIG. 4. Individual profiles of serum PRL over 24 h on Days 3, 6, 9, 12,
and 15 of pregnancy. Asterisks denote significant surge peaks of PRL for
each female (defined as > 70% of the maximum PRL concentration for that
individual). Shaded areas depict the dark phase of the 14L:10D photoperiod.
820
EDWARDS ET AL.
CB-154 Treatment
(N=5)
U.
1.0
'--
~ '0.5
30S
e
*i .
o
--
(N=5)
30
9
E
m
20
Ioo
.............
· ·
II
0
4
8
·
·
12
16
Time (hours)
FIG. 5. Mean + SEM serum concentration of (a) PRL and (b) P4 over
a 24-h period on Day 15 of pregnancy after three daily injections of 50 .g
of CB 154 {(Days 13-15). The number of repeatedly sampled females is shown
in parentheses. The majority of PRL samples are at the lower limit of sensitivity of the assay and may actually be significantly lower than shown.
Shaded areas depict the dark phase of the 14L:10D photoperiod.
slightly above the lower limit of the assay. The number of
peak values varied significantly across gestation days
(ANOVA; F4,34 = 8.60, p < 0.01). Comparisons across days
showed Day 3, 6, and 15 females to be similar, having, on
average, 2.8 + 0.4 peak values per day, while Day 9 and 12
females were significantly lower at 1.6 ± 0.2 and 0.6 + 0.3
peak values per day (Fisher PLSD, p < 0.05).
On Day 3, both dawn and dusk peaks were apparent in
5 of 7 females, and a significant diurnal peak at 1300 ( 2
h) was present in 4 of 7 females. Between females, pulses
were asynchronous and differed in amplitude. The number
of peak values per day ranged from 1 to 3. On Day 6, PRL
peaks were more tightly entrained to dusk and dawn. All 7
females had a dawn surge, 4 females had diurnal peak values at 1300 h (+ 2 h), and 1 female failed to have a dusk
surge. On Day 9 of gestation, 5 of 8 females failed to show
a dusk PRL peak value, with 2 of those females also lacking
a dawn peak value. Those 2 females did, however, have
diurnal peak values. On Day 12 of gestation, 4 of 7 females
had serum PRL at the lower limit of assay sensitivity for the
entire day. The other 3 had either one dawn peak value (n
= 2) or two diurnal peak values (n = 1). On Day 15, peaks
had resumed in all females at dusk and dawn and in 3 of
6 females during the day.
Role of Late-Gestation PRL Surges
PRL as a luteotropin. Figure 5 shows the 24-h serum
profile of PRL and P4 on Day 15 of gestation after the 3-day
CB 154 regimen. PRL surges were abolished by the CB 154
injection regime. Average PRL levels in the serum (0.104 +
0.034 pxg-ml-') were significantly lower than for the untreated Day 15 females shown in Figure 1 (t9 = 3.13, p =
0.01). Although PRL surges were eliminated, mean serum
P4 remained high (24.8 + 0.9 ng ml - l) and were similar to
those of untreated Day 15 females in concentration (Fig. 2;
t = -0.18,p = 0.86) and in temporal pattern (correlation,
R = 0.73). No signs of pregnancy failure (e.g., vaginal
bleeding) were detected after injections. All females gave
birth as expected on the 18th day following mating.
PRL as a mammotropin and/or lactogen. Analyses
failed to detect any difference in P4 concentration at 1900
h on Day 15 between animals that had been injected with
oil (n = 8), 50 Ig CB 154 (n = 8), and 300 jig CB 154 (n
= 5) relative to 1) each other (F,22 = 0.54, p = 0.59) or
to 2) samples taken at 1900 h from repeatedly sampled females as represented in Figures 2 and 5 (F4, 31 = 0.30, p =
0.87). In all cases, the mean P4 concentration was 22.8 +
1.4 ngml-1. Serum PRL concentrations differed significantly
across the three groups (F220 = 6 .95,p < 0.01). For females
treated with 50 g (0.040 + 0.007 jig-ml - ') and those treated
with 300 jig (0.044 + 0.010 [ig-ml-') CB 154, concentrations were similar, low, and repeatedly at the lower limit
of the assay compared to the significantly higher (0.728 +
0.232 VIg-ml - ') level in oil-injected females (Fisher PLSD,
p < 0.05). The latter was similar to levels found in repeatedly sampled females on Day 15 at 1900 h (t1 2 = 0.46,p =
0.65).
