Cerebroside Sulfotransferase Deficiency Ameliorates L-selectin-dependent Monocyte Infiltration in the Kidney after Ureteral Obstruction*

Mononuclear cells infiltrating the interstitium are involved in renal tubulointerstitial injury. The unilateral ureteral obstruction (UUO) is an established experimental model of renal interstitial inflammation. In our previous study, we postulated that L-selectin on monocytes is involved in their infiltration into the interstitium by UUO and that a sulfated glycolipid, sulfatide, is the physiological L-selectin ligand in the kidney. Here we tested the above hypothesis using sulfatide- and L-selectin-deficient mice. Sulfatide-deficient mice were generated by gene targeting of the cerebroside sulfotransferase (Cst) gene. Although the L-selectin-IgG chimera protein specifically bound to sulfatide fraction in acidic lipids from wild-type kidney, it did not show such binding in fractions of Cst-/- mice kidney, indicating that sulfatide is the major L-selectin-binding glycolipid in the kidney. The distribution of L-selectin ligand in wild-type mice changed after UUO; sulfatide was relocated from the distal tubules to the peritubular capillaries where monocytes infiltrate, suggesting that sulfatide relocated to the endothelium after UUO interacted with L-selectin on monocytes. In contrast, L-selectin ligand was not detected in Cst-/- mice irrespective of UUO treatment. Compared with wild-type mice, Cst-/- mice showed a considerable reduction in the number of monocytes/macrophages that infiltrated the interstitium after UUO. The number of monocytes/macrophages was also reduced to a similar extent in L-selectin-/- mice. Our results suggest that sulfatide is a major L-selectin-binding molecule in the kidney and that the interaction between L-selectin and sulfatide plays a critical role in monocyte infiltration into the kidney interstitium.


Mononuclear cells infiltrating the interstitium are involved in renal tubulointerstitial injury. The unilateral ureteral obstruction (UUO) is an established experimental model of renal interstitial inflammation. In our previous study, we postulated that L-selectin on monocytes is involved in their infiltration into the interstitium by
UUO and that a sulfated glycolipid, sulfatide, is the physiological L-selectin ligand in the kidney. Here we tested the above hypothesis using sulfatide-and L-selectin-deficient mice. Sulfatide-deficient mice were generated by gene targeting of the cerebroside sulfotransferase (Cst) gene. Although the L-selectin-IgG chimera protein specifically bound to sulfatide fraction in acidic lipids from wild-type kidney, it did not show such binding in fractions of Cst ؊/؊ mice kidney, indicating that sulfatide is the major L-selectin-binding glycolipid in the kidney. The distribution of L-selectin ligand in wildtype mice changed after UUO; sulfatide was relocated from the distal tubules to the peritubular capillaries where monocytes infiltrate, suggesting that sulfatide relocated to the endothelium after UUO interacted with L-selectin on monocytes. In contrast, L-selectin ligand was not detected in Cst ؊/؊ mice irrespective of UUO treatment. Compared with wild-type mice, Cst ؊/؊ mice showed a considerable reduction in the number of monocytes/macrophages that infiltrated the interstitium after UUO. The number of monocytes/macrophages was also reduced to a similar extent in L-selectin ؊/؊ mice. Our results suggest that sulfatide is a major Lselectin-binding molecule in the kidney and that the interaction between L-selectin and sulfatide plays a critical role in monocyte infiltration into the kidney interstitium.
Blood monocytes that extravasate to sites of inflammation differentiate into macrophages and induce inflammatory response and apoptosis of tubular epithelial cells leading to tubulointerstitial injury in renal inflammation (1,2). In the experimental model of kidney interstitial inflammation (3,4), unilateral ureteral obstruction (UUO) 1 leads to interstitial inflammation, interstitial fibrosis, and tubular atrophy (5). Although infiltration of monocytes/macrophages into the interstitium following UUO is well documented, the mechanism involved in this process remains elusive.
