Department of Medicine, Cedars-Sinai Medical Center, Los Angeles, California 90048
Because pregnancy and estrogens both induce pituitary lactotroph hyperplasia, we assessed theexpression of pituitary cell cycle regulators in two models of murine pituitary hyperplasia. Femalemice were assessed during nonpregnancy, pregnancy, day of delivery, and postpartum. We alsoimplanted estradiol (E2) pellets in female mice and studied them for 2.5 months. Pituitary weightin female mice increased 2-fold after E2 administration and 1.4-fold at day of delivery, comparedwith placebo-treated or nonpregnant females. Pituitary proliferation, as assessed by proliferatingcell nuclear antigen and/or Ki-67 staining, increased dramatically during both mid-late pregnancyand E2 administration, and lactotroph hyperplasia was also observed. Pregnancy induced pituitarycell cycle proliferative and inhibitory responses at the G1/S checkpoint. Differential cell cycle reg-ulator expression included cyclin-dependent kinase inhibitors, p21Cip1, p27Kip1, and cyclin D1. Pi-tuitary cell cycle responses to E2 administration partially recapitulated those effects observed atmid-late pregnancy, coincident with elevated circulating mouse E2, including increased expressionof proliferating cell nuclear antigen, Ki-67, p15INK4b, and p21Cip1. Nuclear localization of pituitaryp21Cip1 was demonstrated at mid-late pregnancy but not during E2 administration, suggesting acell cycle inhibitory role for p21Cip1 in pregnancy, yet a possible proproliferative role during E2
administration. Most observed cell cycle protein alterations were reversed postpartum. Murinepituitary meets the demand for prolactin during lactation associated with induction of both cellproliferative and inhibitory pathways, mediated, at least partially, by estradiol. (Endocrinology153: 5011–5022, 2012)
Robust polyclonal lactotroph proliferation is inducedduring pregnancy (1–5), and in rats pituitary pro-
liferation peaks at delivery and returns to prepartumlevels several days after lactation initiation (1–3). Afterlactation, partial or complete regression of redundantlactotrophs occurs (6, 7). Pharmacological estradiol(E2) treatment also induces early lactotrophic hyper-plastic responses, angiogenesis, and prolactinoma de-velopment in rats, coincident with pituitary tumor-transforming gene (PTTG), basic fibroblast growthfactor, and vascular endothelial growth factor induc-tion (8, 9). Because E2 also potentiates lactotroph pro-liferation and cell cycle modulation during pregnancy,we studied effects of pregnancy and E2 administrationin murine pituitary glands.
Cyclin-dependent kinases (CDK), critical regulators ofcell cycle progression, are modulated by fluctuations inactivators (cyclins) or inhibitors (CDK inhibitors) (10).Little is known regarding endocrine-mediated molecularmechanisms underlying pituitary cell cycle events duringgestation and E2 administration, in part due to severaltechnical limitations. Pituitary glands comprise heteroge-neous populations of lactotroph, somatotroph, gonado-troph, corticotroph, thyrotroph, and folliculostellate cellcomponents (11). Accordingly, studies using whole pitu-itary glands may not reflect events occurring in a single cellsubtype. This would be important because cell cycle pro-tein responses to growth factors likely occur predomi-nantly in a subset of specific cells during the G1 and S
Endocrinology, October 2012, 153(10):5011–5022 endo.endojournals.org 5011
phases. Because only whole pituitary glands can be studiedin vivo, single-cell type alterations in these proteins couldbe masked or diluted by a larger compartment of predom-inantly noncycling pituitary cells. Finally, no mouse orhuman lactotroph cell line is available for in vitro study.Insights into mechanisms involved in lactotroph cycle reg-ulation have been obtained from experiments in CDK4-deficient mice, which exhibit decreased lactotroph mass,function, and postnatal proliferation (12, 13). BecauseCDK4 is an essential regulator of the G1/S checkpoint,lactotroph cell cycle progression is likely dependent oncomponents that converge on this enzyme.
Hence, it was hypothesized that gestation or E2 admin-istration would induce pituitary cell cycle progression atthe G1/S checkpoint and an inhibitory cell cycle responseonly after delivery. This study elucidates a parallel prolif-erative and inhibitory cell cycle response in both models,possibly restricting pituitary hyperplastic growth. BecauseCDK inhibitors are shown to be involved in either pro-gression or arrest of the cell cycle at the G1/S checkpoint,this appears to be a particularly important site of pituitaryproliferative control during gestation or E2 administra-tion. Most of the observed cell cycle alterations duringpregnancy were reversed at lactational and postweaningstages. These results elucidate pituitary cell cycle modu-lation in parallel to pituitary remodeling occurring duringpregnancy or E2 administration.
Materials and Methods
AnimalsWild-type (WT) mice of a C57/BL6 genetic background were
maintained in a light- (12 h light, 12 h darkness cycle) and tem-perature-controlled room. Animals were euthanized using CO2
chambers, and blood withdrawn directly from the heart, andpituitary glands harvested for analysis. Sera were stored at �80C until hormone assay. Experiments were approved by the In-stitutional Animal Care and Use Committee.
Reproductive studyWT virgin females at approximately 8–10 wk of age were
impregnated by caging with a WT male. The presence of a vag-inal mucus plug was considered as d 0.5 of pregnancy. Maleswere removed 2 wk later, before parturition.
Female mice were assigned to six reproductive stages: 1) non-pregnant; 2) 2 wk pregnancy [this stage was determined bycounting 14.5 (range 13.5–15.5) d after observing a vaginal plugand by embryo morphology consistent with this gestational age];3) day of delivery was determined as the 24-h period after pupswere delivered, on average at 20.5 (range 19.5–21.5) d of preg-nancy; 4) 3 wk lactation (the day of parturition was counted asd 0 of lactation, and pups were weaned on d 21 of lactation andmaternal pituitary glands harvested on the same day); 5) 3 wkafter weaning (this stage included mice whose pups were weaned
3 wk earlier, and females in groups 4 and 5 with litters smallerthan three pups were not included for study); and 6) 3 wk post-partum. In this group pups were euthanized on the day of par-turition. Maternal pituitary glands were harvested 3 wk afterdelivery.
