Minireview: Prolactin Regulation of Adult Stem Cells

Abstract

Adult stem/progenitor cells are found in many tissues, where their primary role is to maintain homeostasis. Recent studies have evaluated the regulation of adult stem/progenitor cells by prolactin in various target tissues or cell types, including the mammary gland, the prostate, the brain, the bone marrow, the hair follicle, and colon cancer cells. Depending on the tissue, prolactin can either maintain stem cell quiescence or, in contrast, promote stem/progenitor cell expansion and push their progeny towards differentiation. In many instances, whether these effects are direct or involve paracrine regulators remains debated. This minireview aims to overview the current knowledge in the field.

Stem cells are believed to be present in many, if not all, adult organs. So-called “adult stem cells” are usually quiescent, and one of their primary roles is to maintain and/or repair the tissue in which they are found, making them key players in tissue and organism homeostasis. This property relies on their unique ability to self-renew and to differentiate into multiple specialized cell types of the host tissue. Progenitor cells, in turn, show limited self-renewal and can give rise to 1 (or few) mature cell types (1). Historically, the best characterized stem cells have been those of the hematopoietic lineage; the first review articles referenced in PubMed appeared in the 1960s. Since these pioneering reports, growing evidence for the existence of adult stem cells in several other tissues has accumulated.

One of the models of choice for the study of adult stem cells in epithelial tissue is the crypt-villus system of the small intestine, thanks to the very short life cycle (4–5 d) of its epithelial cell layer that requires permanent renewal (2). Studies of this peculiar system led to the discovery that both slow/noncycling and fast-cycling intestinal stem cells coexist. The fast-cycling stem cells that express Lgr5 (leucine-rich repeat-containing G protein-coupled receptor 5) (3) are the engines of crypt self-renewal: they can generate a population of slow/nondividing daughter cells that can either differentiate into Paneth cells or, in case of damage, be used as “reserve stem cells” that can reacquire the ability to express Lgr5 and give rise to other differentiated intestinal cells (4, 5). In all, it seems that under physiological conditions, certain tissues like the intestine and skin can self-renew constantly via asymmetric division of stem cells. In contrast, other tissues mainly rely on multipotent progenitors for self-renewal (hematopoietic system), or on the replication of differentiated, mature cells (liver and pancreatic β-cells) (6, 7). In addition to these physiological mechanisms of self-renewal, tissue injury or aggression also can activate self-renewal processes, eg, the prostate epithelium after castration and androgen restitution (8). The activation of these stem/progenitor cells eventually leads to tissue repair and regeneration.

Thanks to their regenerating capacities, adult stem cells add potential value to the current therapeutic arsenal, as highlighted for decades by hematopoietic stem cells from bone marrow used for transplantation purposes. The more recent discoveries that adult stem cells also reside in organs long thought to be unable to regenerate, such as the brain or the heart, have opened new routes for developing unsuspected cell-based therapies for neurologic disorders or heart diseases (9). The manipulation of adult somatic cells into induced pluripotent stem cells offers great promise in this field as well (10). Finally, within recent years, stem cells have also emerged as potential drivers of, and hence as new targets for, cancer initiation and perhaps even more cancer recurrence. For example, chemotherapy-resistant breast cancer cells exhibit stem-like properties making them good candidates for initiating breast cancer regrowth upon escape after initial treatment (11). Whether these cells are true “cancer stem cells,” resulting from oncogenic transformation of stem cells, or whether they represent dedifferentiated cells resulting from the phenotypic conversion of transformed epithelial cells (eg, through epithelial-mesenchymal transition [EMT]), remains a matter of debate (12–14), which falls beyond the scope of this minireview.

The microenvironment where stem cells are localized within each tissue provides signals regulating their quiescence, self-renewal, and survival, which are essential for stem cell homeostasis. This microenvironment, called the stem cell niche, includes the stem cells and their progeny, surrounding mesenchymal or stromal cells, extracellular matrix, and other cell types, such as endothelial and neural cells (15). In each tissue, the stem cell niche presents particular properties, which involve regulatory autocrine, paracrine, and/or endocrine signals (15). The main signaling pathways known to regulate stem cell homeostasis involve the TGF-β superfamily, the Wingless-type mammary tumor virus integration site family (Wnt) pathway, and Notch signaling; however, many other factors have been described to play a role (15). In this minireview we will discuss what is known about the effects of the hormone prolactin (PRL) on adult tissue stem or progenitor cells.

The PRL Axis

Intracellular signaling

PRL is a pituitary-secreted polypeptide hormone of 23 000 Da. It elicits its biological functions via a specific receptor, the PRL receptor (PRLR). As an archetype member of the cytokine receptor superfamily, the PRLR is a nonenzyme single-pass transmembrane receptor that requires associated kinases to propagate and translate the hormonal signal into transcriptional activation of target genes. Many recent reviews have covered this topic, which will not be developed here (16–19). Briefly, the main canonical signaling pathway downstream of the PRLR involves the tyrosine kinase Janus kinase 2 (JAK2) and the transcription factor STAT5 (signal transducer and activator of transcription 5), both of which are activated by tyrosine phosphorylation. In many tissues (prostate, breast, pancreas, etc), STAT5 activation, monitored using phospho-specific antibodies and/or its nuclear localization, is used as a molecular hallmark of active PRLR signaling, although one should note that in a tissue-dependent manner, other cytokine receptors can also activate this pathway (20). Other pathways classically activated by the PRLR involve the mitogen-activated protein kinase, Sarcoma (Src) and phosphoinositide 3-kinase/Akt pathways. Of note, the PRLR exists in various isoforms which exhibit different capacities to activate these pathways (21, 22).

