Metabolism and disposition of oseltamivir (OS) in rats, determined by immunohistochemistry with monospecific antibody for OS or its active metabolite oseltamivir carboxylate (OC): A possibility of transporters dividing the drugs’ excretion into the bile and kidney

Abstract Among any drugs, no comparative pharmacological study on how prodrug and its active metabolite behave in animal bodies is available. Immunohistochemistry (IHCs) using newly prepared two monoclonal antibodies, AOS‐96 and AOC‐160, monospecific for oseltamivir (OS) and its metabolite oseltamivir carboxylate (OC) were developed, simultaneously detecting the uptake or excretion of OS and OC in the intestine, liver, and kidney of rats to which OS was orally administered. In the intestine, IHC for OS revealed OS highly distributed to the absorptive epithelia with heavily stained cytoplasmic small granules (CSGs). IHC for OC showed that OC also distributed highly in the epithelia, but without CSGs, suggesting that OS was partly converted to OC in the cells. In the liver, OS distributed in the hepatocytes and on their bile capillaries, as well as on the lumina from the bile capillaries to the interlobular bile ducts. OC distributed in the whole cell of the hepatocytes, but without CSGs nor on any lumina through the interlobular bile ducts. In the kidney, a few levels of OS distributed in the cytoplasm of almost all the renal tubule cells, but they contained numerous CSGs. In contrast, OC distributed highly in the proximal tubules, but very slightly in the lower renal tubules of the nephrons. Thus, it was concluded that the two drugs behave in completely different ways in rat bodies. This paper also discusses a possibility of the correlation of OS or OC levels in tissue cells with their known transporters.


| INTRODUC TI ON
Oseltamivir (OS) is an ethyl ester-type pro-drug of Ro 64-0802 (oseltamivir carboxylate; OC), a potent and selective inhibitor of viral neuraminidase, a key enzyme involved in the release of the influenza virus from infected host cells. 1,2 OS is used for the prevention, treatment, and prophylaxis of epidemic seasonal influenza of Influenza virus A and B. 3 An ethyl ester group of OS is enzymatically cleaved to the active metabolite OC predominantly in the plasma in rats, 4 while in humans (and monkeys), the cleavage takes place in the liver, 5 and both OS and OC are found in the blood. 6,7 The accurate localization of OS and OC in tissue cells would help to develop better understanding of the biological response to the drugs in the pharmacological studies. Previous distribution studies for drugs in animals have been almost exclusively undertaken by autoradiography, as reported by Wiltshire et al 8 using 14C-labeled OS, but without detailed information. We searched for an alternative procedure for immunohistochemistry (IHC), which has the advantage over autoradiography of being able to directly observe the tissue localization of drugs using a common light microscope. Also, in principle, IHC detects a drug molecule itself according to the specificity of the antibody (Ab) and is as highly sensitive of the assay procedure as autoradiography. Furthermore, it does not need any special skills during the operation and can be completed within less than a couple of days. However, there were no reports available on IHC for drugs developed by pharmacologists or histologists until 2005. 9 Prior to our starting to investigate IHC for drugs, we had been engaged in the research for low molecular weight endogenous amines, polyamines, 10 and histamine, 11 by IHC procedure, in which Abs specific for their amines were prepared and used with light and electron microscopy. Our studies revealed that polyamines localized on the free and attached ribosomes in any cell types, 12 and also that histamine in rats is concentrated in the specific cores of cytoplasmic granules in the enterochromaffin-like cells of the stomach. 13 Those studies taught us that for developing the IHC for small molecule compounds, it was critically important to set up the accurate conditions of the fixation, especially for the small water-soluble antigens in situ, 12,13 and also understand how to prepare Ab, 14,15 the antigen masking effect caused by the fixative, 16 its unmasking procedure for reaction with Ab(s), 10,11,16 and how to remove the non-specific binding of Ab. 17 We, therefore, adapted these understandings to the development of IHCs for drugs especially with primary aliphatic amino group(s) in their molecules. 9 In addition, small molecular endogenous amines have been extensively studied with IHCs by others. 18 The pharmacodynamics of pro-drug together with its active metabolite at the cell level would be valuable in the animal bodies because no such knowledge so far exists. Both OS and OC are ideal drug molecule models with an aliphatic amino group as the fixation site in situ for IHC because a simultaneous comparative study by autoradiography might be almost impossible to perform. Also, as the transport systems responsible for OS or OC uptake or excretion a member of the ATP binding cassette (ABC) transport proteins Mdr1 (P-glycoprotein: P-gp), [19][20][21][22][23] the organic anion transporter (OAT1), 24 the OAT3, 1 multidrug resistance-associated protein 4 (Mrp4/abcc4), [25][26][27] and the peptide transporter1 (Pept1) [28][29][30][31] have been demonstrated by pharmacological studies using knockout mice or cell lines expressing such transporters. However, the in vivo role of transporters in drug disposition, and metabolism, has been scarcely established. We have tried a few studies on the correlation of transporters with drug disposition, 32 We now report on the development of two mAbs monospecific for OS and OC, and the two IHCs for the uptake of OS and OC in the small intestine, liver, and kidneys of rats that were administered OS orally.
The finding that both OS and OC which were uptaken into the hepatocytes flow into the bile capillary and the sinusoidal capillary possibly through the P-gp and Mrp4, respectively, suggests that the transporter(s) play a critical role in dividing the excretion of both of the drugs into the bile and kidney. Also, it is suggested that OS partly undergoes the first-pass hydrolysis into OC within the intestinal epithelia, and that OS, but not OC, might be classified into the "Lysosomotropic" cationic amphiphilic drug category. The IHC clearly demonstrated its utility as a new tool for pharmacology and toxicity studies at the cell level.

