Deficiency of the tetraspanin CD63 associated with kidney pathology but normal lysosomal function

Abstract

CD63 is a member of the tetraspanin superfamily that constitutes a main component of the lysosomal membrane. In mice, two CD63 gene loci are present, with only one of these two being functional. We generated and analyzed mice deficient for active CD63. Disruption of CD63 results in a complete loss of CD63 protein expression. Despite its abundance in late endosomes/lysosomes, the lack of CD63 does not cause obvious endosomal/lysosomal abnormalities. CD63 knockout mice are viable and fertile without gross morphological abnormalities in the majority of tissues. No alterations in the populations of immune cells and only minor differences in platelet function were observed. This suggests that the lack of CD63 could be successfully compensated for, most likely by other tetraspanins. However, CD63 deficiency leads to an altered water balance. CD63 knockout mice show an increased urinary flow, water intake, reduced urine osmolality, and a higher fecal water content. In principle cells of the collecting duct of CD63-deficient mice, abnormal intracellular lamellar inclusions were observed. This indicates that the sorting of apical transport proteins might be impaired in these cells. CD63 knockout mice provide an important tool for analyzing the various postulated functions of CD63 in vivo.

FIG. 1.

Comparison between the CD63 gene and pseudogene. (A) Alignment of the mCD63 gene on chromosome (Chrom.) 10 and the pseudogene on chromosome 18. Red or green, amino acid substitutions; blue, N glycosylation sites (N-X-S/T); purple, transmembrane domains; yellow, cytoplasmic loops; orange, extracellular domains; black lines, synthetic peptides used for immunization (DNAstar software; DNASTAR, Inc., Madison, WI). (B) MEF were transfected with pEGFP-mCD63-C1-10 (chromosome 10, gene) (a to c) or with pEGFP-mCD63-C1-18 (chromosome 18, pseudogene) (d to f). Cells were costained against lysosome-associated membrane protein 1 (LAMP-1). The majority of CD63 is colocalizing with LAMP-2 in the lysosomal compartment, whereas the protein arising from the CD63 pseudogene is not localized to the lysosomal compartment. This can be explained by the Y-to-H mutation in the C-terminal sorting motif of the protein. (a, d) Merged (green, eGFP-CD63; red, LAMP-2); (b, e) eGFP-CD63; (c, g) LAMP-2. Scale bar = 10 μm.

FIG. 2.

Targeted disruption of the CD63 gene on mouse chromosome 10. (A) Strategy for the inactivation of the CD63 gene by homologous recombination in ES cells. (a) Partial structure of the genomic locus representing about 12 kb of the CD63 gene region. Coding exons are indicated by light gray boxes, noncoding exons by black boxes, and flanking introns by solid lines. Probe A denotes a DNA probe used for Southern blot analysis. (b) Targeting vector pBluescript KS-CD63-neo/TK, with about 7.4 kb homology to the CD63 gene locus. The neo cassette was inserted as an XbaI fragment into an NheI restriction site at Val3 introduced in exon 2 by mutagenesis PCR. (c) Predicted CD63 gene locus after homologous recombination. (B) Southern blot analysis of ES cell clones. Probe A was hybridized to BglII-, XbaI-, or NheI-digested genomic DNA from ES cell clones E32, E47, and E31. Additional 3.4-kb, 7.2-kb, and 6.0-kb DNA fragments, respectively, indicate a targeted allele. (C) Northern blot analysis of kidney and MEF cells. (D) RT-PCR analysis of CD63 expression. Total RNA was used for reverse transcription, followed by PCR amplification of the CD63 cDNA open reading frame. A 712-bp fragment is amplified in CD63^+/+ and is absent in CD63-deficient MEF. (E) Western blot analysis of CD63 expression using an antibody directed against a synthetic peptide from the second luminal loop E2 of mCD63. (a) Proteins in cell lysates derived from wild-type and CD63-deficient mice were deglycosylated using PNGase F. (Upper) Deglycosylated CD63 molecules were detected in CD63^+/+ cells and absent in CD63-deficient cells. PMN, polymorphonuclear leukocyte; MΦ, macrophage. (Lower) Coomassie staining of the membrane demonstrates equal protein loads. (b) Glycosylation of CD63 impairs antibody binding, leading to more-unspecific binding in MEF cells.