Functional pleiotropy

Functional studies performed in the field since the discovery of PRL in the late 1920s (23, 24) have clearly established that the primary role of this hormone in mammals was to induce mammary gland differentiation during pregnancy and to ensure milk production after delivery (25). However, the phenotypic analyses of mice lacking expression of PRL (26) or its receptor (27) have illustrated the exceptional functional pleiotropy of this hormonal system (21). Indeed, genetic manipulation of PRLR signaling leads to alteration of many functions that extend far beyond mammopoiesis and lactopoiesis, including, among others, female reproduction (28), bone turnover (29), pituitary homeostasis (30), glucose tolerance (31), maternal behavior (32), hair growth (33), and adipogenesis (34, 35). In many tissues, PRLR triggering has been shown to involve locally produced PRL acting via an autocrine/paracrine loop, which adds another level of complexity to our understanding of PRL biology (36). Nowadays, it is well accepted that many functions elicited by the PRL/PRLR system participate in coordinating whole-body homeostasis, both in normal physiological states as well as in specific contexts such as adaptive responses to pregnancy/lactation (37, 38) or to stress (39, 40).

At the cellular level, PRL has been shown to regulate a wide variety of critical cell responses, including proliferation, differentiation, survival, and migration, to cite only a few. Intriguingly, PRL can in some instances regulate these responses in opposite ways depending on the context. For example, PRL can protect pancreatic β-cells from apoptosis (41, 42), whereas it can stimulate apoptosis in the pituitary (43, 44). Even more striking, PRL was shown to regulate human hair growth but its ultimate effects were shown to be opposite depending on the gender and/or the location of the hair follicle within the body (45). Clearly, understanding such a functional versatility requires fully disentangling the intracellular components that regulate PRLR signaling and/or target genes in a tissue-specific manner.

PRLR distribution

The functional pleiotropy of PRL is evidenced by the wide tissue distribution of its receptor. According to the pioneering studies performed by Freemark and coworkers 20 years ago, the PRLR is expressed in a multitude of tissues at fetal stages in rats and humans (46, 47). Rare examples of cell types that do not express this receptor may include late hypertrophic cartilage or calcified bone cells (21). However, it is fair to stress that the identification of tissues, and even more of cell types, that express the PRLR protein has been hampered by the poor quality of anti-PRLR antibodies. There is currently no acknowledged antibody for reliable detection of the mouse PRLR, which is a persistent limitation to address the role of this hormone in the dozens of genetically modified mouse models where it may be relevant. Antibodies directed against (and used by P. Kelly's group for cloning) the rat PRLR (48) can be used in this species but face specificity issues when used in other species, including mouse and human tissues (see below). Importantly, a recent systematic analysis of several commercial anti-PRLR antibodies disclosed that only one of them (clone 1A2B1) specifically and reliably recognized the human PRLR in immunohistochemistry, immunoprecipitation, and immunoblot experiments (49). Unfortunately, based on our experience and in good agreement with the results of that study, this antibody appears to be of low affinity, probably preventing the detection of the PRLR when expressed below a certain threshold. Due to these technical limitations, precise mapping of PRLR-positive compartments or individual cell types remains a challenge whatever the approach used (immunohistochemistry, immunofluorescence, cell sorting, etc), and this particularly applies to stem cells that are usually found in very small amounts in adult tissues. Hence, alternative approaches are often used by researchers to assess PRLR expression in specific tissue/cell types, including functional readouts (eg, STAT5 activation, target gene expression or cell responses after acute PRL stimulation in vitro) or detection of PRLR mRNA in cell populations enriched by cell sorting.

Regulation of Adult Stem Cells by PRL

At different levels, both adult stem cells and PRL intrinsically contribute to maintain tissue/organism homeostasis. It is therefore possible that the latter may participate in the regulation of the former. In fact, a still limited number of studies have evaluated the action of PRL on tissue stem cells, and most of these data have become available fairly recently. This minireview is thus timely to overview the current knowledge in the field. Interestingly, the tissue-specific actions of PRL (mentioned above) might reflect distinct regulatory effects of this hormone on stem/progenitor cells of each tissue. In fact, PRL has been shown to maintain quiescence on particular stem cell types, whereas it can promote proliferation and/or differentiation on others. Table 1 is provided in support of the text below to summarize details of experimental data.

Table 1.

