Measurement of Peptide Secretion
and Gene Expression in the Same
Cell
Kathryn Scarbrough, Nancy G. Weiland, Gregg H. Larson,
Maria Angela Sortino, Sufen Chiu, Anne N. Hirshfield, and
Phyllis M. Wise*
Departments of Physiology (K.S., N.G.W., G.H.L., M A S . , S.C.,
P.M.W.) and Anatomy (A.N.H.)
University of Maryland School of Medicine
Baltimore, Maryland 21201
by experimental evidence from a wide variety of peptides in several tissues that are regulated by many
different factors. Thus, studies on PRL (1-3), GH (4, 5),
POMC (6-8), gonadotropins (9-12), vasopressin (1315), and several peptides in brain tissue (16, 17) suggest that physiologically stimulated peptide biosynthesis and/or secretion are coupled with gene transcription. Moreover, inhibition of peptide secretion often
results in decreased gene transcription and protein
synthesis for a specific peptide (1, 6-8, 18). This coupling of secretion to transcription and protein synthesis
may permit a cell to modulate the level of gene expression and subsequent peptide or protein biosynthesis to
keep pace with its secretory needs. In the past, investigators have analyzed whether the level of gene
expression is coupled to secretion at the tissue level.
The method we have developed allows analysis of this
question at the level of the individual cell.
Combining the reverse hemolytic plaque assay sequentially with in situ hybridization permits quantitation
of the amount of peptide secreted by an individual cell
and quantitation of the level of a specific mRNA in the
same cell under a variety of experimental conditions.
We have tested this new method in a well characterized
endocrine system to answer several questions. First, is
the ability of a lactotroph to secrete PRL, and/or the
amount of PRL secreted, correlated with the amount of
PRL mRNA present in that cell? Second, does the ability
of estradiol and/or TRH to alter PRL secretion correlate
with the amount of PRL mRNA in that lactotroph? Third,
do these stimuli transform other pituitary cells into
lactotrophs or influence the percentage of lactotrophs
secreting PRL?
A combined reverse hemolytic plaque-/n situ hybridization assay was developed to allow analysis of the
relationship between peptide secretion and gene
expression within individual cells. We used the pituitary lactotroph as a model system, but this strategy should be widely applicable. It can be used to
test hypotheses regarding if and when peptide secretion and gene expression are coupled in any
system in which antibodies to the secreted peptide
and probes complementary to the mRNA are available. Using the mRNA hybridization signal to identify
certain cell types, this method may also be useful in
further studies on the biochemical mechanism of
peptide secretion. In addition, questions regarding
whether a cell known to secrete a given peptide
contains other specific mRNAs and the relationship
between these mRNAs and the secretion of the
peptide can be studied using this strategy.
We found striking heterogeneity among lactotrophs in both gene expression and PRL secretion
and a lack of correlation of these parameters within
individual lactotrophs under every treatment examined. We also present the first direct visualization
and quantitation of the percentage of nonsecreting
PRL mRNA-containing cells after estradiol treatment
and in the presence or absence of the PRL secretagogue, TRH. Finally, we found that in ovariectomized rats, nonsecreting lactotrophs exhibited significantly higher levels of PRL mRNA than lactotrophs that were actively secreting PRL during the
assay. (Molecular Endocrinology 5: 134-142, 1991)
It should be emphasized that although we have applied this method to the secretion of PRL and its level
of gene expression in the lactotroph, the strategy employed here can be used to test the relationship between a peptide's secretion and its gene expression in
any system in which antibodies to the secreted peptide
and probes complementary to the mRNA are available.
In addition, it is now possible to determine the relationship between peptide secretion and other specific
INTRODUCTION
It is generally accepted that the level of secretion of a
peptide and/or its concentration in tissue correlate with
its average level of mRNA. This concept is supported
0888-8809/91/0134-0142S02.00/0
Molecular Endocrinology
Copyright © 1991 by The Endocrine Society
134
Secretion and Gene Expression in the Same Cell
mRNAs present in a cell. One can also determine
whether age or pathological changes alter the relationship between secretion and gene expression in individual cells.
RESULTS
The combined plaque-/n situ hybridization method can
clearly detect PRL mRNA-containing cells (i.e. those
covered with silver grains) in a mixed population of
pituitary cells (Fig. 1). In addition, lactotrophs that are
secreting PRL can be differentiated from cells secreting
very little or no PRL by the presence or absence of
clear plaques surrounding mRNA-containing cells. Neill
and his colleagues (19, 20) have demonstrated convincingly that the area of the hemolytic plaque is an
accurate and sensitive index of the amount of peptide
secreted. The smallest detectable plaque depends on
the integrity of the red blood cell lawn. We calculated
our limit of detection to be approximately 400 /tm2,
corresponding to a circular plaque with a radius of only
about 11 juim.
