The ATP-binding cassette transporter (ABC transporter) superfamily is one of the largest and most ancient families with representatives in all existing phyla from prokaryotes to man. ABC transporters utilize the energy of ATP hydrolysis to transport a wide variety of substrates across cellular membranes, including metabolites, lipids, sterols and drugs (Higgins, Reference Higgins1992; Dean & Allikmets, Reference Dean and Allikmets1995). The classification as ABC transporter is based on the sequence and organization of the ATP-binding domains, also known as nucleotide-binding folds (NBFs). ABC transporters play a central role in the development of multidrug resistance (MDR). This is caused by several factors, one of which is increased excretion of the drug from the cell by ABC transporters. For example, ABCB1 and ABCG2, also known as P-glycoprotein and breast cancer resistance protein (BCRP), respectively, give resistance to therapeutically relevant drugs such as topotecan, irinotecan and doxorubicin (Allen & Schinkel, Reference Allen and Schinkel2002; Di Nicolantonio et al. Reference Di Nicolantonio, Mercer, Knight, Gabriel, Whitehouse, Sharma, Fernando, Glaysher, Di Palma, Johnson, Somers, Toh, Higgins, Lamont, Gulliford, Hurren, Yiangou and Cree2005).
More recently ABC transporters were identified that are implicated in the translocation of lipids. They were found either on the basis of homology with known transporters or as causative genes in disease loci. Hence, the function of ABCA1 and ABCG1 is to export excess cellular cholesterol into the HDL pathway and reduce cholesterol accumulation in macrophages (Oram & Vaughan, Reference Oram and Vaughan2006). ABCA7 was demonstrated to mediate a similar reaction as ABCA1 to generate HDL in vitro and it may be involved in lipid metabolism in kidneys and adipose tissues (Wang et al. Reference Wang, Lan, Gerbod-Giannone, Linsel-Nitschke, Jehle, Chen, Martinez and Tall2003; Linsel-Nitschke et al. Reference Linsel-Nitschke, Jehle, Shan, Cao, Bacic, Lan, Wang and Tall2005; Kim et al. Reference Kim, Fitzgerald, Kang, Okuhira, Bell, Manning, Koehn, Lu, Moore and Freeman2005). ABCG5 and ABCG8 are highly expressed in the intestine and liver cells where they form heterodimers that limit the absorption of dietary sterols in the intestine and promote cholesterol elimination from the body through hepatobiliary secretion (Yu et al. Reference Yu, Li-Hawkins, Hammer, Berge, Horton, Cohen and Hobbs2002; Mutch et al. Reference Mutch, Anderle, Fiaux, Mansourian, Vidal, Wahli, Williamson and Roberts2004).
The expression of several ABC transporters is under tight transcriptional regulation and orphan nuclear receptors play an important role in this context. Nuclear receptors comprise a family of transcription factors that act as heterodimers, which bind to promoter elements and induce gene expression. The retinoid X receptor (RXR) is a compulsory partner in the heterodimer. The other partner ascertains the specificity for the activating ligand and for the target gene (Di et al. Reference Di, Okret, Kersten, Gustafsson, Parker, Wahli and Beato1999). Other genes relevant for regulating the expression of ABC transporters involved in lipid and carbohydrate metabolism are the liver X receptors (LXR) α and β, the peroxisome proliferator-activated receptors (PPAR) α and γ and the sterol regulatory element binding proteins (SREBP) 1 and 2.
Whereas ABC transporters play a substantial role in hereditary human diseases, only scarce information is available about their expression and function in food-producing animals. Only few ABC proteins have been identified in Bos taurus (Ambagala et al. Reference Ambagala, Feng, Barletta and Srikumaran2002; Taguchi et al. Reference Taguchi, Saeki and Komano2002; Beharry et al. Reference Beharry, Zhong and Molday2004; Cohen-Zinder et al. Reference Cohen-Zinder, Seroussi, Larkin, Loor, Everts-van der, Lee, Drackley, Band, Hernandez, Shani, Lewin, Weller and Ron2005; Farke et al. Reference Farke, Viturro, Meyer and Albrecht2006; Viturro et al. Reference Viturro, Farke, Meyer and Albrecht2006) and their functions remain mostly speculative.