Teats were not visible through the ventral pelage of CB
154- and oil-treated females on Day 13 of gestation; however, teats were prominent in both groups by Day 15, being
visibly swollen and protruding through the ventral pelage.
No significant difference in mean pup weight for the first
6 days of lactation was found between litters from oil- and
CB 154 (50 and 300 ,ug)-injected females (p > 0.68 in both
cases). Mean pup weight at birth (1. 51
0.05 g) had doubled by Day 4 and reached 3.91 + 0.18 g by Day 6. Milk
was visible in pup stomachs on each of the 6 lactation days
for both oil- and CB 154-treated groups, with no difference
in overall quantity noted.
Mortality rate of the pups, however, differed with experimental treatment. CB 154-injected females cannibalized
more of their young than oil-injected controls (contingency
table analyses; X12 = 4.91, p < 0.05). Pups were observed
either 1) completely missing from the home cage with traces
of blood on the wood chip bedding, 2) partially cannibalized with their remains found on the bedding, or 3) in the
process of being cannibalized. The mean number of pups
cannibalized per litter was 1.83 + 0.48. No female cannibalized her entire litter. The number of litters that were
partially destroyed was significantly greater for CB 154-injected females than for oil-injected controls (X 2 = 5.16, p
< 0.025).
PROLACTIN SECRETION DURING GESTATION IN PHODOPUS
DISCUSSION
The repeated sampling protocol of this study demonstrated a coitally induced PRL surge profile similar to that
obtained when the 24-h pattern on Day 3 of pregnancy was
compiled from females sampled only once [4]. The two daily
surges (dusk and dawn) were sensitive to photoperiod but
not tightly entrained by the light:dark cycle. A significant
midday surge on Day 2 of preimplantation was reported by
Erb and Wynne-Edwards [4] and was thought to be the first
endocrine commitment to a pregnancy. In the present study,
significant midday peak values were detected in 4 of 7 females on Day 3 and as a general pattern on Day 5, with
weak evidence of diurnal surges on Days 3, 6, and 7. Therefore, these results suggest that a midday surge is a feature
of the pituitary PRL secretion pattern during the first half
of pregnancy. However, detection of peak values within a
Djungarian hamster population is complicated by interindividual variability in the precise timing and occurrence of
surges. Such interindividual variability contrasts with findings for pregnant and pseudopregnant rats, in which daily
surges of PRL are synchronous among females within a colony in the presence of a given photoperiod [2, 26]. Thus,
while coital stimulation induces a monophasic PRL surge
pattern (dusk surge only) in the golden hamster [7] and a
biphasic pattern in the rat [2] and mouse [3], it appears that
the general trend in the Djungarian hamster is a triphasic
PRL secretion pattern. However, it remains a possibility that
the other species also have more frequent surges that are
not detected because females are not synchronous with respect to their occurrence. During early pregnancy or an
induced pseudopregnancy in the rat, PRL surges maintain
the integrity of CL P4 production, which in turn renders the
myometrium quiescent and allows for successful implantation [22, 27].
