| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
SPECIAL TOPICS |
* Mammalian NutriPhysioGenomics, Department of Animal Sciences and Division of Nutritional Sciences, University of Illinois, Urbana 61801; and
Department of Animal Sciences, Rutgers, The State University of New Jersey, New Brunswick 08901
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONLITERATURE CITED |
|---|
Key Words: genome • mammary gland • micro ribonucleic acid • small interfering ribonucleic acid • stem cell • systems biology
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONLITERATURE CITED |
|---|
). We have general knowledge on mammary growth regulation via the classic mammogenic and lactogenic hormones and growth factors (Akers, 2006), on aspects of mammary cell proliferation-apoptosis as they relate to lactation persistency (Capuco et al., 2003), and on ways to manipulate the nutrition of the animal to regulate not only milk biosynthesis (Bauman et al., 2006) but potentially rate of mammary growth at a young age (Meyer et al., 2006). Scientists have relied on morphological examination of the tissue, hormonal manipulation of mammary cells in vivo and in vitro, and, most recently, studies of mRNA expression on a large scale (Connor et al., 2007). However, major gaps in knowledge exist for each of these areas in domestic animal species, primarily due to the expense and difficulty in working with large animals.
Given these limitations, the vast majority of information that we have pertaining to the control of specification, proliferation, differentiation, survival, and death of mammary gland cells comes from studies in laboratory animals. Numerous strains of genetically altered mice have allowed the characterization of signaling networks that control the normal development and neoplastic conversion of mammary tissue (Hennighausen and Robinson, 2005). Concepts have emerged as to how mammary cells specifically activate signaling pathways to satisfy the diverse needs of the developing tissue during puberty, pregnancy, lactation, and involution of the gland (Hennighausen and Robinson, 2005). Recent studies on the origin of mammary stem cells and their role in postnatal development have revealed a new layer of complexity in understanding how the mammary gland develops (Lamarca and Rosen, 2008).
The availability of mouse DNA sequence information coupled with species-specific reagents (e.g., antibodies, proteins) have clearly facilitated the characterization of the behavior of molecular networks at multiple points of mammary gland development. It is reasonable to believe that ongoing efforts in the sequencing and annotation of the bovine and swine genomes will, in the near future, allow animal scientists across disciplines to emulate efforts in the rodent lactation biology field. Although experimental mouse models represent an important tool to unravel molecular mechanisms that drive lactation, a long-term goal of the livestock lactation biology scientific community will be to rely less on laboratory animals. This manuscript attempts to provide an overview of techniques, approaches, and concepts that could be valuable in future studies of lactation biology in livestock species. These are put into context through brief discussion of previous work in rodents as well as in domestic animals where available.
The Mouse as a Model for Lactation Biology—Transgenesis and the Omics Revolution
The development of genetically altered mice to study basic questions in mammary gland biology has been a fertile area of research over the last 20 yr. The ability to knock out or overexpress a gene in mice and study the resulting phenotypic change in terms of mammary gland morphology using whole-mount analysis and immunohistochemistry has yielded a wealth of information as to how hormones and growth factors regulate multiple aspects of mammary gland development and lactation. More advanced genetic approaches using crelox technology and mammary-specific promoters have allowed for gene knockout exclusively in the mammary gland at specific times during pregnancy (Brisken and Rajaram, 2006; Howlin et al., 2006). Furthermore, by monitoring changes in downstream targets in genetically altered mice, these approaches have led to the unraveling of signaling and transcriptional cascades that regulate gene expression in response to specific hormones and growth factors (Shillingford and Hennighausen, 2001; Hennighausen and Robinson, 2005).
The ability to manipulate the mouse genome coupled with the availability of genome sequence undoubtedly makes the mouse a unique research tool. Genomic resources for the mouse are increasing at an astounding pace. In the past decade, screening of mammary tissue with high-throughput gene expression technologies (i.e., microarrays) has revealed a greater number of genes involved in mammary development than previously recognized (Master et al., 2002; Rudolph et al., 2003). Work from Master et al. (2002) and Rudolph et al. (2003) provided the first attempts to discern gene expression pathways in mammary tissue across several developmental stages including nulliparous, stages of mammary proliferation (d 3 and 7 of pregnancy) and secretory differentiation during pregnancy (d 12 and 17 of pregnancy), lactation (d 1, 2, and 9), and forced involution (2, 7, or 28 d after pup removal).