The aim of the current study was to detect ABC transporters that are known to play a role in lipid transport in the bovine mammary gland. The demonstration of transporter gene levels in the lactating mammary gland may identify candidate transporters involved in lipid homeostasis in the lactating mammary gland. Furthermore, expression levels of these transporters in lactating as compared with nonlactating mammary gland tissue may identify a subset of transporters involved in lipid and cholesterol transport into milk.
Materials and Methods
Cows and their management
This study was performed according to the requirements of the Bavarian state animal welfare committee (Germany).
Four healthy dairy cows (German Braunvieh), two of them in their first and two in their second lactation, free of clinical udder health problems were used. At the beginning of the experiment, two cows were in early lactation (4–100 d) and two were lactating for more than 300 d. The animals were kept in a loose-housing barn and milked twice daily at 5.00 and 17.00 in a milking parlour. Feeding consisted of maize and grass silage, hay and concentrate according to their individual levels. Water was available ad libitum.
Biopsy procedure
Biopsies were carried out after the morning milking. The experimental quarters were clipped and cleaned. Cows were mildly sedated with an i.v. injection of 0·8 ml xylazine (2%, CP-Pharma, Burgdorf, Germany). Biopsy samples from the rear quarters were taken from a caudal direction and the biopsy site was carefully selected to avoid the cisternal region and larger subcutaneous blood vessels. The area was washed, sterilized with 70% ethanol and then anaesthetized by a s.c. injection of 3·5 ml lidocaine (2%, Chassot, Ravensburg, Germany). Thereafter a puncture incision was made through the skin with a sterile single-use scalpel and one or two biopsies of one single quarter were taken through this incision using a human Bard®Magnum™ Biopsy Instrument (BARD, Covington, USA) and a Bard®Magnum™ Core Tissue Biopsy Needle (12 g×10 cm) (BARD). A core of maximum 20 mg mammary tissue was extracted. Until the next morning a swab was fixed with an adhesive Fixomull® stretch plaster (Beiersdorf AG, Hamburg, Germany) to avoid contamination. Mammary biopsies were taken as described above for a total of nine consecutive days as well as one and two weeks after onset of the dry period. Neither during nor after the experiment did the cows need anti-inflammatory or antibiotic therapy. The wound dressing was removed 2 d after taking the last biopsy.
RNA isolation and cDNA synthesis
The tissues were homogenized with an Ultra-Turrax (IKA®-WERKE, Staufen, Germany). Total RNA of mammary gland biopsies was isolated using TriPure (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's recommendations. To quantify the amount of total RNA extracted, optical density (OD) was determined. The integrity of the RNA was verified by the OD260/OD280 absorption ratio >1·8.
Synthesis of first strand complementary DNA (cDNA) was performed with reverse transcriptase (SuperScript II, Invitrogen, Karlsruhe, Germany) and random hexamer primers according to the manufacturer's instructions.
Oligonucleotide primers and PCR
All primer pairs (Table 1) were designed according to bovine or human nucleic acid data bank sequences and covered one or two exon boundaries to avoid amplification of genomic DNA. To test specificity of the primers and to evaluate the optimal annealing temperature, all primer pairs were initially tested by gradient PCR using mammary gland cDNA. PCR reactions were performed in a PCR thermocycler (Biometra, Göttingen, Germany) and contained 50 ng of cDNA, 12·5 μl 2×PCR Master Mix (Promega, Madison, USA), 0·8 μm of forward and reverse primer (Metabion, Martinsried, Germany) and nuclease-free water in a final volume of 25 μl. PCR products were subjected to gel electrophoresis in 2% agarose gels containing 0·05 μl of GelRed™ (Biotium, Hayward, USA) per ml. The DNA fragments were extracted using the Wizard SV Gel and PCR Clean-Up System (Promega), and both strands were commercially sequenced (Agowa, Berlin, Germany).
