Sulfatide with ceramide composed of phytosphingosine (t18:0) and 2-hydroxy FAs in renal intercalated cells

Diverse molecular species of sulfatide with differences in FA lengths, unsaturation degrees, and hydroxylation statuses are expressed in the kidneys. However, the physiological functions of specific sulfatide species in the kidneys are unclear. Here, we evaluated the distribution of specific sulfatide species in the kidneys and their physiological functions. Electron microscopic analysis of kidneys of Cst-deficient mice lacking sulfatide showed vacuolar accumulation in the cytoplasm of intercalated cells in the collecting duct, whereas the proximal and distal tubules were unchanged. Immunohistochemical analysis revealed that vacuolar H+-ATPase-positive vesicles were accumulated in intercalated cells in sulfatide-deficient kidneys. Seventeen sulfatide species were detected in the murine kidney by iMScope MALDI-MS analysis. The distribution of the specific sulfatide species was classified into four patterns. Although most sulfatide species were highly expressed in the outer medullary layer, two unique sulfatide species of m/z 896.6 (predicted ceramide structure: t18:0-C22:0h) and m/z 924.6 (predicted ceramide structure: t18:0-C24:0h) were dispersed along the collecting duct, implying expression in intercalated cells. In addition, the intercalated cell-enriched fraction was purified by fluorescence-activated cell sorting using the anti-vacuolar H+-ATPase subunit 6V0A4, which predominantly contained sulfatide species (m/z 896.6 and 924.6). The Degs2 and Fa2h genes, which are responsible for ceramide hydroxylation, were expressed in the purified intercalated cells. These results suggested that sulfatide molecular species with ceramide composed of phytosphingosine (t18:0) and 2-hydroxy FAs, which were characteristically expressed in intercalated cells, were involved in the excretion of NH3 and protons into the urine.

In the kidneys, sulfatide is highly expressed in the distal nephron segments and medulla (9). In addition to sulfatide, more complex sulfoglycolipids, such as lactosylceramide sulfate (SM3) and gangliotetraosylceramide-bis-sulfate (SB1a), have been identified in the kidneys (1). In contrast to sulfatide, SM3 is mainly present in the cortex and, to a lesser extent, in the medulla, whereas SB1a is distributed throughout all parts of the kidneys (10). Tissue-specific sulfoglycolipid structures are thought to result from the selective expression of specific glycosyltransferases that synthesize the neutral oligosaccharide backbone. The kidneys of Cgt-deficient mice lack sulfatide because of loss of the precursor GalCer; however, the levels of SM3 and SB1a, whose biosynthetic pathways circumvent the ceramide galactosylation reaction, are increased by 2-3-fold, which may partially compensate for sulfatide (11). By contrast, Cst-deficient mice do not express any sulfoglycolipids in the kidneys (12), and the mutant kidneys were initially thought to appear normal (8). However, Stettner et al. (13) subsequently performed intensive analyses on mice deficient in Cst and UDP-glucose:ceramide glucosyltransferase (Ugcg) in paired box gene 8 (Pax8)-expressing renal cells; they observed a lower urinary pH accompanied by lower ammonium (NH 4 + ) excretion in Cst-deficient kidneys. Using acid load experiments, they found that sulfatides may play important roles in renal ammonium processing, urine acidification, and acid-base homeostasis (13). However, the molecular mechanisms underlying the pathological phenotype have not yet been elucidated.