Nonpregnant females were euthanized at approximately8–10 wk. In other groups, females were euthanized according totheir respective reproductive stages. For the weight analysis, thepituitary glands were fixed in 10% formalin and weighed 24 hafter collection.
Pellet administrationEight- to 10-wk-old WT female mice were surgically im-
planted under isoflurane anesthesia with 17�-estradiol pellet for90 d (1.5 mg/pellet; Innovative Research of America, Sarasota,FL) or placebo pellet. Mice were euthanized after 70–75 d. Freshpituitary glands were collected for assays and weighed immedi-ately after collection.
Protein analysisEach tissue lysate was prepared from four pituitary glands in
radioimmunoprecipitation assay buffer (Sigma, St. Louis, MO)containing protease inhibitor cocktail (Sigma). Protein concen-trations were measured by bicinchoninic assay protein assay(Thermo Scientific, Swdesboro, NJ). Equal amounts (30 �g) ofproteins were separated by NuPage Novex Bis-Tris gels (Invit-rogen, Carlsbad CA) and electroblotted onto polyvinylidene di-fluoride membranes (Millipore, Bedford, MA). Membraneswere incubated with anti-p15INK4b (1:200; ab53034; Abcam,Cambridge, MA); p16INK4a (1:200; sc-1207; Santa Cruz Bio-technology, Santa Cruz, CA); p18INK4c (1:500; 39-3400; Invit-rogen); p19INK4d (1:100; 39-3100; Invitrogen); p21Cip1 (1:200;BD-556431; BD PharMingen, San Diego, CA); p27Kip1 (1:200;sc-528;SantaCruzBiotechnology);p57Kip2 (1:500;P0357;Sigma);proliferating cell nuclear antigen (PCNA; 1:300; sc-56; SantaCruz Biotechnology); p53 (1:300; ab31333; Abcam); Phos-RbTyr356 (1:200; sc-56175; Santa Cruz Biotechnology); cyclinA (1:200; sc-596; Santa Cruz Biotechnology); cyclin B1 (1:500; ab72-100; Abcam); cyclin B2 (1:200; sc-28303; SantaCruz Biotechnology); cyclin D1 (1:200; ab16663; Abcam);cyclin D3 (1:200; sc-182; Santa Cruz Biotechnology); cyclin E(1:200; sc-481; Santa Cruz Biotechnology); glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 1:1,000; sc-25778;Santa Cruz Biotechnology); and actin (1:20,000; MAB1501;Chemicon, Temecula, CA) antibodies overnight, followed bythe corresponding secondary antibodies (1:1,000; GE Health-care, Indianapolis, IN). Proteins were visualized using the en-hanced chemiluminescence Western blotting detection re-agents (GE Healthcare) super signal Western system (Pierce,Rockford, IL). Scanning densitometry of protein bands wasdetermined by pixel intensity using NIH Image J software(National Institutes of Health, Bethesda, MD)and normalizedagainst that of actin or GAPDH.
Immunofluorescent and reticulin stainingsPituitary glands were dissected and fixed in 10% formalin for
paraffin sectioning. For immunofluorescence analysis, 5-�mparaffin pituitary sections were deparaffinized and rehydrated.Antigen retrieval was performed in 10 mM sodium citrate and0.25% Tween 20 by incubating slides at 98 C for 40 min and
5012 Toledano et al. Pituitary Cell Cycle during E2 and Pregnancy Endocrinology, October 2012, 153(10):5011–5022
cooling for 20 min. For nuclear antigens, slides were permeabil-ized with 1% Triton X-100 in PBS for 30 min. Sections wereincubated with blocking buffer for 1 h and then with anti-PCNA(1:50; sc-56; Santa Cruz Biotechnology); Ki-67 (1:1000;ab15580; Abcam); prolactin (1:500; sc-7805; Santa Cruz Bio-technology); GH (1:200; the National Hormone and PeptideProgram, National Institute of Diabetes and Digestive and Kid-ney Diseases, Torrance, CA, and Dr. A. F. Parlow, Harbor-UCLA Medical Center, Los Angeles, CA); and p21Cip1 (1:30;BD-556341; BD PharMingen) antibodies at 4 C overnight. In-cubation with corresponding secondary antibodies conjugatedto Alexa 488 or Alexa 568 fluorescent dyes (all MolecularProbes, Eugene, OR) was performed, and sections counter-stained with 4�,6�-diamidino-2-phenylindole (300 nM; D3571;Invitrogen). The samples were imaged with Leica TCS/SP spec-tral confocal scanner (Leica Microsystems, Mannheim,Germany).
PCNA proliferative index was determined using ImageJ Soft-ware (Image Processing and Analysis in Java; National Institutesof Health) and based on the number of positively stained nucleidivided by the total number of nuclei counted. Three fields eachcontaining approximately 500 cells were counted from each an-imal and two to three animals from each group analyzed. PCNA-immunoreactive cells were similarly examined to determine thepercentage of pituitary cells immunoreactive for both prolactinand PCNA of the total PCNA-immunoreactive cells counted.Silver staining for reticulin fiber detection was performedselectively.
RNA isolation and PCRRNA samples were prepared from four pituitary glands. To-
tal RNA was isolated using the RNeasy minikit (QIAGEN, Va-lencia, CA) and eluted in water. cDNA was synthesized from 500ng total RNA using 1 �g of oligo(deoxythymidine) primer, Su-perScript II reverse transcriptase, and accompanying reagents(Invitrogen). Real-time PCR was amplified in 25-�l reactionmixtures [100 ng template, 0.5 �M of each primer, 10 �l 2�SYBR Green master mix (Applied Biosystems, Foster City, CA)]using 95 C for 1 min, followed by 40 cycles of 95 C for 20 sec and60 C for 40 sec. GAPDH and 18s RNA were assessed as internalcontrols.
