The neuropeptide galanin augments lobuloalveolar development.

Mammary lobuloalveolar development during pregnancy is controlled by ovarian sex steroids and pituitary prolactin release. In organ culture these hormones are incapable of reproducing the density and size of lobuloalveoli seen in mice, suggesting the existence of other undiscovered factors. We showed previously that galanin knockout mice fail to lactate sufficiently for pup survival following their first pregnancy. Here we demonstrate that prolactin treatment of galanin knockout mice allows pup survival but does not completely rescue lobuloalveolar development or reduced milk protein expression. When galanin was used in combination with prolactin in mammary organ culture, larger and more numerous lobules were produced than with prolactin alone. Galanin alone produced sustained activation of STAT5a and the induction of milk protein expression but did not induce lobulogenesis. Examination of the transcriptional interaction between galanin and prolactin using oligonucleotide microarrays demonstrated synergistic and antagonistic modes of interaction between these hormones. These data establish a new role for galanin as a hormone augmenting mammary development during pregnancy in concert with prolactin.

Mammary lobuloalveolar development during pregnancy is controlled by ovarian sex steroids and pituitary prolactin release. In organ culture these hormones are incapable of reproducing the density and size of lobuloalveoli seen in mice, suggesting the existence of other undiscovered factors. We showed previously that galanin knockout mice fail to lactate sufficiently for pup survival following their first pregnancy. Here we demonstrate that prolactin treatment of galanin knockout mice allows pup survival but does not completely rescue lobuloalveolar development or reduced milk protein expression. When galanin was used in combination with prolactin in mammary organ culture, larger and more numerous lobules were produced than with prolactin alone. Galanin alone produced sustained activation of STAT5a and the induction of milk protein expression but did not induce lobulogenesis. Examination of the transcriptional interaction between galanin and prolactin using oligonucleotide microarrays demonstrated synergistic and antagonistic modes of interaction between these hormones. These data establish a new role for galanin as a hormone augmenting mammary development during pregnancy in concert with prolactin.
Development of the mammary gland occurs predominantly after birth and is systemically controlled by the pituitaryovarian axis. A small number of hormones control the different facets of mammary development; for example, estrogen and growth hormone regulate ductal elongation, and progesterone is essential for ductal side branching and alveolar bud formation (1)(2)(3). Prolactin (PRL), 1 via support of ovarian steroid hormone synthesis and direct action on the mammary epithelial cells, is critical for two stages of mammary development, lobuloalveolar development during early pregnancy and lacto-genesis during late pregnancy (4,5). Other factors may remain to be discovered, as models of mammary development in vitro do not reproduce the extent of mammary development seen in vivo.
Galanin is a 29-amino acid peptide originally isolated from porcine intestine (6) that has been implicated in the control of a number of biological processes including cognition, feeding behavior, neuroendocrine responses, mitogenesis, and nociception (7). Galanin signals through a family of three G proteincoupled receptors, galanin receptors (Galr) 1-3 (8 -10). The generation of mice carrying a loss-of-function mutation of the galanin gene has enabled investigation of the functions of galanin in vivo. Galanin regulates the development of sensory and cholinergic neurons, hippocampal excitability, and modulation of the pain response (11)(12)(13)(14). Overexpression of galanin in neurons suppresses epileptic-like-induced seizures (15). Galanin is also a mitogen for the prolactin-secreting pituitary lactotroph cells. Overexpression of galanin in the lactotroph induces hyperplasia and consequent hyperprolactinemia (16), and galanin knockout (Gal Ϫ/Ϫ ) mice display reduced prolactin levels during pregnancy associated with lactational failure (17). Together these data demonstrate an essential role for galanin in the control of neuronal and neuroendocrine function.
The galanin gene is located at chromosome 11q13, and like many genes in this region it is amplified in around 13% of breast cancers (18). Galanin is expressed by a number of breast cancer cell lines, but expression does not correlate with amplification. In contrast, galanin expression correlates with estrogen and progesterone receptor expression and is regulated by estradiol and progesterone (19). In the rat, serum levels of galanin increase during pregnancy and peak at mid-pregnancy with levels 7-fold greater then those observed in nulliparous animals (20). These observations suggest that the role of galanin in mammary gland development may involve more than simple modulation of pituitary prolactin secretion. To further investigate this hypothesis we utilized galanin knockout mice, combined with mammary transplantation, whole organ culture, and transcript profiling, to examine the role of galanin in mammary development.