Primer sequences and amplification product sizes of reference and target genes
Real-time PCR
Quantitative real-time PCR (qPCR) in bovine mammary gland tissue was carried out using LightCycler® DNA Master SYBR Green technology (Roche Diagnostics). PCR reactions were performed in a final volume of 10 μl, using 1 μl of the LC FastStart DNA Master SYBR Green I (Roche Diagnostics), 4 pmol of each primer, 3 mm-MgCl2, and 50 ng of cDNA. Before amplification, an initial high-temperature incubation step was performed to activate the DNA polymerase and to ensure complete denaturation of cDNA. Product-specific PCR conditions are listed in Table 2. Amplified products underwent melting curve analysis after the last cycle to specify the integrity of amplification. Data were analysed using the Second Derivate Maximum calculation described in the LightCycler® Software 3.5. All runs included a negative cDNA control consisting of PCR-grade water, and each sample was measured in duplicate.
PCR conditions for quantitative PCR measurements of reference and target genes
† Fluorescence acquisition was performed at the end of each elongation cycle
Data analysis and statistics
Quantitative real time PCR data were processed using the relative quantification ΔΔCT-method (2−ΔΔCT (Livak & Schmittgen, Reference Livak and Schmittgen2001). Expression changes are shown as relative up- or down-regulation compared with lactation (Table 3). Data were obtained as CT values, which represent the cycle number at which logarithmic plots cross a calculated threshold. For normalization of target gene (TG) expression, the CT values of the reference genes (RGs) glyceraldehyde-3-phosphate dehydrogenase (GAPDH), β-actin and ubiquitin were averaged and the mean value served as reference gene index. The following calculations were performed:
SigmaPlot software (Systat Software Inc., San Jose CA, USA) was used for statistical analysis.
Gene expression differences in nonlactating relative to lactating mammary glands. Values indicate n-fold expression (2−ΔΔCT)±sem in lactation as compared with dry period 1 (D1, 1 week after drying off) and dry period 2 (D2, 2 weeks after drying off). A value of 1 indicates no change in relative expression, values >1 and <1 indicate elevated and decreased mRNA expression levels, respectively
To analyse differences in the expression levels of the RGs, CT values were compared using a t test. For the TGs the expression changes were analysed at the ΔCT levels with a paired t test to exclude potential bias because of averaging data that had been transformed through the equation 2−ΔΔCT. A P value <0·05 was regarded as statistically significant. The linear regression was calculated using the regression wizard included in the SigmaPlot software.
Results
Mammary gland biopsies from nine consecutive days during lactation and after the first and second week after dry off, respectively, were analysed. To determine whether lactation altered transporter gene expression, individual transporter RNA expression levels were compared in lactating and nonlactating bovine mammary glands. For normalization the arithmetic mean of the CT values originating from the three housekeeping genes GAPDH, β-actin and ubiquitin, was used. The mean of the housekeeping genes showed no significant differences between lactation and dry period (P=0·1572, data not shown). However, as β-actin represents not an optimal housekeeper in the mammary gland, all expression profiles were also evaluated without β-actin, i.e., data were normalized to the mean of the two housekeepers GAPDH and ubiquitin. The final outcome, however, remained essentially unchanged (data not shown).
Table 3 summarizes the n-fold differences in target gene RNA expression levels between lactating and nonlactating mammary glands. There was a significant difference in expression between lactation and dry period for ABCA1, ABCA7, ABCG2, SREBP1 and α-lactalbumin (Fig. 1, Table 3). Within the first week after dry off (D1) an increase of ABCA1 gene expression was observed which reached statistical significance (P=0·0439) in the second week (D2). A significant decrease of SREBP1 gene expression was detected in the first (D1, P=0·0204) and second (D2, P=0·0435) week of the dry period. ABCA7 and ABCG2 expression decreased at the beginning of dry period (D1) and declined significantly (P=0·0323 and P=0·0382, respectively) in the second week (D2). Likewise, the expression of α-lactalbumin showed a declining trend within the first (D1) and a significant decrease (P=0·0113) in the second week of dry period (D2).
Changes in mRNA expression levels of ABCA1, ABCA7, ABCG2, α-lactalbumin, LXRα and SREBP1 in bovine mammary glands between lactation and dry periods. L=normalized mean (ΔCT values) for lactation; D1=normalized mean (ΔCT values) for the first week of the dry period; D2=normalized mean (ΔCT values) for the second week of the dry period. Error bars indicate sem. Means without a common letter are significantly different (P<0·05).