RadioimmunoassaySerum concentrations of E2 (picograms per milliliter) were
determined using the DSL-4800 assay kit (Diagnostic SystemsLaboratories, Webster, TX).
Statistical analysisDifferences between groups were analyzed by t test. Values
were considered significant when P � 0.05.
Results
Pituitary mass and lactotroph proliferation duringpregnancy and E2 administration
To address pituitary expansion in the mouse during E2
administration and pregnancy, we first assessed pituitaryweights and expression of two proliferation markers,
PCNA and Ki-67. Female mouse pituitary mass increased2- or 1.4-fold with E2 administration or at day of delivery,compared with placebo-treated or nonpregnant females,respectively (P � 0.0001 and � 0.001, Fig. 1, A and E).Pituitary to body weight ratio at day of delivery increasedby 60% compared with respective weights at nonpregnantstages (0.16 vs. 0.10, respectively, P � 0.0001, Fig. 1F).This expansion did not reverse after the day of delivery.However, rate of pituitary weight increase stabilized be-tween day of delivery and 3 wk after weaning (pituitary tobody weight ratio of 0.16 vs. 0.18, respectively, Fig. 1F).At 3 wk postpartum, 3 wk after pup removal at day ofdelivery, pituitary weight only partially reverted to pre-gestational levels (P � 0.01; see Fig. 2, A and B).
Increased pituitary weight correlated strongly with in-creased PCNA and Ki-67 immunostaining (Fig. 3). Onaverage, in nonpregnant females 0.8 � 0.6% of pituitarycells were PCNA positive (Table 1 and Fig. 3E). Con-versely, PCNA labeling increased to 4.4 � 1.6% of cells atday of delivery (P � 0.001, Table 1 and Fig. 3F). Ki-67antigen detection also confirmed increased proliferationduring E2 administration and at 2 wk of pregnancy andday of delivery stages vs. placebo administration and non-pregnancy, respectively (Fig. 3, P, L, and M vs. Fig. 3, Oand K). By Western analysis, increased pituitary PCNAexpression and Rb phosphorylation were observed duringE2 administration and at 2 wk pregnancy and day of de-livery, compared with placebo-treated and nonpregnantfemales, respectively (Fig. 3, A and B). Furthermore, ele-vated PCNA and Ki-67 labeling in E2-treated and 2 wkpregnant mice closely mirrored the 2- to 8-fold increasesin PCNA and Ki-67 mRNA expression (Fig. 3, C and D).Parallel staining for prolactin and PCNA or Ki-67 showedthat lactotroph cells comprised the main proliferating celltype either at 2 wk pregnancy and day of delivery or duringE2 administration (Fig. 3, L, M, and P, and Table 1). Pro-lactin coexpression in PCNA-positive cells increased from76 � 25 to 93 � 6% during nonpregnant and day ofdelivery stages, respectively (P � NS, Table 1). However,the proportion of somatotrophs in the PCNA-positivecells did not change during either nonpregnant or day ofdelivery stages (P � NS, 12 � 18 and 11 � 3%, respec-tively, Table 1 and Fig. 3, I and J).
The results confirm that marked lactotroph prolifera-tion occurred concurrent with increased pituitary weightduring mid-late pregnancy, and lactotroph stabilizationoccurred after the late phase of lactation. During 2.5-months of E2 administration, pituitary replication was ev-ident by increased protein levels, mRNA induction, andimmunostaining of proliferation markers. Pituitary pro-liferation rates during E2 administration also correlatedwith serum E2 levels (Fig. 1D).
Endocrinology, October 2012, 153(10):5011–5022 endo.endojournals.org 5013
Pituitary expansion during estradiol administrationand pregnancy is related to hyperplasia
Anterior pituitary cells were diffusely distributed at 2wk pregnancy, day of delivery, and 3 wk lactation, and thereticulin fiber network was intact, consistent with the def-inition of lactotroph hyperplasia (Fig. 4, A–D). A similarpattern was detected in specimens derived from E2- vs.placebo-administered mice (Fig. 4, G vs. F). However, thehistological distribution became more nodular at 3 wklactation (Fig. 4D). Correspondingly, reticulin fiber dis-solution accompanied hyperplasia in several fields, butdistinct pituitary tumors were not observed. Reticulin fi-ber distribution at 3 wk postpartum (Fig. 4E) demon-strated a reversal to the nonpregnant pattern (Fig. 4A).
Differentially expressed pituitary cell cycleregulation genes at mid-late gestation
We isolated protein and total RNA samples fromfemale pituitaries at different reproductive stages, in-cluding nonpregnant, 2 wk pregnancy, day of delivery,3 wk lactation, and 3 wk after weaning. We identifiedseveral gene products differentially expressed in pitu-itary lysates compared with nonpregnant controls.The most marked changes were observed at 2 wk ofpregnancy. At this stage, elevated protein levels ofCDK inhibitors p15INK4b and p21Cip1 were detected(Fig. 5, B and E). In contrast, p16INK4a and p27Kip1
levels were decreased, whereas p18INK4c and p57Kip2
levels were unchanged.