Surges of PRL were absent during midgestation and reestablished in the late term. Thus, we reject the hypothesis
that the PRL secretion pattern persists throughout gestation
and that Edwards et al. [1] failed to detect continuing PRL
surges with the once-a-day sample paradigm. Just prior to
parturition, a peak of pituitary PRL in the serum occurs in
the rat and mouse [28,29] and may be analogous to the
proestrous PRL surge. However, that single surge bears little resemblance to the phasic pattern found in the serum
of late-pregnant Djungarian hamsters at least 3 days before
parturition (Day 15) and probably from Days 14 through
17 [1]. A number of neuronal factors appear to play a role
in the stimulation and inhibition of PRL surges [30]. Hypothalamic dopamine (DA) neurons provide the major tonic
inhibition of PRL release, and lowering of DAergic tone is
required for expression of the twice-daily surge pattern of
PRL in the rat [31]. Since the dusk surge terminates 24 h
earlier than the dawn surge during midgestation in the rat
[32], separate neural mechanisms were thought to control
each surge [2]. Further studies have confirmed this hypoth-
821
esis ([6] for review) and have shown that each surge oscillator differentially stimulates the release of several hypothalamic neuromodulators that may inhibit DA neurons
[33] and/or directly stimulate the release of PRL from lactotroph cells [34]. Similar to what is observed in the rat, the
dusk surge (and midday surge) in the Djungarian hamster
terminates, prior to the dawn surge, by Day 9 of pregnancy.
If the analogy to the rat is correct, than at least two separate
control systems are necessary to regulate the coitally induced PRL secretion pattern in the pregnant Djungarian
hamster.
In the mid- and late-pregnant rat, mouse, and golden
hamster, continued secretion of CL P4 requires placental and/
or embryonic secretion of PRL-like factors (i.e., placental
lactogens) and possibly LH-like factors having luteotropic
activity (see [6] for review). Such factors are likely candidates for luteal maintenance in the late-pregnant Djungarian hamster, since reduction of PRL concentrations with CB
154 failed to affect serum P4 concentrations or pregnancy
success. In addition, the responsiveness to PRL of luteal cells
decreases from early pregnancy to mid- and late pregnancy
in the rat [22, 35] and golden hamster [36], suggesting that
the resumption of PRL surges may not be detected by lateterm CL. Nevertheless, while the present study does not
provide any support for a luteotropic role of late-gestation
pituitary PRL, it is still possible that the greatly reduced PRL
levels that remained following CB 154 treatment contributed to luteal function.
In the rat and mouse, the termination and continued
suppression of PRL surges at and beyond midpregnancy is
suggested to result primarily from a short-loop negative
feedback effect of placental lactogens on the control of PRL
release [6,37]. Accordingly, pituitary PRL surges terminate
earlier (2 days) in pregnant than in pseudopregnant rats
[38]. Such pituitary inhibition does not occur in the Djungarian hamster, since the late-pregnancy PRL secretion pattern is not suppressed and the termination of dusk surges
in pregnant and pseudopregnant females occurs with a similar time course [5]. Midgestation termination of PRL surges
may, instead, be an intrinsic property of the circadian pacemaker that drives the PRL secretory oscillators and overall
rhythm. Reactivation of the pituitary pneumonic must
therefore involve a stimulatory signal. Estrogen, secreted by
late-term CL and the preovulatory follicles that appear in
the ovary on Day 12 of Djungarian hamster pregnancy [1],
is one possible candidate. Estrogen increases the magnitude and enhances the duration of the dusk PRL surge in
pregnant rats when elevated serum P4 levels of pregnancy
are maintained [32]. Additionally, in cycling rats, rising estradiol concentrations facilitate the preovulatory surge of
PRL by blocking the inhibitory influence of dopamine on
lactotrope secretion [39-41]. A similar mechanism may be
involved in the late-pregnant Djungarian hamster.
During late pregnancy, the mammary gland undergoes
considerable growth and differentiation in preparation for
822
EDWARDS ET AL.
lactation. PRL and hormones with PRL-like bioactivities participate with steroids (estrogen and corticosterone) and other
peptides in regulating growth responses and synthesis of
milk proteins [8]. Since the CL did not appear to be responsive to pituitary PRL surges during late gestation in the
Djungarian hamster, responsiveness of mammary tissue was
investigated through teat development and lactational ability of the dam. Teat development was advanced in both oiland CB 154-treated females on Day 15, suggesting that the
late-term PRL surges in the Djungarian hamster were not
required for mammary gland development. This does not,
however, preclude a role for other PRL-like proteins that
may be present in the maternal circulation at that time. The
rat, for example, has at least eight different forms of PRLlike proteins that vary temporally in their synthesis and secretion into the maternal circulation during pregnancy [6].