| Chemicals
Tamiflu was from Chugai Pharmaceutical, Tokyo, Japan.

| Preparation of OS-or OC-bovine serum albumin (BSA) conjugates
OS-GA-BSA conjugate was prepared according to our previous method using GA as cross-linking agents. 10 OS (approx. 30 mg) was dissolved in 2.0 mL of 0.12 mol/L borate buffer, pH 10, and 15 µg of GA was mixed and incubated at room temperature (RT) for 30 seconds with stirring, and to the mixture was then added BSA (15 mg) in 1.0 mL of the buffer. This was followed by incubation for 30 minutes before NaBH 4 (5 mg) was added to terminate the reaction. The reaction mixture was further incubated for 30 minutes with slow stirring. The conjugate was then purified by a column chromatography of Sephadex G-75 equilibrated with 10 mmol/L phosphate buffer, pH 7.0 containing 4 mol/L urea. Also, OC-GA-BSA was prepared in the same manner as described above using OC instead of OS. The resulted conjugates OS-GA-BSA and OC-GA-BSA were used as immunization antigens for anti-oseltamivir (AOS) and anti-oseltamivir carboxylate (AOC) monoclonal antibodies (mAbs), respectively, or also as solid-phase antigens or inhibitors for an enzyme-linked immunosorbent assay (ELISA) described below.

| Preparation of AOS and AOC mAbs
As for the AOS mAb, three five-week-old, female BALB/c mice were injected intraperitoneally (i.p.) with 100 µg of OS-GA-BSA conjugate emulsified in complete Freund's adjuvant (Difco). Subsequently, they received three injections of 50 µg of the conjugate alone at twoweek intervals. Following immunization, antisera were collected, and antibody titers were evaluated with ELISA as described below.
The mouse with the best immune response was selected for hybridization. The mouse received a fourth i.p. booster injection and was sacrificed 4 days later. Experiments for the AOC mAb were similarly carried out using the conjugate OC-GA-BSA as the immunogen.

| Cell fusion and cloning
In these experiments for either AOS or AOC, the spleen cells (2 × 10 8 ) from the immunized mouse and 3x10 7 myeloma cells (P3/ NS-1) were fused with the help of polyethylene glycol according to our previous method. 10 Cells suspended in HAT medium were plated out in 96-well tissue culture plates (Corning) at a density of 10 5 cells per well in which 10 5 feeder cells had been plated. From 10 to 20 days postfusion, the wells were screened for reactivity using an ELISA method, as described below. Limiting dilutions of positive cultures were carried out two or three times to obtain monoclonality, and sub-isotyping of the mAbs was performed using a Mouse Monoclonal Sub-isotyping kit (American Qualex Int.). Ascites were raised in BALB/c mice pretreated with Pristene by intraperitoneal injection of 2 × 10 6 hybridoma cells.

