Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

A diffusion model has been established to predict drug transfer into milk by using estimates of drug binding and ionization in milk and serum as well as a lipid partitioning factor for milk (Fleishaker et al., 1987; Fleishaker and McNamara, 1988). Although this model has successfully estimated the milk-to-serum concentration ratios (M/S) for several drugs transferred passively by diffusion (Fleishaker and McNamara, 1988), it does not predict the high M/S observed for several agents. For example, reports have shown that certain agents, including cimetidine, ranitidine, and nitrofurantoin are actively transported into animal or human milk (Oo et al., 1995; McNamara et al., 1996; Kari et al., 1997). However, a concentrative drug transport mechanism in the lactating mammary gland has yet to be fully characterized and identified.

Previous studies have shown evidence for a cimetidine transport system in the rat mammary epithelium that is saturable and inhibited by ranitidine (McNamara et al., 1996). It produces an observed M/S that is 6 times greater than the M/S predicted by the diffusion model (McNamara et al., 1992). Also, investigations by Kari et al. (1997) show that nitrofurantoin, an acidic nitrofuran antibiotic, enters rat milk at concentrations 75 times those predicted by the diffusion model. Because both cimetidine and nitrofurantoin are renally secreted into urine and concentrated into milk (showing M/S values several times greater than predicted), yet structurally different (i.e., cationic versus anionic, respectively), these agents were chosen as in vivo probes to elucidate potential mechanisms of concentrative xenobiotic transport into milk. Although the involved transport mechanisms have not been identified, the activity is termed lactating mammary epithelial drug transport (LMEDT) in this article.

The identities of the genes responsible for cimetidine and nitrofurantoin LMEDT activities are unknown. One or more proteins may be involved in the concentrative transport mechanisms for these agents. Cimetidine and nitrofurantoin are secreted by rat and human kidneys, and both have probenecid-sensitive components (Braunlich et al., 1978; Conklin, 1978; Lin et al., 1988; McEnvoy, 1998). Because probenecid inhibits a part of the renal clearances of cimetidine and nitrofurantoin, it may also inhibit the secretion of cimetidine and nitrofurantoin in the mammary gland.

As organic anions or cations, these agents may be substrates for rat organic anion (rOAT) or organic cation (rOCT) transporters. In the rat, three rOATs (rOAT1, rOAT2, and rOAT3) have been cloned (Sekine et al., 1997, 1998; Sweet et al., 1997; Kusuhara et al., 1999). Also, three rOCTs (rOCT1, rOCT2, and rOCT3) have been isolated (Grundemann et al., 1994; Okuda et al., 1996; Kekuda et al., 1998). All six of these transporters have been detected in rat kidneys, but only rOAT2, rOAT3, and rOCT1 have been detected in rat livers (Grundemann et al., 1994; Okuda et al., 1996; Sekine et al., 1997, 1998; Sweet et al., 1997; Kekuda et al., 1998; Kusuhara et al., 1999). However, it is not known which of these are expressed in the lactating rat mammary epithelium. As a step toward identifying the LMEDT systems for cimetidine and nitrofurantoin, it is important to know which cloned transporters may be expressed. RT-PCR provides a method to quickly rule out certain transporters that are not detectably transcribed in the lactating rat mammary epithelium.

The purpose of this study was to further elucidate the potential transport processes for cimetidine and nitrofurantoin. To determine whether they share a common transport component, the interactions between cimetidine and nitrofurantoin secretion into milk as well as the effects of probenecid were determined. Also, whether probenecid and PAH were actively secreted into milk was determined by testing the diffusion model. Finally, because some of these agents interact with rOATs and rOCTs, the expression of the mRNA transcripts for these transporters was determined by RT-PCR.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. The agents used in the infusions (cimetidine, nitrofurantoin, probenecid, and p-aminohippurate sodium) as well as ranitidine HCl and furazolidone were purchased from Sigma Chemical Co. (St. Louis, MO). Potassium phosphate monobasic and all organic solvents (HPLC grade) were purchased from Fisher (Pittsburgh, PA).