Milk was visible in all pup stomachs following CB 154 treatment. Therefore, this CB 154 treatment regime did not prevent lactation. However, although 3-day injections of 50 or
300 g CB 154 significantly decreased PRL concentrations
throughout Day 15 with no sign of recovery, the total duration of the effect was not known. Djungarian hamsters
may rely on other endocrine signals for the initiation of
lactogenesis; or metabolic clearance of CB 154 prior to parturition on Day 18 may have allowed PRL to rebound.
Another possible role of PRL secretion during late gestation is the regulation of the onset of maternal behavior
towards the pups. It is well documented that the endocrine
system of the pregnant and parturient mammal plays an important role in preparing the female to care for her young
postpartum (see [42] for review). Research using hypophysectomized female rats has indicated that PRL and PRL-like
molecules help stimulate the onset of maternal behavior
[43, 44]. Suppression of endogenous PRL release prevented
the rapid onset of maternal behavior (as measured by pup
retrieval, grouping of pups in nest, and crouching over the
nest) toward rat young and resulted in partial cannibalization of litters [45]. In the current study, CB 154-treated
females also routinely killed a few pups from each litter. A
deficit in overall quantity of milk production may have been
the stimulus promoting this behavior as a means of increasing the chances of survival of the remaining pups.
The absence of midgestation PRL surges and the resumption of PRL surges in late gestation reported by Edwards et al. [1] were clearly confirmed by the present study.
In addition, interindividual variability in the amplitude, timing, and occurrence of PRL surges was demonstrated, with
the dusk and dawn surges most robust and the midday surge
most variable. Reliable evidence for resumption of the Day
6 PRL surge pattern on Day 15 of pregnancy was also found.
However, these late-gestation PRL surges are not critical for
luteal function. Neither could we demonstrate a mammogenic or lactogenic role for the late-pregnancy surges, although weak evidence for a role in maternal behavior was
found.
REFERENCES
1. Edwards HE, Jenkins KL, Mucklow LC, Erb GE, Wynne-Edwards KE. Endocrinology of the pregnant Djungarian hamster, Phodopus campbelli. J Reprod Fertil
1994; 101:1-8.
2. GunnetJW, Freeman ME. The mating-induced release of prolactin: a unique neuroendocrine response. Endocr Rev 1983; 4:44-60.
3. Barkley MS, Bradford GE, Geschwind II. The pattern of plasma prolactin concentration during the first half of mouse gestation. Biol Rprod 1978; 19:291296.
4. Erb GE, Wynne-Edwards KE. Prolactin, follicle-stimulatinghormone, and luteinizing hormone during preimplantation in the Djungariari hamster (Phodopus
campbelli). Biol Reprod 1994; 50:1328-1333.
5. Edwards HE, Wynne-Edwards KE. Spontaneous termination of an induced pseudopregnancy in the Djungarian hamster, Phodopus campbelli. Horm Behav 1994;
28:165-180.
6. Soares MJ, Faria TN, Roby KF, Deb S. Pregnancy and the prolactin family of hormones: coordination of anterior pituitary, uterine, and placental expression. Endocr Rev 1991; 12:402-423.
7. Talamantes F, Marr G, DiPinto MN, Stetson MH. Prolactin profiles during the
estrous cycle and pregnancy in hamster as measured by homologous RIA. Am J
Physiol 1984; 247:126-129.
8. Tucker HA. Lactation and its hormonal control. In. Knobil E, Neill J (eds.), The
Physiology of Reproduction, Vol 2, 2nd Ed. New York: Raven Press, Ltd.; 1994:
1065-1098.
9. Sinha YN, Lewis UJ, Vanderlann WP. Effects of administrating antisera to mouse
growth hormone and prolactin on gain in litter weight and on mammary nucleic
acid content of lactating CH mice. J Endocrinol 1972; 55:31-40.