| Dilution ELISA
ELISA was performed similarly to our previous method for antispermine mAbs. 10 In screening clones for production of antibody against OS-GA-BSA (or OC-GA-BSA), wells in microtiter plates were coated with 10 µg/mL of each of the conjugates for 30 minutes at RT. The wells were then incubated overnight at 4°C with antiserum (diluted 1:3000), hybridoma culture supernatant, or ascites fluid (diluted 1:100 000), followed by goat anti-mouse IgG labeled with horseradish peroxidase (HRP) (diluted 1:2000) for 1 hour at 25°C.
The amount of enzyme conjugate bound to each well was measured using o-phenylenediamine as a substrate, and the absorbance at 492 nm was read with an automatic ELISA analyzer (ImmunoMini NJ-2300, Nalje Nunc Int. Co. Ltd.).

| Inhibition ELISAs 34,35
Wells in a microtiter plate are coated with 100 µL of the OS-GA-BSA

| Binding ELISAs
According to our previous method, 10 the wells in a microtiter plate coated with poly-L-lysine (30 µg/mL) were activated with 2.5% GA in 50 mmol/L borate buffer, pH 10.0, for 1 hour The wells were subsequently incubated with test compounds at various concentrations for 1 hour at RT. Excess aldehyde groups were blocked with 0.5% sodium borohydride. The wells were further incubated for 1 hour with 1% skimmed milk to block nonspecific protein binding sites, and then the plates were incubated overnight at 4°C with the primary AOS-96 (or AOC-160) at 1:50 diluted with phosphate-buffered saline (PBS) containing 0.05% Tween 20 (PBST). The wells were then incubated for 1 hour with HRP-labeled goat anti-mouse IgG (diluted 1:2000).
The bound enzyme activity was measured as described above.

| Animals
Normal adult male Wistar rats (Kyudo Exp. Animals), body weight 200-250 g, were used in this study. The principles of laboratory animal care and specific national laws were observed. The animals were housed in temperature-and light-controlled rooms (21 ± 1°C and 12L: 12D) and starved overnight but with free access to tap water.
Single oral doses of 20 mg OS/kg of body weight were administered to nine Wistar rats. At 1, 3, 12, and 24 hours after the administration three rats for each group were anesthetized and perfused transcardially with 2.0% GA in PBS (10 mmol/L phosphate buffer, pH 7.2, containing 0.15 mol/L NaCl). The small intestine, liver and kidney were quickly excised and postfixed in the same fixative GA overnight at 4°C and was subsequently routinely embedded in paraffin.