In the work of Rudolph et al. (2003), through a coherent set of assumptions, a "refined" data set of >1,300 genes that varied significantly across stage of lactation was further classified into unique clusters according to the direction of change between adjacent time points such that all potential patterns of expression could be clearly discerned. Combined use of trajectory clustering and data mining for molecular functions determined that ~50% of the 1,300 genes were turned off (down-regulated) or had no change in expression during the period of proliferation and secretory differentiation. Among the downregulated transcripts were several putative stromal genes, genes associated with fatty acid oxidation, and genes coding for components of the proteasome, which degrades unneeded or damaged proteins. From a strict developmental standpoint, this analysis allowed the authors to suggest that 1) there is potentially a loss of stromal tissue during milk synthesis, not simply a dilution of expression due to expansion of the epithelial cell compartment; 2) mammary tissue shuts down the use of fatty acids as energy sources during milk fat synthesis; and 3) mammary epithelial cells (MEC) downregulate synthesis of proteasomal proteins in favor of other biosynthetic processes such as milk protein synthesis or metabolic enzymes (Rudolph et al., 2003).
Additional analysis indicated that mouse mammary tissue exhibited coordinate upregulation of milk protein genes (Csn1s1, Csn3, Lalba) from pregnancy through d 1 of lactation, when mRNA expression reached a peak in expression and remained flat through the milk synthesis stage. Genes associated with fatty acid and triacylglycerol synthesis, however, remained flat through pregnancy and were sharply upregulated by d 1 of lactation. This response indicated that mammary lipid synthesis is upregulated coincident with secretory activation. Further, it was proposed that the transcription factor sterol responsive element binding factor 1 (SREBF1) has a central role in the regulation of the lipogenic program (Rudolph et al., 2003, 2007); SREBF1 has been well-established as a regulator of lipogenesis in rodent liver (Horton et al., 2002). The proposed model of metabolic gene activation comprised not only SREBF1 but a network of hormones such as progesterone and prolactin, as well as their potential downstream targets [e.g., prolactin receptor, transforming growth factor-β2, and IGFBP5 (Rudolph et al., 2003)]. Clearly, this integrative approach involving gene expression, clustering analysis, functional classification, and literature data mining provides a suitable framework to allow the generation of hypotheses that can be tested experimentally.
Systems Biology Approach to Lactation
Systems biology is a relatively new field that focuses on the systematic study of complex interactions in biological systems using an approach of integration instead of reduction. In the past, the scientific method has been used primarily toward reductionism (i.e., understanding the specific function of individual genes, proteins, and cells). One of the goals of systems biology is to discover new emergent properties that may arise from examining the interactions between all components of a system to arrive at an integrated view of how the organism functions (Butcher et al., 2004; Bruggeman and Westerhoff, 2007; Naylor et al., 2008).
From a lactation standpoint, systems biology could refer specifically to the interactions between the components of the mammary gland and how these interactions give rise to the function and behavior of the tissue. Work in model organisms during the past 15 yr has clearly demonstrated the applicability of high-throughput methods to discern regulatory and metabolic networks (Feist and Palsson, 2008). Figure 1 outlines a systems biology scheme that would be amenable to high-throughput studies of mammary gland tissue at the level of transcriptomics, proteomics, and metabolomics. Besides generating essential information at each level, subsequent integration of data would allow lactation biologists to define the potential interrelationships that ultimately drive tissue function (e.g., protein-DNA, protein-protein, and metabolite-protein interactions). Such an analysis would encompass enzymes and metabolites in a pathway as well as "growth factors, transcription factors, receptors, intracellular signaling intermediates, and extracellular molecules that must ultimately interact to determine the size of the mature udder and the functional capacity of the mammary gland" (Akers, 2006). Ultimately, these combined approaches might lead to discovery of regulatory targets that could be tested further (i.e., model-directed discovery) or help address a broader spectrum of basic and practical applications including interpretation of phenotypic data, metabolic engineering, or interpretation of lactation phenotypes.
