Overexpression of Branched O-Linked Oligosaccharides on T Cell Surface Glycoproteins Impairs Humoral Immune Responses in Transgenic Mice*

The aberrant expression of core 2O-glycans on T cell surface glycoproteins has been associated with various immunodeficient syndromes such as Wiskott-Aldrich syndrome and AIDS. To determine the effect of this aberrant expression of core 2 O-glycans on immune responses, we previously generated transgenic mice overexpressing core 2 β-1,6-N-acetylglucosaminyltransferase (C2GnT) in T cells, and demonstrated that T cell primary immune responses mediated through interaction between T cells and antigen-presenting cells are impaired in the transgenic mice (Tsuboi, S., and Fukuda, M. (1997)EMBO J. 16, 6364–6373). In this study, we determined whether overexpression of core 2 oligosaccharides on T cells leads to impaired humoral immune responses by B cells using the same transgenic mice. When T cells were activated, both T and B cells from the transgenic and control mice expressed an equivalent amount of CD40L and CD40, which are, respectively, the receptor and counter-receptor for the interaction between T and B cells. However, activated T cells from the transgenic mice induced B cell proliferation less efficiently than those from control mice, regardless of whether B cells were isolated from control or the transgenic mice. This suggests that overexpression of core 2 O-glycans on T cell surface glycoproteins renders T cell-B cell interaction inefficient. Moreover, in the transgenic mice both immunoglobulin isotype switching and germinal center formation were also impaired. Taken together, these results indicate that aberrant expression of core 2 O-glycans on T cell surface glycoproteins results in impaired humoral immune responses due to an impaired interaction between T and B cells.

Among the various cell surface carbohydrates, the biological function of mucin-type O-glycans, which are attached to serine or threonine residues on cell surface glycoproteins, is not well understood. There are, however, several reports suggesting that O-glycans are involved in the modulation of cell adhesion processes. For example P-and L-selectins bind with higher affinity to sialyl Le X and sulfated sialyl Le X expressed in Oglycans than those expressed in N-glycans, possibly due to multivalency of O-glycans (5)(6)(7).
Various pathological conditions, including certain immunodeficient syndromes, are associated with aberrant expression of core 2 O-glycans. Examples include Wiskott-Aldrich syndrome (13,14), leukemia (15,16), AIDS (16 -18), and malignant transformation (19). In these diseases, core 2 O-glycans are highly expressed on resting T cells, while they are not expressed in normal conditions. To examine the influences of aberrant expression of core 2 O-glycans on the immune system, we have previously generated transgenic mice whose T cells overexpress core 2 O-glycans. We demonstrated that T cells from these transgenic mice exhibit reduced primary T cell immune responses, as assessed by delayed-type hypersensitivity, proliferation upon stimulation, and cytokine production (20). We demonstrated that reduced immune responses of T cells are due to impaired interaction between T cells and antigen-presenting cells (20). The next important question is whether overexpression of core 2 O-glycans affects humoral immune responses. In this study, we address this question by examining humoral immune responses in the transgenic mice compared with those of control mice. We demonstrate that overexpression of core 2 O-glycans impairs humoral immune responses to thymus-dependent antigen. We suggest that this reduced B cell immune response is due to inhibition of the interaction between T and B cells (T-B interaction).

EXPERIMENTAL PROCEDURES
Transgenic Mice-Transgenic mice (Lck-mC2GnT) were established in the previous study (20). Briefly, the transgene, mouse C2GnT, was cloned into pBluescript under the control of p56 lck promoter (20,21) and injected into FVB/N mouse zygote pronuclei. Stable lines of mice were generated by breeding founders with FVB/N mice (20). Two founder mice were generated (Lck-mC2GnT-1 and -2), and all experiments were carried out using cells isolated from 6 -8-week-old litter mates, which were generation 4 and 5. Sex-and age-matched littermates were used as wild-type controls. The results obtained from each founder were not designated, as no detectable difference in the results was noted using their offspring.
