Materials and Methods

Materials

Monoclonal rat anti-epithelial basement membrane antigen (IgG_a subtype) was purchased from Harlan SeraLab (Loughborough, England). M-450 sheep anti-rat IgG Dynabeads, magnetic particle concentrator (MPC-L), and Dynal sample mixer were purchased from Dynal Biotech (www.dynal.net; Lake Success, NY). PeCap polyester filters (54-um gauze) were purchased from Sefar America, Inc. (Kansas City, MO). RPMI 1640 medium (containing 25 mM Hepes buffer and 2 mMl-glutamine), IMEM (improved minimal essential medium Zn²⁺ option Richter's modification) medium, media supplements, antibiotic/antimycotic (100×) solution, fetal bovine serum (FBS), Dulbecco's phosphate-buffered saline (no calcium or magnesium) (PBS), trypsin/EDTA (0.05%/0.02% in PBS), cell culture freezing medium, and a superscript preamplification system for first-strand cDNA synthesis kit were obtained from Invitrogen (Invitrogen, Carlsbad, CA). Collagenase type 1A, hyaluronidase type 1-S, and other chemicals were obtained from Sigma-Aldrich (St. Louis, MO). Deoxyribonuclease 1 and LightCycler DNA master SYBR Green I (1×) were supplied by Roche Diagnostics (Indianapolis, IN). Human total RNA from liver, kidney, and placenta was purchased from BD Biosciences Clontech (Palo Alto, CA). All human cell lines were obtained from American Type Culture Collection (www.atcc.org; Manassas, VA). An RNeasy mini kit for total RNA isolation was obtained from QIAGEN (Valencia, CA).

Tissue Sources

The University of Kentucky Medical Institutional Review Board approved all tissue collection and study protocols. The University of Kentucky Tissue Procurement Service provided normal mammary epithelial tissue from four individual surgical reduction mammoplasty specimens (left and right breast). All tissue samples were stored in RPMI 1640 medium containing 5% FBS and 1% antibiotic/antimycotic solution at 4°C and processed within 24 h of initial collection. Milk samples from six healthy, lactating women volunteers (lactation stage breakdown: 1, 1, 1.5, 9, 10, and 11 months) provided the source of lactating mammary epithelial tissue. Expressed milk was stored on ice and processed within 1 h of collection. Individual specimens for each tissue source were pooled before mammary epithelial cell purification procedures and transporter gene RNA expression profiling.

Mammary Epithelial Cell Isolation and Purification Procedures

Heterogeneous Single-Cell Suspensions from Reduction Mammoplasty Tissue.

To achieve a single-cell suspension from normal mammary tissue, breast organoids were first prepared from reduction mammoplasty tissue according to the method of Gomm et al. (1995). Briefly, skin and grossly fatty areas were removed, and the reduction mammoplasty tissue was minced into 0.5-cm cubed pieces with two opposing scalpel blades. The minced tissue was transferred to 50-ml polypropylene conical centrifuge vials (one third the volume of the vial) and digestion mix [RPMI 1640 medium containing 5% FBS, 1% antibiotic/antimycotic solution, hyaluronidase (1 mg/ml), and collagenase (1 mg/ml)] was added such that only a small air space remained to allow for gentle mixing. After a 24-h incubation (37°C) with gentle agitation, breast organoids (ductal and lobuloalveolar fragments obtained from the digestion of reduction mammoplasty tissue) were pelleted by centrifugation (600g at 4°C for 5 min) and aspiration of the fat layer and remaining supernatant. The pelleted organoids were washed three times with 20 ml of 5% RPMI 1640 medium containing 5% FBS and 1% antibiotic/antimycotic solution (5% RPMI 1640) (37°C). To remove blood cells, the organoids underwent three sedimentation steps (1g at 37°C) in 30 ml of 5% RPMI 1640 medium, with careful aspiration of supernatant between each sedimentation step. Following sedimentation, the breast organoids were frozen in 1-ml cell culture-freezing medium and stored at −80°C until sufficient quantities of breast organoids were collected. To achieve a single-cell suspension, thawed breast organoids (37°C) were pooled and washed three times with 30 ml of RPMI 1640 medium (4°C) containing 1% FBS and 1% antibiotic/antimycotic solution, followed by a single wash with 30 ml of PBS (4°C). Addition of trypsin/EDTA containing 0.4 mg/ml deoxyribonuclease 1 (in a 1:2 organoid/trypsin ratio) allowed for digestion (30 min at 37°C) of the breast organoids, and digestion was terminated by the addition of 30 ml of RPMI 1640 medium (4°C) containing 10% FBS and 1% antibiotic/antimycotic solution. The digestion product was pelleted (600g for 5 min at 4°C) and washed three times in 10 ml of 1% RPMI 1640 medium (4°C). The cells were resuspended in 10 ml of 1% RPMI 1640 medium (4°C) and filtered through a 54-μm PeCap polyester filter to achieve a single cell suspension. The cells were counted in a Coulter counter, and a small volume of cell suspension was saved for cytospin and immunocytostaining.