We performed tests to determine whether the level
of mRNA could be quantified under these experimental
circumstances. The effect of increasing concentrations
of PRL cRNA (0.1-0.8 Mg/ml) on the hybridization signal
from estradiol-treated rats, as assessed by two indices,
can be seen in Fig. 2. Triplicate slides from each concentration of probe were analyzed. Both the mean area
of computer-enhanced grains and the mean gray level
within a window of fixed size placed over cells gave
qualitatively similar results; saturation was achieved at
approximately 0.7 nQ/m\ 35S-labeled cRNA. This demonstrates that both methods of quantitation accurately
reflect increased signal. We chose to analyze all remaining data using the area of enhanced grains for
reasons detailed in Materials and Methods. A second
control experiment demonstrated that increasing the
length of exposure to emulsion resulted in a linear
increase in the mean area of enhanced grains for at
least 6 days when hybridizing with low or high concentrations of PRL cRNA (Fig. 3).
To determine the effects of estradiol and/or TRH on
PRL gene expression and secretion by individual cells,
we analyzed the data from five ovariectomized (OVX)
and five estradiol-primed rats. Table 1 summarizes the
data from all animals when we average the results from
approximately 150 cells/slide and 2 slides/treatment.
We defined a lactotroph as any pituitary cell that was
covered with silver grains, indicating that it contained
PRL mRNA. Our finding that lactotrophs comprised
approximately 45% of all pituitary cells agrees well with
immunocytochemical estimates of the proportion of
pituitary cells that contain PRL (21, 22). In addition to
quantitating the amount of PRL secreted by a cell based
upon the size of the hemolytic plaque surrounding it,
we also categorized lactotrophs as either secreting or
nonsecreting. A secreting lactotroph was any cell that
135
formed a detectable plaque during the incubation
period, regardless of the size of the plaque. We categorized all mRNA-containing cells without plaques as
nonsecreting. Therefore, the terms secreting and nonsecreting are used relative to our experimental conditions. Lactotrophs that do not form detectable plaques
during a 1-h incubation may still be secreting PRL at a
very low rate, since other investigators have reported
that the percentage of PRL plaque-forming cells increases somewhat with added incubation time (20).
Compared to OVX rats, 4 days of exposure to estradiol in vivo 1) increased the percentage of pituitary cells
that secreted PRL [(plaque-forming cells/pituitary cells)
x 100; P < 0.05]; 2) increased the percentage of
lactotrophs that secreted PRL, i.e. [(plaque-forming
cells/mRNA-containing cells) x 100; P < 0.05]; and 3)
did not increase the percentage of pituitary cells expressing PRL mRNA [(mRNA-containing cells/pituitary
cells) x 100]. Therefore, under these conditions estradiol stimulated lactotrophs to enter a secretory state,
but did not transform other pituitary cell types into
lactotrophs.
In OVX rats, treatment with TRH in vitro increased
the percentage of lactotrophs that secreted PRL (P <
0.05). Thus, TRH recruited nonsecreting lactotrophs
into the pool of secreting cells. In estradiol-primed rats,
TRH was ineffective in stimulating any further increase
in the percentage of secreting lactotrophs.
As expected, estradiol stimulated PRL secretion per
lactotroph (mean plaque area; P < 0.0001) and increased mRNA levels per cell (mean area of enhanced
grains; P < 0.0001). There was a main effect of TRH
on mean plaque area (P < 0.0002) and a significant
interaction term (P < 0.05); therefore, the effect of TP.H
on PRL secretion depended on the steroidal milieu of
the rats. Exposure to TRH for 60 min in vitro did not
cause a detectable increase in PRL mRNA levels in
individual lactotrophs (mean area of enhanced grains),
and there was no interaction. Thus, when we compared
OVX to estradiol-primed rats and averaged the results
from 150 cells/slide, we confirmed the general relationship between increased levels of mRNA encoding PRL
and increased secretion of PRL.
Surprisingly, in OVX rats, PRL mRNA levels were
significantly higher in nonsecreting lactotrophs than in
secreting lactotrophs (which were grouped together
regardless of plaque size) under basal and TRH-stimulated conditions (Fig. 4; P < 0.0001). Treating rats with
ostradiol substantially increased PRL mRNA in all lactotrophs and abolished the difference in PRL mRNA
levels in secreting compared to nonsecreting cells.
Data from individual cells within a single animal's
pituitary showed marked heterogeneity in both plaque
area and level of mRNA under every experimental condition examined. Figure 5 is a scattergram from one
representative OVX and one representative estradioltreated rat in which we monitored mRNA levels and
hormone secretion in the same cells. Each symbol
represents a single cell. Contrary to the small variation
in mean plaque area and the mean level of mRNA
MOL ENDO-1991
136
Vol 5 No. 1
Fig. 1. Dispersed Pituitary Cells Surrounded by Red Blood Cells (x250)
A, Cell that is covered with silver grains and is surrounded with a hemolytic plaque, indicating that it is a lactotroph containing
PRL mRNA and secreting PRL. B, Cell containing PRL mRNA but not secreting PRL. C, Pituitary cell, presumably a gonadotrope,
thyrotrope, corticotrope, or somatotrope, exhibiting no PRL mRNA or PRL secretion.