Glycosyltransferase Assay-To measure C2GnT activity, T cells were purified from spleens with Mouse T Cell Enrichment Column (R&D Systems, Minneapolis, MN) (22). The resultant preparations were Ͼ95% pure T cells, as determined by flow cytometric analysis for ␣␤ T cell receptor. B cells were purified from spleens by flow cytometric sorting. The resultant preparations were Ͼ99% positive for B220 by flow cytometric analysis. Purified T or B cells were homogenized in 0.1 ml of an extraction solution (0.4% Triton X-100, 150 mM NaCl), and the assay was performed essentially as described using Gal␤133GalNAc␣13p-nitrophenol as a substrate (9). Similarly, UDP-Gal:␤-GlcNAc␤-1,4-galactosyltransferase and UDP-Gal:␣-Gal-NAc␤-1,3-galactosyltransferase were assayed using p-nitrophenylglycosides (23,24). In all of these assays, resultant products were adsorbed to a disposable C18 cartridge (Waters, MA) and eluted by 70% methanol in water as described (9).
Western Blot Analysis-T cells were purified from a pool of four spleens (from two males and two females) using the procedure described above. Whole cell lysates of purified T cells were prepared and blotted as described previously (20). The membranes were incubated with the anti-CD43 monoclonal antibodies as described (20), and with anti-CD44 (IM7, culture supernatant, 1:40 dilution; Ref. 25), anti-CD45RB (23G2, 10 g/ml), and anti-CD2 (RM2-5, 0.5 g/ml). The antibodies except for anti-CD44 were purchased from PharMingen (San Diego, CA).
Analysis of CD40 Ligand (CD40L) Expression on T Cells-CD4 ϩ T cells were purified from spleens with Mouse T Cell CD4 Subset Column Kit (R&D Systems, Minneapolis, MN) (22). The resultant preparations were Ͼ95% pure CD4 ϩ T cells, as determined by flow cytometric analysis for CD4. Purified CD4 ϩ T cells were activated for 6 h in 24-well plates coated with anti-CD3 (142C-11, 20 g/ml) and mouse ICAM-1-IgG fusion protein (10 g/ml) as described (20). In parallel, purified T cells were activated for 6 h with PMA and ionomycin as described (20). Activated T cells were subsequently harvested, washed, stained with PE-conjugated anti-CD40 ligand, and then analyzed by flow cytometer.
B Cell Proliferation Assay-Spleen cell suspension was prepared from a pool of spleens from eight mice (four males and four females). CD4 ϩ T cells were isolated from half of a spleen cell suspension as described above. B cells were isolated from the other half of the cell suspension by a combination of treatment with antibodies to Thy-1.1, CD4, and CD8 (PharMingen), followed by rabbit complement (ICN, Irvine, CA) and incubation on plastic dishes to deplete adherent cells (29). The resultant preparations were Ͼ95% pure B cells, as determined by flow cytometric analysis with B220. For B cell proliferation assays, purified B cells (2 ϫ 10 5 cells/well, 200 l) were cultured in RPMI 1640 medium containing 10% fetal calf serum, 2-mercaptoethanol (50 M), penicillin (100 units/ml), and streptomycin (100 g/ml) for 72 h in 96-well flat-bottomed tissue culture plates. In some experiments, the medium also contained 10 ng/ml IL-4 (PharMingen), 1 g/ml lipopolysaccharide (Sigma), or 5 g/ml anti-CD40 (HM40 -3, PharMingen). In parallel, purified B cells (1 ϫ 10 5 cells/well) were co-cultured with non-activated or T cells activated for 6 h (1 ϫ 10 5 cell/well) in the same medium as described above. The T cells were fixed with 0.4% paraformaldehyde in PBS for 15 min at room temperature, then washed and used to stimulate B cells (30). Cultures were pulsed with 1 Ci/well [ 3 H]thymidine (NEN Life Science Products) for the last 6 h of a 72-h culture period and then harvested onto a glass fiber filter. [ 3 H]Thymidine uptake was measured by liquid scintillation counting.