Monocyte infiltration is a multistep process in which chemokines and adhesion molecules play key roles. The initial step of this process involves the binding of monocytes to the endothelium of venules, mediated by various adhesion molecules such as selectins (6). The selectin family consists of three members, designated P-, E-, and L-selectin. The P-and E-selectins are present on activated endothelial cells, whereas L-selectin is expressed on leukocytes. Results of our previous study in rats with UUO (7) suggested that L-selectin on monocytes mediates their infiltration into the interstitium, based on the observation that a neutralizing antibody against L-selectin inhibited monocyte infiltration.
L-selectin was originally identified as a homing receptor to capture lymphocytes in the flowing bloodstream in lymph nodes (8 -10). However, L-selectin ligands are expressed not only in lymphoid organs but also in other tissues (11)(12)(13). The main L-selectin-binding molecules are sialylated, fucosylated, and sulfated glycans on mucin-like molecules such as Gly-CAM-I and CD34, which are located on lymph node high endothelial venules and are involved in the homing (14). L-selectin also binds to a sulfated glycolipid, sulfatide (15)(16)(17), and chondroitin sulfate and heparin/heparan sulfate proteoglycans (18 -20). We have recently shown that collagen XVIII interacts with L-selectin (21). Furthermore, chemically synthesized 3Ј-sulfo Lewis a and 3Ј-sulfo Lewis x oligosaccharides were found to be potent L-selectin ligands as are sialyl Lewis a /Lewis x determinants or sulfatide (22). Results of our previous study (7) suggested that sulfatide is an L-selectin ligand in the rat kidney and contributes to the interstitial monocyte infiltration following UUO. These conclusions were based on the histochemical distribution of sulfatide, demonstrated by an anti-sulfatide monoclonal antibody that was very close to that of L-selectin ligand detected by an L-selectin-IgG chimera protein before and after ureteric ligation and that exogenously added anti-Lselectin monoclonal antibody or sulfatide considerably inhibited monocyte infiltration (7).
Sulfoglycolipids are comprised of acidic glycolipids containing sulfate esters on their oligosaccharides. They have been implicated in a variety of physiological functions through their interactions with extracellular matrix proteins, cellular adhesion receptors, blood coagulation systems, complement activation systems, cation transporter systems, and microorganisms (23). The distribution of sulfoglycolipids is tissue-specific and they are abundant in myelin sheaths, spermatozoa, renal tubular cells, and epithelial cells of the gastrointestinal tract (23). Sulfation of glycolipids is catalyzed by cerebroside sulfotransferase (CST, EC 2.8.2.11), which is located in the Golgi apparatus (24). Recently, Cst-null mice were generated by gene targeting to elucidate the physiological functions of sulfoglycolipids (25). The Cst Ϫ/Ϫ mice completely lack sulfatide in the brain and seminolipid in the testis and manifest some neurological disorders because of myelin dysfunction as well as spermatogenesis arrest (25,26). These observations emphasize the critical roles of sulfoglycolipids in the intercellular interactions during myelin formation and spermatogenesis. Because sulfatide is rich in renal tubular cells, its involvement in renal function such as ion transport has been suggested (23). However, the Cst-null mice do not exhibit signs of renal failure or apparent histological abnormality in the kidney (25). A detailed analysis with some stress would be desirable to elucidate the biological function of sulfatide in the kidney.
Here we tested the hypothesis that interaction between sulfatide and L-selectin mediates infiltration of monocytes into the kidney interstitium, using sulfatide-and L-selectin-deficient mice. The present results established the critical role of their interaction in monocyte infiltration that occurs in the inflamed kidney.

EXPERIMENTAL PROCEDURES
Animals-Cst Ϫ/Ϫ mice and L-selectin Ϫ/Ϫ mice were generated as described previously (25,27). All experimental protocols described in the present study were approved by the Ethics Review Committee for Animal Experimentation of Okayama University Graduate School of Medicine and Dentistry.