EXPERIMENTAL PROCEDURES
Animals-Gal Ϫ/Ϫ mice (17) used in these studies were of the 129OlaHsd genetic background. Rag1 Ϫ/Ϫ mice (21) on the inbred C57BL/6J background were purchased from Animal Resource Centre, Perth, Australia. All animals were specific pathogen-free and housed with food and water ad libitum with a 12-h day/night cycle at 22°C and 80% relative humidity.
mRNA Isolation-The 4th inguinal mammary gland was frozen in liquid nitrogen before storage at Ϫ80°C prior to use. Total RNA was extracted using TRIZOL reagent (Invitrogen) according to the manufacturer's instructions.
Recombined Mammary Gland Transplantation-Donor mammary tissue (1 mm 3 ) from Gal ϩ/ϩ or Gal Ϫ/Ϫ 12-week-old mice was inserted into the excised fat pad of Gal ϩ/ϩ or Gal Ϫ/Ϫ 3-week-old mice cleared of endogenous epithelium. This recombined mammary epithelium-stroma complex was then grafted between the abdominal cavity and skin, between the 3rd and 4th mammary glands of 3-week-old Rag1 Ϫ/Ϫ mice (22). This procedure resulted in 100% transplant survival with Ͼ95% showing ductal outgrowth. Using this method, recombinations of mammary epithelium and stroma were produced that allowed deletion of the galanin gene from stroma and/or epithelium.
Histological Analysis-Mammary whole mounts were made by spreading the gland on a glass slide and fixing in 10% formalin solution. Glands were defatted in acetone before carmine alum (0.2% carmine, 0.5% aluminum sulfate) staining overnight. The whole mount was dehydrated using a graded ethanol series followed by xylene treatment for 60 min and storage and photography in methyl salicylate. Morphometric analysis was performed by counting the number of side branches, alveolar buds, or lobuloalveoli per mammary gland (n ϭ 5) for mammary gland cultures or from four representative fields of view from whole 4th inguinal glands.
PRL Treatment of Mice-On the morning of the observation of a vaginal plug, 6 -8-week-old mice were implanted with a 0.25-l per h, 28-day mini-osmotic pump (Alzet) containing unmodified PRL prepared as described (23). Either 0.6 or 1.2 g were delivered per 24 h. On the first day post-partum maternal behavior of mothers was observed, pups were examined for the presence of milk, and glands were taken for histological analysis.
Mammary Gland Culture-Four-week-old BALB/c mice were implanted with estrogen, progesterone, and cholesterol pellets (41). Following 9 days of treatment, the whole fourth glands were removed and stretched onto siliconized lens paper and placed into Petri dishes containing 2 ml of Weymouth's 152/1 medium supplemented with penicillin (100 units/ml), streptomycin (100 g/ml), gentamycin sulfate (50 g/ ml), 20 mM HEPES, insulin (5 g/ml), hydrocortisone (100 ng/ml), and aldosterone (100 ng/ml) (basal medium; IAH) to monitor ductal side branching, with and without 100 nM rat galanin (Auspep). To assess lobuloalveolar development ovine PRL (1 g/ml; Sigma) was added to the basal medium with or without galanin. Glands were maintained in a tri-gas incubator at 50% O 2 and 5% CO 2 in air. Medium was changed after 24 h and then every second day for 6 days before morphology and histology were assessed.
Transcript Profiling-Total RNA was extracted using TRIZOL reagent (Invitrogen), purified using RNeasy Mini Kit (Qiagen), cDNA synthesis was performed using Superscript II (Invitrogen), synthesis of Biotin-labeled cRNA was performed using an BioArray HighYield RNA Transcript labeling kit (Enzo Diagnostics) and hybridized to Affymetrix MGU74v2 GeneChips overnight as per the manufacturer's instructions. Arrays were performed in duplicate using four-six glands per treatment group from two separate replicate experiments. Analysis was performed using the Affymetrix GeneChip v5 software (MAS 5), with treatment groups compared back to IAH treatment as the baseline comparison. Principal components analysis was performed using JMP (SAS Institute). Venn diagrams were formed by selecting genes called increasing or decreasing by MAS 5 with a -fold change greater then 1.7 compared with IAH. These groups were further restricted by excluding genes with a magnitude -fold change Ͼ1.2 induced by the other treatments.