The kidneys play central roles in the regulation of blood pH maintenance. In response to metabolic acidosis, excess protons in the blood are buffered by bicarbonate synthesized during the renal glutaminolysis process, and ammonium ions are excreted into the urine (14). This physiologically relevant ammoniagenesis process, which occurs in the proximal tubule (14), is preserved in Cst-deficient kidneys (13). The majority of luminal ammonium ions secreted from the proximal tubule are reabsorbed at the thick ascending limb of Henle's loop and secreted again from the collecting duct, a process that involves parallel proton and NH 3 secretion (15). The epithelium of the collecting duct is composed mainly of two different types of cells: principal cells and intercalated cells (16)(17)(18)(19). The Rhesus glycoproteins, Rhbg and Rhcg, which act as ammonia transporters, are expressed on the plasma membranes of intercalated cells (15). NH 3 in the interstitium is transported across the basolateral membrane through both Rhbg and Rhcg. Most basolateral NH 4 + uptake is mediated by Na + -K + -ATPase, where NH 4 + is substituted for K + . Incorporated NH 4 + is dissociated into NH 3 , and protons and intracellular NH 3 are secreted into the lumen across the apical membrane through Rhcg. Protons secreted by H + -ATPase and H + -K + -ATPase combine with luminal NH 3 to form NH 4 + , neutralizing urinary pH (15). Stettner et al. (13) stated that "transepithelial NH 3 and NH 4 + inward as well as proton outward movements showed similar rates in intracellular alkalinization and pH recovery, suggesting unaltered luminal NH 3 entry and proton secretion by Ugcg/Cst-deficient outer medullary collecting duct (OMCD) intercalated cells," interpreting the result of the in vitro microperfusion experiment using OMCD cells. However, the initial alkalinization phase reflecting NH 3 entry and formation of intracellular NH 4 + is remarkably suppressed in Ugcg/Cst-deficient OMCD intercalated cells and the subsequent acidification phase reflecting NH 4 + entry, and dissociation of NH 3 and H + is preserved in Ugcg/Cst-deficient cells, as seen in the experiment of Rhcg-deficient mice (20), which was demonstrated using the same experimental procedure. Proton secretion is expected to follow the latter phase in which H + is formed. Therefore, we predicted that NH 3 excretion through Rhcg is deteriorated in sulfatide-lacking mice.
Recent advancements in IMS have revealed the characteristic distributions of specific molecular species of lipids among different tissues or cells (24). In fact, we have analyzed sulfated glycolipids in the glial developmental process (25), pharmacokinetics in an Alzheimer's mouse model (26), and phospholipid localization in the retinal layer in an optic nerve injury model (27) using iMScope MALDI-IMS analysis with high spatial resolution. Accordingly, in this study, we used an IMS approach to evaluate the morphological abnormalities in the kidneys of Cst-deficient mice and assess the involvement of specific molecular species of sulfatide in biological processes in the kidneys.

Animal care
All animal experiments were conducted in strict accordance with the institutional guidelines of the National Institutes of Health and the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health Animal Care. Maintenance and surgeries were performed in full compliance with the regulatory standards for the Animal Research Facilities of the Animal Ethics Committee of Kansai Medical University (Hirakata, Japan; approval ID: 20-086). C57BL6J and TgH (CST-neo) mice (8) were housed in plastic cages with standard bedding and continuous access to food and water. The temperature was maintained at 22 • C under standard light conditions with a 12 h light/dark cycle.

Tissue preparation
Kidneys of 10-week-old female C57BL6 and TgH (CST-neo) mice (8) were quickly dissected. Unfixed tissues were embedded in cold 2% carboxymethylcellulose on dry ice and sliced to a thickness of 10 μm using a cryostat (CM3050 S; Leica, Nussloch, Germany). The sections were mounted on glass slides coated with indium tin oxide (SI0100N) for IMS and microslide glass (CRE-02; Matsunami Glass, Kishiwada, Japan) for immunofluorescent labeling. Hoechst 33258 was purchased from Nacalai Tesque. 1% Triton X-100 for 12 h. After rinsing with PBS, the sections were incubated with Alexa Fluor 594-or Alexa Fluor 488-conjugated secondary antibodies (1:1,000 dilution) for 1 h, washed with 0.1 M PBS, and mounted with a medium containing 100 mM dithiothreitol, 5 μg/ml Hoechst 33258, 50% glycerol, and PBS (pH 7.4). Images were obtained using a confocal laser-scanning microscope (LSM700; Carl Zeiss, Jena, Germany). Pearson's correlation coefficient (r) for the quantitative analysis of colocalization was obtained by performing pixel intensity correlations over space using ImageJ (Fiji)'s Coloc 2 system (National Institutes of Health, Bethesda, MD).