Secretion and Gene Expression in the Same Cell
137
400-1
200-i
~
300-
160"
120-
200
80R 2 . 0.71
100-
R2 - 0.81
40"
0.1
0.3
0.5
[cRNA] ng/ml
0.7
0.9
0.1
0.3
0.5
[cRNA]
0.7
0.9
ng/ml
Fig. 2. The Effect of Increasing Concentrations of 35S-Labeled PRL cRNA on the Hybridization Signal
Each point represents the mean of approximately 150 cells quantitated per slide. Rats were OVX for 14 days and treated with
capsules containing estradiol (180 ^g/ml) for 4 days before pituitary cells were dispersed and used in the reverse hemolytic plaquein situ hybridization assay. Saturation of the endogenous PRL mRNA occurs at approximately 0.7 ng/m\ labeled cRNA regardless
of the method used to quantitate the hybridization signal.
600 n
8
1.0 ng/ml probe:
R 2 =0.92
450-
3000.1 ng/ml probe:
R 2 =0.86
150
1
2
3
4
5
6
Days of Exposure
Fig. 3. The Effect of Increasing Exposure to Emulsion on the
Hybridization Signal
Each point represents the mean of 100-150 PRL mRNAcontaining cells/slide. Rats were OVX and treated with estradiol as described in Fig. 2.
among animals in each group shown in Table 1, we
observed strikingly large variations in PRL secretory
activity and PRL gene expression among individual cells
from each rat's pituitary. For example, in the OVX TRHtreated panel in Fig. 5, the plaque area ranged from 0 70,000 jum2, and the area of enhanced grains ranged
from 0-500 nm2. In addition, we detected no correlation
between the amount of PRL secretion by an individual
cell and its level of PRL mRNA under the experimental
conditions we tested. That is, the level of PRL mRNA
and secretory activity vary independently of one another
within individual cells from any pituitary in any treatment
group. This is in dramatic contrast to the comparison
between experimental treatments, in which estradiolinduced increases in PRL gene expression were correlated with increased hormone secretion.
DISCUSSION
This combined plaque and hybridization assay enables
investigators, for the first time, to differentiate between
nonsecreting and secreting lactotrophs and to quantitate accurately the effects of steroids and neurohormones on recruitment from the nonsecreting into the
secreting pool of cells. Whereas previous methods were
able to demonstrate an effect of estrogen on PRL
secretion and mean mRNA levels (23, 24), they were
unable to distinguish among potentially heterogeneous
pools of lactotrophs or to monitor secretion and gene
expression in the same individual cell.
Perhaps the most interesting observation arising from
this study is that under all experimental conditions
tested, we found no correlation between the amount of
secretory activity and the level of gene expression
exhibited by individual cells. There are several possible
explanations for this lack of correlation within individual
cells. Protein secretion occurs through both constitutive
and regulated pathways (25, 26). If a cell secretes only
via the constitutive pathway we would expect to observe some relationship between mRNA level and
amount of secretion, since peptide would be secreted
immediately upon synthesis. On the other hand, if a cell
secretes via the regulated pathway we would be less
likely to observe a correlation between a cell's level of
secretion and its level of mRNA, since regulated release
occurs from a storage pool of peptide sequestered in
secretory vesicles. Endocrine cells may use both pathways in the secretion of hormone (26, 27). It is not
surprising that under TRH-stimulated conditions there
is no correlation between secretion and gene expression, since TRH increases the secretion of older stored
hormone (28), and cells have different amounts of
stored hormone (29). Pulse-chase experiments have
shown that newly synthesized PRL is secreted rapidly
under basal conditions (2b, 30), suggesting that lactotrophs may use the constitutive pathway for secretion
under these conditions. Our finding of a lack of any
correlation between mRNA and secretion under basal
conditions in lactotrophs from both OVX and estradiolprimed rats raises the possibility that even in the absence of classical secretagogues, lactotrophs may secrete stored hormone to a significant extent. Recent
data from Chen et al. (31) support this hypothesis. By
Vol 5 No. 1
MOL ENDO-1991
138
Table 1. Effects of Estradiol and/or TRH on Mean Levels of PRL mRNA and Secretion by Individual Cells
ovx (n = 5)
Estradiol-Primed (n = 5)
27.8 ± 1.9
33.6 ± 2.2
33.4 ± 2.3
30.8 ± 2.8
60.0 ± 4.6
74.4 ± 4.4
78.0 ± 4.9
76.5 ± 4.0
46.4 ± 1.0
45.2 ±1.7
43.0 ± 1.0
43.4 ± 1.4
3329 ± 344
5829 ± 698
18,983 ±2212
29,530 ±1269
63.5 ± 4.8
247.3 ± 14.6
228.9 ± 8.7
1
% Pituitary Cells Secreting PRL
vehicle-treated
TRH-stimulated
% Lactotrophs Secreting PRL2
vehicle-treated
TRH-stimulated
% Lactotrophs3
vehicle-treated
TRH-stimulated
Mean Plaque Area (/im2)
vehicle-treated
TRH-stimulated
Mean Area Enhanced Grains (^m*
vehicle-treated
TRH-stimulated
63.1 ±4.1
# plaque-forming cells
# pituitary cells
# plaque-forming cells
' # mRNA-containing cells
# mRNA-containing cells
# pituitary cells
Estradiol treatment of ovariectomized rats for 4 days increased significantly the percent lactotrophs secreting PRL, P < 0.05;
however, it did not increase the percentage of lactotrophs in the population of pituitary cells (P > 0.05). TRH recruited more
lactotrophs from ovariectomized rats to secrete PRL (P < 0.05) but was ineffective in increasing further the percentage of
lactotrophs from estradiol-primed rats that secreted PRL. Estradiol increased PRL secretion (mean plaque area, P < 0.0001) and
the abundance of PRL mRNA (mean area of enhanced grains, P < 0.0001). There was a main effect of TRH on mean plaque area
(P < 0.0002) and a significant interaction term (P < 0.05); therefore, the effect of TRH depended on the steroidal milieu of the rats.