Quantitative RT-PCR-Quantitative PCR was performed using LightCycler technology (Roche Applied Science). Primers were designed on the basis of mismatch to other genes for WDMN1, ␤-actin, WAP, ␤-casein, ␦-casein, Elf5, Glycam1, IGF-1, GHR, SPOT 14, and prolactin receptor. PCR reactions were performed in 10-l volume with 1 l of cDNA, 5 pmols of each primer, and FastStart DNA Master SYBR Green I enzyme mix (Roche Applied Science) as per the manufacturer's instructions. Relative quantitation of the product was performed by comparing the crossing points of different samples normalized to an internal control (␤-actin). Each cycle in the linear phase of the reaction corresponds to a 2-fold difference in transcript levels between samples. Each reaction was performed in triplicate using pooled RNA from the four-six mammary glands or the treatment groups utilized for transcript profiling.

Prolactin Supplementation of Gal Ϫ/Ϫ Mice Allows Lactation
That Is Sufficient for Pup Survival, but Lobuloalveolar Development Is Not Completely Rescued-Targeted disruption of the galanin gene resulted in failure of ductal side branching during puberty and lactational failure following the first pregnancy. We have previously ascribed this effect to the reduced levels of serum prolactin seen in Gal Ϫ/Ϫ animals, but this hypothesis has not been tested, and the defects in the mammary glands of Gal Ϫ/Ϫ have not been investigated during pregnancy (17). At day 12 of pregnancy the size and density of lobuloalveoli were decreased in Gal Ϫ/Ϫ mammary glands compared with that in galanin wild type (Gal ϩ/ϩ ) mice (Fig. 1A). This defect continued throughout pregnancy, and at the 1st day post-partum, Gal Ϫ/Ϫ mammary glands showed reduced lobuloalveolar density compared with Gal ϩ/ϩ mice (Fig. 1B). Histological examination showed that although the lobuloalveoli had formed, lactation had not commenced in Gal Ϫ/Ϫ mice (Fig. 1C). Examination of the stomach contents of the pups for milk showed that 11 of 12 knockout females were unable to lactate following their first pregnancy (Fig. 1D), despite the observation of normal maternal behavior and suckling of pups. Differentiation of the mammary epithelium was assessed by quantitative analysis of the mRNA levels of several milk protein genes. Mid (␤-casein), and late (WAP) stage markers of epithelial cell differentiation were all decreased in Gal Ϫ/Ϫ mammary glands compared with Gal ϩ/ϩ littermates (Fig. 1E). Epithelial content was assessed by quantitative measurement of keratin 18 mRNA levels and showed similar levels in Gal ϩ/ϩ and Gal Ϫ/Ϫ glands (Fig. 1E). This finding, combined with the histological findings, indicate that the lobuloalveoli had formed in Gal Ϫ/Ϫ mammary glands but that differentiation and lactogenesis had failed. Thus the reduced area of epithelium apparent in the Gal Ϫ/Ϫ whole mounts and histology at term is due to a failure of lobuloalveolar engorgement because of failed onset of milk secretion, but not to a detectable decrease in epithelial cell number. A similar defect is seen in prolactin receptor heterozygous ϩ/Ϫ mice (4), and interestingly this effect was lost following their second pregnancy, as seen in Gal Ϫ/Ϫ mice (24).