IMS analysis with MALDI by iMScope
Mouse kidneys were embedded in 2% carboxymethylcellulose on dry ice and sliced to 10 μm thickness using a cryostat. The sections were mounted on glass slides coated with indium tin oxide (Matsunami Glass). Sliced samples were vapordeposited using 9AA as a negative mode matrix and CHCA as a positive mode matrix under dry vacuum conditions. Sublimation of 9AA and CHCA was performed at 220 • C for 9 min and 250 • C for 15 min, respectively, using SVC-700TMSG/7PS80 vacuum vapor deposition equipment (Sanyu Electron, Tokyo, Japan).
IMS was performed using an iMScope MALDI-mass spectrometer (Shimadzu, Kyoto, Japan) in the negative and positive reflection mode. To acquire mass spectra, the samples were scanned using a 1,000 Hz Nd:YAG laser with 50 accumulating shots. The detector voltage and sample voltage were 1.66 and 3 kV, respectively. We collected spectra over the m/z range from 600 to 1,000 at a scan pitch of 10-25 μm, and the intensity of the iMScope laser was set to 46. The data were analyzed using the IMS Solution software package (Shimadzu). After IMS analysis, the section was washed by PBS to remove the deposited matrix and stained with hematoxylin and eosin according to the general method.

Electron microscopy
Mice were deeply anesthetized by intraperitoneal injection of a mixture of anesthetics (0.15 mg/kg medetomidine hydrochloride, 2 mg/kg midazolam, and 2.5 mg/kg butorphanol tartrate) and perfused with 0.1 M PB followed by 2% formaldehyde and 2% glutaraldehyde in 0.1 M PB. Kidneys were sectioned at 2 mm thickness, immersed in the same fixative for 6 h at 4 • C, and postfixed with 2% osmium tetroxide in 0.1 M PB at 4 • C. Tissues were then embedded in epoxy resin, and sections of 80 nm thickness were prepared. The sections were mounted on silicon wafers, stained with 1% uranyl acetate for 15 min, and then stained with Sato's lead staining solution for 5 min (28). The cells were imaged using scanning electron microscopy (JSM-7800F; JEOL Ltd, Tokyo, Japan). Backscattered electrons were detected, and black and white were reversed.

Fluorescence-activated cell sorting
Eight-week-old C57BL/6 mice were perfused with PBS for 5 min, and blood was removed completely. The kidneys were dissected out, and the capsule was removed and cut into small pieces in HBSS. The tissues were then treated with 0.5% collagenase and DNase I by shaking at 37 • C for 2 h. The suspension was centrifuged for 5 min at 1,500 rpm at 4 • C. After removal of the supernatant, the remaining tissue was washed twice in HBSS and suspended in 2 ml HBSS. The tissues were then dissociated by trituration. The resulting cell suspension was filtered through a 70 μm nylon cell strainer. To isolate the intercalated cells, fluorescence-activated cell sorting (FACS) was performed using a FACS Aria instrument (BD Biosciences, Franklin Lakes, NJ) with antibodies targeting ATP6V0A4, a marker of intercalated cells. After treatment with mouse kidney-derived collagen, the cells were collected and labeled with anti-ATP6V0A4 antibodies (1:200 dilution) on ice for 15 min. The cells were then collected by centrifugation for 5 min at 15,000 rpm and 4 • C, washed with PBS, and colabeled with Alexa Fluor 488-labeled anti-rabbit IgG on ice for 15 min. The cells were sorted by forward scatter and side scatter. After centrifugation for 5 min at 1,500 rpm at 4 • C, methanol was added, and the cells were mechanically disrupted using a BioMasher (Nippi Incorporated, Tokyo, Japan). The lysed solution was dropped onto a metal plate and analyzed using IMS.

RT-qPCR
Total RNA was extracted from FACS-purified vacuolar H + -ATPase (V-ATPase) 6V0A4-positive cells using an RNeasy Mini Kit. Complementary DNA was reverse-transcribed using a ReverTra Ace qPCR RT master mix with genomic DNA remover. qPCR was then performed using a Rotor-Gene platform and Thunderbird qPCR Mix. Primer sequences corresponding to each gene used for PCR are listed in Table 1. A thermal cycler, with the following settings: one cycle at 95 • C for 1 min, followed by 40 cycles of denaturation at 95 • C for 10 s and extension at 60 • C for 60 s. The 2 −ΔΔCt method was used to calculate relative target gene expression levels (29) using hypoxanthine phosphoribosyltransferase 1 as a housekeeping gene. PCR with all complementary DNA samples was performed in duplicates.