|
). The core principle of these techniques is utilizing computer resources to solve problems on scales of magnitude far too great for human discernment. Major research efforts in this field include DNA sequence alignment, gene prediction, genome assembly, protein structure alignment, protein structure prediction, prediction of gene expression, and protein-protein interactions (Butcher et al., 2004). Bioinformatics techniques were recently applied (Lemay et al., 2007) to murine mammary transcriptomic data derived from the work of Rudolph et al. (2003). Initially, a principal component analysis was used to subdivide the data set (>4,800 genes) into components (i.e., sets of genes), each describing a proportion (%) of the variance in the data (Figure 2). Through this analysis, it became obvious that the behavior of a small subset of genes (592 of ~4,800 genes, panel A, Figure 2) accounted for ~50% of the variation in expression patterns throughout lactation. This major trend was a gradual increase in expression between late pregnancy through parturition that remains high during lactation and decreases at involution (Lemay et al., 2007). Clearly, this result supports the view that in the mouse, the transcriptional changes necessary to prepare the gland for secretion occur well in advance of parturition (Lemay et al., 2007); that is, this might represent a crucial developmental window where genes can be targeted to study function.
|
to discern the biological functions in mammary tissue represented by the top 3 principal components (Figure 2
). These biological functions are easily mined through the use of gene ontology (GO) analysis, which classifies genes into a molecular function, cellular component, or biological process based on published experimental evidence (http://www.geneontology.org; last accessed Dec. 9, 2008). One of the aims of the GO Consortium is to provide a controlled vocabulary that can be used to describe any organism. However, it is intuitive that many functions, processes, and components are not common to all life forms. Despite this limitation, GO analysis of the first principal component by Lemay et al. (2007) revealed functions that are consistent with known functions of MEC such as calcium-mediated signaling, transferase activity, and cell proliferation. Some genes that have been annotated with respect to immune cells also appear to be important for development of the murine mammary gland during late pregnancy (Lemay et al., 2007).
Annotation with respect to the biological context of lactation in livestock would be an important undertaking. A recent paper in the Journal of Animal Science provides an in-depth overview of the development of an animal trait ontology, which is essential for annotating genes-proteins to biological functions (Hughes et al., 2008). In essence, animal scientists have had to rely on data from studies of the murine mammary gland to identify genes and pathways that are relevant to milk production traits in cattle (Ron and Weller, 2007). Ongoing efforts to characterize bovine (Bionaz et al., 2007) and swine (Tramontana et al., 2008) mammary transcript profiles during the transition from late pregnancy through subsequent stages of lactation will help define stage-specific gene sets as already done in the mouse. The National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/sites/entrez?db=genome; last accessed Dec. 9, 2008) and the National Animal Genome Research Program (http://www.ani-malgenome.org; last accessed Dec. 9, 2008) provide online access to several livestock genome-related tools, including bioinformatics software, microarray resources, and ongoing gene annotation projects.
Transcriptional Networks and Gene Expression Plasticity
The process of gene expression is often the origin and effector of a response, wherein the information contained within a genome is interpreted and then ultimately used to produce the building blocks (proteins) required for a given biological response (Wittkopp, 2007). Modifications of tissue regulatory networks can partly be driven by changes in gene expression (Wittkopp, 2007). Because transcription is controlled through multiple mechanisms (e.g., posttranslational modification of factors, specific interactions with coactivators, thermodynamics of protein-protein, and protein-DNA interactions), it is obvious that any gene regulation network fits into a network of networks (or global network) that represents not only transcription factor-DNA interactions but also the factors that modulate these interactions (Wittkopp, 2007). High-throughput technologies and associated bioinformatics tools are especially well-suited for studying the complex regulation of transcriptional networks in tissues (Naylor et al., 2008).
Due to the complexity underlying biological processes such as lactation, computational prediction of gene networks will remain a challenging process. However, there are several commercially available software packages that have been widely adopted by the life sciences research community (e.g., Ingenuity Pathway Analysis, GeneGo; Ingenuity Systems Inc., Redwood City, CA; GeneGo bioinformatics software Inc., San Diego, CA). By necessity, most of these are user-friendly and do not require prior training in programming or mathematics. Experimental results from high-throughput technologies (e.g., microarrays, proteomics) are interpreted in the context of greater biological systems through the use of manually curated repositories of molecular interactions, regulatory events, gene-to-phenotype associations, and chemical knowledge. These proprietary "knowledge bases" provide the building blocks for pathway construction.