Because homozygous disruption of the galanin gene results in decreased levels of plasma PRL during pregnancy, we determined whether treatment of Gal Ϫ/Ϫ mice with PRL would rescue the defect in lobuloavleolar development and lactation. Using a mini-osmotic pump, treatment with either 0.6 or 1.2 g of PRL per 24 h throughout the duration of pregnancy restored lactation and allowed pup survival (Fig. 1D). Mammary gland whole mounts showed a partial rescue of lobuloalveolar density (Fig. 1B), but histological examination showed that many alveoli had not commenced lactation, identified by their retention of highly proteinaceous (pink staining) contents (Fig. 1C). Analysis of milk protein gene expression revealed that PRL treatment did not even partially rescue the defect in lobuloalveolar differentiation measured by the expression of the milk proteins WDMN1 and ␤-casein, despite allowing pup survival (Fig. 1F). Keratin 18 levels were unchanged by treatment with PRL. Thus, although PRL treatment of Gal Ϫ/Ϫ mice restored lactation to a level sufficient for pup survival, it failed to produce any detectable rise in milk protein gene expression. If the effects on lactation seen in Gal Ϫ/Ϫ mice were only mediated via pituitary prolactin secretion then we would have expected to see some rise in milk protein levels given the rescue of lactation. Thus it is unlikely that our failure to rescue milk protein expression is solely because of insufficient administration of prolactin, and we conclude that galanin acts to influence mammary differentiation via a mechanism additional to the regulation of prolactin secretion. We have investigated this hypothesis further.

Galanin and Galanin Receptors Are Differentially Expressed
in the Mammary Gland-The failure to rescue milk protein expression with PRL treatment of Gal Ϫ/Ϫ mice indicated that galanin may act by an additional mechanism to regulate mammary epithelial cell differentiation. Expression of mRNA for galanin and Galr 1-3 was examined by RT-PCR using mouse mammary glands collected at various stages of development (Fig. 2). The galanin transcript was expressed at all time points, from estrous in nulliparous mice through to lactation, but was not detected during involution. Expression of transcripts for the galanin receptors was tightly regulated and coordinated. Transcripts for all three receptors were most highly expressed at day 7 of pregnancy. Galr1 transcripts were only detected at this time, whereas Galr2 mRNA was also detected at lower levels throughout the later stages of pregnancy and involution, and Galr3 mRNA was also detected during estrous and diestrous in the nulliparous mice. Very low expression of Galr3 mRNA could also be detected at 5 days of involution with longer exposure (data not shown). The coordinated regulation of galanin receptors in the mammary gland and increase of galanin in serum during pregnancy suggested a possible endocrine role for galanin, whereas the expression of galanin in the mammary gland also raised the possibility of an autocrine or paracrine mechanism.
An Autocrine or Paracrine Mechanism of Galanin Action Is Not Essential for Mammary Gland Development-To determine whether galanin produced by the mammary gland was necessary for normal development, we employed mammary epithelial transplantation. Donor mammary epithelium from mature Gal ϩ/ϩ or Gal Ϫ/Ϫ mice was inserted into the excised fat pad of 3-week-old Gal ϩ/ϩ or Gal Ϫ/Ϫ mice cleared of endogenous epithelium. The recombined mammary epithelium-stroma complex was grafted between the skin and abdominal cavity of 3-week-old Rag1 Ϫ/Ϫ mice (22). This allowed deletion of the galanin gene from the stroma and/or epithelium in the context of a normal endocrine background including normal circulating prolactin and galanin levels.
Ablation of galanin from the stroma, epithelium, or from both did not recapitulate the failure of ductal side branching observed in nulliparous Gal Ϫ/Ϫ mice nor the impaired lobuloalveolar development seen on the 1st day post-partum ( Fig. 3) (data not shown). These data demonstrate that an autocrine/ paracrine role for galanin is not essential for mammary development in the context of normal levels of circulating galanin. Thus endocrine galanin is sufficient for normal mammary gland development in the absence of mammary-produced galanin. Mammary-produced galanin could, however, play a role in pathological situations where endocrine galanin levels become deficient.
Galanin Can Act Directly on the Mammary Gland to Induce Lobuloalveolar Development-Galanin may act in an endocrine manner via mammary galanin receptors to induce lobuloalveolar development. As galanin treatment in vivo would indirectly induce mammary development via endocrine regulation of PRL and progesterone, we utilized an in vitro mammary gland culture model of mammogenesis (25).