Human samples
The study of human samples was approved by the Institutional Review Board of Kansai Medical University, and informed consent was obtained from each patient (protocol number: 2020308). All study protocols were consistent with the recommendations of the Declaration of Helsinki as a statement of ethical principles for medical research involving human subjects.

Statistical analysis
The average signal intensity of the region of interest (250 × 250 μm) in three different areas of the outer medulla from each animal was analyzed using ImageJ software. Comparison of the signal intensity of immunostaining in wild -type and CST-null mice was performed by Student's t-test. Values represent mean standard deviations of signal intensity from three adult animals per group.

Cst-deficient mice showed abnormalities in intercalated cells of the collecting duct
To examine morphological changes caused by the loss of sulfatide, we compared kidney tissues from wildtype and Cst-null mice by electron microscopy. No morphological abnormalities were observed in the slit structure of podocytes that made up the glomerulus (Fig. 1A, B), structures of microvilli and basal lamina in the proximal tubules (Fig. 1C, D), Henle's loop (Fig. 1E, F), basal infolding in the thick ascending limb (Fig. 1G, H), and distal tubule ( Fig. 1I-L). The cortex collecting duct was composed of principal cells with cilia characterized by low infolding near the base of the cell ( Fig. 2A, E, and P in G) and intercalated cells protruding toward the lumen without cellular interdigitation and basal infolding ( Fig. 2B-D, F-I). It has been known to appear from the connective tubules to the outer medullary collecting ducts (16)(17)(18)(19). Electron microscopy analysis of the cortex collecting duct of Cst-null mice ( Fig. 2E-I) revealed abnormalities in the intercalated cells protruded into the lumen, which accumulated many vesicles and lipid droplets in the cytoplasm compared with that of wild-type mice ( Fig. 2B-D). Because accumulation of intracellular vesicles was found in the intercalated cells of Cst-null mice, the distribution of the intercalated cell-specific V-ATPase, which is localized in the apical plasma membrane in intercalated cells (16,17), was investigated using immunohistochemistry. V-ATPase 6V0A4 subunit-positive granules were predominantly localized in the cytoplasm of Cst-null intercalated cells, whereas these were confined to the apical side in wild-type intercalated cells (Fig. 3A). The intracellular signal intensity of V-ATPase 6V0A4 was significantly higher in Cst-null mice than in wild-type mice (Fig. 3B). These sections were stained for a lysosomal marker LAMP2 to ascertain whether the accumulated V-ATPase-containing vesicles represent lysosomes. A part of intracellular V-ATPase was overlapped with LAMP2 in Cst-null mice, while no overlapping was observed in wild-type mice (Fig. 3C). The colocalization between V-ATPase and LAMP2 was analyzed using the Pearson's correlation coefficient. The Pearson's correlation coefficient (r) was 0.32 for Cst-null mice and 0.02 for wild-type mice. Therefore, a portion of V-ATPase colocalizes with lysosomes in Cst-null mice, suggesting that the increased V-ATPase-positive vesicles in Cst-null mice are lysosomes.

Seventeen sulfatide molecular species were detected in mouse kidneys
To investigate the specific sulfatide molecular species present in mouse kidneys, iMScope MALDI-IMS analysis was performed on wild-type and Cst-null mice. Based on a previous report on sulfoglycosphingolipids in renal cells (23), analyses were performed in negative mode with 9AA as a matrix. The sulfatide composition of the whole mouse kidney was compared with that described in previous studies (21)(22)(23) and the Human Metabolome Database (http://www.hmdb.ca/). The observed mass value was compared with the mass value <0.04 differences to reference m/z.  Table 2 and Fig. 4). Because of the absence of their spectra in the Cst-null kidney, all 17 signals were assigned as sulfatide species (Fig. 4). We then further characterized these 17 sulfatide species in our subsequent analyses.