Signaling and Metabolic Gene Network Motifs in the Mammary Gland
As discussed above, genetically altered strains of mice have allowed the elucidation of an integrated network of signaling events controlling normal and neoplastic development of the murine mammary gland. Hormones and growth factors activate a myriad of receptors including steroid receptors as well as cell surface receptors belonging to both the tyrosine kinase and cytokine superfamilies, in addition to other receptor families. Engagement of these receptors then sets in motion complex intracellular signaling networks, ultimately leading to activation of transcription factors to regulate gene expression. Recent data suggest that classical steroid hormones may also reside in the plasma membrane, where they mediate rapid signaling responses (i.e., nongenomic effects) via crosstalk with tyrosine kinase receptors such as the IGF and epidermal growth factor receptors (Losel and Wehling, 2003; Song and Buttgereit, 2006; Song et al., 2006). This is an area that has not been investigated in mammary cells derived from domestic animals. Recent information has also highlighted the role of inhibitors such as suppressors of cytokine signaling, integrins, and lipid rafts as components of these networks, which essentially serve to stop-control the flow of information emanating from cell-surface receptors in mammary cells (Hennighausen and Robinson, 2005).
Although a network may be modeled to describe all possible regulatory interactions occurring under any condition (e.g., Figure 1), it is often practical to study in greater detail smaller portions of the network that can be considered autonomous. In the case of the mammary gland, such a subnetwork unit (i.e., a module) could encompass the reconstruction of metabolic networks associated with milk synthesis. Through the use of mRNA expression and relative abundance, milk composition and yield, and published data from the nonruminant literature, a recent study has attempted to reconstruct the milk fat synthesis and secretion networks in bovine mammary tissue through the lactation cycle (Bionaz and Loor, 2008). A total of 45 genes with potential roles in mammary de novo fatty acid synthesis, triacylglycerol synthesis, phospholipid synthesis, fatty acid import and trafficking, and lipid droplet secretion were selected to reconstruct the mammary lipid synthesis network in Holstein cows (Figure 3). Messenger RNA expression was evaluated in mammary biopsies obtained at –15, 1, 15, 30, 60, 120, and 240 d relative to parturition (Bionaz and Loor, 2008). Results from this exercise reinforced the view that a network of genes participates in coordinating milk fat synthesis and secretion [i.e., results challenged the proposal that a single transcription factor (SREBF1) is central for milk fat synthesis regulation]. Data suggested a pivotal role for a concerted action among nuclear receptors (peroxisome proliferator-activated receptor gamma), co activators (peroxisome proliferator-activated receptor gamma, coactivator 1
), and their downstream targets. The proposed network also provides several candidate genes for future hypothesis-driven studies of mammary lipid synthesis regulation.
|
As highlighted in the preceding sections, current approaches to analyzing gene expression data can successfully identify groups of co-expressed genes (Lemay et al., 2007). In the livestock postgenomic era, however, attention will need to focus on ways of studying the functions of genes and how they are regulated. A major first step toward comprehensively understanding the differential control of gene expression in mammary tissue at different physiological stages is mapping the functional regulatory sequences in DNA that are responsible for transcriptional regulation. These regions of DNA are largely composed of transcription-binding sites (TBFS) and also nuclear receptor-binding sites (i.e., hormone response elements; Odom et al., 2004). Identification and categorization of the entire repertoire of TBFS are among the greatest challenges in systems biology (Gerstein et al., 2007). The ENCODE (Encyclopedia of DNA Elements) project was specifically developed to begin the process of defining TBFS in the human genome (Gerstein et al., 2007).
Large-scale, genome-wide transcription factor-nuclear receptor binding analysis, which identifies physical interactions between regulators and the regulatory DNA regions they bind to, provides direct evidence of regulatory relationships (Buck and Lieb, 2004). The development of the chromatin immunoprecipiation (ChIP) assay was a defining event in the study of protein-DNA interactions in vivo (Buck and Lieb, 2004). The technique allows the identification of DNA elements bound by proteins such as transcription factors or nuclear receptors under a specific cellular context. To date, ChIP-based approaches have been used to investigate the binding of 7 different nuclear receptors [e.g., estrogen receptor- in a cancer mammary cell line (MCF7)] to genomic DNA in the human and mouse (Deblois and Giguere, 2008). The use of the ChIP-real-time quantitative PCR (qPCR)-microarray technique is now considered the gold standard to confirm the presence of a transcription factor or nuclear receptor at a putative regulatory region (Deblois and Giguere, 2008).