Milk and Serum Protein Binding and Milk Fat Partitioning Determinations. Blank fresh milk and serum were harvested from lactating female rats under ketamine/acepromazine anesthesia. Protein binding in milk and serum was determined by equilibrium dialysis against a phosphate buffer adjusted to the pH of milk or serum, as described previously (McNamara et al., 1992). Milk fat partitioning was determined by spiking whole milk and then centrifuging aliquots to obtain skim milk, as described previously (McNamara et al., 1992). Values for pH of rat serum and milk used for the calculation of un-ionized fractions were 7.46 and 6.67, respectively, as measured previously in our laboratory (D. E. Burgio and P. J. McNamara, unpublished data).

Infusion Studies. The protocols 86-0217 M and 97-0052 M were approved by the University of Kentucky Institutional Animal Care and Use Committee. Sprague-Dawley lactating female rats (dams, 250-350 g) were received from Harlan (Indianapolis, IN) and allowed to acclimatize to their environment. The jugular and femoral veins were cannulated under ketamine/acepromazine anesthesia on day 10 to 15 post partum. After a day for recovery, the dams were separated from the pups and the infusions began. Each dam (n = 6) was randomized to receive an intravenous infusion of the probe drug either alone or with the inhibitor on the first day and crossed over on the second day to complete both phases (Table 1). The dams were separated from the pups before the beginning of the infusion. The infusions commenced, lasting 5 h for nitrofurantoin and 8 h for cimetidine, PAH, and probenecid, with blood samples being drawn hourly for the last 3 (or 4) h of the infusion. The blood samples were protected from light, allowed to clot, centrifuged to harvest serum, and frozen until analysis. Milk samples were obtained by manual milking under light anesthesia (ketamine/acepromazine) at the end of the infusion, protected from light, and frozen (20°C) until analysis.

Cimetidine HPLC. Cimetidine concentrations in rat serum and milk were determined as reported previously (McNamara et al., 1992). Briefly, ranitidine was added as an internal standard to each 100 µl of serum or milk aliquot. The sample was alkalinized with 5 N NaOH, extracted into 1 ml of methylene chloride, evaporated under nitrogen, and reconstituted in mobile phase. A 50 µl volume was injected onto a C18 column using a Shimadzu HPLC system as previously described (McNamara et al., 1992) and eluted with 6% acetonitrile in water containing 0.25 µM acetic acid and 0.2 µM triethylamine. The UV absorbance was measured at 228 nm. The milk samples were diluted if the concentration was greater than the linear range for the standard curve.

Nitrofurantoin HPLC. The 50 µl aliquots from the nitrofurantoin M/S observed versus predicted determinations (Table 3) were precipitated with 125 µl of acetonitrile, vortexed, and then centrifuged for 10 min at 4°C, and 50 µl of the supernatant was injected onto the HPLC as described below. To maximize column life, all other samples were extracted using a modified method of Pons et al. (1990). Briefly, to each 50 µl aliquot of serum or milk, 25 µl of a 5 µg/ml furazolidone HCl was added as an internal standard. The sample was acidified and proteins precipitated by adding 1 N HCl and vortexing for 30 s. The sample was extracted into 1 ml of methylene chloride, vortexed, and centrifuged. The organic layer was decanted, evaporated under nitrogen, and reconstituted with mobile phase. The sample was injected onto the same Shimadzu HPLC system with a Lichrosorb 5 RP18 125 × 4.0-mm column (Phenomenex, Torrance, CA) and eluted with 10% acetonitrile 90% 25 mM potassium phosphate buffer (pH 3.00) at 1.0 ml/min. UV absorbance was measured at 366 nm. Peak height ratios (nitrofurantoin/furazolidone) were used for comparison with the standard curve. The milk samples were diluted if the concentration was greater than the linear range for the standard curve (0.2 to 6.25 µg/ml).

p-Aminohippurate HPLC. Each 100 µl of serum or milk sample was precipitated with 500 µl of acetonitrile, vortexed, centrifuged 10 min, and the supernatant was decanted and evaporated under nitrogen. The residue was reconstituted in mobile phase (0.06 M potassium phosphate monobasic, pH 6.4) and injected onto a C18 column as described above for nitrofurantoin. Detection of UV absorbance was determined at 254 nm. The mean recovery was 95% and the standard curves were linear from 0.78 to 100 µg/ml.