Effect of PRL on Stem/Progenitor Cells From Various Tissues

Tissue	Species	Model	Cell Type Involved	PRLR Expressed	PRL Effect(s)	Comment(s)	Reference(s)
Mammary gland	Mouse	In vivo	Luminal progenitors	Undetermined	Orients cells towards alveolar fate	Paracrine effect mediated by RankL secreted by mature sensor luminal cells stimulated by PRL and Pg (synergystic effect) (see Figure 1)	Lee et al (66)
							Schramek et al (70)
Prostate	Mouse	In vivo	Basal/stem cells	Undetermined	Stimulates proliferation	Effects observed under forced expression of PRL in the prostate, likely mediated by paracrine factors (see Figure 2)	Rouet et al (98)
							Sackmann-Sala et al (102)
			LSC-med (putative luminal progenitors)	Undetermined	Stimulates proliferation	Effects observed under forced expression of PRL in the prostate, likely mediated by paracrine factors (see Figure 2)	Sackmann-Sala et al (101, 102)
Brain
Forebrain	Mouse (female)	In vitro	NSCs	Short isoform in immunoblot (controversial)	Stimulates proliferation and self-renewal (neurosphere)	No effect of PRL alone, cooperation of PRL (700 ng/mL) with EGF	Shingo et al (106)
	Mouse (male and female)	In vivo	NSCs	Short isoform in immunoblot (controversial)	Stimulates proliferation and differentiation into olfactory neurons		Shingo et al (106)
							Mak and Weiss (113)
	Human	In vitro	NSCs	Long isoform (mRNA)	Stimulates proliferation	hGH or hPRL effects observed in the absence of EGF or bFGF	Pathipati et al (109)
					Modulates migration	Stimulation at high concentration (500 ng/mL), no or inhibitory effect at lower concentrations	Pathipati et al (109)
Hippocampus	Mouse	In vitro	Hippocampal precursor cells	Short PRLR (blot, data not shown)	Increases neurosphere number	PRL effects observed at 1–2 ng/mL, no effect at higher concentration; PRL null mice show reduced number of hippocampal-derived neurospheres.	Walker et al (114)
	Mouse (male)	In vivo	Hippocampal progenitor cells	Long and short (S2, S3) (qPCR, IF)	Stimulates proliferation and differentiation into olfactory neurons		Mak and Weiss (113)
Bone marrow
MSCs	Human	In vitro	MSCs	Intermediate isoform in undifferentiated MSCs (mRNA)	Stimulates proliferation	PRL effects observed at 1–50 ng/mL, no effect at higher or lower concentrations; no autocrine PRL (but PRL identified in human synovial fluid using proteomic approaches)	Ogueta et al (117)
				Long isoform upon differentiation (mRNA)	Stimulates differentiation	Effects only observed when PRL and glucocorticoid are added together; differentiation induces autocrine PRL expression (mRNA)	Ogueta et al (117)
	Rat	In vitro	MSCs	Long isoform	Induces pancreatic endocrine markers	PRL effects observed at 500 ng/mL	González et al (119)
Hematopoietic	Human/mouse	In vivo	Myeloid and erythroid progenitors	Undetermined	Increase proliferation and/or decrease apoptosis	Effects assumed to be indirect	Welniak et al (134)
	Human	In vitro	CD34+ progenitors	Undetermined	Activates Stat5 pathway	Effects assumed to be direct	Cwikel et al (136)
Hair follicles	Mouse (female)	In vivo	Hair follicle stem cells	Yes (mRNA)	Maintain quiescence		Goldstein et al (147)
	Human female	Explant cultures	Epithelial stem cell-rich hair follicle bulge region	Yes in outer root sheath keratinocytes (mRNA, IHC)	Stimulates stem cell associated keratins 15 and 19	Autocrine PRL effect demonstrated using PRLR antagonists. Exogenous PRL effects at 400 ng/mL	Ramot et al (152; and references therein)
Colon cancer	Human	In vitro	Colon cancer cell lines	Yes, increased vs normal colon cells (mRNA, blot)	Increases colosphere number and size	PRL effects observed at 500–1000 ng/mL mediated by Notch pathway	Neradugomma et al (157)

Methodologies used to assess stem/progenitor properties

With some exceptions (eg, Lgr5+ intestinal stem cells), a tissue stem cell is assumed to be quiescent in physiological contexts. It is capable of self-renewal and is multipotent, meaning that it can give rise to multiple (all) cell types within the tissue. To do this, the stem cell will undergo asymmetric division, replicating itself while giving rise to a more differentiated daughter cell (called progenitor cell). Progenitor cells can be unipotent, bipotent, and in some cases, multipotent but have limited self-renewal capacity (1). Many in vitro assays have been developed along the years to reveal these properties in miniaturized and controlled experimental systems. One of the most often used methodologies involves 3-dimensional-culture systems for spheroid formation in suspension or semisolid matrices like collagen or matrigel (50). These spheres are usually named per reference of their tissue of origin (neurospheres, mammospheres, prostaspheres, etc). The survival of cells through several passages (self-renewal) in spheroid culture is accepted as a measure of stemness, whereas their multipotency can be tested by addition of specific prodifferentiation cocktails in diverse culture conditions (1). Miniorgan (organoid) cultures are another means to demonstrate the ability of putative stem cells to generate multiple lineages in vitro. Initially developed for studies of intestinal stem/progenitor cells, this approach is now used for many other tissues (mammary gland, prostate, liver, etc) (51–53). In addition to these in vitro assays, cell grafting into immunocompromised mice (54, 55) and lineage tracing studies (56) can be used to establish the tissue regeneration and multipotent capacity of particular stem/progenitor cells. In terms of cell isolation, specific stem cell markers are continually described that permit their enrichment by cell sorting (FACS, fluorescence-activated cell sorting analysis) and facilitate their identification within a specific tissue (57, 58). One of the limitations in this matter is that different research groups may use distinct (home-defined) markers to identify stem/progenitor cell (sub)populations, making it hard to compare the results reported in these publications.

As described below, PRL has been proposed to exert both direct and indirect effects on stem and/or progenitor cells of different tissues. Naturally, the question of direct vs indirect effects of factors regulating stem cells is easier to address in vitro, where the microenvironment can be better controlled. However, culture conditions might affect the expression of important receptors/signaling molecules on stem/progenitor cells, producing biased results. In essence, the determination of PRLR expression in tissue stem cells in vivo would best establish the likelihood of a direct effect of PRL on these cells. Unfortunately, as mentioned above, antibodies displaying good affinity and specificity for the human and even more the mouse PRLR represent a recurrent limitation in this matter.

The order of the sections below was established according to the amount of data available for each tissue in the physiological context. Based on these criteria, we first discuss the mammary gland (the main PRL target tissue), then the prostate (the current focus of our lab) and the brain. Next, we discuss tissues for which available data are either less recent (bone marrow), more limited (hair follicle), or only concern the pathological context (colon cancer).