Ductal side branching similar to that seen during puberty was produced when mammary glands were cultured in IAH (Fig. 4). The addition of 100 nM galanin to the IAH-containing medium did not alter ductal or lobuloalveolar development measured by quantitative morphology and histology. When PRL was added to the culture medium, lobuloalveolar development was observed (Fig. 4), although as noted previously, not to the extent observed during pregnancy. The addition of 100 nM galanin to IAH ϩ PRL medium resulted in a 3.8-fold increase in the number of lobuloalveoli per gland (8.6 Ϯ 2.1 IAH ϩ PRL versus 33.0 IAH ϩ PRL ϩ galanin; p ϭ 0.005), causing the glands to resemble those observed during pregnancy. Additionally, the size of individual lobuloalveoli in IAH ϩ PRL ϩ galanin-treated glands was also greater than in IAH ϩ PRL-treated glands (Fig. 4). These data show that galanin can act directly on the mammary gland to augment PRL-mediated lobuloalveolar development, establishing galanin as a new endocrine factor active during this phase of development.
Galanin Action on Mammary Gland Differentiation Results in Activation of STAT5-To investigate the mechanisms behind the induction of lobuloalveolar development by galanin we examined activation of the JAK/STAT, MAP kinase and PI 3-kinase signaling pathways by PRL and galanin. As expected, in mammary glands treated with PRL we saw an increase in total STAT5a and a dramatic increase in phosphorylated STAT5 in these glands (Fig. 4). Similarly, PRL treatment resulted in the sustained activation the MAP kinase pathway. Although the level of total ERK1/2 decreased, the levels of phosphorylated ERK dramatically increased in mammary glands exposed to PRL. Examination of PI 3-kinase signaling revealed decreased mobility but no increase in total Akt in glands receiving PRL. PRL did not increase phosphorylation of Thr-308 and Ser-473 residues of Akt, those most commonly associated with Akt activation. The decrease in mobility may represent phosphorylation of other sites on the Akt molecule.
Surprisingly, galanin treatment alone resulted in activation of the JAK/STAT pathway, similar to PRL (Fig. 4), but in stark contrast to PRL, galanin did not induce sustained activation of the MAP kinase pathway or alter the mobility of Akt. When mammary glands were treated with galanin and PRL there were no dramatic changes to the effects produced by either hormone alone. The apparent slight diminution in pERK and increase in Akt and pAkt (Thr-308) in the figure were not consistent between experimental replicates.
We examined markers of mammary epithelial cell differentiation by Western blot. Again as expected, mammary glands treated with PRL synthesized the milk proteins WAP and ␣and ␤-casein. Strikingly, galanin alone produced induction of milk protein synthesis despite failing to induce lobuloalveolar development (Fig. 4).
These results show that galanin can induce epithelial cell differentiation, as measured by milk protein synthesis and the sustained activation of the JAK/STAT pathway. Whether galanin acts directly via its receptors to activate STAT5 or whether this effect is indirect remains a question for further investigation. The salient point is that galanin treatment caused sustained activation of the STAT5 pathway and cell differentiation measured by milk protein expression. In contrast, prolactin caused sustained activation the JAK/STAT and the additional activation of the MAP kinase and possibly PI 3-kinase pathways (and is known to activate these pathways directly) and induced both epithelial cell differentiation and epithelial cell proliferation. Together these hormones have a synergistic effect and allow lobuloalveolar development to proceed in vitro to a level beyond that achievable by PRL alone.
In conclusion, Gal Ϫ/Ϫ mammary glands show failed differentiation despite prolactin supplementation, and wild type glands in organ culture show differentiation in response to galanin alone. These observations establish galanin as a hormone with potent ability to produce differentiation of the mammary epithelium.
Transcriptional Profiling of Galanin and Prolactin Induced Mammary Development-To examine the nature of the transcriptional interaction between prolactin and galanin that controls mammary gene expression during lactogenesis, we measured the transcriptional response of the cultured mammary glands shown in Fig. 4 to galanin, PRL, and PRL ϩ galanin using the Affymetrix microarray suite and MGU74Av2 oligonucleotide GeneChips. A principal components analysis of this data is shown in Fig. 5, using a Venn diagram approach to define sets of genes that were regulated by the three treatments. The numbers of genes and identities of the genes within these sets are shown in Fig. 6.