10. Tomogane H, Ota K, Yokayama A. Decrease in litter weight gain and in progesterone secretion in lactating rats treated with antiserum to rat prolactin. J Reprod
Fertil 1976; 47:347-349.
11. Akers RM, Bauman DE, Cupuco GT, Goodman GT, Tucker HA. Prolactin regulation of milk secretion and biochemical differentiation of mammary epithelial
cells in periparturient cows. Endocrinology 1981; 109:23-30.
12. Wynne-Edwards KE, Huck UW, Lisk RD. Influence of pre- and post-copulatory
pair contact on pregnancy success in Djungarian hamsters, Phodopus campbelli.
J Reprod Fertil 1987; 80:241-249.
13. Rosser AE, RemfryCJ, Keverne EB. Restricted exposure of mice to primer pheromones coincident with prolactin surges blocks pregnancy by changing hypothalamic dopamine release. J Reprod Fertil 1989; 87:553-559.
14. Kenney AM, Evans RL, Dewsbury DA Postimplantation pregnancy disruption in
Microtus ochrogaster,M. pennsylvanicus and Peromnjscus maniculatus. J Reprod
Fertil 1977; 49:365-367.
15. Bronson FH. Mammalian Reproductive Biology. Chicago: University of Chicago
Press; 1989.
16. Erb GE, Edwards HE, JenkinsKL, Mucklow LC,Wynne-Edwards KE. Induced components in the spontaneous ovulatory cycle of the Djungarian hamster (Phodopus campbelli). Physiol Behav 1993; 54:955-959.
17. Wynne-Edwards KE, Terranova PF, Lisk RD. Cyclic Djungarian hamsters, Phodopus campbelli,lack the progesterone surge normally associated with ovulation
and behavioral receptivity. Endocrinology 1987; 120:1308-1316.
18. Flecknell PA. Laboratory Animal Anesthesia. Toronto, Ont.: Academic Press; 1987.
19. Raper SE, Barker ME, Burwen SJ, Jones A. Isoflurane as an anesthetic for experimental surgery. Anat Rec 1987; 218:116-122.
20. Rao MR, Bartke A, Parkening TA, Collins TJ. Effect of treatment with different
doses of bromocriptine on plasma profiles of prolactin, gonadotrophins and testcsterone in mature male rats and mice. Int J Androl 1984; 258-268.
21. Narburgh LJ,Turner J, Freeman SJ. Evaluation of the teratogenic potential of the
dopamine agonist bromocriptine in rats. Toxicol Lett 1990; 50:189-194.
22. Gafvels M, Bjurulf E, Selstam G. Prolactin stimulates the expression of luteinizing
hormone/chorionic gonadotropin receptor messenger ribonucleic acid in the
rat corpus luteum and rescues early pregnancy from Bromocriptin-induced abortion. Biol Reprod 1992; 47:534-540.
23. Newkirk KD, Silverman DA, Wynne-Edwards KE. The ontogeny of thermoregulation in the Djungarian hamster Phodopus campbelli. Physiol Behav 1995; 57:117124.
24. Erb GE, Wynne-Edwards KE. Pre-implantation endocrinology in the Djungarian
hamster (Phodopus campbelli): progesterone, estrogen, corpora lutea, and embryonic development. Biol Reprod 1993; 49:822-830.
25. Butt WR The iodination of follicle stimulating and other hormones for radioimmunoassay. J Endocrinol 1972; 55:453-454.
26. Yogev L, Terkel J. Effects of photoperiod, absence of photic cues, and suprachiasmatic nucleus lesions on nocturnal prolactin surges in pregnant and pseudopregnant rats. Neuroendocrinology 1980; 31:26-34.
PROLACTIN SECRETION DURING GESTATION IN PHODOPUS
27. Hodgen GD, Itskovitz J. Recognition and maintenance of pregnancy. In. Knobil
E, Neill J (eds.), The Physiology of Reproduction. New York: Raven Press, Ltd.;
1988: 1995-2021.