Heterogeneous Single-Cell Suspensions from Milk.

Milk samples were centrifuged in 50-ml conical polypropylene vials at 600g for 20 min. The fat layer was removed with a spatula, and the remaining skim milk layer was aspirated with a pipette. The cell pellet was washed with 30 ml of PBS (4°C), pooled, and resuspended in 20 ml of 5% RPMI 1640 medium. The cells were counted in a Coulter counter, and a small volume of cell suspension was saved for cytospin and immunocytostaining.

Isolation of a Pure Population of Mammary Epithelial Cells.

Immunomagnetic separation techniques were used to isolate a pure population of mammary epithelial cells according to the method of Gomm et al., (1995) with some modifications. Briefly, M-450 Sheep anti-rat IgG Dynabeads were coated with primary rat monoclonal antibody to epithelial membrane antigen (EMA) specific for human mammary epithelial cells before incubation with mammary tissue cell suspensions. A 12.5-μl (5 × 10⁶ beads or 18.72-mg beads) aliquot of Dynabead suspension (4 × 10⁸beads/ml) was added to six 15-ml polypropylene round-bottomed tubes. To remove the preservative, each aliquot of Dynabeads was washed three times with 7 ml of RPMI 1640 medium (4°C). The medium was removed each time after placement of the tubes on the magnetic particle concentrator (MPC) for 2 min. The Dynabeads were resuspended in 1 ml RPMI 1640 medium (4°C) and transferred to 1.5-ml microcentrifuge tubes. To each 12.5-μl aliquot of washed second antibody-coated bead suspension, 187.2 μl of rat monoclonal antibody to EMA (500 μg/ml) was added to achieve a concentration of 5 μg of antibody/mg of beads. An overnight incubation (4°C) with rotation allowed the coating of the Dynabeads with the primary antibody to EMA. Excess unbound antibody was removed following the aspiration of the supernatant after placement of the Dynabead-antibody suspension on the MPC for 2 min. The coated Dynabeads were resuspended with 1 ml of RPMI 1640 medium (4°C) and transferred to new 15-ml tubes. The primary antibody/secondary antibody bead complexes were washed 4 times with 7 ml of RPMI 1640 medium (4°C) by rotation for 30 min, and placed on the MPC to remove any further unbound antibody. The coated Dynabeads were resuspended in 100 μl of RPMI 1640 medium (4°C).