Anti-KLH Antibodies Analysis by ELISA-Mice were intraperitoneally immunized with 100 g of KLH (Sigma) in complete Freund's adjuvant. Mouse anti-KLH IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA levels in the serum were determined by an isotype-specific ELISA. Briefly, 96-well plates were coated with 10 g/ml KLH for 24 h. Plates were then washed and blocked with 1% bovine serum albumin in PBS. Diluted serum samples were then added to wells, and plates were incubated for 1 h at 37°C. Anti-KLH IgM antibody titer was determined with biotin-conjugated goat anti-mouse IgM and streptavidinconjugated horseradish peroxidase (PharMingen). The other antibody titers were determined using alkaline phosphatase-conjugated goat anti-mouse IgG1, IgG2a, IgG2b, IgG3, or IgA (PharMingen). ELISAs were developed by using 1-Step ABTS substrate (Pierce) for peroxidase and p-nitrophenyl phosphate substrate (PharMingen) for alkaline phosphatase (31,32).
Immunocytochemical Detection of Germinal Centers-Mice were immunized with KLH as described above. Spleen flat sections (5 m thick), 7 days after the immunization, were fixed with 4% paraformaldehyde in PBS. The sections were incubated with antibody or lectin in a humidified box (30). The sections were suspended in 10 mM HEPES buffer containing 0.15 M NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 3% fetal calf serum, and stained with biotinylated peanut agglutinin (Vector Laboratories, Burlingame, CA) for 1 h at room temperature (33). Parallel sections were suspended in PBS supplemented with 3% fetal calf serum and stained with biotinylated goat anti-mouse IgM (R6 -60.2, PharMingen). Biotinylated peanut agglutinin and anti-mouse IgM were stained with PE-conjugated streptavidin (PharMingen). The slides were evaluated and photographed under a Zeiss Axioplan fluorescence microscope.

Generation of Transgenic Mice Overexpressing Branched
Hexasaccharides on T Cells-As shown previously, we have generated transgenic mice that express the mouse C2GnT gene under the control of the proximal promoter of p56 lck . This promoter allowed increased activity of C2GnT in lymphoid tissues where T cells are present (20). However, the previous results were obtained on a mixture of T and B lymphocytes, and we could not exclude the possibility that B lymphocytes also contained increased C2GnT activity in the transgenic mice. To determine whether only T cells contain increased C2GnT activity, T and B cells were purified from spleens, and glycosyltransferases were assayed in cell lysates of purified T and B cells. As shown in Table I, T cells derived from the transgenic mice (Lck-mC2GnT) contained C2GnT activity approximately 50 times higher than did T cells derived from control mice. In contrast, the activities of ␤-1,3-galactosyltransferase, which forms core 1 in O-glycans, and ␤-1,4-galactosyltransferase, which forms N-acetyllactosamines, were almost equal in control and the transgenic mice (Table I). Moreover, no difference in C2GnT activities was found in B cells between the transgenic and control mice (Table I). These results indicate that C2GnT is specifically increased in T cells of the transgenic mice without an accompanying change in the activities of other glycosyltransferases.
In the above experiments, as well as those that follow, we found that results obtained from two independently generated transgenic lines (Lck-mC2GnT-1 and -2) were practically identical. Therefore, we conclude that clonal variations play little or no role in determining the phenotype of either line.