Distribution of specific sulfatide species classified into four patterns Histology-directed IMS of the mouse kidney demonstrated that sulfatide species were present in the cortex, outer medulla, and inner medullary regions (Fig. 5). The distributions of specific sulfatide species were classified into four patterns (Table 3) The most abundant sulfatide species, including m/z 878.60, 892.61, and 906.63, containing ceramide composed of sphingosine (d18:1) and hFAs, were highly expressed in the outer medulla and some tubules in the cortex (Fig. 5G, J, N), whereas sulfatide species containing hFAs were not expressed in the inner medulla (Fig. 5C, E-H, J, K, M-O). Sulfatide species containing phytosphingosine (t18:0) were present over the entire length of the collecting ducts (Fig. 5L, R). Sulfatide species with C20-sphingosine were clearly detected in the inner medulla collecting ducts (IMCDs) (Fig. 5I, P) in accordance with a previous report (30).
Sulfatide species of m/z 896.61 and 924.64 were expressed in intercalated cells There are two types of cells along the collecting duct, that is, principal cells and intercalated cells. The ratio of  principal to intercalated cells varies between tubule segments, with a ratio of 2:1 in the outer medullary collecting duct and 3:1 in the cortical collecting duct (16)(17)(18)(19). To investigate which types of cells expressed the sulfatide species of m/z 896.61 and 924.64, we evaluated IMS images and biomarker expression (Fig. 7). The lectin DBA binds to the apical aspect of the collecting duct (31,32). The collecting ducts stained with DBA were distinguished from the distal convoluted tubules, which express calbindin (Fig. 7A). String-shaped tissues were stained from the cortex through the outer medulla to the inner medulla with DBA ( Fig. 7A-C). Principal cells stained with DBA and intercalated cells expressing V-ATPase 6V0A4 (16,17) were exclusive to each other along the same collecting duct (Fig. 7C, D). V-ATPase-positive intercalated cells were scattered around the collecting duct (Fig. 7C-F), similar to images of sulfatide species with m/z 896.61 and 924.64 (Fig. 6C,  F, H, 7G, H).
On the other hand, when IMS images were compared, m/z 906.63 of pattern II and m/z 924.64 of pattern III were exclusive to each other from the cortex to the outer medulla and the population of m/z 924.64 ion signal positive cells were less than m/z 906.63 (Fig. 7G), indicating a similar localization to that of immunostaining. When histological images stained with hematoxylin-eosin overlapped with the IMS image, sulfatide species with m/z 896.61 and 924.64 of pattern III were found to be localized along a series of   Sulfatide species of m/z 896.61 and 924.64 were the predominant species in intercalated cells To confirm whether sulfatide species of m/z 896.61 and 924.64 were expressed in intercalated cells, purified intercalated cells obtained using FACS were analyzed with MS. Collagenase-digested cells from murine kidneys were stained with primary antibodies recognizing V-ATPase 6V0A4, a marker for intercalated cells, stained with a fluorescently tagged secondary antibody, and then subjected to FACS analysis ( Fig. 8A-F). The cells were sorted using the forward scatter and side scatter (Fig. 8A). And then, the sorted cells were divided into two populations-ATPase− (ATPase low) and ATPase+ (ATPase high) (Fig. 8B). The ATPase+ population were purified by eliminating ATP-positive doublet cells (Fig. 8C-F). Methanol extracts of purified V-ATPase 6V0A4-positive cells (Fig. 8D) were analyzed for sulfatide species using iMScope (Fig. 8G). The predicted sulfatide ions were detected at m/z 850.57, 878.60, 892.61, 896.61, 906.63, and 924.64. The relative ratio of ion intensity for each sulfatide species showed that the V-ATPase 6V0A4positive intercalated cells predominantly contained sulfatide species with m/z 896.61 and 924.64 (Fig. 8G). This result was consistent with findings of IMS image analysis (Fig. 7).