The antibody/bead complexes were used to purify mammary epithelial cells from the heterogeneous cell populations isolated from reduction mammoplasty tissue and breast milk. The cell suspensions from these tissues were divided into equal volumes and pelleted at 600gfor 5 min (4°C). The supernatant was aspirated, and the pelleted cells were resuspended in 1.5 ml of 1% RPMI 1640 medium (4°C). Aliquots of the antibody/bead complexes were added to achieve a ratio of 1:1 beads per target cell (approximately 10–30% of starting population of the heterogeneous cell suspensions according to cytospin results; Fig. 1). The antibody/bead complexes and cell suspension were incubated at 4°C for 15 min, with gentle inversion of the tubes after 4, 8, and 12 min of incubation. At the end of this period, 2 ml of 1% RPMI 1640 medium (total volume 3.5 ml) (4°C) was added to each vial, and the specifically bound cells were collected by the placement of vials into the MPC (2 min) and aspiration of the supernatant. The supernatant was collected in separate 15-ml polypropylene vials for RNA isolation and for cytospin preparation and immunocytochemistry of this cell population. The cells bound to Dynabeads were washed three times with 3.5 ml of 1% RPMI 1640 medium (4°C), and the complex was resuspended in 10 ml of 1% RPMI 1640 medium (4°C). The specifically bound cells were detached from the Dynabeads by mechanical dissociation (repeated pipetting through a 10-ml pipette). The detached cells were counted with a Coulter counter, and a small volume was saved for cytospin preparations and immunocytostaining.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Immunocytochemical staining with cytokeratin cocktail, CK22, of cells collected from fresh human breast milk and reduction mammoplasty tissue before and after purification procedures. A, section through a normal mammary gland duct; luminal mammary epithelial cells stain positive; B, cells collected from breast milk before purification; C, cells collected from breast milk depleted of mammary epithelial cells; D, mammary epithelial cells collected from breast milk after purification; E, mammary epithelial cells collected from reduction mammoplasty tissue after purification (magnification, 25×).

Immunocytostaining

Cytospins of each cell population aliquot were made. Cytocentrifuge preparations were air dried and fixed with alcohol. The cell preparations were stained using the Vectastain ABC kit (Vector Laboratories, Inc., Burlingame, CA) and a monoclonal mouse anti-human (IgG₁) cytokeratin cocktail CK22 (40–68 kDa) (Biomeda Corp. Foster City, CA) specific for simple epithelial cells.

Total RNA Extraction and Reverse Transcription.

Total RNA was extracted from purified lactating and nonlactating mammary epithelial cells using an RNeasy mini kit following the manufacturer's protocol. Pelleted cells were resuspended and counted using a Coulter counter before RNA extraction to assess cell yields following cell isolation and purification procedures and to prevent overloading of the RNeasy columns. Cells were homogenized using a roto-stator tissue homogenizer. Total RNA concentration was determined by the measurement of optical density at 260 nm with an ultraviolet spectrophotometer (Beckman spectrophotometer; Beckman Coulter, Inc., Fullerton, CA). Total RNA integrity was verified by an OD₂₆₀/OD₂₈₀ absorption ratio greater than 1.7. From each sample preparation, 1 μg of total RNA was used to synthesize first-strand cDNA by reverse transcription with Invitrogen's Superscript RT kit following the manufacturer's instructions.

Transporter Gene-Specific Primer Design

Gene sequences for primer design were obtained from the National Center for Biotechnology Information's GenBank. Primer pairs for each transporter gene, EMA, common acute lymphocytic leukemia antigen (CALLA), β-actin, β-casein, and α-lactalbumin were chosen with assistance from the primer design software Oligo 4.0 (National Biosciences, Inc., Plymouth, MN) to give amplification products between ∼250 to 800 base pairs. Forward and reverse primer sequences for each gene and their corresponding amplicon size are provided in Table1. The primer sequences for EMA (EMA forward primer: 5′-3′: TACCACCCTTGCCAGCCATAG; EMA reverse primer: 5′-3′: CGCAACCAGAACACAGACCAG) and for CALLA (CALLA forward primer: 5′-3′: TAAGCAGCCTCAGCCGAACCTACA; CALLA reverse primer: 5′-3′: GACCAAGACCTCCATTATCAGCAA) gave amplification products of 511 and 831 base pairs, respectively.