| Immunohistochemistry (IHC)
The IHC method for OS or OC was carried out essentially according to our previous methods. 33,36,37 While they were under sodium pentobarbital (60 mg/kg; Abbott Laboratories) anesthesia, the rats were perfused intracardially with PBS at 50 mL/min for 2 minutes at RT and then with a freshly prepared solution of 2% GA in 10 mmol/L phosphate buffer, pH 7.2, for 6 minutes. Kidney, liver, and intestine During each process of the treatment, the specimens were washed three times with TBS containing 0.9% sodium metabisulphite (SMB).
Next, the specimens were blocked with a protein solution containing 10% normal goat serum (NGS) supplemented with 0.1% NaN 3 , In the present studies, the optimal conditions for immunostaining with respect to protease digestion differed in the different cell tissues. These differences may reflect differences in the binding and masking of OS or OC by proteins and other macromolecules in different cellular compartments, which may thus constitute different obstacles to the penetration of the antibody into such cell tissues.
In this study, the three different IHC procedures were mainly conducted by changing the periods for the protease (0.003%) digestion: namely, IHCs with no protease digestion (IHC-N), with mild protease digestion for 15-30 minutes (IHC-M), and with strong protease digestion for 2 hours (IHC-S), in the conditions of which the former one detects non-masked drug and the latter two detect masked drug. 38,39 Importantly, unmasking of tissue cell specimens by the protease digestion are commonly essential, especially in the IHC when GA or Karnovsky fixative (1965) 40 etc is used as a fixative for covalently fixing in situ the mobile small molecular compounds such as the endogenous amines, polyamines, 41 histamine, 11 serotonin, 42 dopamine, 43 GABA, 44 glutamic acid, 45 and drugs, 46 since these molecules are differently fixed in situ within cells via their amino group which reacts with the one of the two aldehyde groups of the bifunctional cross-linker GA, forming Schiff base (C = N). The other aldehyde group at the other end of the GA may be also bound in the same way with an amino group of some protein(s) etc, present in situ. Then, the two Schiff bases are reduced with NaBH 4 , giving the chemically stable C-N bonds, as described above. Antibody (Ab) a high molecular immunoglobulin (IgG) cannot come through the membrane into the cytoplasm, nuclei, mitochondria, and lysosomes etc within cells in tissues, unless the cell specimens are pre-treated with protease, 38 with the exception of an antigen drug on or near the cell surface. In contrast to this, we have previously observed that in drug IHC-N significant immunostaining occasionally occurred over the whole cell and the cells got abnormally large or swollen. 36 Perhaps the cells were affected or broken down by the excessive uptake of the drug.
Accordingly, Ab freely comes through within the cells and tissues. In addition, it is generally known that in IHC the stronger the fixation is, the more strongly the protease excavation for unmasking antigen is 10,11,16 . In these cases, however, the cell morphology gets poorer due to the disruption of the cell structures, even to the point of cases where the superficial antigen drug nearly completely washed away.
Also, the smaller the Ab (particularly, the secondary Ab) is made, for example, F(ab')2, Fab', or Fab etc, the better it enters cells. 47 The present IHCs for OS or for OC had characteristic advantages of using the tissue sections from the same paraffin block, and therefore would give this comparative study for drugs' tissue localization more accurate data, because it is free from experimental errors which may occur from the use of different animals and the effects of differing skills in the transcardial fixation procedures.
Control experiments: In the IHC for OS or OC in the small intestine, liver and kidney of rats to which OS was administered, the specificity of immunostaining was ascertained by incubating sections with the secondary antiserum alone (the secondary level controls), and AOS-96 mAb and AOC-160 mAb preabsorbed with OS-GA-BSA (100 µg/mL) and OC-GA-BSA conjugate (100 µg/mL), respectively.

| Immunoelectron microscopy (IEM)
When immune-positive reactions occur in the light microscopic studies of IHC-N, it is the sign of a possibility of the development of IEM due to preservation of the intact ultrastructure, being free from protease digestion. 12,13,46 The post-fixed specimens of the liver were cut into 50-μm-thick sections with a Microslicer (Dosaka EM; Kyoto, Japan), which were then applied to a free-floating procedure of the pre-embedding method, 12,13,46 in principally the same manner as was The regions to be studied were cut with a 2-mm diameter punch, mounted to Epon blocks, and sectioned into ultrathin sections (LKB 8800 Ultrotome ® III), which were then immediately observed in a 100CX Jeol electron microscope. Binding ELISA for the AOC mAb: Immunoreaction occurred only with OC ranging from 10 µmol/L to 1 mmol/L in dose-dependent manner in the tested compounds, sharply in contrast to that for the AOS ( Figure 1C and Cb).  Figure 2C) as well as in the wondering F I G U R E 2 (A-I) Immunostaining for OS or OC in small intestine of OS-administered rats. Rats were orally administered OS at 20 mg/kg, and then sacrificed 3 hours later. Five µm paraffin sections from small intestine block from OS-administered rats were prepared. Then, the section specimens were digested with protease at 0.003% at 30°C either for 30 minutes (D, I