Mammary gland

Together with estrogens and progesterone, PRL is one of the main regulators of mammary gland differentiation during pregnancy (25, 59). Although PRLR knockout mice displayed normal ductal outgrowth during puberty, no functional alveolar compartment was formed during pregnancy in knockout females, even when supplied with progesterone to partially overcome their nonfunctional corpora lutea (27, 28, 60). In fact, PRL and progesterone cooperate to allow ductal side branching (28). Mammary epithelium transplantation experiments demonstrated that the mammopoiesis/lactopoiesis defects observed in PRLR knockout mice were autonomous to the epithelial compartment and not due to the altered hormonal environment (60). Genetic manipulations of various PRLR signaling components in mice have evidenced the essential role of the PRLR/STAT5A pathway in the proliferation and differentiation of the mammary epithelium (25). In fact, PRL and progesterone cooperate to induce the commitment of luminal progenitors into alveolar (secretory) cells to prepare the mammary gland for lactation (for review, see Ref. 61) (Figure 1). Mechanistic evidence for this notion features a role for downstream paracrine signals, including receptor activator of nuclear factor κ-B (Rank) ligand (RankL) and Wnt4 (62–65). After estradiol and progesterone treatment, RankL and Wnt4 expression was shown to markedly increase in the sorted luminal population, whereas their receptors (Rank and Low-density lipoprotein receptor-related protein 5 (LRP5), one of the Wnt4 receptors) concomitantly increased in the basal population enriched in mammary stem cells; similar observations were made in pregnant mammary glands (65). Subsequent studies showed that RankL, secreted by mature luminal sensor cells in response to PRL and progesterone stimulation, induced expression of E74-like factor 5 (Elf5) in luminal progenitors (Figure 1) (66). Elf5, a member of the Erythoblastosis virus E26 transformation specific (ETS) transcription factor family, is a master regulator of cell fate decisions during embryogenesis (for review, see Ref. 67). Accordingly, Elf5 is expressed in all luminal progenitors and promotes their differentiation into alveolar cells during pregnancy (68). Of note, Elf5-positive cells were shown to be predominantly estrogen receptor negative (69). Because Rank is also present on the membrane of mammary stem cells, this and other signals (eg, Wnt4) may favor proliferation of stem and progenitor cells, leading to an increased cellular flux towards the secretory lineage (for review, see Ref. 67). The PRLR is required for the establishment of this functional loop (Figure 1) (70). Interestingly, enforced expression of Elf5 is capable of restoring lobuloalveolar development and milk production in PRLR knockout mammary epithelium, demonstrating the essential role of this transcription factor as a mediator of PRL physiological actions in the mammary gland (71). Such an effect of Elf5 is consistent with its capacity to increase transcription of STAT5 (72). In adults, Elf5 is expressed in luminal progenitors and in mature luminal cells of the mammary gland but not in the stem cell enriched fraction (73). Genetic Elf5 deficiency during pregnancy led to an accumulation of luminal progenitors (and enforced Elf5 expression had the opposite effect), whereas no variations were observed in virgin females (69). Accordingly, self-renewal capacities of the progenitor population, as estimated by their capacity to generate mammospheres in vitro, were increased when Elf5 was absent (69) and decreased when it was overexpressed (74). Interestingly, mice lacking STAT5-induced microRNA-193b displayed precocious mammary differentiation similar to that observed in mice overexpressing Elf5, suggesting that STAT5-regulated developmental stages of the mammary gland, including stem/progenitor cell fate and alveolar differentiation involve cross talks with miRNAs (75).

Figure 1. PRL promotes alveolar fate of mammary epithelial cells via paracrine mechanisms.

PRL is necessary to sustain progesterone synthesis by the corpus luteum in the ovary (bottom dotted square). Together, progesterone and PRL synergistically regulate mammary epithelial hierarchy during pregnancy (upper dotted square). This involves the production of RankL by mature luminal sensor cells, which in turn induces Elf5 expression in luminal progenitors and forces them towards the secretory lineage (large red arrow). In mature luminal cells, STAT5 phosphorylation (P in yellow circle) induced by PRLR activation promotes the secretory phenotype (milk production). Although genetic studies showed that STAT5 is required for the maintenance of progenitor cell populations (159), a direct role of PRL on these cells remains elusive. Symbols used for ligands and receptors are explained at the bottom of the figure.

Definitive data concerning the expression of PRLR in mammary stem or progenitor cells are lacking. Regarding stem cells, data from 2 reports suggest that they do not express the PRLR. One of these studies used human primary mammary epithelial cells (obtained from lactating breast and cultured in matrigel) to show that PRL stimulation had no effect on cluster of differentiation (CD) 49f-positive cells (basal cells positive for tumor protein p63 and cytokeratin (CK)-14, thought to be enriched in stem cells), whereas CK-18⁺ (luminal) cells showed increased proliferation in response to PRL and appeared to express the PRLR by immunofluorescence (76). The other study showed by quantitative PCR that the luminal and not the basal subpopulation expressed the PRLR in adult virgin mouse sorted mammary cells (77). Although the mechanistic studies described above suggest that PRL action on luminal progenitors is indirect, the PRLR expression status in this luminal subpopulation remains to be established.

Interestingly, the prodifferentiation properties of the PRL/STAT5 cascade that orient the mammary cell fate towards the secretory lineage have been proposed to underlie the protective role of this cascade in breast cancer patients. Indeed, the loss of STAT5A phosphorylation was shown to correlate with breast cancer aggressiveness and with resistance to antiestrogen therapy (78, 79). At the cellular level, the PRL/STAT5 pathway was shown to favor homotypic breast cancer cell adhesion and to prevent EMT, thereby reducing invasive properties in vitro (80, 81). Accordingly, Elf5 was shown to promote epithelial characteristics by repressing transcription of Snail2, a master regulator of EMT (82, 83). Also, a recent study showed that PRL inhibited the induction of a CK-5-positive population with tumor stem cell characteristics (84). Of note, enforced expression of Elf5 suppressed estrogen sensitivity in estrogen receptor-positive luminal breast cancer cells and promoted basal characteristics in basal breast cancer cells, supporting the involvement of this pathway in the acquisition of resistance to antiestrogen therapy (82). Whether such effects also may occur when the Elf5 pathway is activated by PRL/STAT5 is unknown. Clearly, the connections between the regulation of mammary cell hierarchy by PRL (and its downstream targets) and its role in breast cancer initiation/progression require additional efforts to be fully disentangled (85).