There was no significant effect of TRH on the mean area of enhanced grains and there was no interaction.
• 300
300"
E
3
(f>
c
O
XJ
CD
O
c
(0
sz
c
UJ
T
250"
• 250
200-
200
150-
- 150
100
100
50
50
0
Sec
Non-Sec
Vehicle
Sec
Non-Sec
TRH
Ovariectomized
Sec
Non-Sec
Vehicle
Sec
Non-Sec
0
TRH
Estradiol-primed
Fig. 4. PRL mRNA Levels (Area of Enhanced Grains) in Secreting vs. Nonsecreting Lactotrophs
Rats were either OVX for 18 days or OVX for 14 days and treated with estradiol for 4 days (n = 5/group). All cells containing
computer-enhanced grains were considered secreting if plaque areas were greater than approximately 400 ^m 2 or nonsecreting if
plaques were absent or less than 400 Mm2. In OVX rats, nonsecreting lactotrophs exhibited significantly higher PRL mRNA levels
than lactotrophs that secreted measurable amounts of PRL (P < 0.0001). Estradiol treatment of OVX rats abolished the difference
in mRNA content in secreting compared to nonsecreting lactotrophs.
Secretion and Gene Expression in the Same Cell
139
Estradiol - Primed
Vehicle - Treated
OVX
Vehicle - Treated
600
E
Nonsecretory
cells
600-
400^
400-
200"
200-
Nonsecretory
cells
(0
c
2
o
10 3
10 4
10 s
103
10 6
104
105
106
10 5
10 6
•o
o
c
n
Estradiol - Primed
TRH - Stimulated
OVX
TRH - Stimulated
UJ
600-
Nonsecretory
cells
Nonsecretory
cells
400-
400"
200-
10 3
10 4
10*
10 6
10 3
10 4
Plaque Area ((j.m )
Fig. 5. Scattergram of a Representative Slide from an OVX and an OVX Estradiol-Primed Rat
Half of the slides from each animal were treated with TRH (10~7 M) for 60 min. Cells containing mRNA (100-150 per slide) with
or without plaques were quantified. Each point represents a single cell. Within the population of secreting cells, there is no
correlation between an individual cell's secretion of PRL (plaque area) and the level of PRL mRNA (area of enhanced grains) after
a 1-h treatment with or without TRH or after a 4-day treatment with physiological concentrations of estradiol.
using cycloheximide to effectively eliminate the constitutive pathway of peptide secretion, these investigators
found that 50% of PRL- and GH-secreting cells release
stored hormone in the absence of stimulation by a
secretagogue. The release of stored hormone under
basal conditions by even a percentage of lactotrophs
would mask any relationship between hormone biosynthesis and secretion.
Alternatively, the lack of correlation between the level
of mRNA encoding a hormone and that hormone's
secretion by a cell under basal conditions may result
from a temporal desynchronization between hormone
secretion and biosynthesis. Changes in the rate of
secretion occur more quickly than changes in the level
of mature mRNA, since the latter depends upon the the
half-life of the mRNA in addition to the level of gene
transcription (32). A dissociation between the two
measures could develop if basal synthesis and secretion of PRL are periodic, having successive phases of
high and low activity. A complete lack of correlation
between the level of mRNA and the amount of secretion
would result, since a single pituitary would most likely
contain cells at all different stages of the cycle. Neill
and his colleagues (19, 20) have shown that the secretory activity of a given cell varies from day to day.
Whether secretory activity varies over a period of hours
has not yet been determined.