| Drug uptake in liver
In the liver at 1-h post administration, in the lower magnification,  Figures 3E, F and 4E and F). Also, the parts seem to get morphologically somewhat wounded, possibly due to the protease digestion in the IHC (Figures 3E and 4E). Thus, cell morphology of the hepatocytes around the Glisson's capsules was compared with ones around the central veins by the HE staining, OS-or OC staining using the protease-treated or nontreated specimens of the liver from OS-administered rats or controls (Figures 3F and 4F; Figure S1A-F).
The HE tests showed the abundant cytoplasmic colorless gaps existed in the hepatocytes around the Glisson's capsules, but not the central vein ( Figure S1D-F). It is therefore suggested that the specific gaps of the cytoplasm of the hepatocytes is where the immunoreaction occurred by the IHCs-N for OS or for OC. Furthermore, in

| Immunoelectron microscopy (IEM) for OS or OC in liver
The IEM was developed essentially by the same procedure as the IHC but using 50 µm Microslicer sections of the liver, which, however, had not been pretreated with protease. In the liver specimens at 1-h post administration, the IEM for OS showed that the immuno-

| Drug uptake in kidney
In the kidney at 3-h post administration, the IHC-N for OS or OC produced no staining in any cell types of the nephrons except extraordinarily huge and swollen cells, which were heavily stained as the whole cell, mainly in the collecting ducts especially around the entrance to the inner medulla (data not shown). This staining pattern was quite similar to the results obtained by IHC-S for OS as well as OC, as described below (Figures 6D and E and 7D and E). The IHC-S for OS produced no or only very faint staining on the microvilli and occasionally in the nuclei, and moderate staining on abundant CSGs in the S1 and S2 segments of the proximal tubule cells in all which the cytoplasm showed very slight staining ( Figure 6A and B). Then, moderate staining occurred in the cytoplasm of the The IHC-S for OC clearly showed no CSG in any cell types of the kidney ( Figure 7A-D). No or only a very slight staining for OC distributed in some cells of the glomeruli and the capillaries ( Figure 7B). On the other hand, surprisingly, in striking contrast to the OS staining ( Figure 6A and B), very strong staining for OC occurred in the microvilli, cytoplasm, and nuclei of the S1 and S2 segment cells of the proximal tubules ( Figure 7A-C). Following weak or moderate staining in the S3 segments, almost no staining suddenly appeared in the thin Henle's loops followed by the distal tubules and collecting ducts. In the similar distribution patterns to those for OS staining ( Figure 6D and E), strongly stained huge or swollen cells were found mainly in the collecting ducts and occasionally in the distal tubules, especially near the entrance to the inner medulla ( Figure 7D). Absorption controls and secondary label control showed no reaction in any case ( Figure 7E and F). F I G U R E 6 (A-G) Immunostaining for OS in the kidney of OS-administered rats. Rats were orally administered OS at 20 mg/kg and then sacrificed 3 hours later. OS IHC was carried out with protease digestion of sections at 0.003% at 30°C for 2 hours (IHC-S) prior to immunoreaction with AOS-96 mAb. This severe protease digestion was inevitably needed for the immunostaining, especially in the S1 and S2 segment cells of the proximal tubules. In the process, however, their cell morphology seemed to be extensively damaged, resulting in their microvilli being almost completely disappearing (B). Lower (A) and higher magnification (B-G): (a) A wide variety of immunostaining was observed in the nephrons composed of the segments or parts, as clearly classified by their specific colors. (B, C) Immunoreactivity was very faint in the microvilli (blue arrowheads), cytoplasm, nuclei of the S1 and S2 segment cells (black arrows), in which, however, the CSGs are still recognized, and moderate staining in the S3 segment cells (red arrowheads) with their damaged villi, which are only very slightly stained (blue arrowheads). The S3 segment was identifiable by the presence of periodic acid-Schiff-positive brush borders. Very weakly stained blood capillaries (green arrows) were also observed. Strong immunostaining was exclusively concentrated in the numerous CSGs at the apical sites or all around the nuclei in the cytoplasm of the straight and convoluted distal tubules (red arrows), followed by the collecting ducts (black arrowheads), although the cytoplasm of all the three cell types seem