Optimization of LightCycler PCR Conditions

For reliable quantification of transporter gene expression, LightCycler PCR conditions specific for each gene were optimized with respect to primer concentration, MgCl₂concentration, annealing temperature (T_M), and cycle number using the LightCycler (Roche Molecular Biochemicals, Indianapolis, IN) and positive control tissues. Positive control tissues consisted of the human hepatocellular carcinoma cell line HepG2, the human breast carcinoma cell line ZR-75-1, the human renal adenocarcinoma cell line 786-0, the human colon adenocarcinoma cell line HT-29, and kidney tissue. Conditions were established to give specific melting peaks following a melting curve analysis. To differentiate desired from undesired amplification products, specific melting peaks were correlated with gel electrophoresis results. Amplification products were separated by 1.5% agarose gel electrophoresis using a 1× Tris borate-EDTA running buffer (0.09 M Tris-borate, 0.002 M EDTA, pH 7.8) and analyzed with a photoimager following ethidium bromide staining. A specific band corresponding to the appropriate size of the amplicon for each gene and to a specific melting peak was positive identification of the appropriate amplification product. These optimized conditions were used for transporter gene expression profiling.

LightCycler PCR Master Mix

A SYBR Green I DNA master kit was used for LightCycler PCR. Table 2 gives the concentrations of the reaction components in the master mix used for each transporter gene-specific LightCylcer reaction and the corresponding annealing temperature and total cycle number. A total volume of 18 μl of the LightCycler master mix was filled into LightCycler glass capillaries under reduced light conditions, and 2 μl of cDNA product (from 1 μg of total RNA) was added as PCR template. The capillaries were capped, briefly centrifuged, and placed in the LightCycler carousel with minimal delays.

LightCycler PCR Time and Temperature Profiles

A standard LightCycler amplification cycle protocol was established for the transporter genes. Thermal cycling was carried out as follows. The first segment of the amplification cycle consisted of a denaturation program at 95°C for 30 s. The second segment consisted of a four-step primer annealing, amplification, and quantification program repeated for a set number of cycles (see Table2) with a ramp rate set at 20°C/s. The conditions were 95°C for 0 s, annealing T_M for the specific transporter (see Table 2) for 10 s, elongation at 72°C for 40 s, and a single fluorescence acquisition point at the end of the elongation phase. A higher temperature fluorescence acquisition point was required in the amplification cycle program for several transporter genes (see Table 2) to improve the specificity of SYBR green I quantification. The third segment consisted of a melting curve program. The conditions were 95°C for 0 s and 72°C for 10 s with a linear temperature transition at 0.1°C/s from 72°C to 94°C (except for OATP-E, which required the melting curve analysis to proceed to 98°C) with continuous fluorescence acquisition. The melting curve analysis is necessary to differentiate fluorescence associated with specific amplification products from nonspecific amplification products. The last segment consisted of a final cooling program to 40°C.

Data Analysis and Calibration Curves

The LightCycler analysis software allowed quantitative analysis of the PCR reactions. A calibration curve for each transporter gene and for the housekeeping gene β-actin was generated from serial dilutions of template cDNA (based on 1 μg of total RNA) from respective control tissues. To remove background fluorescence, an arbitrary threshold fluorescence level in the exponential phase of the amplification curve was set to exceed the mean baseline fluorescence emissions (noise level) of each amplification plot. A crossing point defined the cycle number at which the best-fit line through the log-linear portion of each amplification plot intersected the threshold level. The calibration curve was constructed from the plot of crossing points against the log value of serially diluted standard samples. Only crossing points associated with a specific melting peak were used in the construction of the calibration curve and in the assessment of unknown samples. Interpolation of the standard curve using the crossing points calculated for each transporter gene and β-actin in the unknown samples allowed quantification of the respective genes. A nontemplate negative control was incorporated in all PCR reactions.

Results

Purity of Mammary Epithelial Cell Population.

Before purification procedures, immunocytostaining of cytospins from cells isolated from breast milk samples and from reduction mammoplasty tissue demonstrated only 10 to 20 and 10 to 30% staining, respectively, of cells by cytokeratin antibody cocktail specific for mammary epithelial cells (Fig. 1). Consequently, purification procedures for the respective cell populations were necessary before transporter expression profiling experiments. Following purification, cell populations isolated from breast milk and from reduction mammoplasty gave >90% purity for mammary epithelial cells in all preparations (Fig. 1). Cell purity was confirmed by normal RT-PCR assessment of EMA (an epithelial cell surface membrane antigen) and CALLA (a myoepithelial and fibroblast cell surface membrane antigen) expression as specific markers of each cell population. Moreover, β-casein and α-lactalbumin expression, specific molecular markers of mammary gland differentiation (Levine and Stockdale, 1985; Hahm et al., 1990), in heterogeneous and purified populations were assessed. Electrophoresis gels indicated strong positive bands for EMA, β-casein, and α-lactalbumin in the purified populations and with a very slight band for CALLA (data not shown). Strong positive bands for EMA, CALLA, β-casein, and α-lactalbumin were demonstrated in the cell populations isolated from milk and reduction mammoplasty before purification procedures (data not shown). These results suggest the purity of the preparations was sufficient to allow a comparison of transporter expression profiles between the respective tissues.