Prostate

The physiological role of PRL in the prostate gland is not clearly established. Beyond its stimulatory role on epithelial secretion, energy metabolism, and citrate production (86), explant organ cultures of human or rodent tissue have evidenced a proliferative and antiapoptotic role of PRL on the prostate epithelium (87–89). Otherwise, animal models of disrupted PRLR signaling (through disruption of the PRL, PRLR, or STAT5A genes) display only minor prostate histological defects (90–92). For the most part, local rather than circulating PRL levels are associated to prostate pathology, suggesting a more important autocrine/paracrine regulation of the gland by this hormone (93).

Prostate stem cells are believed to be contained within the basal cell compartment of the prostate epithelium and to display the typical flattened morphology of basal cells. These cells are proposed to give rise to all 3 prostate epithelial lineages, ie, basal, luminal, and neuroendocrine cells during development (Figure 2A) (94). In contrast to luminal cells, which are largely dependent on androgens for survival, basal/stem cells are resistant to androgen ablation and can regenerate the prostate epithelium upon androgen restoration (8). Importantly, basal/stem cells have been identified as cells-of-origin for prostate cancer in humans and mice (95, 96). In addition, because of their capacity to survive androgen deprivation, they are hypothesized to participate in the generation of castration-resistant tumors after initial prostate cancer treatment by androgen ablation (94). Of note, a subpopulation of androgen-independent luminal cells called castration-resistant NK3 homeobox 1-expressing cells (CARNs), which might represent luminal progenitors, also have been described to regenerate the regressed epithelium when testosterone was restored (97). Although less well characterized, luminal progenitors also may participate in tumor development and resistance to treatment.

Figure 2. PRL induces prostate basal/stem cell amplification via paracrine mechanisms.

A, Proposed prostate epithelial lineage hierarchy, including stem cells and bipotent progenitors of basal morphology, which can give rise to mature basal cells or luminal cells, the latter through intermediate cell and luminal progenitor (eg, CARN and LSC-med) states. B, The Pb-PRL and WT mouse prostate epithelium. Model of PRL action in the Pb-PRL mouse prostate showing that PRL induces the phosphorylation of STAT5 (P in yellow circle) in luminal cells. This possibly induces the secretion of paracrine factors that will promote the proliferation of the basal/stem cell compartment. Paracrine signals also may originate from stromal cells in response to PRL. Activation of STAT5 in this compartment remains to be clearly established, as does the expression of the PRLR (marked “?”). Symbols used for secreted molecules and receptors are explained at the bottom of the figure. Inset, FACS profiles for WT and Pb-PRL prostates illustrate the amplification of basal/stem cells (encircled in yellow) and LSC-med cells (light-orange full arrows) in the Pb-PRL epithelium. Luminal cells are encircled in dark orange, stromal cells in blue.

Using the probasin-PRL (Pb-PRL) mouse model, we have previously reported that PRL overexpression in the prostate leads to the amplification of the basal/stem cell compartment (Figure 2B) and that this can be prevented by a PRLR antagonist (98). Pb-PRL mice, which present prostate-specific expression of PRL under the control of a short probasin promoter, display normal serum testosterone levels but marked prostate hyperplasia and prostatic intraepithelial neoplasia lesions (99). The amplification of basal/stem cells in this model closely correlates to the activation of STAT5, which is the main signaling molecule activated by PRL in the prostate tissue (100, 101). This tight correlation could suggest that STAT5 is directly responsible for the amplification of the basal/stem cell compartment. However, STAT5 activation in Pb-PRL prostates seems to be restricted to luminal cells. Our previous immunofluorescence experiments for phosphorylated STAT5 and the basal marker tumor protein p63 have shown no colocalization of these 2 molecules on Pb-PRL prostate tissue slides, indicating that basal/stem cells do not show activation of the STAT5 pathway (98). These data suggested that basal/stem cells might not be direct targets of PRL stimulation and that paracrine mechanisms might mediate their amplification in response to PRL/STAT5 signaling in other cell compartments (ie, luminal or stromal cells) (Figure 2B). Our ongoing functional analyses aim to determine whether basal/stem cells grown as prostaspheres can respond to PRL stimulation, and whether they express the PRLR.

In addition to the amplification of basal/stem cells in Pb-PRL prostates, we have recently reported the amplification of a putative luminal progenitor population, as evidenced by FACS (Figure 2B) and immunohistochemistry/immunofluorescence analyses of prostate samples from Pb-PRL mice compared with wild-type (WT) controls (101). Similarly to basal/stem cells, these cells expressed stem cell antigen 1 (Sca-1) (a stem/progenitor cell marker in the prostate). However, they were positive for the luminal CK-8 and not the basal CK-5. By analogy to basal/stem cells referred to as Lin⁻Sca1⁺CD49f^high (LSC) according to their FACS staining profile, we have called this population Lin⁻Sca1⁺CD49f^med (LSC-med) (Figure 2B). In vitro differentiation assays showed that LSC-med cells could differentiate into mature (Sca-1-negative) luminal cells upon dihydrotestosterone stimulation. In addition, LSC-med cells display some stem-like characteristics in that they can generate prostaspheres in 3-dimensional culture. These spheres are larger in size but markedly fewer in number compared with spheres originating from prostate basal/stem cells (101). Finally, our preliminary data suggest that LSC-med cells have low self-renewal properties. Whether LSC-med cells and CARN cells represent the same luminal progenitor population still needs to be determined.