Unilateral Ureteral Obstruction-Cst Ϫ/Ϫ mice (n ϭ 5, female) were subjected to complete UUO as described previously (3). Briefly, the right ureter was ligated with a silk ligature under anesthesia at the junction of the upper third and lower two-thirds. Wild-type mice (n ϭ 5, female) from the same litter were used as control. At day 2, the mice were sacrificed, and the obstructed and non-obstructed kidneys were removed. Portions of the tissue were processed for TLC blotting and immunohistochemistry. L-selectin Ϫ/Ϫ mice (n ϭ 5, female) and wildtype mice (n ϭ 5, female) were operated on in the same way as Cst Ϫ/Ϫ mice and were sacrificed and the kidneys dissected out for immunohistochemistry.
Histopathological Examination-Each kidney specimen was divided into two parts. One portion was fixed in 10% buffered formalin and embedded in paraffin. Four-micrometer-thick sections were stained with periodic acid Schiff for morphological examination. The other portion of each specimen was rapidly frozen in liquid nitrogen and cut with a cryostat. Four-micrometer-thick acetone-fixed frozen sections were used for immunohistological evaluation. Ligands for L-selectin and sulfatide were detected by the indirect immunofluorescence method as described previously (7). Briefly, the frozen sections were fixed with cold acetone for 3 min and stained with LEC-IgG or anti-sulfatide monoclonal antibody (GS5) (7,30) for 24 h at 4°C. Recombinant Ig chimera or mouse IgM was used as a control for nonspecific staining (7,11). Then, the sections were stained with fluorescein isothiocyanate (FITC)-labeled goat anti-human IgG antibody (ICN Pharmaceuticals, Inc., Aurora, OH) or FITC-labeled goat anti-mouse IgM antibody (ICN Pharmaceuticals, Inc.) for 30 min at room temperature. The sections were washed in phosphate-buffered saline, mounted with PermaFluor (Shandon, Pittsburgh, PA), and examined under a fluorescence microscope (LSM-510; Carl Zeiss, Jena, Germany).
Monocyte/macrophage infiltration into the renal interstitium was estimated by immunostaining, using a specific rat monoclonal antibody against mouse monocyte/macrophage (F4/80; Serotec, Oxford, UK). In brief, the frozen sections were fixed with cold acetone for 3 min, and nonspecific protein binding was blocked by incubation with normal goat serum and avidin for 20 min. The sections were first incubated with F4/80 for 60 min at room temperature. Rat IgG was used as a control for nonspecific staining. Then the sections were incubated with biotinlabeled goat anti-rat IgG antibody (Jackson Immunoresearch Laboratories, West Grove, PA) for 30 min at room temperature. Endogenous peroxidase activity was blocked by incubating the sections in methanol containing 0.3% H 2 O 2 for 30 min. After that, the sections were stained with a Vectastain ABC kit (Vector Laboratories, Burlingame, CA). The sections were then counterstained with Mayer's hematoxylin.
Cell Counts-The nuclei of F4/80-positive cells were counted by examining 10 randomly selected areas near the borderline of the cortex and medulla under high magnification (ϫ400). Cell number per mm 2 was expressed according to the method described by Saito and Atkins (31).
Statistical Analysis-All values are expressed as mean Ϯ S.E. Differences between groups were examined for statistical significance using one-way analysis of variance followed by Scheffe's test. A p value less than 0.05 denoted the presence of a statistically significant difference.

RESULTS
Sulfatide Is Absent in Cst-null Mice-Because the profiles of the kidney glycolipids were noticeably different between normal male and female mice (Fig. 1, a and c) as described before (28), total glycolipids from male and female kidneys of Cst-null mice were separately investigated. As shown in Fig. 1, spots corresponding to authentic sulfatide SM4s were absent in both cases. Instead, spots corresponding to GalCer, which is the precursor of sulfatide in terms of CST, increased. These results indicate that CST is responsible for the biosynthesis of sulfatide in the kidney.