We used principal components analysis to distinguish patterns of altered gene expression (Fig. 5). Here, genes with a decreased expression relative to IAH are colored red, and those FIG. 4. Galanin acts directly on the mammary gland to induce lobuloalveoli development. Whole mounts of mammary glands following whole organ culture in vitro after culture in the presence of IAH, with or without galanin and PRL as indicated. Arrows indicate lobuloalveoli. H&E, hematoxylin-and eosin-stained 5-m sections from the same glands. Western blot analysis of the expression of milk proteins, STAT5, ERK, and Akt in mammary glands following IAH, IAH ϩ galanin, and/or PRL treatment is shown. Milk protein (␣-casein, ␤-casein, and WAP) expression in explant mammary glands demonstrates that milk protein levels are increased following galanin ϩ PRL, PRL, or galanin treatment alone. Increased levels of phosphorylated STAT5 was observed in mammary glands following treatment with galanin and/or PRL. Galanin alone was not able to induce activation of the MAP kinase pathway. Phosphorylated ERK1/2 was increased in mammary glands treated with PRL or PRL ϩ galanin despite a decrease in the total levels of ERK. This demonstrates marked specific activation of MAP kinase signaling in those glands treated with PRL. Examination of the PI 3-kinase pathway revealed decreased mobility but no increase in total Akt in explants receiving PRL. This decrease in mobility was not because of phosphorylation of the two residues most commonly associated with Akt activation.

FIG. 5. Transcriptional interaction between galanin and prolactin revealed by transcript profiling.
Transcript profiling of the cultured mammary glands detailed in Fig. 4 using Affymetrix U74A2 chips is shown. Principle components analysis with genes colored according to MAS 5 calls of increasing (green) or decreasing (red) gene expression in response to treatment with galanin (G), PRL (P), or galanin ϩ PRL (PG), compared with IAH alone is shown. Groups correspond to the sets shown in Fig. 6A, and the identity of set members is shown in Fig. 6B. Galanin treatment induces transcriptional changes that are also induced by prolactin (i). A set of mainly increasing transcriptional changes was identified that requires both prolactin and galanin (ii). The transcriptional activity of Prolactin independent of galanin was robust (iii), but in contrast the activity of galanin independent of prolactin was virtually nonexistent (iv). The reason for this is demonstrated in v, where clear antagonism of the transcriptional effects of galanin by prolactin was seen. In contrast galanin antagonized only a small proportion of prolactin-induced transcriptional changes (vi). with increased expression are colored green. No arbitrary -fold change value was applied in this analysis. It is apparent from this analysis that a strikingly complex and non-symmetrical transcriptional interaction exists between galanin and prolactin during lobuloalveolar differentiation. Three major sets of genes showed a robust response.
The first major set of genes we identified comprises genes that changed expression in response to all three treatments: PRL, galanin, and PRL ϩ galanin (Fig. 5i). The genes are thus regulated independently by galanin or PRL. Use of an arbitrary -fold change value of 1.7 reveals 136 genes (40% of all regulated genes) in this set (Fig. 6A). Genes with increased expression in this set (Fig. 6B) include markers of mammary epithelial cell differentiation, such as the milk proteins (WAP, WDMN-1, and five casein family members). Others here include CIS and SOCS2, negative regulators of the JAK/STAT signaling pathway, providing functional demonstration of activation of the JAK/STAT pathway by galanin and prolactin. Genes with demonstrated roles in mammary development are also independently regulated by both galanin or PRL. These include E74-like factor 5 (Elf5), GHR, IGF-1, IGF-binding protein 5 (IGFBP-5), and helix-loop-helix protein Id2 (2,(27)(28)(29). Galanin did not induce PRL or prolactin receptor gene expression, and prolactin did not regulate galanin or its receptors, excluding this simple mechanism for these transcriptional effects.