28. Grattan D, Averill RLW. Effect of ovarian steroids on a nocturnal surge of prolactin secretion that precedes parturition in the rat Endocrinology 1990; 126:11991205.
29. Markoff E, Talamantes F. Serum placental lactogen in mice in relation to day of
gestation and number of conceptuses. Biol Reprod 1981; 24:846-851.
30. Kordon C, Drouva SV, Martinez de Ia Escalera G, Weiner RI. Role of classic and
peptide neuromediators in the neuroendocrine regulation of LH and prolactin.
In. Knobil E, Neill J (eds.), The Physiology of Reproduction, Vol 1, 2nd Ed. New
York: Raven Press, Ltd.; 1994: 1621-1682.
31. Arey BJ, Freeman ME. Hypothalamic factors involved in the endogenous stimulatory rhythm regulating prolactin secretion. Endocrinology 1989; 124:878-883.
32. Freeman ME, Sterman JR. Ovarian steroid modulation of prolactin surges in cervically stimulated ovariectomized rats. Endocrinology 1978; 102:1915-1920.
33. Sagrillo CA, Voogt JL. Mechanisms for the stimulatory effects of opiodergic and
serotonergic input signals on prolactin in pregnant rats. Life Sci 1992; 50:14791489.
34. Arey BJ, Freeman MF. Oxytocin, vasoactive-intestinal peptide, and serotonin regulate the mating-induced surges of prolactin secretion in the rat. Endocrinology
1990; 126:279-284.
35. Niswender GD, JuengelJL, McGuire WJ, Belfiore CJ, Wiltbank MC. Luteal function:
the estrous cycle and early pregnancy. Biol Reprod 1994; 50:239-247.
36. Yuan W, Greenwald GS. Lactotropic effects of follicle stimulating hormone (FSH):
I. FSH has in vitro luteotropic and synergistic effects with luteinizing hormone
37.
38.
39.
40.
41.
42.
43.
44.
45.
823
and prolactin on progesterone production by hamster luteal cells during pregnancy. Biol Reprod 1994; 51:43-49.
Tonkowicz P, Robertson M, Voogt J. Secretion of rat placental lactogen by the
fetal placenta and its inhibitory effect on prolactin surges. Biol Reprod 1983;
28:707-714.
Gorospe WC, Freeman ME. An ovarian role in prolonging and terminating the
two surges of prolactin in pseudopregnant rats. Endocrinology 1981; 108:12931302.
Freeman ME. The neuroendocrine control of the ovarian cycle of the rat. In.
Knobil E, Neill J (eds.), The Physiology of Reproduction, Vol 2, 2nd Ed. New
York: Raven Press, Ltd.; 1994: 613-709.
DeGreef WJ, Klootwijk W, Karels B, Visser TJ. Levels of dopamine and thyrotropin-releasing hormone in the hypophysial stalk blood during an estrogen-stimulated surge of prolactin in the ovariectomized rat. J Endocrinol 1985; 105:107119.
NeillJD, Nagy GM. Prolactin secretion and its control. In: Knobil E, NeillJ (eds.),
The Physiology of Reproduction, Vol 1, 2nd Ed. New York: Raven Press, Ltd.;
1994: 1833-1860.
Numan M. Maternal behavior. In. Knobil E, Neill J (eds.), The Physiology of Reproduction, Vol 2, 2nd Ed. New York: Raven Press, Ltd.; 1994: 221-302.
Bridges RS, DiBiase R, Loundes DD, Doherty PC. Prolactin stimulation of maternal behavior in female rats. Science 1985; 227:782-790.
Bridges RS, Dunckel PT. Stimulation of maternal behavior in rats after treatment
with ectopic pituitary grafts and progesterone. Biol Reprod 1987; 37:518-525.
Bridges RS, Ronsheim PM. Prolactin (PRL) regulation of maternal behavior in rats;
bromocriptine treatment delays and PRL promotes the rapid onset of behavior.
Endocrinology 1990; 126:837-848.