FIG. 1. Murine pituitary expansion after 2.5 months E2 administration and at different reproductive stages. Pituitary weights (milligrams) andpituitary to body weight ratios (milligrams per gram) in placebo- and E2-treated female mice (n � 20–25/group) (A and B) and female mice at fivereproductive stages (E and F): non-P, P2w, DOD, L3w, and 3w post-W (n � 7–13/group). Results are expressed as means � SD. For calculatingweight ratios of females, body weight at the day of mating, i.e. pregestational weight, was considered. C, Pituitaries isolated from female micetreated with placebo (left) and E2 (right) for 2.5 months. Scale bar, 5 mm. D, Serum E2 concentrations (picograms per milliliter) in placebo- andE2-treated female mice. Levels are expressed as means � SD. *, P � 0.05; **, P � 0.01; †, P � 0.001; ‡, P � 0.0001. N, Number of mice pergroup; non-P, nonpregnant; P2w, 2-wk pregnancy; DOD, day of delivery; L3w, 3 wk lactation; 3w-post-W, 3 wk after weaning.
5014 Toledano et al. Pituitary Cell Cycle during E2 and Pregnancy Endocrinology, October 2012, 153(10):5011–5022
Quantitative PCR confirmation of selected differen-tially expressed genes concurred with Western blottingresults. Specifically, p15INK4b and p21Cip1 increased 2.3-and 5.9-fold, respectively, at 2 wk pregnancy (Fig. 6B),and immunostaining demonstrated increased pituitarycell p21Cip1 expression at 2 wk pregnancy, concurrentwith increased Ki-67 labeling (Fig. 6D). Nuclear Ki-67labeling is reflective of proliferating cells, and p21Cip1 wasnot colocalized in these cells (Fig. 6D) but did localize inthe nuclear compartment of lactotroph and nonlactotrophcells (Fig. 6F). p21Cip1 staining was decreased at day ofdelivery, with less nuclear and more cytoplasmic expres-sion (Fig. 6E).
Cyclin D1, important for G1 progression, was also dif-ferentially expressed with highest expression observed inmid-late pregnancy (Fig. 5, D and E). Levels of cyclin E andcyclins B1 and B2, important for G1/S and G2/M progres-
sion, respectively, were stable acrossthe reproductive stages (Fig. 5D).
As shown above (Fig. 3B), pituitaryPCNA protein expression reversed topregestational levels at 3 wk lactationand 3 wk after weaning. Moreover, el-evated PCNA expression observed atthe day of delivery also decreased 3 wkpostpartum (Fig. 2C), i.e. 3 wk afterpup removal at day of delivery. Corre-spondingly, specific pituitary stainingfor Ki-67 was undetectable at 3 wkpostpartum (Fig. 3N). PCNA expres-sion decreased at all these stages com-pared with expression observed in non-pregnant mice (Fig. 3B). Generally,expression of pituitary cell cycle regu-lators reversed to nonpregnant levels atday of delivery or at 3 wk lactation (Fig.5B). Moreover, p15INK4b and p16INK4a
expression revealed a dual pattern after2 wk pregnancy. After p15INK4b levelsdecreased to a trough at 3 wk lactation,they increased at 3 wk after weaning.Conversely, p16INK4a peaked at day ofdelivery and then decreased again at 3wk lactation (Fig. 5B). Cyclin D1 ex-pression decreased after day of deliveryto a lower level than nonpregnancy(Fig. 5D), similar to the observed re-bound decrease pattern of PCNA.
The general reversibility of CDK inhibitor levels afterdelivery was concordant with their expression at 3 wkpostpartum, i.e. without a 3-wk lactation period (Fig. 2C).Furthermore, p16INK4a expression was not reversed butrather elevated at 3 wk postpartum, in contrast to thepattern observed at lactation (Fig. 5B).
Pituitary cell cycle regulation during E2
administrationUnlike short-term serum E2 level peaks observed during
estrus in female mice, longer sustained increased E2 levelsare present during the second half of pregnancy (14). Wetherefore studied long-term E2 administration as a modelfor lactotroph hyperplasia. As shown above, pituitary cellproliferation increased with E2 administration and at re-productive stages 2 wk pregnancy and day of delivery.Some pituitary cell cycle regulators differentially ex-
FIG. 2. Pituitary gestational expansion, proliferation, and cell cycle alterations reverse partiallyor completely within 3 wk of pup removal on DOD. A and B, Pituitary weights (milligrams)and pituitary to body weight ratios (milligrams per gram) in female mice at the indicatedreproductive stages: non-P, DOD, and 3w-PP (n � 4–12/group). Data are expressed as means� SD. C, Protein expression of PCNA and the indicated CDK inhibitors that are altered at DODreverse partially or completely at 3w-PP back to the expression level of the nonpregnantstage. Western blot analysis of the indicated proteins in total protein extracts (30 �g each)from pituitaries of four wild-type mice per time point. **, P � 0.01; †, P � 0.001; ‡, P �0.0001. Non-P, Nonpregnant; DOD, day of delivery; 3w-PP, 3 wk postpartum and after pupremoval on day of delivery.
Endocrinology, October 2012, 153(10):5011–5022 endo.endojournals.org 5015
pressed during E2 administration were comparably ex-pressed at reproductive stage 2-wk pregnancy, includingp15INK4b, p16INK4a, p21Cip1, and p27KIP1(Fig. 5, A, B, andF). p57KIP2 expression levels were not changed with eitherE2 administration or at reproductive stage 2 wk preg-nancy. Concordant with Western blotting results, quan-titative PCR detected comparable p15INK4b and p21Cip1
mRNA elevations both during E2 administration and at 2
wk pregnancy (1.9- to 2.3-fold and 4.6- to 5.9-fold, re-spectively, Fig. 6, A and B). p53 expression was unchangedduring pregnancy or E2 administration (data not shown).Whereas cyclins B1 and B2 increased with E2 administra-tion, levels of cyclin D1 did not vary (Fig. 5, C and F). Incontrast, cyclin D1 increased at 2 wk pregnancy, but cyclinB1 and B2 levels were not altered at different reproductivestages (Fig. 5D).
In summary, pituitary cell cycle regulation during E2
administration partially recapitulated findings observedduring pregnancy, specifically at the reproductive stage of2 wk of pregnancy.