Sulfatide species containing ceramide of t18:0-C22:0h and t18:0-C24:0h were present in human kidneys Next, we investigated whether sulfatide molecular species of m/z 896.6 (t18:0-C22:0h) and 924.6 (t18:0-C24:0h) specifically expressed in intercalated cells were common in human. All 17 sulfatide molecular species found in the mouse kidney were also observed in the human kidney (Fig. 9A). In addition, IMS images of specific sulfatide species in the human kidney ( Fig. 9C-F) were similar to those in the mouse kidney (Figs. 5-7). Sulfatide species with m/z 896.6 and 924.6 were colocalized, showing punctate structures from the cortex through the outer medulla to the inner medulla (Fig. 9C). However, the distribution patterns of sulfatide species of m/z 890.6, 918.6, and 920.6 were different from those of m/z 896.6 and 924.6 ( Fig. 9D-F). These results suggested that sulfatide species containing ceramide composed of t18:0-C22:0h and t18:0-C24:0h were specifically localized in intercalated cells in the human kidney.
Precursors of SM4s t18:0-C22:0h and SM4s t18:0-C24:0h did not accumulate in the collecting duct To investigate whether abnormalities of the intercalated cells in Cst-deficient mice occurred because of accumulation of the precursor of specific sulfatide species, we measured GalCer species in Cst-null mice using IMS (Fig. 10). In the GalCer reference derived from the bovine brain, three GalCer molecular species,  (Fig. 10A). No GalCer molecular species (Table 4) were detected in the wild-type mouse kidney (Fig. 10B) (Fig. 10C). Imaging analysis showed that these GalCer species accumulated in the inner medulla in the Cst-null mouse kidney (Fig. 10G, H). Neither precursor GalCer species with ceramide t18:0-C22:0h or t18:0-C24:0h was detected in the Cst-null mouse kidney (Fig. 10C, G, I). These results suggest that abnormalities in Cst-null intercalated cells are not because of the accumulation of precursors of sulfatide species (Table 4).

DISCUSSION
In this study, we demonstrated that there were morphological abnormalities in intercalated cells along the collecting duct in Cst-null kidneys and that sulfatide molecular species whose ceramide was composed of phytosphingosine (t18:0) and a very long-chain hFA were expressed in intercalated cells. Intercalated cells can be classified into three types: type A, type B, and non-A/non-B cells (15)(16)(17). Type A intercalated cells are present throughout the collecting duct and are morphologically characterized by numerous microprojections on a bulging apical surface, numerous mitochondria, a centrally located nucleus, and moderate basal infolding. V-ATPase is localized in the apical plasma membrane. Type B cells are mainly observed in the initial collecting duct and are characterized by intact nuclei and moderate basal infolding. The apical surface is less swollen than that in type A cells, and short microprojections are detected. V-ATPase is localized to the basolateral membrane. Finally, non-A/ non-B cells are tall cuboidal cells and are mainly found in the connecting segment. These cells showed a surface covered with protruding microprojections, small vesicles, and mitochondria distributed in the cytoplasm. V-ATPase is localized in small vesicles and the apical membrane. However, in this study, we did not evaluate the specific type of intercalated cells expressing such unique sulfatide species. A study by Stettner et al. (13) suggested that NH 3 excretion was suppressed through Rhcg in the apical aspect of type A intercalated cells, whereas other studies suggested that all types of cells along the collecting duct, including principal cells, collaborated functionally in the excretion of NH 3 and protons to regulate acid-base homeostasis (15)(16)(17). Therefore, we could not exclude the possibility that other types of intercalated cells may also be involved in the suppression of NH 3 excretion in the Cst-null kidney.