The combined plaque-/n situ hybridization assay allowed us to make a second important observation:
nonsecreting lactotrophs from OVX rats contained, on
the average, a greater amount of PRL mRNA than
secreting cells. It is possible that spontaneous periodic
changes in secretion rate, as discussed above, may
explain this finding. That is, lactotrophs may appear
nonsecreting if we monitored them during the low point
of their secretory cycle. The level of mRNA within a
given cell may depend on the length of time elapsed
since it made the transition from high to low secretory/
synthesis activity. To test the hypothesis of bursts of
secretory activity and its relationship to levels of mRNA
it may be necessary to perform sequential plaque assays over the course of several hours, followed by in
situ hybridization.
Estradiol is a potent stimulator of PRL gene expression (24, 33), PRL storage (29), and basal secretion of
PRL (23, 34). Our data suggest that estradiol exerts its
genomic effect on all lactotrophs, regardless of whether
they are secreting or nonsecreting; thus, it substantially
increased the level of PRL mRNA in these cells and
obliterated the difference in PRL mRNA levels between
Vol 5 No. 1
MOL ENDO-1991
140
secreting and nonsecreting lactotrophs. It is remarkable
that even after 4 days of constant estradiol stimulation,
a small number of lactotrophs did not secrete a detectable amount of PRL during the 60-min incubation
period. Perhaps estradiol stimulates both the rate of
secretion and the rate of synthesis, but does not fully
eliminate the secretory cycle. It is also possible that
individual lactotrophs vary dramatically in their sensitivity to estradiol, and the nonsecreting cells have just
begun increased PRL mRNA synthesis in preparation
for increased secretion. Future application of this
method may include in situ hybridization using a probe
complementary to the estrogen receptor coupled with
the plaque assay.
The presence of nonsecreting PRL mRNA-containing
cells was not unexpected. Increased numbers of cells
form plaques with increased periods of incubation, although in previous studies most PRL-secreting cells
were detected after a period of 1 h (20). These data
indicate that some lactotrophs secrete PRL at a very
low rate and may require prolonged incubation periods
to be detected. Experiments employing a sequential
plaque assay (35) demonstrated that the amount of
PRL secreted by a given cell varies from day to day
(19, 20). In addition, 5% of the lactotrophs did not form
plaques on a given day. These nonsecreting cells had
not simply died during the experiment, because they
were able to secrete measurable amounts of PRL on
following days. Therefore, the lack of plaque formation
around some lactotrophs after a 1-h incubation does
not necessarily reflect on their potential for future secretion. Alternatively, it is possible that PRL mRNAcontaining cells that appear nonsecretory may be secreting a variant form of PRL not recognized by our
antiserum. This possibility seems unlikely for two reasons. First, our antibody is polyclonal. It was raised
against purified PRL from the National Pituitary Hormone Program, which was presumably a mixture of
PRL variants. Therefore, the antiserum is likely to be
directed at multiple variants. Second, although previous
data (36) have been interpreted to suggest that lactotrophs secrete different variant forms of PRL, we know
of no data that clearly demonstrate that individual cells
secrete a single variant of PRL.
In conclusion, this combination of the plaque assay
and in situ hybridization allows investigators to approach questions of the relationship between gene
expression and secretion at a new level of resolution.
Because it labels a particular cell type regardless of its
secretory activity, this method will be important for
further studies on functional differences between cells
synthesizing and secreting a particular peptide. This
strategy should also prove extremely useful to researchers studying diverse issues in endocrinology,
biochemistry, and eukaryotic molecular biology.
MATERIALS AND METHODS
Sprague-Dawley rats (Zivic-Miller, Allison Park, PA), weighing
200-250 g, that showed two consecutive estrous cycles were
anesthetized with ether and OVX (day 0). On day 14, an
estradiol-containing Silastic capsule that produces plasma
concentrations of estradiol of 10-20 pg/ml (37) was implanted.
Four days later rats were decapitated, and cells from anterior
pituitaries were prepared for the reverse hemolytic plaque
assay (19,20,38). Pituitary and ox red blood cells were infused
into Cunningham chambers and equilibrated for 60 min, then
antirat PRL antibody (39) with and without TRH (10~7 M) was
added to slides and incubated at 37 C for an additional 60
min. Guinea pig serum, a source of complement, was added
for 30 min to induce lysis of red blood cells in the area
surrounding the cells secreting PRL. Cells were fixed for 1-3
days in RNase-free phosphate-buffered 4% paraformaldehyde
at 4 C before in situ hybridization was performed.
On the day that in situ hybridization was performed, Cunningham chambers were dried in air for approximately 2 h, and
the glass coverslips were removed while the slides were
immersed in RNase-free phosphate buffer (0.1 M; pH 7.4).
Slides were washed sequentially in phosphate buffer, diethylpyrocarbonate-treated water, and triethanolamine buffer (0.08
M; pH 8.0); acetylated with 0.25% acetic anhydride in triethanolamine buffer; and set in 2 x SSC (1 x SSC is 150 ITIM NaCI
and 15 ITIM trisodium citrate) until probe was applied. We
transcribed PRL cDNA (provided by R. A. Maurer and W. W.