In all, these results indicate that enforced PRL expression induces the amplification of stem cells and putative luminal progenitor cells in the prostate epithelium. In agreement, a higher prevalence of intermediate CK-5⁺CK-8⁺ cells also was evidenced in the Pb-PRL epithelium (101), maybe suggesting an increased rate of basal-to-luminal differentiation in response to PRL. However, the specific underlying mechanisms remain to be established. The amplification of prostate stem and progenitor cells by PRL may play an important role in the development of prostate tumors and/or in recurrence after androgen-ablation therapies (102).

Brain

Neural stem cells (NSCs) have been described in 2 regions in the adult brain: the subgranular zone in the subventricular zone adjacent to the lateral ventricles and the hippocampal dentate gyrus. It is now well accepted that NSCs can regenerate astrocytes, oligodendrocytes (nonneural cells), and neurons (103). As described below, recent animal studies have established PRL as a potent mediator of neurogenesis in these 2 regions, making this hormonal system a central regulator of both maternal and paternal behavior via the control of olfactory cues. These recent findings add value to earlier behavioral studies that used pharmacological (104, 105) or genetic (32) approaches to reveal the role of PRL in the regulation of rodent maternal behavior.

Forebrain

In 2003, a pioneering paper in the field showed that forebrain olfactory neurogenesis observed in early pregnancy was mediated by PRL, providing a potential mechanism for the altered maternal behavior earlier reported for PRLR-deficient female mice (32). Using various molecular and genetic approaches, Shingo et al showed that PRL stimulated the proliferation and self-renewal of adult NSCs of the female subventricular zone and their differentiation into olfactory interneurons (106). Subsequently, the same group reported that PRL also mediated dominant male pheromone-induced neurogenesis in the olfactory bulb (107). Of note, such PRL effects were not observed in the female hippocampus (see below). These actions on olfactory neurogenesis were initially proposed to be direct, given that they were demonstrated using in vitro stimulation assays on cultured NSCs (putative in vitro equivalent of the precursors to new olfactory interneurons in vivo); interestingly, a cooperation with epidermal growth factor signaling was shown in NSC proliferation and neurosphere generation in in-vitro assays (106). In agreement with these data suggesting that NSCs express the PRLR, an immunoreactive band potentially corresponding to the short PRLR isoform was identified in subventricular zone extracts as well as cultured NSCs by Western blotting. The expression of this specific isoform suggested that PRL effects may be STAT5 independent, because this signaling cascade is not activated by short PRLR isoforms (21). However, a subsequent study by the group of D. Grattan failed to identify the short and long PRLR mRNAs as well as STAT5 activation in the subventricular zone, leading them to suggest that PRL regulation of neurogenesis in the female subventricular zone may be indirect, potentially through PRL-sensitive afferent neurons (108). Such a discrepancy probably underscores the limitations of mechanistic studies involving commercial anti-PRLR antibodies in mice.

In humans, the expression of long PRLR mRNA was demonstrated in forebrain-derived NSCs by PCR (109); no expression of PRL or of GH (another PRLR ligand in primates) (110) could be observed, suggesting the absence of an autocrine loop. Both exogenous GH and PRL were able to promote NSC proliferation independently of epidermal growth factor. Also, both stimulated NSC migration particularly at high concentrations (Table 1), and the use of receptor-specific antagonists supported that the PRLR probably more than the GH receptor was involved in this effect. Finally, both ligands induced neuronal progenitor proliferation, but only PRL had proliferative effects on glial progenitors.

Overall, although functional in vitro assays suggest that NSCs may be direct PRL targets, quantitative and qualitative expression of the PRLR in these cells needs to be further clarified. Whatever the PRLR status, there are now several experimental arguments supporting that the physiological rise of circulating PRL levels that is observed very early in pregnancy is essential to direct adaptive postpartum maternal behaviors (ie, ∼20 d later), including reduction of anxiety and stimulation of maternal care (111; for review, see Ref. 112). This timing nicely fits with the fact that newly generated granule cells become morphologically mature 2 weeks after initial cell division in the subventricular zone.

Hippocampus

In males, in sharp contrast to females, PRL reportedly induces neurogenesis in both the subventricular zone and the hippocampus (dentate gyrus) (113). A tentative explanation for such sex-specific effects involved different expression pattern of the PRLR in the brain (113), although this hypothesis awaits confirmation by other groups. Based on the co-labeling of cells for PRLR and stem cell lineages markers, the same authors suggested that the PRLR was expressed in male neuronal progenitor cells (stained with doublecortin) but not in NSCs. Despite the limitations of expression data involving anti-PRLR antibodies, the fact that intact PRLR signaling was required for increased neurogenesis in males interacting with pups was unambiguously provided by the absence of such an effect in PRLR knockout males. In another study, PRL was shown to activate neurosphere generation from mouse primary adult hippocampal precursor cells in vitro (114). Interestingly, the neurospheres grown in the presence of PRL could not be maintained long-term, indicating a limited self-renewal capacity. The authors concluded that the observed actions of PRL probably were produced on a more restricted population of hippocampal precursor cells rather than on NSCs (114), in agreement with the report by Mak and Weiss (113).

Bone marrow

Mesenchymal stem cells (MSCs)

MSCs were isolated initially from the bone marrow, but later, it was demonstrated that MSC or MSC-like cells also were present in other tissues, such as adipose tissue, umbilical cord blood, or even muscle (115). These cells are able to self-renew and differentiate into multiple cell types (chondrocytes, osteocytes, adipocytes), although to a lesser extent than embryonic stem cells and induced pluripotent stem cells. This has opened new hopes for using MSCs for regenerative medicine purposes (116).