For LXRα and lactoferrin an increase in gene expression within the dry period was observed while for ABCG5 a decreased expression level, from the lactating to nonlactating state was demonstrated. These trends, however, were not statistically significant.
All genes tested, except ABCG1 and PPARγ for which no apparent changes in gene expression during lactation and involution were observed, showed a clear trend towards significance in the second week of dry period (Table 3).
As LXRα, PPARγ and SREBP1 are known to be implicated in the regulation of ABC transporters involved in lipid homeostasis, we performed regression analysis with the relative gene expression values of the ABC transporters and these factors. We observed a similar trend for ABCA1 and LXRα (Fig. 2A), ABCA7 and SREBP1 (Fig. 2B) and ABCG5 and SREBP1 (Fig. 2C) gene expression, but for none of the other genes tested (data not shown). The relationship between ABCA1 and LXRα, ABCA7 and SREBP1 and ABCG5 and SREBP1 showed a correlation coefficient (r) of 0·82 (Fig. 2A, insert), 0·91 (Fig. 2B, insert) and 0·95 (Fig. 2C, insert), respectively.
Relative mRNA expression levels (ΔΔCT values) of (A) ABCA1 and LXRα, (B) ABCA7 and SREBP1 and (C) ABCG5 and SREBP1 during nine consecutive days in lactation (L1–L9) and after dry off (D1, D2) with regression lines and correlation coefficients (inserts). L=lactation; D1=first week of the dry period; D2=second week of the dry period; r=correlation coefficient. Error bars indicate sem.
Discussion
The active involution process in cows begins with the cessation of regular milk removal, either by drying off or by weaning the calf, and is probably complete by 21–30 d after dry off. It is a transition phase of the mammary gland from the lactating to the nonlactating state; milk continues to accumulate for a couple of days after drying off. A significant reduction in fluid volume in the gland occurs between days 3 and 7 of involution (Hurley, Reference Hurley1989). However, it is also known that some cows involute very slowly after non-milking (C Farke, HHD Meyer, RM Bruckmaier and C Albrecht, unpublished observations). These factors may explain our observation of marked differences in gene expression levels predominantly in the second week of the dry period.
Concentrations of milk-specific components such as caseins, α-lactalbumin, β-lactoglobulin, and milk fat decline during the first 2–3 weeks of the dry period (Hurley & Rejman, Reference Hurley and Rejman1986). In agreement with these findings we observed a significant decrease in α-lactalbumin gene expression, confirming adequate biopsy sampling procedures and mRNA measurements and demonstrating that our samples are representative for analysing expression differences between the lactation and dry period (Fig. 1). However, it cannot be completely excluded that, owing to the biopsy sampling technology, varying proportions of peripheral tissues may be included and that the limited numbers of biopsies collected may not completely reflect the majority of the tissue because of local variations. These factors may indicate that very subtle variations in gene expression may not be significantly identified.
The expression of lactoferrin, another control gene in this experiment, is regulated differently from that of other milk proteins. An increase of lactoferrin gene expression was observed during the first and second week of the dry period, which is in concordance with reports showing that lactoferrin is very low in bovine milk during mature lactation and is markedly elevated during mammary involution (Goodman & Schanbacher, Reference Goodman and Schanbacher1991). This trend, however, did not reach statistical significance presumably owing to the low number of animals included in this study and a high interindividual variability between the cows. Indeed, when lactoferrin expression profiles were analysed in the context of every single cow, three of the four animals showed the expected increase in lactoferrin expression during the dry period while one animal exhibited an opposite trend (data not shown). This could indicate that also for other genes measured in this study a closer association between expression changes might be detected if the results across the genes examined were compared within each animal separately as opposed to using combined averaged data.
Having principally confirmed that our sampling procedures and RNA measurements were adequate and reliable, we focused on the expression pattern of selected ABC transporters involved in lipid, phospholipid and cholesterol transport. We hypothesized that the major structural, morphological and functional differences which occur during and after the onset of the involution period might lead to marked differences in the gene expression levels of candidate ABC transporters implicated in lipid homeostasis in the mammary gland. We therefore compared gene expression levels of ABCA1, ABCA7, ABCG1 and ABCG5 in lactating and nonlactating mammary glands.