A second interesting subset of genes (Fig. 5ii) with 14 members with a -fold change Ͼ1.7 was regulated by treatment using PRL ϩ galanin but not by treatment with either hormone FIG. 6. Identity of the genes that are regulated by prolactin and galanin. A, Venn diagram showing the total number of genes found to be increasing or decreasing at least 1.7-fold in response to treatment of mammary explants with galanin, PRL, or galanin ϩ PRL, in comparison to IAH only. These sets correspond to the principle components analysis shown in Fig. 5. B, selected genes identified by the Venn diagram approach shown in Fig. 6A. Labels and colors indicate their position in the Venn diagram. -Fold change was determined using MAS 5, and selected candidates were verified using quantitative RT-PCR. alone, identifying a synergistic effect of these two hormones that is consistent with the synergistic effect of galanin and PRL on lobuloalveolar development. Almost all genes in this set showed an increase in expression indicating an overwhelmingly positive transcriptional effect of these hormones in synergy. Genes in this group include platelet-derived growth factor receptor ␤ (PDGFR␤), interleukin 1 receptor antagonist, and steroidogenic acute regulatory protein (Fig. 6B).
A third major set (Fig. 5iii) contains 154 regulated genes that change greater than 1.7-fold and is found at the intersection of the PRL and PRL ϩ galanin treatment groups. This group of genes are regulated by PRL regardless of the presence of galanin. Genes in this group include procollagen I␣ 1 and 2, nuclear factor I/X, claudin 5, and zinc finger protein 125. In striking contrast, the reciprocal set of genes at the intersection of the galanin and PRL ϩ galanin treatment groups (Fig. 5iv) contains just one gene with a change of 1.7-fold or more, although 30 genes are regulated by galanin when PRL is not present (Fig. 5v). This asymmetry indicates first that prolactin has a much greater unique transcriptional influence than galanin, regulating 160 genes that are not regulated by galanin, compared with 31 genes that are regulated by galanin but not prolactin. Second, prolactin acts to antagonize almost all (30/31 genes) of the unique influence of galanin. Galanin does not have the same effect on prolactin-induced gene expression, as only six of 160 genes show prolactin-regulated expression that was antagonized by galanin (Fig. 5vi). Genes found in these asymmetric sets include IGFBP-6, PDGFR␣, dermatopontin, and glucose phosphate isomerase 1 (Fig. 6B). These transcript profiles show that galanin and prolactin interact in the mammary gland to control gene expression via mechanisms that are independent, common, antagonistic, and synergistic. DISCUSSION Experiments detailed in this study show that mammary epithelial cell differentiation was impaired in galanin knockout mice and could not be rescued by prolactin supplementation that was sufficient to allow pup survival. Experiments with mammary glands in culture demonstrated that galanin alone could act directly on the mammary gland to greatly enhance cell differentiation. The combination of galanin and prolactin produced lobuloalveolar development in vitro resembling that seen in whole animals, the first demonstration in vitro of development to this extent. Galanin caused sustained activation of the STAT5 pathway whereas prolactin caused sustained activation of the STAT5, MAP kinase, and possibly Akt pathways. Analysis of the transcriptional results of this differential activation of signaling pathways showed the interaction between prolactin and galanin had independent, common, antagonistic, and synergistic components.
A major finding from these experiments is that galanin exerts an endocrine effect on mammary gland development. A model of the proposed endocrine role for galanin in mammary gland development is presented in Fig. 7, where galanin is proposed to exert both indirect and direct effects. The indirect effects stem from galanin's action as a growth factor to the lactotroph, the PRL producing cells of the pituitary. Via this mechanism galanin controls the level of circulating PRL (17), which in turn acts both indirectly and directly to control ductal side branching and lobuloalveolar development, respectively (4,5,30). The direct effects of galanin on the mammary gland arise from its presence in the circulation, derived from both the pituitary and the placenta (20). Placental production of hormones represents a mechanism by which the developing fetus can "hijack" the maternal endocrine system to ensure its nourishment and survival (31). Our results suggest that this may be extended to the preparation of the mammary gland for lacta-tion via the influence of placental galanin.
A second major finding from these experiments is that galanin can produce sustained activation of the STAT5 signaling pathway. This pathway is essential for lobuloalveolar development and differentiation (32) and is directly activated by receptors for prolactin, growth hormone, and epidermal growth factor (2). STAT5 also shows decreased DNA binding activity in the developmentally deficient mammary glands of Id2 Ϫ/Ϫ mice (28). It is clear that activation of STAT5 represents the convergence point of many different pathways critical for lobuloavleolar development. Galanin, via mechanisms that remain to be fully defined, is capable of producing sustained activation of this pathway.