In addition to ChIP assays, there are open source-Web-accessible computational tools (Bailey et al., 2006) that allow researchers to find motifs in DNA or protein sequences that serve as binding sites for transcription factors or nuclear receptors. This approach could be helpful in analyzing microarray or proteomics data sets, which often uncover large numbers of seemingly co-regulated genes as has been the case in previous murine mammary transcript studies (Lemay et al., 2007). The software searches for statistically significant motifs within user-provided DNA sequences that may be present in some or all of the input sequences (Bailey et al., 2006). A recent study used this computational approach to search for putative TFBS in microarray data from bovine cardiac muscle (Zadissa et al., 2007). Because the bovine genome sequence is currently incomplete, authors used human genome sequence information as a reference set for prediction of motifs in the cattle gene expression data set. The expected evolutionary conservation in TBFS among species is expected to be high (Zadissa et al., 2007). Our complete understanding of transcription factor and nuclear receptor action will depend on the construction of regulatory networks (e.g., Figure 3). Although several different approaches are required to build complex regulatory networks and connect them with biological functions (Figure 3), the application of ChIP-based approaches in combination with systematic functional and bioinformatics analysis of the data will be crucial to achieve this goal. Future studies in this area will be instrumental to improving our understanding of the molecular mechanisms of gene regulation in the mammary gland and its direct links with physiology and disease.
Cattle and Swine Genomes—The Lactation Context
Completion of cattle (http://www.animalgenome.org/cattle/; last accessed Dec. 9, 2008) and swine (http://piggenome.org; last accessed Dec. 9, 2008) genome projects is important, first to study putative biological functions of the DNA sequences in the context of lactation and second to assess the variations in DNA sequence that are directly responsible for specific traits (e.g., milk production capacity, greater milk protein:fat) or are very closely associated with a functional change (Davis, 2005; Tellam, 2007). Availability of DNA sequence in combination with high-throughput technologies (e.g., transcriptomics, proteomics, metabolomics) and bioinformatics will allow for the identification of the entire catalog of proteins synthesized by cattle and swine mammary gland. Currently, there are swine and bovine microarray platforms (>20,000 transcripts) available to the animal science community through commercial companies (Affymetrix Inc., Santa Clara, CA), USDA-funded consortiums (e.g., http://bovineoligo.org; last accessed Dec. 9, 2008), or US institution-investigator-led collaborations (e.g., http://pigoligoarray.org; last accessed Dec. 9, 2008). Proprietary microarrays also have been developed (Loor et al., 2007; Singh et al., 2008). To date, the use of this technology by the animal science community on a routine basis has been modest. High costs, particularly with industry-developed platforms, have most likely hampered wider use. However, as discussed below, studies using genome-enabled approaches to answer basic questions in lactation biology in domestic species are beginning to increase in number.
MicroRNA: Regulators of the Regulators
It has been estimated recently that ca. 50% of genomic DNA in humans is transcribed into RNA transcripts, of which only 2% is translated into proteins with 98% representing noncoding RNA (Zhang, 2008). A recently recognized mechanism that contributes to this posttranscriptional control is RNA interference (RNAi). Ribonucleic acid interference is a normal regulatory mechanism of eukaryotic cells that uses small double-stranded RNA molecules complementary to their targets to silence a specific gene (Almeida and Allshire, 2005; Sen and Blau, 2006; Kim and Rossi, 2008). Ribonucleic acid interference relies on small double-stranded RNA molecules (18 to 500 nucleotides in length), well below the size of the majority of mRNA species; thus, they are termed small noncoding RNA. Two novel classes of short noncoding RNA have been discovered, termed microRNA (miRNA) and small interfering RNA (Zhang, 2008). Both molecular species have strong inhibitory effects on mRNA translation and represent an important, novel layer that regulates gene expression (Zhang, 2008). MicroRNA act primarily through translational repression, although message degradation may also occur (Bartel, 2004; Bagga et al., 2005). To date, miRNA have been identified through genome sequence analysis algorithms resulting in the identification of potential stem-loop structures containing miRNA in many species with available genome sequence. There are currently >5,000 miRNA that have been sequenced, and microarray platforms containing these miRNA probes for different organisms are commercially available (Zhang, 2008). As of July 2008, there were 117 bovine and 55 swine miRNA sequences identified (miRBase; http://microrna.sanger.ac.uk/sequences/; last accessed Dec. 9, 2008). MicroRNA abundance and expression differ among different tissues in many species, including cattle (Coutinho et al., 2007; Gu et al., 2007). Studies have identified several miRNA in cattle adipose and mammary gland tissue (Gu et al., 2007), most of which are homologous to known mammalian miRNA. The next 10 yr will likely see vigorous work in elucidating the differences in miRNA and their target mRNA expression in the livestock mammary gland. This information will be valuable in providing an additional level of understanding of the gene expression regulatory networks in the mammary tissue during lactation (Figure 1).