Sulfatide Is a Ligand for L-selectin in Mouse Kidney-To investigate biochemically the ligands for L-selectin in mouse kidney, acidic lipids were extracted from the kidney and their reactivity to L-selectin was examined by TLC blotting analyses. As shown in Fig. 2B, the LEC-IgG bound to a material comigrating with sulfatide in the acidic lipid fraction from the wild-type kidney as well as the authentic bovine brain sulfatide, whereas it did not react to ganglioside GM3, consistent with the results of Suzuki et al. (15). In Cst Ϫ/Ϫ kidney, sulfatide disappeared from the acidic lipid fraction ( Fig. 2A), consistent with the results in Fig. 1, and the reactivity to LEC-IgG was also completely abolished (Fig. 2B). These results clearly indicate that the L-selectin ligand in lipid fraction of mouse kidney is sulfatide.
Interstitial Changes in UUO-No abnormalities were detected in Cst Ϫ/Ϫ or wild-type mice before UUO (Fig. 3, A and B). UUO resulted in tubular dilation and atrophic changes in the interstitium of the obstructed kidney (Fig. 3, C and D). In contrast, the non-obstructed kidney showed no signs of tubular dilation or atrophy (Fig. 3, E and F). The non-obstructed kidneys of L-selectin Ϫ/Ϫ and wild-type mice were also similar to the normal control kidneys (data not shown).
Distribution of L-selectin Ligands and Sulfatide in Mouse Kidney-LEC-IgG reacted exclusively with the distal tubules where sulfatide is expressed in wild-type mouse kidney (Figs. 4A and 5A). This reactivity was almost absent in the kidney of Cst Ϫ/Ϫ mice (Figs. 4B and 5B), supporting the results of bio-chemical experiments showing that the major L-selectin ligand in the kidney is sulfatide. Following ureteric ligation, the reactivity of LEC-IgG in the distal tubules disappeared and, instead, a new reactivity emerged in the interstitium and peritubular capillaries (Figs. 4C and 5C). This reactivity of LEC-IgG after ureteral obstruction was also almost abolished in the interstitium of Cst Ϫ/Ϫ mice (Figs. 4D and 5D). Distribution of the reactivity with LEC-IgG showed a similar pattern with that of anti-sulfatide antibody before and after ureteral obstruction. No staining was observed in the kidney when irrelevant immunoglobulin was used as a negative control.

FIG. 1. Two-dimensional thin-layer chromatograms of total lipid extract from the kidneys of male (a) and female (c) wildtype mice and male (b) and female (d)
Cst ؊/؊ mice. Total lipid extract corresponding to 2.5 mg kidney was separated on a TLC plate with chloroform/methanol/water containing 0.2% CaCl 2 (65:35:8 by volume) for the first direction (I) and chloroform/methanol/acetone/acetic acid/water (7:2:4:2:1 by volume) for the second direction (II). Glycolipids were visualized with the orcinol-sulfuric acid reagent. Asterisks (*) indicate unidentified constituents that moved close to SM4s but appeared grayish with the orcinol reagent. A mixture of acidic lipids of rat brain and standard SM4s was applied as a reference in the upper and left sides, respectively, in each plate.

FIG. 2.
Binding of L-selectin to sulfatide in mouse kidney on TLC blotting. Standard sulfatide from bovine brain and gangalioside GM3 from human kidney (Std) as well as acidic lipid fractions corresponding to 5 mg protein extracted from Cst Ϫ/Ϫ (Ϫ/Ϫ) and wild-type (ϩ/ϩ) kidneys were developed on a TLC plate, transferred to a polyvinylidene difluoride membrane, and stained with orcinol-sulfuric acid (A) or reacted with LEC-IgG (B) as described under "Experimental Procedures."

FIG. 3. Morphologic alteration of the interstitium induced by UUO. No abnormality was detected before UUO in wild-type (A) and
Cst Ϫ/Ϫ mice (B). Tubular dilation and atrophy were observed at 2 days after UUO in wild-type (C) and Cst Ϫ/Ϫ mice (D). However, no abnormalities were detected in the non-obstructed kidneys of the wild-type (E) and Cst Ϫ/Ϫ mice (F).