Chin) using [35S]UTP and SP6 RNA polymerase to yield a
riboprobe with a specific activity of 1.1-1.7 x 108 dpm/^g.
35
S-Labeled cRNA (0.6 /xg/ml) was mixed with hybridization
buffer [final concentrations, 50% deionized formamide, 10%
dextran sulfate, 0.3 M NaCI, 10 mM Tris (pH 8.0), 1 ITIM EDTA,
0.02% BSA, 0.02% Ficoll type 400, 0.02% polyvinylpyrrolidone, and 10 ITIM dithiothreitol] and with excess (500 Mg/ml)
unlabeled tRNA. This probe solution (23 n\) was applied to
each slide, followed by a Parafilm coverslip. Slides were hybridized overnight in humidified petri dishes at 53 C. This
temperature is 20 C below the theoretical Tm for the probe.
The Tmwas calculated based on the known length and nucleotide content of the probe using the formula Tm = 69.3 + 0.41 (G
+ C)% - (650/average length of probe in nucleotides) (40)
with modifications for RNA and our hybridization conditions
(41). We confirmed this was the correct hybridization temperature for the probe under our experimental conditions by
performing a melting experiment (data not shown). After hybridization, Parafilm coverslips were floated off, and slides
were washed in 4 x SSC containing 4 mM dithiothreitol. All
subsequent buffers contained 2 mM dithiothreitol; slides were
treated for 30 min with RNase-A (20 (ig/m\) at 37 C and
washed sequentially in RNase buffer [10 mM Tris (pH 8.0), 1
mM EDTA, and 0.5 M NaCI; 37 C; 30 min], 2 x SSC (room
temperature; 60 min), 0.1 x SSC (60 C; 30 min), and 1 x SSC
(room temperature; 10 min). Slides were dehydrated in graded
alcohols containing 300 mM ammonium acetate, dipped in
photographic emulsion, exposed for 1-6 days, and developed
using conventional photographic methods. Slides were stained
for 15 sec with 4% methyl green.
Quantitation of Hybridization Signal
Quantitation of mRNA was performed using a BioQuant IV
Image Analysis system (R & M Biometrics, Bethesda, MD). A
window of fixed size, large enough so that all grains on a cell
with a very high signal were included, was placed over each
cell. The relative amount of mRNA hybridized was measured
within this window using two indices: 1) gray level (which is
related to light transmission through the cell), and 2) area
covered by silver grains. For the latter index of hybridization
signal, a gray level threshold was chosen such that only
exposed silver grains were computer enhanced and detected
by the image analysis system; histologically stained red blood
cells and pituitary cell bodies did not contribute to the measurement. This measure of hybridization signal is reported as
the area of enhanced grains. The area of enhanced grains is
less prone than measurements of gray level to variation due
to instability of the light level transmitted by the video camera
Secretion and Gene Expression in the Same Cell
and slight variations in histological staining from slide to slide.
In addition, unlike gray level, the area of enhanced grains is
not affected by the size of the cell or whether the mRNAcontaining cell is surrounded by a clear plaque or by red blood
cells. This measurement is also a more sensitive measure than
gray level when cells exhibit low signal intensity. Because
contributions to the gray level from the stained cell nucleus
and cytoplasm are removed in the measurement of area of
enhanced grains, cells with low signal are more easily distinguishable from background (data not shown).
Methodological Considerations
We performed several preliminary experiments to optimize the
preservation of plaques around secreting lactotrophs and
mRNA in the cells. We reasoned that living cells should exclude
exogenous RNases, and the fixation procedure should inactivate RNases present on the slides at the end of the reverse
hemolytic plaque assay; therefore, the plaque assay was
performed without special precautions to combat RNases. We
found that the integrity of plaques was maintained best when
the reverse hemolytic plaque assay was terminated by submersion and storage of the assay chambers in 4% paraformaldehyde at 4 C. The alternative of infusing paraformaldehyde
through the chambers and/or dehydrating the cells and storing
the slides in ethanol resulted in loss of red blood cells from the
slides. Storage in paraformaldehyde at 4 C for up to 1 week
did not alter the level of PRL mRNA signal compared to the
signal detected after overnight storage (data not shown). The
red blood cell lawn became progressively thinner with time,
but storage for 1 week did not compromise our ability to
quantitate plaque areas. Furthermore, since the cells continue
to be osmotically active during the in situ hybridization steps,
red blood cells were distorted or disrupted if posthybridization
rinses did not include a final rinse in an iso-osmotic buffer (1
x SSC). Simultaneous quantitation of plaque area and silver
grains required that the red blood cells and pituitary cells be
stained, but that the stain not interfere with the computerassisted quantitation of grains. Therefore, we modified the
reverse hemolytic plaque assay procedure to stain slides for
only 15 sec in 4% methyl green with no counterstain.