From Omics to Protein Function
The above discussion highlights recent advances in genome-enabled technologies that will likely shape basic research in mammary gland biology and lactational physiology in the livestock industries in the next 10 yr. Because qPCR and microarray data provide information exclusively on changes in gene expression, key questions for the 21st century will be how to use this information to (1) understand protein function and (2) formulate hypotheses that can be tested using both in vitro and in vivo techniques to answer basic biological questions that will ultimately increase the efficiency of livestock production. One component of dairy production that has been identified as key to maximizing lactation efficiency is the persistency of lactation. Several management practices including bovine ST administration (Bauman, 1999), increased frequency of milking (Wall and McFadden, 2007), photoperiod manipulation (Dahl, 2008), continuous milking (Hale et al., 2003), or a combination of these treatments (Annen et al., 2007, 2008) have recently been explored in attempts to increase overall lactation yield and persistency. However, delineation of the molecular and cellular biology that underlies these improvements is in its infancy.
Milk yield at any given time is a function of the number of MEC and their secretory activity over time. In addition, loss of MEC occurs during involution after cessation of milking as well as during early lactation in the dairy cow. The actual changes in cell number that occur over each phase of the lactation cycle in modern, high-producing dairy cows is extensively underinvestigated due to the difficulty and expense of conducting such in vivo experiments in large animals. Results cannot necessarily be extrapolated from rodent studies, given the extended lactation that occurs in dairy cows and the fact that most are pregnant during a good part of lactation as well as the dry period. Capuco et al. (1997) have shown that cell proliferation is significantly increased from the early to late dry period in pregnant dry dairy cows. Additional in vivo studies by this group showed that in nonpregnant, lactating cows milked twice daily, the increase in milk yield from parturition to peak lactation is due to increased secretory activity per cell rather than to cell proliferation. However, after peak lactation, declining milk yield is due to loss of MEC by apoptosis (Capuco et al., 2003). In sharp contrast to the wealth of information available in rodents and humans, little is known relative to the molecular mechanisms that regulate cellular proliferation and apoptosis in cells of mammary origin in domestic livestock over the course of the lactation cycle.
Recently, Sorensen et al. (2006) examined changes in proliferation and apoptosis across the lactation cycle in lactating, pregnant cows by collecting mammary biopsies and examining paraffin-embedded sections for Ki-67 (i.e., proliferating cells) or TUNEL staining (i.e., apoptotic cells). A significant fraction of alveolar cells stained positive for Ki-67 (8.6%) during the dry period compared with a small fraction (0.4%) during lactation. The fraction of apoptotic cells was greater immediately after dry-off and in early lactation than during other periods. In a subsequent analysis of the samples, changes in expression of a variety of genes known to play roles in cell proliferation or apoptosis, or both, were examined using qPCR (Norgaard et al., 2008). The authors used these changes in gene expression to begin to form hypotheses relative to what might be occurring at the molecular level to explain the observed changes in rates of proliferation and apoptosis across physiological state.
In a similar vein, genomic approaches similar to those used in the rodent field (described above) have recently been published. Expression profiles of approximately 23,000 gene transcripts in tissue from mammary gland biopsies collected 5 d before parturition and 10 d after calving in dairy cows have been recently published (Finucane et al., 2008). Using GO analysis, they showed that genes associated with transport activity, lipid and carbohydrate metabolism, and cell signaling factors were upregulated, whereas the main downregulated genes were associated with cell division and protein and RNA degradation. They went on to further analyze the increased expression of solute carrier family 2 (facilitated glucose transporter), member 1 (also known as GLUT1) mRNA at the protein level. Interestingly, immunohistochemistry showed that before parturition, GLUT1 protein was primarily localized in ductal epithelial and blood vessel endothelium, whereas during lactation, it was predominantly localized in the basolateral and apical membrane of mammary alveolar epithelial cells. This is an excellent example of how gene expression studies will need to be used in combination with basic biological tools such as immunohistochemistry and microscopy to ultimately study protein localization and function of genes involved in mammary gland biology in domestic animals. A similar approach was used recently (Singh et al., 2008) to examine genes involved in cell-to-extracellular matrix (ECM) communication during forced involution in mid lactation, nonpregnant dairy cows. These efforts have provided initial evaluations of genes affected by change of physiological state in lactating dairy cows.