Real-Time RT-PCR Assay Validation.

Table 2 provides the specific set of PCR conditions for each transporter gene. Verification of appropriate RT-PCR products for each transporter gene was determined by PCR product separation on 1.5% agarose gel electrophoresis. Amplified PCR products showed a single band and the expected amplicon size (Table 2) for each gene assessed. Specificity of the RT-PCR product was also confirmed with melting curve analysis using LightCycler software. A specific melting curve and annealing temperature for each transporter gene analyzed was associated with a specific band of appropriate size on gel electrophoresis. The calibration curves gave linear quantification ranges over 2 to 4 orders of magnitude, depending upon the transporter gene and level of expression in the control tissue, as indicated by a correlation coefficient of >0.99. To ensure high PCR efficiency, every attempt was made to achieve calibration curve slopes close to the theoretical optimum (−3.322) and y-intercepts near the threshold cycle value for the negative control. Replicates of each transporter gene in the various tissues were not performed due to template limitations. Several transporter genes required fluorescence acquisition at higher temperatures (see Table 2) than the elongation cycle temperature during the amplification program to eliminate nonspecific fluorescence signals due to primer dimers and to ensure a more accurate quantification of the desired product. The relative expression level for each gene in each tissue sample was interpolated from the calibration curve. The expression level for the housekeeping gene β-actin was similarly calculated in each tissue, and all expression data were normalized to β-actin to control for systematic variation in sample preparation and analysis. Hence, the final results were expressed as the ratio of relative transporter expression levels/β-actin expression levels.

RNA Expression Analysis of Transporter Genes in Mammary Epithelial Cells.

Table 3 summarizes the relative RNA (transporter gene/β-actin ratios) expression levels of representative members of the OCT, OCTN, OAT, OATP, MRP, MDR, nucleoside, nucleobase, and peptide transporter families. All values recorded fall within the calibration curve range. RNA levels that were detectable (specific gel electrophoresis band and specific melting peak following melting curve analysis) but below the calibration curve range are denoted as below the level of quantification. Since their values fall outside the calibration curve range, the ratios were not determined. RNA levels defined as below the limit of detection show no specific melting peak and no band on gel electrophoresis. Since the principal purpose of the study was to compare the relative expression levels of the various transporters in lactating mammary epithelial cells with tissues known for their secretory activity as a means of identifying candidate transporters in drug accumulation into milk, the LightCycler assays were not optimized to achieve maximum sensitivity of quantification. Therefore, entries below the limit of detection do not preclude possible expressions at low levels. Entries with “E” refer to positive controls and tissues that demonstrated a specific melting curve and appropriately sized band on gel electrophoresis, but primer-dimer interference precluded accurate quantification in some tissues. Limitation in template quantity prohibited rerun analyses resulting in incomplete data for some of the transporters.

Approximately two-thirds of the evaluated transporters were expressed in MEC. Expressed members of the OCT family include OCT1 and OCT3, but OCT2 was neither expressed in nonlactating MEC nor lactating MEC. OCTN1 was expressed only in lactating MEC, but OCTN2 was expressed in all MEC. OAT1, OAT2, OAT3, and OAT4 were not expressed in MEC. In the OATP family, OATP-A, OATP-B, OATP-D, and OATP-E, but not OATP-C, were expressed in all MEC. MDR1, MRP1, MRP2, and MRP5 were expressed in MEC, but MRP3 and MRP4 lie below the level of detection. PEPT1 was expressed in both lactating and nonlactating MEC, but PEPT2 was expressed only in lactating MEC. CNT3, ENT1, and NCBT1 were expressed in all MEC, but CNT1 and ENT1 were only expressed in lactating MEC. CNT2, ENT2, and NCBT2 were not expressed in MEC.