FIG. 3.

Immunocytochemical analysis of CD63-deficient MEF. CD63^+/+ and CD63^−/− MEF were fixed with 4% paraformaldehyde-PBS and immunostained against CD63 (green) and LAMP-2 (red) (A to F), cathepsin D (red) and LAMP-2 (green) (G to L), MPR-46 (M, N), and MPR-300 (O, P). CD63 was readily detected in CD63^+/+ but not in CD63^−/− MEF with the antiserum directed against native CD63. No other differences in the amounts or intracellular distributions of the lysosomal marker proteins and mannose 6-phosphate receptors were detected. Scale bar = 10 μm. Blue indicates DAPI (4′,6-diamidino-2-phenylindole) staining of the nuclei.

FIG. 4.

Analysis of CD63^−/− platelets. (A) CD63-deficient mice possess normal platelet numbers (n = 9). MFI, mean fluorescence intensity; wt, wild type. (B) CD63 is not essential for platelet adhesion to collagen under flow. Whole-blood samples from the indicated mice (n = 15) were perfused at wall shear rates of 1,000 s⁻¹ (4 min) over a collagen-coated surface. The thrombus area, meaning the area covered with platelets, was determined for phase-contrast images. Similar results have been obtained under low shear rates of 150 s⁻¹ (10 min). (C) Heparinized platelet-rich plasma samples from wild-type and CD63-deficient (ko) mice were stimulated with the indicated concentrations of ADP, U46619, CRP, or collagen (coll.), and light transmission was recorded with a standard aggregometer. Representative results for each genotype are shown (n = 3). CD63-deficient platelets seem to respond more strongly to collagen or CRP. (D, E) Diluted whole-blood samples were stimulated with ADP (10 μM) ± U46619 (U46; 1 μM), thrombin (0.02 U/ml, 0.002 U/ml, or 0.0002 U/ml), or CRP (5 μg/ml or 3 μg/ml) for 2 min each; subsequently incubated with anti-activated GPIIbIIIa (D) or anti-P-selectin (E) for 10 min; and analyzed directly. Platelets were gated by forward-scatter/side-scatter characteristics. No changes in activation were detected between wild-type and CD63-deficient platelets. rest., resting/unstimulated.

FIG. 5.

Histology of CD63^−/− mice. (A to J) One-micrometer-thick semithin sections, stained with Toluidin blue. In cerebellar Purkinje cells (A, B), the liver (C, D), splenic white pulp (E, F), and the lung (G, H), no morphological differences between CD63^+/+ (upper row) and CD63^−/− (lower row) mice can be seen. The dense inclusions seen in hepatocytes correspond to lipid droplets. (I, J) Kidney, outer medullary collecting ducts. PC, principal cells; IC, intercalated cells. In the CD63^−/− mouse (J), the PCs (arrowheads) show dense cytoplasmic inclusions, which are absent from PC in the CD63^+/+ mouse (I). (K) Low electron-microscopic magnification showing PCs of a CD63^−/− mouse with abnormal inclusions. (L) At high resolution, the abnormal inclusions show a concentrically lamellar pattern. Scale bars = 10 μm (A to D), 25 μm (E to H), 10 μm (I and J), 3 μm (K), and 0.1 μm (L).

FIG. 6.

Metabolic analysis of CD63-deficient mice. Eight 11- to 12-month-old female mice (black, CD63^+/+; gray, CD63^−/−) in metabolic cages were analyzed for 24 h. (A) CD63^−/− mice urinate three times more within 24 h and per gram of body weight (Bw). (B) CD63-deficient mice drink significantly more as well. (C) The osmolality of CD63 urine is significantly reduced. (D) Fecal droppings of CD63-deficient mice are bigger than those of wild-type animals. (E) The total amount of feces per gram body weight is significantly increased as well. (F, G) The increased fecal secretion is due not to increased dry weight but to a significantly higher water content.