Additional information on changes in relative rates of cell proliferation and apoptosis across physiological state has been gained from studies examining the effect of omitting the dry period (i.e., continuous milking) with or without bovine ST treatment (Annen et al., 2007; Fitzgerald et al., 2007). Recently, these responses have been correlated with changes in gene expression using qPCR (Annen et al., 2008). A limitation of these studies as well as those described above in terms of interpreting changes in gene expression comes from the fact that whole mammary tissue was used for analysis. The mammary gland is complex, composed of both supportive stroma as well as secretory parenchyma. Within the stromal and parenchymal tissue compartments, multiple cell types exist, as well as different populations of cells (e.g., proliferating, apoptotic, stem cells). Therefore, future studies on changes in gene expression as they relate to the control of MEC turnover will require separation of stromal and parenchymal compartments, or methodologies such as laser capture dissection or in situ hybridization.
This concept is highlighted by recent work using microarrays to examine the role of estrogen in the mammary gland of prepubertal heifers. Ovariectomized or intact heifers were treated with or without estradiol for 3 d, and gene expression was analyzed in both the parenchyma and fat pad of the gland using a high-density bovine oligonucleotide microarray representing greater than 45,000 potential transcripts (Li et al., 2006). As summarized recently (Connor et al., 2007), initial evaluation of the microarray data revealed that 76% of the 124 estrogen-responsive genes identified were reported to be unresponsive to estrogen in other species. Interestingly, substantial regulatory differences occurred between the parenchyma and stroma, because only 17 genes were commonly regulated in both compartments. Evaluation of changes within each compartment indicated that estrogen upregulated genes associated with MEC proliferation and ECM turnover in the parenchyma and genes involved in energy-corrected milk deposition and remodeling in the fat pad, indicating a primary role for estrogen in regulating crosstalk between the 2 compartments during prepubertal growth. Additionally, comparison of estrogen-responsive genes across the bovine, humans, and rodents revealed several differences, suggesting that comparative approaches using microarray analysis will be useful in the future to identify differences in regulatory networks across species. Further analysis of these data using Ingenuity Pathway Analysis (Table 1) indicated that 23 regulatory networks were identified as estrogen-responsive (Li and Capuco, 2008).
|
As discussed above, specific gene knockout technologies have been extensively used to identify the function of individual genes during puberty, pregnancy, involution, and lactation in rodents. At the present time, it is not possible to apply such technology to domestic animal species. Therefore, knock-down of specific genes in vitro using small interfering RNA (siRNA) offers a viable option for studying gene function in mammary gland cells from ruminants and swine. Short interfering RNA are short double-stranded RNA molecules that consist of approximately 21 to 22 bp and possess 3' overhangs of 2 nucleotides that allow for recognition by the enzymatic RNAi machinery of the cell, eventually leading to homology-dependent degradation of the target mRNA. This endogenous activity can be taken advantage of by introducing synthetic siRNA molecules directly into cells. Several different approaches can be used for this. Preformed siRNA duplexes can be directly transfected into cells. This direct approach is simple and usually results in efficient gene knock-down; however, the effect is relatively transient, lasting several days at most. A second approach involves introducing a DNA plasmid or retrovirus that codes for short hairpin RNA that are transported to the cytoplasm after nuclear synthesis via the miRNA export pathway and are processed into siRNA by Dicer, a component of the RNA-induced silencing complex (Kim and Rossi, 2008). This allows for the development of stable cell lines and the use of inducible promoters, which may be more amenable to studying longer-term events such as differentiation in MEC.
Transient knock-down of specific signaling molecules using siRNA has been used to delineate the signaling pathways used by IGF-I and epidermal growth factor to regulate IGFBP3 mRNA and protein expression as well as cell growth in the bovine MEC line MAC-T (Sivaprasad et al., 2004; Fleming et al., 2006). Recently, Leibowitz and Cohick (2008) used siRNA oligos designed against bovine IGFBP3 to show that stress-activated apoptosis is inhibited when IGFBP3 is knocked down in these cells. Protein function is regulated by many posttranslational mechanisms including phosphorylation-dephosphorylation, cellular localization, and ubiquitination. Therefore, knock-down of genes with RNAi can be used to further study the mechanisms by which these proteins affect important processes in lactation biology of domestic animals. Limiting factors for these studies include the availability of mammary cell lines of domestic animal origin. The development of such cell lines would contribute to our ability to develop mechanistic information comparable to what has been achieved in other species. However, considerable progress has been made in the development of commercial reagents that allow for transfection of primary cells, making gene knock-down using siRNA feasible in primary mammary cells isolated from domestic animals in the immediate future.