RNA Gene Transcript Levels in Lactating Mammary Epithelial Cells Relative to Renal, Hepatic and Placental Tissues.

To assess the relevance of transporter RNA expression levels in lactating MEC, individual transporter RNA levels were compared with kidney and liver tissues in which these transporters are known to have an important function in drug disposition. Figure 2summarizes the RNA expression levels for transporter genes as a ratio of expression levels in lactating MEC relative to liver or kidney expression levels. Lactating MEC expression levels of most transporter genes compare reasonably well with liver and/or kidney levels. Notable exceptions include OCT1, OCT3, OATP-B, and CNT1, which demonstrate >10-fold higher expression levels in the liver, and OCTN2 and MDR1, which demonstrate >10-fold higher levels of expression in the kidney. Neither the liver nor the kidney expressed CNT3, but significant CNT3 RNA levels were demonstrated in lactating MEC. PEPT2 and OATP-A were not expressed at quantifiable levels in the liver and kidney, respectively, in contrast to high levels of expression in lactating MEC. Placental expression of transporter genes relative to lactating MEC shows more variable results, with less than 7.5-fold differences in RNA levels for most transporter genes. However, MDR1, PEPT1, and ENT3 demonstrate much greater (>15-fold) expression levels than lactating MEC (Table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Ratio of actin normalized transporter RNA expression levels in lactating mammary epithelial cells to expression levels in liver or kidney. a, ratio determined from extrapolated values (RNA expression levels lay outside the calibration curve range due to excessive dilution of cDNA template). b, below the level of detection; ratio determined from the value of negative control. c, unable to quantitate; no ratio determined.

Comparison of Transporter RNA Expression Levels between Nonlactating and Lactating Mammary Epithelial Cells.

To determine whether lactation alters transporter gene expression, individual transporter RNA expression levels were compared in nonlactating and lactating MEC. Table 4 gives the -fold difference in transporter RNA expression levels between lactating and nonlactating MEC. In general, a stable expression pattern was observed with a 2- to 3-fold difference or less in transcript levels between lactating and nonlactating MEC. This difference may reflect normal physiological variation and interindividual differences in the levels of transporter expression. A greater difference in RNA expression levels, however, was observed for OCT1, OCTN1, OCTN2, MDR1, PEPT2, CNT1, CNT3, and ENT3. Lactating MEC demonstrated ∼8-, ∼6-, ∼28-, ∼9-, ∼7-, and ∼4-fold higher RNA levels for OCT1, OCTN1, PEPT2, CNT1, CNT3, and ENT3, respectively, compared with nonlactating MEC. Conversely, transporter RNA expression levels of OCTN2 and MDR1 were ∼4- and ∼52-fold lower in lactating MEC relative to nonlactating MEC.

Discussion

Quantitative Analysis of 30 Transporter Proteins in Mammary Epithelial Cells.

Detection of transporter gene expression in the human lactating mammary epithelium is a necessary first step in the identification of candidate transporter proteins involved in drug transfer across the lactating mammary epithelium. This study is the first to describe the parallel analysis of RNA expression levels for 30 different representative members of these transporter families in lactating and nonlactating MEC. LightCycler RT-PCR analysis revealed lactating MEC express members from the OCT, OCTN, OATP, MRP, MDR, nucleoside, nucleobase, and oligopeptide transporter families at levels generally comparable to liver and/or kidney expression levels. OCT2, OAT1, OAT2, OAT3, OAT4, OATP-C, MRP3, MRP4, CNT2, ENT2, and NCBT2 were not expressed at levels above the detection threshold of the LightCycler PCR assay. Although examples of RNA expression without protein expression exist (Chesterman et al., 2001), such data may still identify a potential role for these expressed transporters in drug disposition at the mammary gland.

Comparison of Transporter Gene RNA Levels Between Lactating Mammary Epithelial Cells and Liver and Kidney Tissue.