Human bone marrow-derived pluripotent MSCs displayed increased proliferation when incubated with PRL during a 48-hour starvation period in monolayer culture (117). A direct action of PRL on these cells was proposed, because mRNA expression of the PRLR intermediate form was detected in cultured MSCs using RT-PCR (117). Upon cell aggregation-induced chondrogenic MSC differentiation, PRLR expression shifted from the intermediate to the long isoform, and specific extrapituitary PRL transcripts arose (150 nucleotides longer than the pituitary transcript), suggesting the establishment of an autocrine loop. Interestingly, cell aggregates treated with PRL combined with glucocorticoids exhibited enhanced cell-to-cell contacts. Taken together, these data suggest that MSCs treated with PRL alone adopt phenotypes resembling proliferative osteochondrocytes, whereas cells treated with PRL and glucocorticoid resembled differentiated osteoblasts (118). In addition, a more recent study showed that PRL treatment of rat bone marrow MSCs cultured in differentiation medium induced the expression of various pancreatic endocrine-specific genes, including somatostatin and long PRLR (119).

Thus, depending on culture conditions, PRL appears to induce proliferation or differentiation of MSCs. Such a dual potency is reminiscent of the effects mediated by the PRLR/STAT5 pathway in the breast (see above), although to our knowledge the intracellular mechanisms involved downstream of PRLR activation in MSCs remain to be elucidated.

Hematopoietic lineages

According to its classification as a class I hematopoietic cytokine (120, 121), there is a large body of experimental evidence supporting the role of PRL in regulating immune and hematopoietic cell growth and differentiation (for reviews, see Refs. 21, 122–124). Consistently, expression of the PRLR in lymphoid cells has been demonstrated in rodent (125) and in human (126) cells. PRL expression has been reported in T cells (127), suggesting that PRL effects on some hematopoietic cells may be mediated via autocrine/paracrine mechanisms.

Initial experiments involving rodents in which pituitary PRL secretion was abrogated surgically (hypophysectomy) or pharmacologically (treatment with the dopamine analog bromocryptine) revealed the occurrence of deregulations in the hematopoietic system that could be rescued by injection with recombinant PRL (128, 129). These effects were verified in vitro by addition of PRL to the culture medium. In this way, a role of PRL was demonstrated in the proliferation of T lympocytes upon activation by mitogenic factors or antigen presentation (130). In addition, PRL promotes the proliferation of B cells in response to cytokines and favors antibody production (131). In vivo studies using human lymphoid cells (isolated from peripheral blood and transplanted into SCID mice, which are deficient in T and B cells), showed that PRL injection favored successful engraftment and/or proliferation of these cells in the recipient mice (132, 133). Concerning hematopoiesis, PRL also seems to play a role, as evidenced by the fact that PRL injection promotes a significant increase of granulocyte/macrophage colony-forming unit and erythroid burst forming unit precursors in mice, leading to increased hematocrit values. This was further confirmed in experiments of bone marrow reconstitution after irradiation where PRL-injected mice showed an increase in granulocyte/macrophage colony-forming unit and erythroid burst forming unit progenitor populations in the bone marrow and the spleen (133). Similarly, these experiments showed that PRL-injection in irradiated mice induced an increase in B and T cell progenitors, as well as enhanced proliferation of mature B and T cells (133).

Despite these results arguing for a role of PRL on the hematopoietic system, there is currently very little information available regarding potential regulation of hematopoietic stem cells by PRL. Even though PRL and human GH were shown to increase myeloid and erythroid progenitors in vivo (for review, see Ref. 134), it is not clear whether these actions are direct or indirect, in other words whether the PRLR is expressed in stem/progenitor hematopoietic cells. Using biotinylated antibodies directed against the rat PRLR and FACS analysis, PRLR expression was detected in a minority of immature CD4⁻ CD8⁻ double-negative thymocytes (126). Although many precautions and controls were used in that study to validate signal specificity, the recent evidence that the same antibody (called U5) did not cross-react with the human PRLR in breast tissue (49) raises the need for confirmation of these findings using antibodies validated for the human receptor. Human CD34⁺ bone marrow cells have been proposed to express the PRLR, potentially supporting the fact that PRL stimulation in colony assays enhances the growth of granulocytic and erythrocytic progenitors (135). Hematopoietic CD34⁺ progenitors isolated from umbilical cord blood responded to PRL stimulation by activating the STAT5 pathway, supporting a direct effect (136). It is fair to mention that the few data summarized above have been collected 15 or more years ago. Clearly, understanding the actual role of PRLR signaling in hematopoietic stem/progenitor cell biology requires the expression of this receptor to be reevaluated using tools that have been validated since then.

Finally, it is noteworthy that PRL has been associated with various malignancies of the hematopoietic system, eg, lymphomas (137), acute myeloid leukemia (138), and malignant B lymphocytes (139). In agreement with these experimental studies, a recent epidemiologic study identified female hematopoietic cancers as the second-most frequent cancer type linked to hyperprolactinemia (140), which was not correlated to any change in PRLR expression (141). To which extent these effects reflect regulation of hematopoietic (cancer) stem/progenitor cells by PRL remains to be established.

Hair follicle

Several early reports indicated that PRL was a potent hair growth modulator and that annual changes in hair growth of several species were modulated by seasonal PRL (142–145). The delayed hair growth cycle observed in PRLR knockout mice suggested that regulation of epidermal maintenance and regeneration was one of the intrinsic physiological functions of PRL (33). In humans, hair loss has been documented in hyperprolactinemic women, although discriminating the actual role of PRLR signaling from that of other skin/hair regulator(s) (eg, estrogens) remains debated (146).

Recently, a mechanism supporting PRL actions on mouse hair growth regulation was elucidated by showing that hair follicle stem cells responded to PRL stimulation with increased quiescence, resulting in a delay in their activation and thus stalling hair growth (147). Messenger RNA expression of the PRLR has been shown by microarray and in situ hybridization in these cells; again, confirmation of PRLR expression at the protein level was prevented due to antibody limitations. STAT5 activation was also demonstrated in vivo upon local PRL injection and during pregnancy and lactation, when circulating PRL (or placental lactogen) levels were high (147). These results suggest that PRL (or lactogens in general) can stimulate hair follicle stem cells directly.