Recent studies have shown that under stress conditions such as metabolic acidosis, intercalated cells are adaptive and switch between type A and type B cells (35). In addition, transcriptional profiling analysis of single cells showed that the collecting duct generates various cell types via newly identified transitional cells (36). This transitional cell type expressed both intercalated cell and principal cell markers, suggesting that the cell types in the collecting duct undergo cell transitions that are altered by environmental influences (37,38). Lithium treatment was reported to induce cellular remodeling of the collecting duct, ultimately increasing the ratio of renal intercalated cells to renal principal cells (39,40). de Groot et al. (41) reported that a significant proportion of the proliferating cells of nuclear antigen-positive principal cells are arrested in the G2 phase of cell division. Himmel et al. (42) further suggested that the plasticity of renal IMCD cells may play an important role in lithium-induced renal IMCD remodeling. Therefore, further study using inducers of changes in the intercalated cell population such as The sulfatide species in methanol extracts of the sorted ATP6V0A4-high population were analyzed using iMScope (G). Relative ratios of ion intensity for each sulfatide species in the ATP6V0A4-high population to that of the ATP6V0A4-low population, normalized to that of the PI ion (m/z 885.50), are shown (G). The ratio of m/z 896.61 and 924.64 ions in this ATP6V0A4-high population were increased in three independent experiments. The expression levels of Dsgs2 (encoding sphingolipid delta(4)-desaturase) and Fa2h (encoding FA 2-hydroxylase) in the ATP6V0A4-high population were analyzed using RT-qPCR (H). Atp6v0a4 (encoding ATP6V0A4 in intercalated cells), Slc26a4 (encoding pendrin in type B-intercalated cells), and AE1 (encoding SLC4A1 in type A intercalated cells) were used as positive controls. B2m (encoding b2microglobulin) and Gapdh (encoding glyceraldehyde-3-phosphate dehydrogenase) were used as housekeeping genes. chronic lithium exposure may directly demonstrate the consistent changes in the m/z 896.6 and 924.6 species of intercalated cells.
The existence of unique sulfatide species with ceramides consisting of a combination of 2-hFAs and phytosphingosine (t18:0) as the long-chain base has been reported in the kidneys using two-dimensional nuclear magnetic resonance (43) and MS (21). However, their distribution in the kidneys remains unknown. In this study, we evaluated the localization of these molecules in particular cells for the first time. As the scan pitch of IMS was 10-25 μm, it is difficult to confirm that these two molecules are completely expressed in the same cell. However, the ion signal in Fig. 6 clearly shows that the two molecules are in close proximity to each other in a large proportion of cases. Ceramides of these sulfatide species possess three hydroxy groups at the base of the lipid moiety corresponding to the interface between the plasma membrane and the extracellular environment under physiological circumstances. The hydroxylation of ceramides occurs in only a few types of glycosphingolipids, including GalCer and sulfatide. Because the GalCer synthase CGT is localized at the endoplasmic reticulum and prefers hFA-ceramides to non-hFA-ceramides, hFA-ceramides generated in the endoplasmic reticulum tend to be incorporated into GalCer (34). Subsequently, hFA-GalCer was sulfated to produce hFA-sulfatide by CST. Hydroxylation of glycosphingolipids facilitates lipid packing and influences the stability and permeability of the membrane by increasing the amount of hydrogen bonding at the interfacial region of the membrane (44). In fact, the hydroxylation of ceramides has been reported to affect membrane microdomains by strengthening the lateral interactions between neighboring proteins and lipids (45). Because strict control is required for the transport of small molecules, such as NH 3 and protons, across the cell membrane, leak-proof structures of the plasma membrane are desirable.
Sulfation is essential for the NH 3 transport activity of Rhcg, as demonstrated in a study of Cst-null kidneys (13); however, it is unknown whether hydroxylation of ceramide is necessary for it. The phytosphingosine structure is synthesized by the sphingolipid C4-monooxygenase activity of DEGS2 (33), and FA 2-hydroxylation is catalyzed by FA2H (34). Fa2h-deficient mice lack 2-hydroxylated sphingolipids in the central and peripheral nervous systems (46). Although normal compact myelin was formed in these mice, defects in 2-hydroxylated GalCer in the myelin Fig. 9. Mass spectra of the major sulfatide molecular species in the kidneys of humans and localization of the molecular species of m/z 896.6 and 924.6. MALDI-MS spectra of normal kidney tissues analyzed in negative ion mode with 9AA. The averaged mass spectra are shown. The red line shows the same spectral peaks as the 17 major sulfates found in mouse kidneys (A). Optical image of normal human kidney tissue (B). Sulfatide molecules of m/z 896.6 (pattern III, magenta) and 924.6 (pattern III, green) were colocalized with each other in the IM and OM (C). Molecules of m/z 896.6 (pattern III, magenta) and 890.6 (pattern I, green) showed different localizations in the IM and OM (D). Imaging of m/z 920.6 (pattern I, magenta) and 918.6 (pattern IV, green) showed colocalization in the IM and OM (E). Molecules of m/z 896.6 (pattern III magenta) and 918.6 (pattern IV, green) showed different localizations in the IM and OM (F). The scale bar represents 700 μm. C, cortex; IM, inner medull; OM, outer medulla. membrane lead to loss of the long-term stability of myelin and eventual demyelination (46). 2-Hydroxylated sphingolipids are abundant, and FA2H is highly expressed in mammalian skin. In contrast to the nervous system, biosynthesis of most 2hydroxylated sphingolipids was not influenced in Fa2h-null skin (47). In murine skin, FA2H is only expressed in sebaceous glands, and only a subset of 2hydroxylated FAs (C20:0h, C22:0h, C23:0h, and C24:0h) incorporated into glucosylceramide is synthesized by FA2H (47). The set of 2-hydroxylated long-chain FAs synthesized by FA2H in the skin is utilized in sulfatide species in the kidneys. To the best of our knowledge, there are no reports on the physiological functions of the 4-hydroxylation of the sphingoid base. Although the DEGS2 gene general function in the nervous system is unknown, a genome-wide association study of cognitive dysfunction in schizophrenia patients found an association with missense mutations in the DEGS2 gene (48). The DEGS2 gene is expressed in the brain, lungs, intestines, skin, and kidneys (49).