Probenecid HPLC. Each 100 µl sample was spiked to 2.5 µg/ml with antipyrine as the internal standard, acidified with 1 N HCl, and vortexed. The sample was extracted into methylene chloride, vortexed, and centrifuged for 10 min. The organic (bottom) layer was decanted, dried under nitrogen, and reconstituted with mobile phase (75% aqueous 0.06 M potassium phosphate monobasic, pH 6.7, 0.1% triethylamine, and 25% acetonitrile). The sample was injected onto a C18 column as described above for nitrofurantoin and detected by UV absorbance at 275 nm and peak height ratios (probenecid/antipyrine) were compared with the standard curve to determine concentrations. The mean recovery was 78.5% and the linear range for the standard curve was 2.5 to 400 µg/ml.

Pharmacokinetic Analysis. The pharmacokinetic parameters were determined as reported previously (Fleishaker et al., 1987). Observed M/S was determined by the quotient of the milk and serum concentrations at steady state as in eq. 1:

(1)

Predicted M/S was determined as in eq. 2,

(2)

where M/S_predicted is the predicted milk to serum ratio; f^u_s and f^u_m are the un-ionized fractions of the drug in serum and milk, respectively; f_s and f_m are the unbound fractions in serum and milk, respectively; and S/W is the skim-to-whole milk drug concentration ratio. Systemic clearance was determined by dividing the infusion rate by the steady-state serum concentration. Only the parent compounds were analyzed in these studies.

Statistical Analysis. In the infusion studies, two-tailed paired t tests with  = 0.05 were performed on observed M/S and systemic clearance of the measured drug alone or in the presence of the inhibitor. The sample size of six paired determinations was calculated to achieve 80% power to detect a 50% difference in means, expecting a coefficient of variation of about 40% from previous studies (Bolton, 1990; McNamara et al., 1992, 1996). To detect active transport in comparing the observed and predicted M/S, 95% confidence intervals were established (Bolton, 1990) to compare the observed M/S with the range encompassing 50 to 200% of the predicted M/S. If the 95% confidence interval for the observed M/S did not include values encompassed by the 95% confidence interval of 50 to 200% of the predicted M/S, then the null hypothesis (that passive diffusion predominated) was rejected and the alternate hypothesis (that active transport predominated) was accepted.

Detection of mRNA Transcripts by RT-PCR. Total RNA was isolated from the livers, kidneys, and mammary glands of three lactating females rats using the RNeasy Midi kit (Qiagen, Valencia, CA). First strand cDNA synthesis was performed in a volume of 20 µl with 2 µg of total RNA using the SuperScript Preamplification System (Life Technologies, Rockville, MD). The PCR primers (Table 2) were either designed with Oligo 4.0 (Molecular Biology Insights, Cascade, CO) using sequences formatted from GenBank or were from other sources as indicated. The primers were synthesized by IDT (Coralville, IA). Rat -actin PCR was performed according to the method of Serazin-Leroy et al. (1998). The other PCR reactions were performed as follows. The cDNA template (2 µl) was combined with PCR buffer (Life Technologies), 1.5 mM MgCl₂, 1 mM dNTP, 0.3 µM PCR primers, and amplified with 2.5 U of Taq polymerase (Roche Molecular Biochemicals, Indianapolis, IN) in a volume of 50 µl using a PE Applied Biosystems GeneAmp PCR System 9700 (PerkinElmer Inc., Boston, MA). After an initial hot start at 94°C for 3 min, each of 35 cycles consisted of 94°C for 30 s, 58°C (except for rOAT3: 65°C) for 60 s, and 72°C for 60 s. Final elongation was at 72°C for 10 min and samples were held at 4°C until analysis. PCR products were detected by electrophoresis on a 1.5% agarose gel in Tris borate-EDTA buffer. Gels were stained with ethidium bromide, visualized by UV light, and the images were electronically captured using a MultiImage Light Cabinet with ChemiImager 4000 v.3.3b (Alpha Innotech Corp., San Leandro, CA). The electronic files were enhanced using Scion Image beta release 4 (Scion Corp., Frederick, MD) and imported into Microsoft Powerpoint 2000 (Microsoft Corp., Redmond, WA) for presentation.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Table 3 provides the variables used in the predictions of M/S according to the diffusion model and the observed M/S in vivo. The 95% confidence intervals for M/S observed for cimetidine, nitrofurantoin, and probenecid did not overlap the range encompassing the 95% confidence intervals of 50 to 200% of the predicted M/S values for each of these drugs. The respective M/S observed values were 103, 9.3, and 4.7 times predicted for nitrofurantoin, cimetidine, and probenecid. However, the confidence intervals overlapped for the predicted and observed M/S of p-aminohippurate.