Discussion
Whereas monoclonal lactotroph proliferation underliesprolactinoma formation, the prototype of extensive poly-clonal lactotroph proliferation, culminating in pituitaryhyperplasia, occurs during pregnancy (4, 5, 15, 16). Ges-
TABLE 1. Pituitary PCNA labeling and anterior lobehormone coexpression at different reproductive stages inWT female mice
PCNA, prolactin, and GH immunoreactivity were assessed by confocalmicroscopy in mouse pituitary specimens (see Fig. 3). Non-P,Nonpregnant; DOD, day of delivery.
FIG. 3. Pituitary cell proliferation increases in WT mice during 2.5 months E2 administration and at mid-late pregnancy. Western blot analysis andreal-time PCR of proliferation markers in pituitary lysates derived from placebo- and E2-treated female mice (A and C) and from females atdifferent reproductive stages (B and D) (n � 4 animals per group) are shown. Specifically, lactotroph proliferation is induced at pregnancy andduring E2 administration. PCNA (green, E and F), PCNA and prolactin (red and green, respectively, G and H), PCNA and GH (red and green,respectively, I and J), Ki-67 and prolactin (red and green, respectively, K-P). Image size, 250 � 250 �m; inset size, 40 � 40 �m. mRNA values areexpressed as means � SE. *, P � 0.05; **, P � 0.01. non-P, Nonpregnant; P2w, 2-wk pregnancy; DOD, day of delivery; L3w, 3 wk lactation; 3w-post-W, 3 wk after weaning, 3w-PP, 3 wk postpartum.
5016 Toledano et al. Pituitary Cell Cycle during E2 and Pregnancy Endocrinology, October 2012, 153(10):5011–5022
tational lactotroph proliferation may be mediated by hor-monal stimuli (4), including E2 (8, 9). However, E2- andpregnancy-associated pituitary hyperplasia confers no in-creased risk of pituitary adenoma (17). In addition, mostprolactin tumors in women who achieved pregnancy donot enlarge or become clinically significant (18, 19). Thereis scant knowledge of molecular regulators stimulatingand reversing pregnancy-induced lactotroph replication.In this study we showed that as pituitary weight increasesduring pregnancy or E2 administration, the hypophysealcell cycle response also includes activating antiprolifera-tive mechanisms. Specifically, pituitary p15INK4b andp21Cip1 are up-regulated during mid-late gestation, coin-cident with induced proliferation. E2 administration par-tially recapitulates gestation-induced pituitary cell cycleexpression profiles.
Pituitary expansion at mid-late pregnancy strongly cor-relates with increased PCNA and Ki-67 expression, andimmunostaining of these two proliferation markers dem-onstrates increased replication from the nonpregnantstage to day of delivery. Previous analyses of pituitaryproliferation during rodent and human pregnancy werebased mainly on morphological criteria (1–3) includingDNA synthesis and PCNA and 5-bromo-2�-deoxyuridinelabeling. Proliferative labeling indices showed low lac-totroph proliferative activity during pregnancy and lacta-tion and high activity at estrus and the day of parturition.Suckling stimuli in pup-removed lactating rats elicited amarked increase in 5-bromo-2�-deoxyuridine labeling in-dices in the presence of E2 (3). Accordingly, we found that
lactotroph proliferation also peaked at day of delivery.However, the kinetic analysis reveals that this prolifera-tion starts as early as the end of second week of pregnancy.Indeed, several studies have reported increased lactotrophmitosis in women at earlier stages of human pregnancy (4,5) in concordance with our murine results.
In addition to gestational pituitary expansion, our re-sults show elevated pituitary proliferation markers duringE2 administration. Pituitary hyperplasia was confirmed bythe presence of intact reticulin fibers during either mid-latepregnancy or E2 administration. As predicted, parallelstaining for prolactin and either PCNA or Ki-67 indicatedthat lactotroph cells comprise the main proliferating pi-tuitary cell type during E2 administration, pregnancy (2wk pregnancy), and immediate peripartum stage (day ofdelivery).
The present results also confirm reversal of prolifera-tive pituitary activity postpartum or after the late lactativephase, possibly related to low postpartum serum E2 levels(14, 20). Surprisingly, decreased molecular proliferationmarker expression, including PCNA and Ki-67, after theday of delivery did not translate to pituitary mass reorga-nization. The striking increase in pituitary weight duringpregnancy does not reverse but rather stabilizes through-out lactation and postweaning stages. Nonproliferativeprocesses including cytoplasmic accumulation may com-prise the main component of gestational pituitary weightgain, whereas pituitary cell proliferation may be a minorcontributor. In support of this, Castrique et al. (21) re-ported that in prolactin-Cre/ROSA-YFP transgenic mice,
FIG. 4. Reticulin silver staining in female WT mouse specimens at the indicated reproductive stages (A–E) and during 2.5 months placebo (F) andE2 (G) administration. Representative sections at �100 magnification are shown. Non-P, Nonpregnant; P2w, 2-wk pregnancy; DOD, day ofdelivery; L3w, 3 wk lactation; 3w-PP, 3 wk postpartum.
Endocrinology, October 2012, 153(10):5011–5022 endo.endojournals.org 5017
cell volumes of prolactin-positive pituitary cells increasedsignificantly by midlactation.