In the present study, liver and kidney transporter expression profiles were consistent qualitatively with the literature, which affirms the validity of the results of the present study (Cass et al., 1998; Koepsell, 1998; Ritzel et al., 1998; Hogue and Ling, 1999; Klein et al., 1999; Borst et al., 2000; Sekine et al., 2000; Lahjouji et al., 2001; Pennycooke et al., 2001; Shen et al., 2001). In general, transporter RNA levels in lactating MEC compared reasonably well to liver and/or kidney expression levels. Although the liver expressed OCT1, OCT3, OATP-B, and CNT1 at levels considerably higher than lactating MEC, kidney expression levels of these transporters were comparable. High liver expression levels of such transporters are consistent with the liver's physiological role in the clearance of organic ions and hydrophobic molecules and, hence, does not necessarily preclude a role for these transporters at the lactating mammary epithelium. Significantly higher hepatic and renal MDR1 RNA levels relative to lactating MEC suggests MDR1 may have a limited role in drug disposition in the lactating mammary gland. An absence of CNT3 expression in the liver and kidney suggests CNT3 may have an important and specialized role in nucleoside accumulation in milk. Lactating MEC also exhibited increased PEPT2 RNA levels relative to the liver and increased OATP-A RNA levels relative to kidney suggesting these transporters may have a significant role in lactating MEC.

Lactating MEC expression levels of most transporters compared reasonably well with transporter expression in placental tissues. Interestingly, placental MDR1, PEPT1, and ENT3 RNA levels greatly exceeded expression levels in lactating MEC. These data suggest an important, but unidentified, endogenous role for these transporters in the placenta.

Drugs that accumulate into milk demonstrate diverse chemical compositions. To account for the accumulation of widely divergent drug classes in breast milk requires representation from several transporter families. To illustrate, cimetidine, an organic cation, accumulates in milk (Oo et al., 1995). OCT1, OCTN1, and OCT3 expression of OCT3 during lactation may contribute to organic cation disposition in the lactating mammary gland. Studies, however, identify cimetidine as a selective inhibitor, but not a substrate, for several organic cation transporters (Okuda et al., 1996; Zhang et al., 1997; Kekuda et al., 1998; Ohashi et al., 1999; Yabuuchi et al., 1999). These data suggest cimetidine accumulation in breast milk may involve transporters other than OCT1, OCT3, or OCTN members.

Several studies suggest organic anion transporters concentrate the organic anion N₄-acetylatedpara-aminohippurate into bovine milk (Rasmussen, 1969a) and transport the organic anions 4-aminoantipyrine,N₄-acetylated sulfanilamide, and nitrofurantoin across the bovine, caprine, and rat mammary epithelium, respectively (Rasmussen, 1969b; Kari et al., 1997). Additionally, acyclovir accumulates into milk (Taddio et al., 1994), and rOAT1 mediates acyclovir transport across membranes (Wada et al., 2000). Lack of OAT expression in lactating MEC precludes their involvement in the milk accumulation of drugs such as nitrofurantoin and acyclovir. Expressed members of the OATP and MRP transporter families, however, may contribute to organic anion accumulation in milk. Consequently, high transcript levels of OATP-A, OATP-D, OATP-E, MRP1 and MRP5 suggest these transporters may contribute to organic anion transfer across the mammary epithelium. In particular, the basolateral localization of MRP1 in epithelial tissues may impart a protective role (Borst et al., 2000) to the lactating mammary gland extruding potentially toxic drugs from the epithelium and reducing milk levels of drugs. An apical localization of MRP2 (Schaub et al., 1997) implies MRP2 may have some modest contribution to the accumulation of organic anions in breast milk.

Expression of CNT1, CNT3, ENT1, ENT3, and NCBT1 at the lactating mammary epithelium implicates a role for these transporters in the transfer of acyclovir, nitrofurantoin, and other nucleoside or nucleobase drug analogs into milk. Finally, lactating MEC expression of PEPT1 and PEPT2 suggest these transporters may transport peptide drugs or peptidomimetics across the lactating mammary epithelium. Knowledge of transporter expression at the lactating mammary epithelium and their apical or basolateral localization has two functions: first, to identify a priori which drugs have the potential to accumulate in breast milk based upon known transporter substrate profiles, and second, to help narrow the potential list of candidate transporter proteins involved in the accumulation of drugs into breast milk.