to show only very faint immunostaining. Almost no staining in the glomeruli (G). (D, E) In the area close to the entrance to the inner medulla, many huge or swollen cells, which are strongly stained as a whole cell, are found in the collecting duct cells. The round beadslike structure of such uncommon cells was also observed in the collecting ducts  I G U R E 7 (A-F) Immunostaining for OC in the kidney of OS-administered rats. This was carried out in the same manner as that for IHC for OS except using AOC-160 mAb instead of AOS-96 mAb, with protease digestion process for 2h (IHC-S). (A, B) In striking contrast to the faint immunostaining for OS in the S1 or S2 segment of the proximal tubule cells ( Figure 6A, B), the very strong staining for OC was observed all over the cells in their segments (black arrows), followed by the less strong staining in the subsequent S3 segment cells of the proximal tubules (red arrowheads). Almost no staining in the glomeruli (G). (C) Immunostaining ranged from very weak to the moderate on the microvilli (blue arrowheads), in the cytoplasm and nuclei of the S3 segment cells (red arrowheads). Almost no or very slight staining in the blood capillary (green arrows). (A, D) Immunostaining almost disappeared suddenly in the renal tubule cells starting from the thin limb of Henle's loop via the straight and convolution distal tubules to the collecting ducts. However, in a manner similar to the findings in IHC-S for OS ( Figure 6A, D), huge or swollen cells, which are strongly stained as a whole cell, or cells in which only the nuclei were strongly stained, scattered in the collecting ducts (black arrowheads). Such uncommon cells were numerous, occurring in the collecting ducts especially around the entrance to the inner medulla. in general, believed that prodrug is absorbed in the former mode, and P-gp is highly expressed on the microvilli of the epithelia. 19,20 In addition, it remains unclear if the OS absorption (influx) is correlated with PEPT1 expressed on the microvilli of the intestinal epithelia, because it is under debate if OS is or not the substrate for PEPT1. 30,31 The high OC levels in the epithelia might imply that the OS uptaken from the intestinal lumina is cleaved to OC within the cells, possibly by the enzyme carboxylesterase 1 (CES1) of the main enzyme involved in intestinal first-pass hydrolysis, 51 and persists at high OC levels in the epithelia. This seems more probable than that the high OC levels originate from the circulation bearing OC, although it has been reported that in rats, OS is hydrolyzed by CES1 predominantly in the plasm, but scarcely in the intestine. 52 Also, the differences in OS levels in the epithelia and in the lamina propria scattering the free cells seemed smaller than those in OC levels (compare Figure 2A and B to F and G, respectively). This suggests that OS more easily penetrates the connective tissues through the basement membrane out of the epithelia, flowing into the blood capillary, during the course where OS more significantly influenced the free cells (lymphocytes, macrophages, plasma cells etc) in the lamina propria.
Liver: The IHCs-N showed OS distributed on the surface of the lumina ranging from the bile capillaries to the interlobular bile ducts ( Figure 3B and C), but OC did not ( Figure 4B and C). The IHCs-S showed OS localized in the lines along the bile capillaries leading to the interlobular bile ducts ( Figure 3E-G), but OC did not do so ( Figure 4E). These findings show that OS is excreted via the bile capillary to the bile flows, but OC is virtually not. The bile capillaries' membrane of the hepatocytes are specific sites, where P-gp is highly expressed 22,53,54 strongly suggesting that OS is actually and actively excreted at these sites, possibly through P-gp, since OS is among substrates for P-gp. 21 Previously, we have observed the same bile flow pattern in the IHC for doxorubicin, 32 the amphiphilic cationic drug known as the substrate for P-gp. 55 Figure 4E). Perhaps OC predominately originates from circulating OC in the liver artery branches winding the interlobular bile ducts. 