Alterations in cell cycle regulators such as p27Kip1,p18 INK4c, and high mobility group protein A-2 can lead topituitary neoplasia or hyperplasia development (10, 22–24).CDK4 is critical for regulation of lactotroph mass and func-tion, especially postnatal cell proliferation (12, 13). We nowdemonstrate the span of cell cycle regulators in the mousepituitary gland during E2 administration and during preg-nancy, lactation, and the postweaning period. In E2-admin-istered mice, regulators were shown to be differentially ex-pressed, including INK4 inhibitors, CIP/KIP inhibitors, andcyclins. Generally, the protein expression profile of cell cycleproteins, including p15INK4b, p16INK4a, p21Cip1, andp27Kip1, in the expanding pituitary during E2 administrationpartially mimics effects of pregnancy on pituitary cell cycle,specifically at 2 wk pregnancy. Pituitary CDK4 activity dur-ing pregnancy may be positively regulated by association withcyclin D1 and by decreased p16INK4a. Noteworthy, expressionoftheseproteins is frequentlyalteredinhumanpituitarytumors(13, 25–28). Furthermore, up- or down-regulation of pituitary
cell cycle regulatorsdetectedatmid-latepregnancyare reversedto pregestational levels postpartum. Serum E2 levels duringmurine gestation are low in early pregnancy, increase duringmid-latepregnancyuptod17andthendecrease rapidlyafterdelivery and throughout lactation (14, 20). Thus, expressionpatterns of pituitary cell cycle regulators during the last weekof pregnancy suggests a mechanistic role for E2.
Both p15INK4b and p21Cip1 act as negative regulators ofCDK4 and CDK2, respectively, impacting cell prolifera-tion in vitro and in vivo. p21Cip1 also blocks S phase pro-gression by inhibiting PCNA activity, thus causing cellcycle arrest (29). p15INK4b induction was associated withdecreased proliferative activity in rat lactosomatotrophGH3 cells treated with TGF-�1 (30). Elevated p15INK4b
and p21Cip1 expression in pituitary hyperplasic states invivo may be important for the cell cycle response to E2
administration and to pregnancy. The role for p21Cip1 inregulating pituitary mass in the adult mouse and humanGH- and prolactin-secreting pituitary adenomas (31, 32)implies that p21Cip1 may regulate pituitary trophic ho-
FIG. 5. Protein expression of CDK inhibitors and cell cycle regulators is differentially expressed in female mice administered placebo or E2 for 2.5months (A, C, and F) and in females at the indicated reproductive stages (B, D, and E). Protein expression profile during E2 administration partiallymimics reproductive stage P2w. Western blot analysis of the indicated proteins in total protein extracts (30 �g each), each isolated from fourpituitary glands per time point. Levels of the indicated proteins are determined by densitometry (E and F). Values are expressed as means � SE.*, P � 0.05; **, P � 0.01. Non-P, nonpregnant; P2w, 2-wk pregnancy; DOD, day of delivery; L3w, 3 wk lactation; 3w-post-W, 3 wk afterweaning; cyc D1, cyclin D1.
5018 Toledano et al. Pituitary Cell Cycle during E2 and Pregnancy Endocrinology, October 2012, 153(10):5011–5022
meostasis by constraining tumorigenesis when exposed topro-proliferative stimuli.
E2 activates p21Cip1 expression through direct promoterregulation in MCF-7 breast cancer cells (33). We show ele-vated pituitary p21Cip1 expression during E2 administrationandgestation,withoutelevatedp53expression.Whereascellgrowth-inhibiting activity of p21Cip1 correlates with nuclearlocalization, p21Cip1 relocalization to the cytoplasm pro-motes cell survival and proliferation (29). Estrogen stimu-lates p21Cip1 and uterine cell proliferation, whereas cotreat-ment with progesterone prevents both effects (34). In ourstudy, intranuclear p21Cip1 localization was evident mainlyat 2 wk pregnancy. Furthermore, Ki-67, a marker of cyclingcells, did not coexpress with nuclear p21Cip1, suggestingp21Cip1-growth-inhibiting activity. Progesterone levels in-crease throughout pregnancy (14, 20) and are not associated
with lactotroph replication. Thus, the gestational estrogenmilieu may potentiate pituitary proliferation and p21Cip1
transcriptional induction at 2 wk pregnancy. Concomi-tantly, other gestational-related factors mediate p21Cip1 nu-clear location restraining pituitary proliferation. At the dayof delivery, with the altered hormonal milieu, and effects ofunopposed serum E2, p21Cip1 may be transported to the cy-toplasm to mediate suckling and E2-induced pituitary pro-liferation. This may explain higher pituitary proliferationrates at the day of delivery compared with 2 wk pregnancy.During E2 administration, we demonstrated increased pitu-itary proliferation and nonnuclear p21Cip1 expression, sug-gesting that the E2-induced p21Cip1 indeed promotes ratherthan inhibits replication. Simultaneously, thep16INK4a levelsincrease at day of delivery, potentially attenuating the pitu-itary proliferative drive at this stage.
FIG. 6. p21Cip1 and p15INK4b mRNA levels are induced during E2 administration and at 2 wk pregnancy in WT female mice. Real-time PCR ofp21Cip1 and p15INK4b mRNA was performed with extracts derived from placebo- and E2-treated mice (A) and from females at non-pregnant andP2w (B) (n � 4 animals per group). By histological analysis of p21Cip1 at P2w and DOD reproductive stages and during 2.5 months placebo and E2
administration, p21Cip1 is up-regulated and localized to the nucleus at P2w. p21Cip1 and Ki-67 (green and red and arrows and closed arrowheads,respectively, C–E), p21Cip1 and prolactin (red and green and open arrows and open arrowheads, respectively, F), p21Cip1 (green, G and H). mRNAvalues are expressed as means � SE. *, P � 0.05; **, P � 0.01. Non-P, Nonpregnant; P2w, 2-wk pregnancy; DOD, day of delivery.
Endocrinology, October 2012, 153(10):5011–5022 endo.endojournals.org 5019
This study has several limitations. First, we recognizethat pituitary cell cycle control may reflect species andstrain specificity. Second, the lack of adequate functionaldifferentiated mouse or human lactotroph cell culturesmakes comprehensive cell cycle analysis difficult, andmuch insight into human pituitary cell proliferation is de-rived from in vivo animal studies. Third, cell cycle regu-lators are likely only some of several target genes requiredfor pituitary hyperplasia and partial postpartum involu-tion. Unique growth factor regulation of lactotroph cellgrowth and function also likely contributes to pituitarymodeling (35). Furthermore, additional pituitary, ovar-ian, fetal, and placental interrelationships are also likelyoperative during gestation (36).