It has been reported that milk cholesterol is partially synthesized in the mammary gland but that the major proportion is mainly derived from serum cholesterol (Long et al. Reference Long, Patton and McCarthy1980). The mechanism by which serum cholesterol is transferred into the milk is still unclear. In human physiology it is well established that ABCA1 represents a cholesterol efflux regulatory gene that plays a major role in the biosynthesis of high density lipoprotein and reverse cholesterol transport (Oram & Vaughan, Reference Oram and Vaughan2006). ABCA1 is highly expressed in tissue macrophages (Lawn et al. Reference Lawn, Wade, Couse and Wilcox2001) and it has been reported that ABCA1 transcripts are up-regulated in macrophages involved in the engulfment and clearance of apoptotic cells (Luciani & Chimini, Reference Luciani and Chimini1996). We have demonstrated that in bovine tissues ABCA1 is predominantly expressed in lung, oesophagus, uterus, spleen and muscle (Farke et al. Reference Farke, Viturro, Meyer and Albrecht2006). Our present study revealed that ABCA1 gene expression was significantly up-regulated in bovine mammary glands during the dry period. It is possible that this increase of ABCA1 expression in the nonlactating mammary gland could be associated with the reported immigration of macrophages during involution (Monks et al. Reference Monks, Geske, Lehman and Fadok2002). To test this hypothesis we measured mRNA levels of the macrophage-specific CD14 gene in our mammary gland samples. We observed high interindividual variation in CD14 expression especially in the dry period and found a trend towards an elevated expression in the first week of the dry period (D1; data not shown). This tendency, however, did not reach statistical significance and should be tested not only with more animals but should also be evaluated on protein level. Interestingly, macrophages from involuting sheep mammary glands apparently have phagocytic vacuoles containing casein micelles, lipid droplets and cellular debris (Tatarczuch et al. Reference Tatarczuch, Philip, Bischof and Lee2000). This suggests that these cells play a role in clearance of residual milk and fragmented death cells. Whether ABCA1 could be implicated in cholesterol and phospholipid transport or intracellular trafficking in the mammary gland is currently unclear. Fong et al. (Reference Fong, Norris and MacGibbon2007) recently identified apolipoprotein (apo) E and apoAI, a key acceptor of cholesterol effluxed by ABCA1 (Oram et al. Reference Oram, Lawn, Garvin and Wade2000) in bovine milk fat globule membranes. These findings indicate that potential molecular acceptors for ABCA1-meditated cholesterol efflux are present in bovine milk. However, to shed light on the physiological role of ABCA1 in the mammary gland, it is crucial to determine its cellular localization and to investigate whether ABCA1 is expressed in milk fat globules or other intracellular compartments.
Surprisingly, the expression patterns of ABCA1 and ABCA7 in the bovine mammary gland showed opposite trends (Fig. 1). While ABCA1 was up-regulated, ABCA7 expression decreased during the dry period. ABCA1 is induced by cholesterol through the LXR system (Venkateswaran et al. Reference Venkateswaran, Laffitte, Joseph, Mak, Wilpitz, Edwards and Tontonoz2000) whereas ABCA7, which is highly homologous to ABCA1, is negatively regulated by cellular cholesterol (Iwamoto et al. Reference Iwamoto, be-Dohmae, Sato and Yokoyama2006). Wang et al. (Reference Wang, Lan, Gerbod-Giannone, Linsel-Nitschke, Jehle, Chen, Martinez and Tall2003) demonstrated that, in contrast to ABCA1, ABCA7 shows moderate basal mRNA and protein levels in macrophages and no induction by LXR. These studies also show that ABCA7 has the ability to bind apolipoproteins and promote efflux of cellular phospholipids without cholesterol, suggesting a possible role of ABCA7 in cellular phospholipid metabolism in peripheral tissues. This indicates that the high homology between ABCA1 and ABCA7 may not be extrapolated to physiological functions. The physiological role of ABCA7 in the mammary gland currently remains elusive.
Similarly to ABCA7, ABCG5 showed a decreased expression in the dry period (Table 3) which, however, did not reach statistical significance. Our laboratories have previously demonstrated high mRNA expression of ABCG5 and ABCG8 in the bovine liver and digestive tract and, interestingly, in the lactating mammary gland (Viturro et al. Reference Viturro, Farke, Meyer and Albrecht2006). However, in the present set of samples ABCG5 expression was significantly lower, with CT values mostly ranging between 30 and 35. Thus the data gained in these experiments should be interpreted with caution and do not currently allow us to postulate an important role for these genes in the mammary gland.