View this table: [in this window] [in a new window]   TABLE 3 M/S predicted and observed Physicochemical properties, M/S predictions (see bold text), and M/S observations (see bold text) for cimetidine, nitrofurantoin, p-aminohippurate, and probenecid. pKa values are from McEnvoy (1992). Calculated log P values are the result of ACD/LogP v3.6 obtained using the ACD/I-lab service. Data are given as mean (S.D.). Data for cimetidine are taken from McNamara et al. (1992, 1996). Asterisks indicate nonoverlapping confidence intervals (see Materials and Methods).

Table 1 provides the M/S and Cls for the agents studied in the absence or presence of an inhibitor. Steady state was achieved in each infusion for each agent (data not shown). Figures 1 and 2 provide steady-state plots of the serum and milk concentrations for a group of rats receiving cimetidine and nitrofurantoin. The M/S for cimetidine was significantly decreased by coadministration of nitrofurantoin. Also, nitrofurantoin decreased cimetidine Cls. The coadministration of cimetidine did not significantly alter the M/S of nitrofurantoin. However, cimetidine inhibited the Cls of nitrofurantoin.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Effect of nitrofurantoin on cimetidine. Cimetidine (CM) serum and milk concentrations when infused at 0.5 mg/h in the absence or presence of nitrofurantoin 3.75 mg/h (CM or CM + NF, respectively). Serum and milk concentrations designated by circles or squares, respectively, and the absence or presence of nitrofurantoin, respectively, indicated by open or filled symbols (, CM milk; , CM + NF milk; , CM serum; , CM + NF serum). Error bars indicate S.D.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Effect of cimetidine on nitrofurantoin. Nitrofurantoin (NF) serum and milk concentrations when infused at 0.5 mg/h in the absence or presence of cimetidine 30 mg/h (NF or NF + CM, respectively). Serum and milk concentrations designated by circles or squares, respectively, and the absence or presence of cimetidine, respectively, indicated by open or filled symbols (, NF milk; , NF + CM milk; , NF serum; , NF + CM serum). Error bars indicate S.D.

Probenecid increased the M/S of cimetidine but did not alter cimetidine Cls. Probenecid did not significantly change either nitrofurantoin M/S or Cls. The M/S of PAH was not significantly changed by probenecid. Furthermore, it was not significantly different from the predicted M/S. However, probenecid decreased the Cls of PAH.

Figure 3 shows the detection of mRNA transcripts in lactating rats by RT-PCR. -Actin was found to a similar extent in all tissue samples, demonstrating an equivalent amount of RNA was examined for each specimen. All of the transporters were detected in the kidney, and rOAT2, rOAT3, and rOCT1 were detected in the liver. However, only rOCT1 and rOCT3 were detected in the lactating rat mammary gland.

View larger version (62K): [in this window] [in a new window]   Fig. 3.   Transcription of OATs and OCTs in lactating rat livers, kidneys, and mammary glands. Lanes 1 to 3 show results from lactating rat livers. Lanes 4 to 6 show results from lactating rat kidneys. Lanes 7 to 9 show results from lactating rat mammary glands. The cDNA detected is listed along the right margin.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Cimetidine (McNamara et al., 1992, 1996), nitrofurantoin, and probenecid are actively transported into rat milk (Table 3). Cimetidine active transport into milk has been reported previously for both rats and humans (McNamara et al., 1992, 1996; Oo et al., 1995). Our study yielded an M/S for nitrofurantoin over 100 times predicted, and Kari et al. (1997) reported a milk-to-plasma ratio after a single oral dose of nitrofurantoin 75 times predicted. Both the present study and the results of Kari et al. (1997) affirm the tenacity of the nitrofurantoin lactating mammary transport process. Probenecid was actively transported into milk, with an M/S 4.7 times greater than predicted. This is the first report showing active secretion of probenecid by the mammary gland. In contrast, p-aminohippurate transfer into milk was governed largely by diffusion. The approach used in this article would not detect subtle active transport processes into milk because at least a 2-fold difference in magnitude of M/S predicted and observed values would be required for the values to be considered different.