Comparison of two lactotroph hyperplasia paradigms,E2-induced and gestational/suckling-induced, suggests di-verse mechanisms used by the pituitary gland to expandmass. Some genes investigated are not concordantly in-duced in the two models, suggesting that pituitary mass
modulation is activated by diverse signaling mechanisms,depending on physiological and pharmacological condi-tions. In support of this notion, several cell cycle compo-nents likely control pituitary mass homeostasis throughboth pro-proliferative and inhibitory signals. Gestational,lactational, and postweaning pituitary cell cycle progres-sion and arrest likely follow an orchestrated temporal cas-cade determined by the hormonal milieu including E2
levels.We propose a model for this cascade in gestational and
E2-induced pituitary hyperplasia (Fig. 7). Pregnancy andE2 stimuli are required to induce pituitary proliferationthrough the down-regulation of p16INK4a and p27Kip1.Increased p15INK4b expression may attenuate pituitary ex-pansion and delay its kinetics. E2 may also trigger thep21Cip1 promoter, but p21Cip1 is not localized to the nu-cleus during E2 administration, suggesting a role in pro-moting proliferation. At mid-late pregnancy, i.e. 2 wkpregnancy, nuclear p21Cip1 localization regulated by
FIG. 7. Proposed model for pituitary cell cycle regulation during mid-late pregnancy (P2w) and during E2 administration in WT female mice.Pregnancy and E2 administration induce alterations in levels of pituitary cell cycle regulators in a G1/S-dependent manner. The altered levels caneither stimulate or inhibit pituitary cell cycle progression and proliferation. Some alterations induced by E2 administration are mimicked bypregnancy, and others are unique for either one of these stimuli. p21cip1exhibits a dual role. E2 administration also induces cyclins B1 and B2,which have a role in cell cycle progression distal to the G1/S point. Macroscopically, the pituitary is remodeled from a normal to a hyperplasticgland, without formation of pituitary tumors. Cell cycle alterations are reversible after pregnancy, at postpartum, and postlactational stages. P2w,2-wk pregnancy.
5020 Toledano et al. Pituitary Cell Cycle during E2 and Pregnancy Endocrinology, October 2012, 153(10):5011–5022
nonestrogenic factors may limit pituitary proliferation.p21Cip1 possibly represents a rate-limiting pathway, pre-venting a more robust proliferative response to pregnancy.We show that p15INK4b and p21Cip1 levels closely mirrorproliferative profiles throughout pregnancy, further sug-gesting that these components of cell cycle control deter-mine pituitary homeostasis, also in the setting of pharma-cological E2 administration. We hypothesize that pituitaryp21Cip1 up-regulation has unique functions, either pre-venting or promoting excessive pituitary cell replication.Thus, p21Cip1 may be a promising target for investigatingtherapeutic manipulations of pituitary neoplasms, specif-ically of lactotroph and/or somatotroph origin.
Acknowledgments
We thank Lihua Xia for her excellent technical assistance.
Address all correspondence and requests for reprints to:Shlomo Melmed, M.D., Academic Affairs, Cedars-Sinai MedicalCenter, Room 2015, 8700 Beverly Boulevard, Los Angeles, Cal-ifornia 90048. E-mail: [email protected].
This work was supported by Grant CA-75979 from the Na-tional Institutes of Health and The Doris Factor Molecular En-docrinology Laboratory.
Disclosure Summary: The authors have declared that no con-flict of interest exists.
References
1. Carretero J, Rubio M, Blanco E, Burks DJ, Torres JL, Hernández E,Bodego P, Riesco JM, Juanes JA, Vázquez R 2003 Variations in thecellular proliferation of prolactin cells from late pregnancy to lac-tation in rats. Ann Anat 185:97–101
2. Jahn GA, Burdman JA, Deis RP 1984 Regulation of pituitary DNAsynthesis during different reproductive states in the female rat: roleof estrogens and prolactin. Mol Cell Endocrinol 35:113–119
3. Yin P, Arita J 2000 Differential regulation of prolactin release andlactotrope proliferation during pregnancy, lactation and the estrouscycle. Neuroendocrinology 72:72–79
4. Scheithauer BW, Sano T, Kovacs KT, Young Jr WF, Ryan N, Ran-dall RV 1990 The pituitary gland in pregnancy: a clinicopathologicand immunohistochemical study of 69 cases. Mayo Clin Proc 65:461–474
5. Stefaneanu L, Kovacs K, Lloyd RV, Scheithauer BW, Young Jr WF,Sano T, Jin L 1992 Pituitary lactotrophs and somatotrophs in preg-nancy: a correlative in situ hybridization and immunocytochemicalstudy. Virchows Arch B Cell Pathol Incl Mol Pathol 62:291–296
6. Asa SL, Penz G, Kovacs K, Ezrin C 1982 Prolactin cells in the humanpituitary. A quantitative immunocytochemical analysis. Arch PatholLab Med 106:360–363
7. Haggi ES, Torres AI, Maldonado CA, Aoki A 1986 Regression ofredundant lactotrophs in rat pituitary gland after cessation of lac-tation. J Endocrinol 111:367–373