Statistical Analyses
The effects of estradiol and/or TRH on mean levels of PRL
mRNA, PRL secretion, percentage of pituitary cells secreting
PRL, percentage of lactotrophs secreting PRL, and percentage of lactotrophs in the total population of pituitary cells were
determined using two-way analysis of variance. A P value less
than 0.05 was considered significant. There was heterogeneity
of variance in plaque areas; therefore, the data were log
transformed before statistical analysis. The ability of TRH to
recruit quiescent cells from either OVX or estradiol-primed rats
into the secreting pool of cells was evaluated using t tests.
PRL mRNA levels in secreting vs. nonsecreting lactotrophs
under basal and TRH-stimulated conditions were compared
using two-way analysis of variance.
Acknowledgments
PRL cDNA was generously provided by Drs. R. A. Maurer and
W. W. Chin. Excellent technical help was provided by Klara
Margaretten and Katherine Rosewell.
Received July 13,1990. Revision received October 1,1990.
Accepted October 9,1990.
Address requests for reprints to: Dr. Phyllis M. Wise, Department of Physiology, University of Maryland School of
Medicine, Baltimore, Maryland 21201.
This work was supported by NIH Grants AG-02224, HD-
141
15955, HD-13642 and Training Grant HD-07170. The experiments used animals in accordance with University of Maryland
guidelines and federal, state, and local laws. The University of
Maryland's animal facility is accredited by the AAALAC and
complies with the NIH policy governed by the Principles for
Use of Animals and the Guide for the Care and Use of Laboratory Animals.
* NIH MERIT Awardee.
REFERENCES
1. Maurer RA 1980 Dopaminergic inhibition of prolactin synthesis and prolactin messenger RNA accumulation in cultured pituitary cells. J Biol Chem 255:8092-8097
2. Murdoch GH, Potter E, Nicolaisen AK, Evans RM, Rosenfeld MG 1982 Epidermal growth factor rapidly stimulates
prolactin gene transcription. Nature 300:192-194
3. Murdoch GH, Franco R, Evans RM, Rosenfeld MG 1983
Polypeptide regulation of gene expression. J Biol Chem
258:15329-15335
4. Evans RM, Birnberg NC, Rosenfeld MG 1982 Glucocorticoid and thyroid hormones transcriptionally regulate
growth hormone gene expression. Proc Natl Acad Sci
USA 79:7659-7663
5. Barinaga M, Yamonoto G, Rivier C, Vale W, Evans R,
Rosenfeld MG 1983 Transcriptional regulation of growth
hormone gene expression by growth hormone-releasing
factor. Nature 306:84-85
6. Schachter BS, Johnson LK, Baxter JD, Roberts JL 1982
Differential regulation by glucocorticoids of proopiomelanocortin mRNA levels in the anterior and intermediate lobes
of the rat pituitary. Endocrinology 110:1442-1444
7. Gagner J-P, Drouin J 1985 Opposite regulation of proopiomelanocortin gene transcription by glucocorticoids
and CRH. Mol Cell Endocrinol 40:25-32
8. Eberwine JH, Jonassen JA, Evinger MJQ, Roberts JL
1987 Complex transcriptional regulation by glucocorticoids and corticotropin-releasing hormone of proopiomelanocortin gene expression in rat pituitary cultures. DNA
6:483-492
9. Alexander DC, Miller WL1982 Regulation of ovine folliclestimulating hormone j8-chain mRNA by 17/3-estradiol in
vivo and in vitro. J Biol Chem 257:2282-2286
10. Nilson JH, Nejedlik MT, Virgin JB, Crowder ME, Nett TM
1983 Expression of a subunit and luteinizing hormone 0.
J Biol Chem 258:12087-12090
11. Boothby M, Kukowska J, Boime I 1983 Imbalanced synthesis of human choriogonadotropin « and /? subunits
reflects the steady state levels of the corresponding
mRNAs. J Biol Chem 258:9250-9253
12. Milsted A, Silver BJ, Cox RP, Nilson JH 1985 Coordinate
regulation of the messenger ribonucleic acids encoding
the a- and /3-subunits of human chorionic gonadotropin in
HeLa cells and butyrate-resistant variants. Endocrinology
117:2033-2039
13. Majzoub JA, Rich A, van Boom J, Habener JF 1983
Vasopressin and oxytocin mRNA regulation in the rat
assessed by hybridization with synthetic oligonucleotides.