In parallel with the above mentioned lipid transporters, we also measured ABCG2 expression in our mammary gland samples. Jonker et al. (Reference Jonker, Merino, Musters, van Herwaarden, Bolscher, Wagenaar, Mesman, Dale and Schinkel2005) demonstrated that the ABCG2 transporter is strongly induced in the mammary gland of mice, cows and women during lactation and that it is responsible for the active secretion of clinically and toxicologically important substrates into mouse milk. They observed that during involution, ABCG2 expression declined rapidly. In agreement with these data and other studies (Cohen-Zinder Reference Cohen-Zinder, Seroussi, Larkin, Loor, Everts-van der, Lee, Drackley, Band, Hernandez, Shani, Lewin, Weller and Ron2005) our investigations revealed a significant decrease of ABCG2 expression from the lactating to the non-lactating state in the bovine mammary gland (Fig. 1). It is currently unclear why and to what extent ABCG2 is functionally active in the mammary gland. Therefore it is essential to identify physiological ligands for ABCG2 and to investigate which of them may account for the high expression during lactation. In this context, van Herwaarden et al. (Reference van Herwaarden, Wagenaar, Merino, Jonker, Rosing, Beijnen and Schinkel2007) recently demonstrated that ABCG2 not only secretes drugs but also riboflavin (vitamin B2) into milk, implying that vitamin B2 might represent an endogenous ligand for ABCG2 in the mammary gland. Interestingly, a missense mutation in the ABCG2 gene was recently found to affect milk yield, milk fat and protein concentration in cattle (Cohen-Zinder, Reference Cohen-Zinder, Seroussi, Larkin, Loor, Everts-van der, Lee, Drackley, Band, Hernandez, Shani, Lewin, Weller and Ron2005) suggesting a functional role for ABCG2 in milk secretion.
Expression of several ABC transporters, especially those implicated in lipid homeostasis, is regulated by transcription factors such as nuclear receptors and SREBPs. We analysed the expression of PPARγ, LXRα and SREBP1 to investigate a potential correlation between the transporters and their regulators. We observed a similar expression pattern for ABCA1 and LXRα (Figs 1 and 2A) suggesting that LXRα is involved in the regulation of ABCA1 expression in the bovine mammary gland. Indeed LXRα was 3·84±1·45 fold increased in the second week of the dry period (Table 3). However, probably owing to the low number of animals in our experiments and the high interindividual variation, the differences in LXRα expression between lactation and dry period did not reach statistical significance (Fig. 1). Associations between the mRNA expression of transporters and genes involved in their regulation were also observed for ABCA7 and SREBP1, as well as for ABCG5 and SREBP1 (Fig. 2B and 2C). These correlations suggest that SREBP1 is involved in the regulation of ABCA7 and ABCG5 in the bovine mammary gland. These findings are in agreement with previous reports indicating that ABCA7 is regulated by SREBPs but not by LXR (Iwamoto et al. Reference Iwamoto, be-Dohmae, Sato and Yokoyama2006).
In summary, we gained insight into gene expression patterns in the bovine mammary gland during lactation and subsequent involution by analysing candidate genes that are associated with lipid homeostasis. We found significant differences in the expression patterns of a subset of ABC transporters and potential relationships between selected candidate transporters and their regulatory genes. The better understanding of these transporters and the pathways involved in mammary gland lipid transport may help to elucidate novel molecular mechanisms underlying cholesterol and lipid transfer into milk. To unravel the physiological role and underlying regulatory mechanisms of these lipid transporters in the mammary gland, additional cellular localization studies based on immunohistochemistry or in-situ hybridization are essential. Moreover, the identification of endogenous ligands and substrates in the mammary gland are of fundamental importance.
We thank Dr Claudia Werner-Misof for her collaboration. This study was supported by grants from the Vereinigung zur Foerderung der Milchwissenschaftlichen Forschung (Germany) and from the Leonhard-Lorenz-Stiftung (Germany).