8. Heaney AP, Horwitz GA, Wang Z, Singson R, Melmed S 1999 Earlyinvolvement of estrogen-induced pituitary tumor transforming gene
and fibroblast growth factor expression in prolactinoma pathogen-esis. Nat Med 5:1317–1321
9. Heaney AP, Fernando M, Melmed S 2002 Functional role of estro-gen in pituitary tumor pathogenesis. J Clin Invest 109:277–283
10. Quereda V, Malumbres M 2009 Cell cycle control of pituitary de-velopment and disease. J Mol Endocrinol 42:75–86
11. Melmed S 2003 Mechanisms for pituitary tumorigenesis: the plasticpituitary. J Clin Invest 112:1603–1618
12. Moons DS, Jirawatnotai S, Parlow AF, Gibori G, Kineman RD,Kiyokawa H 2002 Pituitary hypoplasia and lactotroph dysfunctionin mice deficient for cyclin-dependent kinase-4. Endocrinology 143:3001–3008
13. Jirawatnotai S, Aziyu A, Osmundson EC, Moons DS, Zou X, Ki-neman RD, Kiyokawa H 2004 Cdk4 is indispensable for postnatalproliferation of the anterior pituitary. J Biol Chem 279:51100–51106
14. Barkley MS, Geschwind II, Bradford GE 1979 The gestational pat-tern of estradiol, testosterone and progesterone secretion in selectedstrains of mice. Biol Reprod 20:733–738
15. Goluboff LG, Ezrin C 1969 Effect of pregnancy on the somatotrophand the prolactin cell of the human adenohypophysis. J Clin Endo-crinol Metab 29:1533–1538
17. Coogan PF, Baron JA, Lambe M 1995 Parity and pituitary adenomarisk. J Natl Cancer Inst 20:1410–1411
18. Shibli-Rahhal A, Schlechte J 2011 Hyperprolactinemia and infer-tility. Endocrinol Metab Clin North Am 40:837–846
19. Bronstein MD, Paraiba DB, Jallad RS 2011 Management of pitu-itary tumors in pregnancy. Nat Rev Endocrinol 7:301–310
20. Smith MS, Neill JD 1977 Inhibition of gonadotropin secretion dur-ing lactation in the rat: relative contribution of suckling and ovariansteroids. Biol Reprod 17:255–261
21. Castrique E, Fernandez-Fuente M, Le Tissier P, Herman A, Levy A2010 Use of a prolactin-Cre/ROSA-YFP transgenic mouse providesno evidence for lactotroph transdifferentiation after weaning, orincrease in lactotroph/somatotroph proportion in lactation. J En-docrinol 205:49–60
22. Lazzerini Denchi E, Attwooll C, Pasini D, Helin K 2005 DeregulatedE2F activity induces hyperplasia and senescence-like features in themouse pituitary gland. Mol Cell Biol 25:2660–2672
23. Fedele M, Visone R, De Martino I, Troncone G, Palmieri D, BattistaS, Ciarmiello A, Pallante P, Arra C, Melillo RM, Helin K, Croce CM,Fusco A 2006 HMGA2 induces pituitary tumorigenesis by enhanc-ing E2F1 activity. Cancer cell 9:459–471
24. Wierinckx A, Auger C, Devauchelle P, Reynaud A, Chevallier P, JanM, Perrin G, Fèvre-Montange M, Rey C, Figarella-Branger D,Raverot G, Belin MF, Lachuer J, Trouillas J 2007 A diagnosticmarker set for invasion, proliferation, and aggressiveness of prolac-tin pituitary tumors. Endocr Relat Cancer 14:887–900
25. Hibberts NA, Simpson DJ, Bicknell JE, Broome JC, Hoban PR,Clayton RN, Farrell WE 1999 Analysis of cyclin D1 (CCND1) allelicimbalance and overexpression in sporadic human pituitary tumors.Clin Cancer Res 5:2133–2139
26. Woloschak M, Yu A, Post KD 1997 Frequent inactivation of the p16gene in human pituitary tumors by gene methylation. Mol Carcinog19:221–224
27. Simpson DJ, Bicknell JE, McNicol AM, Clayton RN, Farrell WE1999 Hypermethylation of the p16/CDKN2A/MTSI gene and lossof protein expression is associated with nonfunctional pituitary ad-enomas but not somatotrophinomas. Genes Chromosomes Cancer24:328–336
28. Seemann N, Kuhn D, Wrocklage C, Keyvani K, Hackl W, Buch-felder M, Fahlbusch R, Paulus W 2001 CDKN2A/p16 inactivationis related to pituitary adenoma type and size. J Pathol 193:491–497
29. Abbas T, Dutta A 2009 p21 in cancer: intricate networks and mul-tiple activities. Nat Rev Cancer 9:400–414
Endocrinology, October 2012, 153(10):5011–5022 endo.endojournals.org 5021
30. Frost SJ, Simpson DJ, Farrell WE 2001 Decreased proliferation andcell cycle arrest in neoplastic rat pituitary cells is associated withtransforming growth factor-�1-induced expression of p15/INK4B.Mol Cell Endocrinol 176:29–37
31. Chesnokova V, Kovacs K, Castro AV, Zonis S, Melmed S 2005Pituitary hypoplasia in Pttg�/� mice is protective for Rb�/� pitu-itary tumorigenesis. Mol Endocrinol 19:2371–2379
33. Mandal S, Davie JR 2010 Estrogen regulated expression of the p21
Waf1/Cip1 gene in estrogen receptor positive human breast cancercells. J Cell Physiol 224:28–32
34. Lai MD, Jiang MJ, Wing LY 2002 Estrogen stimulates expression ofp21Waf1/Cip1 in mouse uterine luminal epithelium. Endocrine 17:233–239
35. Fukuoka H, Cooper O, Mizutani J, Tong Y, Ren SG, Bannykh S,Melmed S 2011 HER2/ErbB2 receptor signaling in rat and humanprolactinoma cells: strategy for targeted prolactinoma therapy. MolEndocrinol 25:92–103
36. Karaca Z, Tanriverdi F, Unluhizarci K, Kelestimur F 2010 Preg-nancy and pituitary disorders. Eur J Endocrinol 162:453–475
5022 Toledano et al. Pituitary Cell Cycle during E2 and Pregnancy Endocrinology, October 2012, 153(10):5011–5022