FIG. 4. Immunohistochemical analysis of the kidneys of wildtype and Cst ؊/؊ mice with LEC-IgG. Kidneys of wild-type (A, C) and
Cst Ϫ/Ϫ (B, D) mice before (A, B) and after (C, D) UUO treatment were subjected to immunohistochemical analysis with LEC-IgG as described under "Experimental Procedures." A, LEC-IgG reacted exclusively with the distal tubular epithelial cells in the kidney of wild-type mice. Glomeruli, peritubular capillaries, and arteries were negative for such reaction. C, the reactivity with LEC-IgG shifted from the distal tubules to the interstitium and peritubular capillaries (arrowheads) after UUO. Inset, a high magnification of the dashed line area. B and D, the reactivity with LEC-IgG was almost abolished in Cst Ϫ/Ϫ kidney irrespective of UUO treatment.

Monocyte/Macrophage Infiltration in the Renal Interstitium of Cst
Ϫ/Ϫ Mice-To elucidate whether sulfatide is responsible for the monocyte infiltration into the renal interstitium induced in our ureteral obstruction model, the extent of increased number of monocytes/macrophages in the interstitium after UUO was compared between Cst-null and wild-type mice. Because monocyte infiltration commences as early as 4 -12 h after UUO (3), it was evaluated at 48 h after UUO in the present study. Monocytes/macrophages were stained with a specific anti-monocyte/macrophage monoclonal antibody, F4/80 (Fig.  6). No staining was observed in the kidney when irrelevant immunoglobulin was used as a negative control. There was no difference in the number of monocytes/macrophages between Cst Ϫ/Ϫ mice and wild-type mice before UUO (Fig. 7). After UUO, there was a 4-fold increase in the number of these cells compared with pre-UUO in wild-type mice (Fig. 7). However, the increased number of monocytes/macrophages induced by UUO was suppressed to almost 50% of that before UUO in Cst Ϫ/Ϫ mice (Figs. 6 and 7), suggesting that sulfatide is highly responsible for the infiltration of monocytes into the renal interstitium.
Monocyte/Macrophage Infiltration in the Renal Interstitium of L-selectin Ϫ/Ϫ Mice-To determine the contribution of Lselectin to the monocyte infiltration into the interstitium, the number of monocytes/macrophages into the interstitium after UUO was compared between L-selectin Ϫ/Ϫ and wild-type mice. Although there was no difference in the number of monocytes/ macrophages between these mice before UUO, the increased number of these cells induced by UUO was significantly reduced in L-selectin-deficient mice (Fig. 8). The suppressed rate of the increment of infiltrated monocytes/macrophages induced by UUO was almost the same between Cst Ϫ/Ϫ and L-selectin Ϫ/Ϫ mice, suggesting that sulfatide on the endothelial cells of peritubular capillaries is the major ligand for L-selectin on the monocytes when they enter the interstitium. The residual elevation in the number of monocytes/macrophages after UUO in Cst Ϫ/Ϫ and L-selectin Ϫ/Ϫ mice indicates the contribution of other as yet unknown factors. DISCUSSION Our previous results indicated that interaction between sulfatide in the kidney and L-selectin on monocytes mediates Immunoperoxidase staining of kidneys of wild-type (ϩ/ϩ) and Cst Ϫ/Ϫ (Ϫ/Ϫ) mice with or without UUO treatment was performed as shown in Fig. 6. The nuclei of F4/80-positive cells were counted in 10 randomly selected areas near the borderline of the cortex and medulla under high magnification (ϫ400). All values are expressed as mean Ϯ S.E. *, p Ͻ 0.01, **, p Ͻ 0.05, by one-way analysis of variance followed by Scheffe's test.