Human skin and its appendages express the receptors of, and are regulated by, a wide number of hormones and neuro-mediators. PRL is one of them (148). In addition, because many of these factors (including PRL) are secreted and regulated locally (148–150), autocrine/paracrine regulation loops are likely, leading Paus and coworkers (151) to consider human skin as a neuroendocrine organ. In organ-cultured hair follicles from human females, PRL stimulated the expression of CK-15 and CK-19, which are acknowledged stem cell-associated markers (152). In vitro assays suggested that these effects emanate from direct actions of PRL on keratinocytes isolated from the outer root sheath, which is a source of stem cells (152). The inhibitory action of PRLR antagonists (153) on CK-15 expression in the absence of added PRL suggests that endogenous, intrafollicular PRL is actually secreted in the stem cell niche and exerts a protective effect on the stem compartment by maintaining normal levels of CK-15 (152).

Colon cancer

The PRLR is expressed all along the gastrointestinal (GI) tract, although its role remains to be elucidated (154). We are not aware of any publications reporting PRL actions on (gastro)intestinal stem cells in a physiological context. Interestingly, cancers affecting the upper GI tract are among those showing the highest risk in hyperprolactinemic patients (140); such epidemiological evidence is lacking for lower GI cancers, including colon cancer (155). Of note, a systematic analysis of 60 National Cancer Institute cell lines (NCI60 panel) failed to demonstrate particularly high levels of PRLR expression in 7 colon cancer cell lines compared with other cancer cell lines (156). However, when compared with normal colon, colon cancer cells (harvested from patients or human cell lines) exhibited higher PRLR mRNA and protein levels, suggesting a role for PRLR signaling in colon cancer cell regulation (157).

Signaling studies revealed that in colon cancer cells, the PRLR activated mainly STAT3 and Erk1/2 but not STAT5 (157). Although there was no evidence for a mitogenic effect of PRL on these cells, PRL stimulation promoted spheroid generation (colosphere). Both spheroid number and size were increased by PRL stimulation in 3 different colon cancer cell lines (157). Interestingly, this effect appeared to be mediated via Notch signaling, because PRL stimulation increased expression of the Notch ligand Jagged1, release of Notch intracellular domain, and up-regulation of various Notch signaling target genes. Treatment with JAK2 or Erk1/2 inhibitors, individually and even more when combined, reduced all these effects, suggesting that both Erk1/2 and JAK2/STAT3 cooperate in mediating PRL-induced Notch pathway activation (157). Whether PRL stimulates colon stem cell activation in vivo remains yet to be determined.

Concluding Remarks

PRL is a pleiotropic and versatile hormone that has multiple target cells/tissues, in which it can induce many, and sometimes opposite, biological responses. Strikingly, such functional diversity is also observed at the level of stem/progenitor cell regulation. For example, in the hair follicle, PRL contributes to maintain the stem cell pool in quiescence, whereas in the mammary gland and in the brain, it stimulates progenitor proliferation and pushes the progeny towards differentiated (mature) states. Although opposite, these effects nicely correlate with, and provide a mechanistic rationale for, the physiological actions of PRL observed on these tissues during pregnancy, ie, arrest of hair growth and promotion of mammopoiesis/lactopoiesis and of parental behavior. In pathological contexts, the amplification of the tumor-initiating cell pool (prostate basal/stem cells) and/or the potentiation of stem-like properties (colon) need to be investigated further to understand whether and how these effects actually contribute to the PRL-induced progression of cognate tumors.

Due to the poor quality of anti-PRLR antibodies, especially when mouse studies are involved, statements on whether PRL regulation of stem/progenitor cells involves direct or indirect mechanisms should be made with much caution. Functional in vitro studies involving exogenous PRL have in some instances been used to support a direct effect, but in this case, the relevance of in vitro (primary) cell cultures regarding the biology of stem/progenitor cells in vivo, and/or the impurity of cell populations enriched by sorting procedures, represents a potential bias. In this respect, the good understanding of the mammary cell hierarchy, the availability of appropriate genetic mouse models, and the evidence that PRL effects are essential for the functional development of this tissue during pregnancy have contributed to provide a clear picture of the paracrine mechanisms by which this hormone regulates mammary cell fate. Clearly, the identification of specific cell population markers and the availability of reliable antimouse PRLR antibodies are needed to achieve such level of understanding in other tissues.

In many instances (see Table 1), the effects of PRL on stem/progenitor cell populations in vitro have been shown to require, or at least to be potentiated by, the addition of other factors (epidermal growth factor, glucocorticoids, etc). This underlines another aspect of PRL biology. Although PRL has many target tissues, PRLR knockout mice have revealed that with the exception of mammary gland development and female reproduction, the functions of PRL on all other tissues are that of a modulator rather than a master regulator (21, 158). Finally, according to the fact that PRL production has been documented in several nonpituitary tissues (36), the involvement of locally produced (as opposed to endocrine) PRL in stem/progenitor cell regulation has been suggested in some instances. Definitive demonstration of such a mechanism would require the analysis of stem/progenitor cell biology in the context of tissue-specific PRL deficiency in vivo (conditional knockout) and/or the use of specific PRLR inhibitors (antagonists, neutralizing antibodies) in vitro.

Funding Statement

This work was supported by La Ligue Contre le Cancer Grants 13/75-63 and 14/75-23, the Cancéropôle Ile de France and Institut National du Cancer Grant INCa_6672, and Fondation ARC Grant PDF20101202327 (to L.S.-S.).