Comparison of Transporter Gene RNA Levels Between Lactating and Nonlactating Mammary Epithelial Cells.

The expression of members of the OCT, OATP, MDR, MRP, nucleoside, nucleobase, and peptide transporter families in lactating MEC may relate to a physiological role for these transporters in milk production. An endogenous function in the mammary gland suggests drug transport across the mammary epithelium may interfere with the normal physiological function of the transporter. This inhibition may adversely affect milk composition and infant health. Any consideration of drug disposition in the mammary gland must acknowledge the exposure risk associated with both the potential for drug accumulation in breast milk and the potential changes in milk composition associated with inhibition of endogenous transporter function.

The transcriptional up-regulation of OCT1, OCTN1, PEPT2, CNT1, CNT3, and ENT3 and down-regulation of MDR1 and OCTN2 may suggest a significant endogenous role for these transporters during lactation. Furthermore, demonstration of lactation-induced alteration in transporter gene expression may identify potential players (i.e., hormones and extracellular matrix components) in transcriptional regulation of these genes. A role for OCT1 and OATP transporters at the mammary gland remains unknown. Nevertheless, given their basolateral membrane localization (http://www.med.rug.nl/mdl/) and their function in organic cation and organic anion uptake, respectively, these transporters may transport various minor constituents of milk (Picciano, 2001).

OCTN1 and OCTN2 transport l-carnitine into cells (Lahjouji et al., 2001). The transport and provision of l-carnitine into milk is important for normal growth and development of the suckling infant (Penn et al., 1980). Functional studies have postulated the presence of a carnitine transporter at the mammary epithelium (Shennan et al., 1998; Zammit et al., 1998). Lactating MEC expression of OCTN1 and OCTN2 corroborates these functional studies. The down-regulation of OCTN2 expression observed in the present study may be related to lactational stage. The activity of Na⁺-dependent l-carnitine transport decreases with progressing lactation and maturation of an infant's capacity to synthesize l-carnitine (Shennan et al., 1998). In the present study, mammary epithelial cells were collected from pooled early and late lactation cycle milk. Possibly, late lactation stage MEC predominated accounting for the apparent down-regulation of OCTN2 expression.

MDR1 expression on apical membranes of epithelia with a protective or excretory function suggests MDR1 has a role in detoxification processes in various organ systems (Schinkel, 1997). Down-regulation of MDR1 RNA expression with lactation may be an adaptive and protective response to reduce milk levels of toxic compounds.

Up-regulation of PEPT2 RNA expression in lactating MEC corroborates a recent study that identified PEPT2 RNA and protein expression in MEC collected from human milk (Groneberg et al., 2002). PEPT2 expression in the lactating mammary epithelium may have an important function as a scavenger uptake system for the short-chain peptide products of milk protein hydrolysis. Enhanced expression of CNT1, CNT2, and ENT1 with lactation may accommodate the enhanced requirements of a highly active secretory epithelium and the need to meet the nutritive demands of the suckling infant for nucleoside precursors (Picciano, 2001).

Conclusions

Mammary epithelial RNA expression of members from the OCT, OCTN, OATP, MDR, MRP, nucleoside, nucleobase, and oligopeptide transport families identifies candidate transporters involved in drug disposition across the mammary epithelium. RNA expression profiling of transporters in the mammary epithelium, however, is only a first step toward elucidating the molecular mechanisms underlying substrate transfer across the mammary epithelium. Further investigations into transporter activity and membrane localization in the polarized mammary epithelium will define an unequivocal role for transporter function in mammary epithelial cells. This will facilitate identification of drugs in which their disposition in the mammary gland may be governed by transporter proteins, especially as substrate binding characteristics and substrate specificities of these transporters become further elucidated. In addition, functional studies are needed to identify the physiological and pharmacological role of these transporters in the mammary gland.