FIG. 8. Quantitative analysis of UUO-induced infiltrating monocytes/macrophages in the renal interstitium of L-selectin ؊/؊ mice. Immunoperoxidase staining of the kidneys of wild-type (ϩ/ϩ) and L-selectin Ϫ/Ϫ (Ϫ/Ϫ) mice with or without UUO treatment was performed. Cell count and statistical analysis were carried out as described in Fig. 7. *, p Ͻ 0.01. monocyte infiltration into the kidney interstitium (7). This conclusion was based on the following results: blockade of Lselectin function by a neutralizing antibody protects against monocyte infiltration after UUO in rats (7); sulfatide binds specifically to L-selectin through the sulfated sugar chain (15); L-selectin ligand relocates from the distal tubules to the interstitium and peritubular capillaries by UUO as sulfatide does (7); and exogenously administrated sulfatide inhibits monocyte infiltration after UUO (7). The present study further strengthens the above results and shows that L-selectin ligand is absent in sulfatide-deficient mice and that monocyte infiltration after UUO is reduced in sulfatide-deficient mice to the same extent as in L-selectin-deficient mice. During the preparation of this report, Lange-Sperandio et al. (32) reported that inhibition of macrophage recruitment to the obstructed kidneys with Lselectin deficiency is associated with a reduction of apoptosis of tubular epithelial cells and interstitial fibrosis. This finding adds strong support to our conclusion.
Where does L-selectin on the monocytes contact with sulfatide in the kidney? Sulfatide is expressed in the distal tubules in normal kidney, where no leukocyte traffic is seen. In addition, leukocytes rarely enter the renal interstitium under normal conditions. It is speculated that sulfatide mediates rolling and/or migration after UUO. First, after UUO, sulfatide relocates from the distal tubules to the interstitium and peritubular capillaries, where monocytes are considered to extravasate into the interstitium. Thus, L-selectin and sulfatide seem to make contact on the peritubular capillary walls. Second, sulfatide, which relocates from the distal tubules to the interstitium, mediates extravasation and migration of monocytes into the kidney interstitium. Recently, it was demonstrated that the number of emigrated leukocytes and the distance of extravascular migration was significantly reduced in L-selectin-deficient mice (33). The same phenomenon might occur in sulfatide-deficient mice. Then why does the distribution of sulfatide change following UUO? There are also two possible mechanisms. One is that sulfatide is shed from injured tubular epithelial cells and moves into the interstitium and peritubular capillary walls. A similar event has been observed in Tamm Horsfall protein (4). The other is that sulfatide is newly synthesized in the interstitium and peritubular capillary walls following ureteric ligation.
In addition to obstructive nephropathy, interstitial infiltration of mononuclear cells is also observed in tubulo-interstitial nephritis, severe glomerulonephritis, and rejection of transplanted kidney. Selectins and their ligands have been shown to mediate monocyte infiltration in a variety of kidney inflammation. For instance, induced E-selectin on the peritubular capillary is implicated in monocyte infiltration in diabetic nephropathy (34). Induction of L-selectin ligands on peritubullar and venous endothelium is suggested to be involved in acute kidney allograft rejection (35). E-and P-selectin have been shown to mediate leukocyte infiltration during ischemia/reperfusion-induced acute renal failure (36,37). Thus, all members of the selectin family may contribute to monocyte infiltration in the kidney. In the present study, monocyte infiltration induced by UUO was considerably, but not completely, suppressed in Lselectin-deficient mice. The residual recruitment activity might be attributed to E-and P-selectins (38). However, there was no significant difference in the number of infiltrated monocytes/ macrophages induced by UUO between triple selectin-deficient mice and L-selectin-deficient mice (32). This finding strongly suggests that E-and P-selectins scarcely contribute to monocyte infiltration in obstructive nephropathy.
Sulfatide has been demonstrated to bind specifically to Lselectin and P-selectin (15)(16)(17)39). In fact, sulfatide appears to regulate inflammation and tumor metastasis mediated by Pselectin. Mulligan et al. (40) showed the preventive effect of sulfatide in P-selectin-dependent lung injury. On the other hand, Borsig et al. (41) reported that sulfatide on tumor cells serves as a P-selectin ligand and facilitates tumor metastasis by interaction with platelets leading to tumor microemboli. The present study provided in vivo evidence that sulfatide plays a critical role as an L-selectin-binding molecule during the process of inflammation.
In conclusion, we demonstrated that sulfatide is a major L-selectin-binding molecule in the kidney and mediates monocyte infiltration in obstructive nephropathy. Our results suggest that blockade of the interaction of sulfatide and L-selectin could be a potentially useful strategy for the treatment of interstitial nephritis.