Mammary Stem Cells
The ability of the mammary gland to undergo extensive remodeling and return to a fully functional state after involution suggests that a cell population with self-renewal potential exists in the mammary gland. Several different approaches have been used to identify mammary stem cells (Molyneux et al., 2007). Early work in the field identified cells at the ultrastructural and light microscopy levels as pale staining cells, and later small light cells, that were proposed to be stem-early progenitor cells based on their small size, mitotic activity, and lack of organelles (Capuco and Ellis, 2005). The ability of potential mammary progenitor-stem cells that were identified by various means to regenerate the tissue when transplanted into the cleared fat pad has been used extensively to identify stem cell potential in rodents (Smalley and Ashworth, 2003). Recently, multiparameter cell sorting based on specific cell surface markers and limiting dilution transplant analysis has been used to identify a rare subset of adult mouse mammary cells that are able individually to regenerate an entire mammary gland in vivo within 6 wk while simultaneously going through repeated rounds of symmetrical self-renewal division (Shackleton et al., 2006; Stingl et al., 2006). Additional studies in rodents have determined that stem cells with true self-renewing potential are estrogen receptor (ER)-negative and reside in the basal epithelial cell layer (Asselin-Labat et al., 2006; Sleeman et al., 2007). However, the possibility still exists for a population of ER positive progenitor cells in the rodent gland (Asselin-Labat et al., 2006; Sleeman et al., 2007).
Given the potential role of progenitor and stem cells in mammary growth and cell turnover, stem cell biology presents tremendous opportunities for increasing the efficiency of milk production. However, research in this area is limited to 2 publications to date. In 2002, it was demonstrated that lightly staining epithelial cells are the primary proliferative population during prepubertal bovine mammogenesis (Ellis and Capuco, 2002). Recently, these label-retaining epithelial cells (LREC; i.e., putative stem cells) from calves were characterized (Capuco, 2007). The greatest percentage of LREC was found in the lower region of the gland, near the cistern, and was decreased toward the periphery of the parenchymal regions, where the ducts were invading the mammary fat pad. In this later region, the percentage of heavily labeled LREC was 0.24%, consistent with estimates of stem cell frequency in the murine gland (Smith and Boulanger, 2003). Because this population of cells was found to contain both ER-negative and ER-positive cells, Capuco (2007) suggested that it contained both ER-negative stem cells as well as ER-positive progenitors.
In contrast to work with rodent stem cells, a hurdle for the study of both bovine as well as human stem cell biology is the lack of an in vitro stem cell transplantation model for normal cells. However, researchers have recently described a "human-in-mouse" model in which the mouse fat pad is cleared of endogenous epithelium and repopulated with human stromal fibroblasts (Proia and Kuperwasser, 2006). Implantation of co-mixed human epithelial and stromal cells results in functionally normal tissue. Given the similarities of the human and bovine glands (Capuco and Ellis, 2005), this exciting innovation may be applicable to studies with bovine mammary stem cells.
Summary
The mammary transcriptomics experiments described in this review have demonstrated the potential of high-throughput technologies for identifying genes involved in regulating mammary development and function in livestock species. An important goal of the future will be to apply additional experimental tools (e.g., gene silencing, chromatin immunoprecipitation) and bioinformatics (e.g., transcription factor binding site identification, lactation-specific ontology) to studies focused not only on MEC but also stem cells. Using a systems biology approach to integrate data generated at the mRNA, protein, metabolite, and tissue level with existing knowledge of enzyme kinetics, biochemistry, and hormone action can in the next 10 to 20 yr allow us to assemble all the components needed to model the mammary gland. Such models will prove useful in determining how we can manipulate complex processes that will have significant long-term economic effect (e.g., lactation persistency, efficiency of conversion of feed to milk, modifying milk fat composition, and increasing protein:fat ratio of milk; Donovan et al., 2001; Davis, 2005; Tellam, 2007). The systems approach represents a major challenge but will undoubtedly further our understanding of basic mammary biology to achieve additional gains in efficiency. Of paramount importance to our success will be the training of future students interested in lactation biology of domestic animals in all of the areas encompassed in this review.
1 Corresponding author: jloor{at}illinois.edu
Received for publication August 4, 2008. Accepted for publication September 23, 2008.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONLITERATURE CITED |
|---|
in breast cancer. Endocr. Relat. Cancer 13(Suppl. 1):S3–S13.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