J Biol Chem 258:14061-14064
14. Uhl GR, Zingg HH, Habener JF 1985 Vasopressin mRNA
in situ hybridization: localization and regulation studied
with oligonucleotide cDNA probes in normal and Brattleboro rat hypothalamus. Proc Natl Acad Sci USA 82:55555559
15. Sherman TG, McKelvy JF, Watson SJ 1986 Vasopressin
mRNA regulation in individual hypothalamic nuclei: a
Northern and in situ hybridization analysis. J Neurosci
6:1685-1694
16. Morris BJ, Feasey KJ, ten Bruggencate G, Herz A, Hollt
V 1988 Electrical stimulation in vivo increases the expres-
Vol 5 No. 1
MOL ENDO-1991
142
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
sion of proenkephalin mRNA and decreases the expression of prodynorphin mRNA in rat hippocampal granule
cells. Proc Natl Acad Sci USA 85:3226-3230
Zoeller RT, Wolff RS, Koller KJ 1988 Thyroid hormone
regulation of messenger ribonucleic acid encoding thyrotropin (TSH)-releasing hormone is independent of the
pituitary gland and TSH. Mol Endocrinol 2:248-252
Maurer RA1981 Transcriptional regulation of the prolactin
gene by ergocryptine and cyclic AMP. Nature 294:94-97
Smith PF, Luque EH, Neill JD 1986 Detection and measurement of secretion from individual neuroendocrine cells
using a reverse hemolytic plaque assay. Methods Enzymol 124:443-465
Neill JD, Smith PF, Luque EH, Munoz de Toro M, Nagey
G, Mulchahey JJ 1987 Detection and measurement of
hormone secretion from individual pituitary cells. Recent
Prog Horm Res 43:175-229
Dada MO, Campbell GT, Blake CA 1984 Pars distalis cell
quantification in normal adult male and female rats. J
Endocrinol 101:87-94
Wilfinger WW, Larsen WJ, Downs TR, Wilbur DL 1984
An in vitro model for studies of intracellular communication
in cultured rat anterior pituitary cells. Tissue Cell 16:483497
Maurer RA, Gorski J 1977 Effects of estradiol-17/3 and
pimozide on prolactin synthesis in male and female rats.
Endocrinology 101:76-84
Maurer RA 1983 Use of recombinant DNA procedures to
study mechanisms controlling synthesis of hypophyseal
hormones. Biol Reprod 28:27-35
Kelly RB 1985 Pathways of protein secretion in eukaryotes. Science 230:25-31
Burgess TL, Kelly RB 1987 Constitutive and regulated
secretion of proteins. Annu Rev Cell Biol 3:243-293
Stirling RG, Shin SH 1990 A high concentration of dopamine preferentially permitted release of newly synthesized
prolactin. Mol Cell Endocrinol 70:65-72
Walker AM, Farquhar MG, Peng B 1980 Preferential release of newly synthesized prolactin granules is the result
of functional heterogeneity among mammotrophs. Endocrinology 107:1095-1104
Nogami H 1984 Fine-structural heterogeneity and morphologic changes in rat pituitary prolactin cells after estro-
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
gen and testosterone treatment. Cell Tissue Res
273:195-202
Swearingen KC 1971 Heterogeneous turnover of adenohypophysial prolactin. Endocrinology 89:1380-1388
Chen TT, Kineman RD, Betts JG, Hill JB, Frawley LS
1989 Relative importance of newly synthesized and
stored hormone to basal secretion by growth hormone
and prolactin cells. Endocrinology 125:1904-1909
Hargrove JL, Schmidt FH 1989 The role of mRNA and
protein stability in gene expression. FASEB J 3:23602370
Shull JD, Gorski J 1984 Estrogen stimulates prolactin
gene transcription by a mechanism independent of pituitary protein synthesis. Endocrinology 114:1550-1557
Lu KH, Koch Y, Meites J 1971 Direct inhibition by ergocornine of pituitary prolactin release. Endocrinology
89:229-233
Frawley LS, Boockfor FR, Hoeffler JP 1985 Identification
by plaque assays of a pituitary cell type that secretes
both growth hormone and prolactin. Endocrinology
116:734-737
Frawley LS, Clark CL, Schoderbek WE, Hoeffler JP,
Boockfor FR 1986 A novel bioassay for lactogenic activity:
demonstration that prolactin cells differ from one another
in bio- and immuno-potencies of secreted hormone. Endocrinology 119:2867-2869
Wise PM, Camp-Grossman P, Barraclough CA 1981 Effects of estradiol and progesterone on plasma gonadotropins, prolactin, and LHRH in specific brain areas of ovariectomized rats. Biol Reprod 24:820-830
Sortino MA, Wise PM 1989 Effects of age and long term
ovariectomy on prolactin secretion, as assessed by the
reverse hemolytic plaque assay. Endocrinology 124:9096
Sortino MA, Cronin MJ, Wise PM 1989 Relaxin stimulates
prolactin secretion from anterior pituitary cells. Endocrinology 124:2013-2015
Maniatis T, Fritsch EF, Sambrook J 1982 Molecular Cloning-A Laboratory Manual. Cold Spring Harbor Laboratories, Cold Spring Harbor, p 388
Cox KH, DeLeon DV, Angerer LM, Angerer RC 1984
Detection of mRNAs in sea urchin embryos by in situ
hybridization using asymmetric probes. Dev Biol
101:485-502