The agents used in the studies in this article may interact with a variety of transporters. Cimetidine interacts with several transporters such as the OCTs (Koepsell, 1998), rOAT3 (Kusuhara et al., 1999), and P-glycoprotein (Collett et al., 1999). Nitrofurantoin transport has not been well characterized, and specific transport mechanisms remain unknown. As an organic anion, nitrofurantoin may be transported by one or more of the known organic anion transporters, such as the OATs, the multidrug resistance-associated proteins, and the organic anion-transporting polypeptides, known as Oatps (Pritchard and Miller, 1993; Roch-Ramel, 1998). PAH is transported by rOAT1, rOAT2, and rOAT3 (Sekine et al., 1998; Uwai et al., 1998; Kusuhara et al., 1999). Probenecid could be a substrate for multidrug resistance-associated proteins or the uric acid transporter (Roch-Ramel, 1998), but not OAT1 (Uwai et al., 1998). Probenecid also inhibits rOAT3 (Kusuhara et al., 1999), but whether it is a substrate is unknown. The degree of expression of these and other drug transporters in the lactating rat mammary epithelium has not been determined in the literature.

Among the drugs studied in this article, several drug-drug transport interactions have been reported. For cimetidine, Lin et al. (1988) showed that probenecid 40 mg/kg i.v. every 40 min decreases cimetidine renal clearance by approximately 20%. Other studies using rat proximal tubular cells to examine uptake of cimetidine yielded an IC₅₀ for probenecid of 708 µM (Boom and Russel, 1993). Weiner and Roth (1981) showed that cimetidine 48 mg/h i.v. administered to rats had no effect on the renal tubular secretion of [¹⁴C]PAH, while abolishing the secretion of the prototypical organic cation [¹⁴C]tetraethylammonium. Regarding nitrofurantoin interactions, the renal excretion rate of nitrofurantoin within 3 h of a 10 mg/kg intraperitoneal dose given to adult Wistar rats was decreased by 60% after a single intraperitoneal dose of probenecid 200 mg/kg (Braunlich et al., 1978). Concerning PAH transport, probenecid inhibits the PAH transporters rOAT1 and rOAT3 (Sekine et al., 1997; Kusuhara et al., 1999). No literature reports any interactions between cimetidine and nitrofurantoin.

M/S Interactions. This study showed that the LMEDT mechanism for cimetidine was inhibited by nitrofurantoin. The type of inhibition (i.e., competitive versus noncompetitive) is unknown. If the interaction is competitive, it could suggest that nitrofurantoin competes with cimetidine for a common transporter. However, this study did not show any inhibition of nitrofurantoin milk secretion due to cimetidine at a rate sufficient to inhibit the M/S of ranitidine (McNamara et al., 1996). Another difference exists in the degree of active transport, in which the difference between observed and predicted M/S for nitrofurantoin was 100-fold and for cimetidine was 9-fold. Although both cimetidine and nitrofurantoin are actively transported into rat milk, they may not share a common transporter, or cimetidine may have a lower affinity for it.

Cls Interactions. The systemic clearances of cimetidine reported here were similar to those previously reported in male or lactating female Sprague-Dawley rats (Weiner and Roth, 1981; McNamara et al., 1992). Weiner and Roth (1981) showed that after i.v. administration to adult male rats, cimetidine renal clearance is 80% of systemic clearance, and that cimetidine undergoes net tubular secretion. Our study also showed a significant decrease in cimetidine Cls in the presence of nitrofurantoin. Nitrofurantoin may inhibit cimetidine clearance by inhibiting one or more renal transport mechanisms or through another mechanism.

Detection of mRNA Transcripts. The RT-PCR reactions detected the expression or absence of the mRNA transcripts for the rOATs and rOCTs in the livers and kidneys as expected (Grundemann et al., 1994; Okuda et al., 1996; Sekine et al., 1997, 1998; Sweet et al., 1997; Kekuda et al., 1998; Kusuhara et al., 1999). Of these transporters, only rOCT1 and rOCT3 were seen in the lactating rat mammary gland. This suggests that rOAT1, rOAT2, rOAT3, and rOCT2 are not responsible for the high M/S ratios of cimetidine and nitrofurantoin observed in lactating rats. Furthermore, because PAH is a substrate for rOAT1, rOAT2, and rOAT3 (Sekine et al., 1998; Uwai et al., 1998; Kusuhara et al., 1999), the lack of PAH active transport into rat milk is functional evidence for their absence. The role of rOCT1 and rOCT3 in the transport of these drugs into rat milk remains to be elucidated.