PRL but not galanin caused sustained activation of the MAP kinase signaling pathway and reduced total MAP kinase levels. MAP kinase has a role in the regulation of cell proliferation and coordinates the mitogenic response of many growth factorreceptor tyrosine kinase-induced signaling events, many of which have a role as regulators of proliferation in the mammary gland (33).
The sustained activation observed in these experiments may reflect the changes in cellular differentiation state elicited by these hormones in addition to direct effects, and so we cannot exclude the possibility that galanin causes transient MAP kinase activation or an indirect action of galanin on Stat5 phosphorylation, but regardless of this caveat these studies demonstrate that galanin alone, like prolactin alone, can cause sustained activation of STAT5 but cannot produce the sustained activation of MAP kinase produced by prolactin. Galanin therefore does not produce an ongoing and major proliferative stimulus but does produce a strong differentiation signal. This is consistent with the effects of galanin on milk protein expression and with the failure of differentiation seen in Gal Ϫ/Ϫ animals. This raises the possibility that galanin may act as a tumor suppressor gene in the mammary gland. Several studies have suggested a similar role for galanin in experimental models of gastric, colon, and pancreatic cancer (34 -37). Further studies are currently aimed at determining whether galanin acts as a tumor suppressor in mammary carcinoma.
GeneChip microarrays were used to examine changes in gene expression in the mammary gland following exposure to galanin, PRL, and galanin ϩ PRL. A striking pattern of gene regulation was observed. The majority of regulated genes fell into three major groups. The first group showed regulation of expression by all three treatments (PRL, galanin, PRL ϩ galanin), indicating that galanin and PRL both act to control the expression of genes in this group without interaction. From our analysis of signal transduction pathways we would expect this set to contain genes predominantly regulated via the STAT5 pathway, and this set contained the milk protein genes, markers of mammary epithelial differentiation, and known JAK/STAT target genes. This group also includes members of the growth hormone/IGF axis, GHR, IGF-1, and IGFBP-5. GHR and IGF-1 have well documented roles in the regulation of ductal growth and milk protein expression (2,38). In whole animals galanin may regulate pituitary growth hormone synthesis and release (39,40) with potential for the regulation of both systemic and local IGF-1 production. IGFBP-5 is a negative regulator of IGF-1 and controls apoptosis in the mammary gland (27).
The second group of genes showed regulation only in response to galanin and prolactin, demonstrating the synergistic regulatory action of prolactin and galanin. Presumably galanin action via its G protein-coupled receptors, combined with the prolactin receptor-stimulated pathways, is responsible for this synergy. Of particular interest is the synergistic induction of PDGFR␤. Although the role of platelet-derived growth factor in normal mammary gland development is unclear, platelet-derived growth factor is a potent mitogen for a variety of different cells including some mammary cells suggestive of a proliferative role for PDGFR␤ in the mammary gland (26).
The third main group of genes (154 genes) showed regulation of expression by PRL independently of galanin. From our analysis of the signaling pathways activated we would expect this group to be transcriptional targets of sustained activation of the MAP kinase and/or PKB signaling pathways. Genes in this group include cell adhesion molecules (procollagen I␣ 1 and 2), transcription factors (nuclear factor I/X), tight junction proteins (claudin 5), and DNA binding molecules (zinc finger protein 125).
In contrast to the large number of genes regulated by PRL regardless of whether galanin was present (154 genes) only one gene was regulated by galanin regardless of the presence of prolactin, and 30 genes were regulated solely by galanin. We conclude that PRL antagonizes the regulation of a significant number of galanin-regulated genes. For example galanin reduces the expression of PDGFR␣, whereas the addition of PRL prevents galanin from reducing PDGFR␣ expression, particularly interesting given the synergistic induction of PDGFR␤. From these studies we can conclude that the transcriptional basis for the interaction of galanin and prolactin has independent, common, antagonistic, and synergistic components.
In summary we have shown that circulating galanin can influence mammary epithelial differentiation. Galanin may be a new member of the small list of systemic hormones that control mammary gland development.