58 The IHC-M for OS showed a few large cells with high OS levels, lying in the sinusoids ( Figure 3D). Thus, the cells might be the Kupffer cells, which show active phagocytosis, swallowing large amounts of OS and growing large in size, which was in contrast to that of the IHC-M for OC, which showed ordinary small Kupffer cells with lower OC levels ( Figure 4D). The results of the IEMs for OS and OC agree well with those of the respective light microscopic IHCs-N for OS and OC. The IEM showed that immunoreaction for OS or OC occurred in the specific wide fields in the cytoplasm (Figure 5Aa and Ba), which is responsible for the colorless gaps under the light microscopy ( Figures 3F and 4F), holding the lower electron density (Figure 5Ad and Bd) and placing the other cell organelles near the cell membranes (Figure 5Aa and Ba). This may be where the immunoreaction by IHC-N for OS or OC occurred (Figures 3A-C and 4A-C). It might, therefore, be reasoned that the drugs in these parts did not undergo any masking effect, probably due to the fact that they were free from the other cell organelles, and were easily washed away in the protease digestion process of the IHCs-S for both drugs. Also, the space of the specific gaps seemed rather inflexible, since no significant morphological changes occurred in time-different specimens of the livers from OS-administered rats or controls (Figure S1A-F).
Kidney: At the 3-h post-administration, OS distributed only slightly in the cytoplasm of the S1 and S2 proximal tubules, except for the S3 segment with some OS levels ( Figure 6A-C). This continued in the cytoplasm of the consecutive renal tubules from the thin limb of Henle's loop to the collecting ducts, while OS were highly concentrated only in the CSGs in those cells ( Figure 6B-E). Taking into account that P-gp are expressed on the microvilli of the proximal tubules and also the Mdr-1b mRNA in almost entire nephrons, 59 it might be assumed that OS is sooner excreted by the efflux P-gp than it takes time to OS is metabolized in the CSGs. 60 Meanwhile, it remains unclear if PEPT1 is involved in re-absorption of OS. 61 The best way to elucidate the correlation of transporters with drug disposition is to histochemically demonstrate their co-localization, but it is impossible, because different fixatives for diffusible drug and protein are inevitably needed for each other. The CSG might appear responsible for the lysosome in the tubules, which have been demonstrated by gentamicin autoradiography, 62 since in our IHCs for gentamicin, 36 vancomycin 46 and OS all the same types of granules as those of the autoradiography were observed. This seems quite reasonable since the lysosome is the subcellular organelle active for the phagocytosis.
Moreover the CSGs were observed in any cell types in the different tissues. This was first demonstrated by the present IHC in which the surfactant saponin was contained in the secondary antibody solution, as well, expecting the better immunoreaction with drug in situ. 63 All data together with the chemical structure of OS suggest that OS, but not OC, is among the "Lysosomotropic" cationic amphiphilic drugs. 64 OC highly distributed in the S1 and S2 proximal tubules, and then immediately disappeared in entire cells from the thin limb Additionally, IHCs-N as well as IHCs-S for OS or OC showed that both the drugs brought about all the similar phenomena in the kidney that huge or swollen cells with the extraordinary high levels of each of the drugs localized mainly in the collecting ducts, especially around the entrance to the inner medulla ( Figures 6D and E and 7D), suggesting that most of these cells might be damaged. Also, the number of such unusual cells for OS or OC seemed comparable to each other ( Figure 6A and 7A). This might mean that both drugs almost equally possess the renal toxicity in rats. The collecting ducts might be more fragile for the drugs than any other renal tubules since we have previously observed the similar phenomena in the IHCs for gentamicin, 36 amoxicillin, 33 peplomycin, 37 and vancomycin, 46 and others. 68,69 This might possibly happen because ca. 5-times larger dosage than that for the clinical trials per a day was administered rats in this study. 3

| DATA REP OS ITORY LINK
The authors declare that at the time of this study was published, data repository link was not available for the authors at Sojo University or Jikei University.

E TH I C S S TATEM ENT
Ethics approval for this study was obtained from the Sojo University Animal Experiment Ethics Committee.

ACK N OWLED G EM ENTS
We are grateful to Y Yoshizaki to help prepare the monoclonal antibodies; and M Shin, Y Yamamoto and K Tanaka for their technical assistance.

D I SCLOS U R E
The authors have no conflict of interest to declare.

AUTH O R CO NTR I B UTI O N S
Fujiwara participated in research design. Fujiwara, Yamamoto, and Saita conducted experiments. Fujiwara, Yamamoto, Saita, and Matsufuji performed data analysis. Fujiwara wrote the manuscript.