In intercalated cells of the kidneys, proton secretion is regulated by the V-ATPase complex, which is present in the lysosomal membrane and the apical aspect of the plasma membrane (50). In the current study, V-ATPasepositive vesicles were found to accumulate in intercalated cells. Therefore, the intracellular vesicles accumulated in intercalated cells were considered V-ATPase-containing lysosomes. NH 3 accumulated in the cells may impair lysosomal function by increasing the intralysosomal pH, resulting in accumulation of lysosomes in the cells, similar to the pharmaceutical effects of chloroquine (51,52). Alternatively, transportation of the V-ATPase protein to the apical membrane may be hampered owing to defects in sulfatide. When the trafficking and regulation of V-type ATPase are impaired by suppression of basolateral anion exchange activity, the size and number of type A intercalated cells are altered, and accumulation of lysosome-like vesicles and multilamellar vesicles is observed (53).
Several studies have demonstrated the involvement of specific sulfatide species in a variety of biological contexts. C18:0 and C18:0h sulfatide species are increased in the white matter of the neonatal rat brain after nitric oxide inhalation as a therapy for brain injury (54,55). These C18:0 and C18:0h sulfatide species also appear in defined regions of oligodendrocyte generation, whereas sulfatide species with longer FAs, such as C24:0 and C24:0h, appear during oligodendrocyte maturation (25). The length of the FA chains of glycosphingolipids affects the fluidity of cell membranes (56) and possibly their biological functions. For example, the longer acyl chains of glycolipids are more readily recognized by antibodies because of increased exposure of the carbohydrate head group on the membrane surface (57). Aggregation based on sugarsugar interactions between GalCer-containing liposomes and sulfatide-containing liposomes depends on hydroxylation and the length of the FA composition of sulfatide species (58). By contrast, a short-chain FA C16:0-containing sulfatide is predominant in the pancreas (59). C16:0 sulfatide inhibits glucose-induced insulin secretion by reducing the K + channel sensitivity to the ATP block, whereas C24:0 sulfatide does not affect this process (58).
In conclusion, our findings showed that V-ATPasepositive vesicles accumulated in the cytoplasm of intercalated cells in the collecting duct of Cst-null kidneys. IMS analysis suggested that sulfatide molecular species with ceramide composed of t18:0-C22:0h and t18:0-C24:0h, which were characteristically expressed in intercalated cells, were involved in the excretion of NH 3 and protons into the urine.

Data availability
All data are contained within the article and available from the corresponding author upon reasonable request. assistance with the histochemical research, Kenya Yamazaki, a medical student at Kansai Medical University, for assistance with IMS analysis, Takayuki Kawaura of the Department of Mathematics of Kansai Medical University for assistant with IMS analysis programming, Akira Saitou for assistance with MALDI-IMS and Hiroyuki Gonda for assistance with the FACS Aria at Central Research Center of Kansai Medical University. They also thank Editage (www. editage.com) for English language editing.