T Cell C2GnT Overexpression Resulted in a Change in Glycosylation of T Cell Surface
Glycoproteins-CD43 is one of the most abundant T cell surface glycoproteins. There are two distinct glycoforms of CD43: the 115-kDa form seen on resting T cells, which carries simple tetrasaccharides and can be detected by S7 antibody, and a 130-kDa glycoform seen on activated T cells, which carries branched hexasaccharides and can be detected by the 1B11 antibody (34). Previously, we have shown that CD43 expressing the hexasaccharide was increased in the whole cell populations of the transgenic mice (20). As shown in Fig. 1, the population of T cells (␣␤ T cell receptor positive cells) displaying the CD43-hexasaccharide, which carries core 2 O-glycans, increased in both spleens and lymph nodes derived from the transgenic mice (Fig. 1). In contrast, the number of B cells, which are positive for B220 staining, was similar in control (49 -52%) and the transgenic (47-50%) mice (Fig. 1). Fig. 2 shows the results of Western blotting of T cell extracts. On the T cells from control mice, most of CD43 was the 115-kDa tetrasaccharide form, while on T cells from Lck-mC2GnT transgenic mice, most CD43 was the 130-kDa hexasaccharide form carrying core 2 O-glycans ( Fig. 2A). Similar results were obtained for other O-glycosylated proteins such as CD44 and CD45RB. CD44 was detected as an 85-kDa protein on T cells from control mice, whereas T cells from the transgenic mice displayed CD44 as an 88-kDa band ( Fig. 2A). CD45RB is an isoform of CD45. Western blotting using a monoclonal antibody, 23G2, identified CD45RB as a 200-kDa band on T cells from control mice, whereas on T cells from the transgenic mice, this 200-kDa band was faint ( Fig. 2A). The antibody 23G2 reacts with a peptide portion that is heavily glycosylated with O-glycans (35,36). It is likely that 23G2 cannot react efficiently with the same polypeptide when it is glycosylated with core 2 O-glycan containing oligosaccharides. In fact, the reactivity of 23G2 was weaker against CD45 expressed on CHO-leu⅐C2 cells, which express core 2 O-glycans (27), than it was against the parent CHO cells, which do not express core 2 O-glycans (lanes 5 and 6 in Fig. 2B). In contrast, almost equal reactivity was observed when another antibody, 30F-11, which reacts with all isoforms of CD45 (28), was used (lanes 3 and 4 in Fig.  2B). On the other hand, mouse C2GnT overexpression did not affect CD2, because CD2 does not contain O-glycans (Fig. 2A). These results strongly suggest that overexpression of mouse C2GnT in Lck-mC2GnT transgenic mice causes these T cell surface glycoproteins to carry substantially more core 2 Oglycans than do glycoproteins derived from control mice. These observations are consistent with the results obtained by Western blotting of C2GnT-overexpressing cell lines (37).
No Difference in CD40 Ligand and CD40 Expression Levels between Lck-mC2GnT Transgenic and Control Mice-We then asked whether overexpression of core 2 O-glycans affects T-B interaction. To address this question, we first examined whether overexpression of core 2 O-glycans affects expression levels of CD40 ligand (CD40L), since T-B interaction is mainly mediated by CD40L-CD40 interaction (26, 27, 38 -40). When T cells were incubated without activators, no CD40L expression was observed in either transgenic or control mice (Fig. 3A, top panel). When T cells were activated for 6 h with anti-CD3 and mouse ICAM-1-IgG fusion protein, T cells from the transgenic mice expressed CD40L on the cell surface at a level equivalent to that of control mice (Fig. 3A, middle panel, 13-17% and 13-15%, respectively). T cells activated with PMA and ionomycin from either the transgenic or control mice expressed similar amounts of CD40L as well (Fig. 3A, bottom panel, 64 -69% and 63-69%, respectively). These results show that similar amounts of CD40L were induced on in vitro activated CD4 ϩ T cells derived from the transgenic mice and control mice (Fig. 3A). In addition, B cells purified from both the transgenic and control mice expressed comparable amounts of CD40 (Fig. 3B).

TABLE I Glycosyltransferase activities in purified T cells and B cells
Glycosyltransferase activities (nmol/h/mg) in the cell extracts of T and B cells isolated from spleens were measured as described under "Experimental Procedures" (mean ϮS.E.). In each assay, two spleens from one male and one female were tested. The activities in the cell extracts from three independent assays are shown. The expression of a CD43-hexasaccharide form carrying core 2 O-glycans in spleen and lymph node was analyzed by flow cytometry. Splenocytes and lymph node cells were stained with PE-anti-CD43-hexasaccharide form carrying core 2 O-glycans (activation-associated form, 1B11) and FITC-conjugated anti-␣␤ T cell receptor (TcR) (H57-597). Splenocytes were also stained with FITC-conjugated anti-B220 antibody (RA3-6B2). Spleens and lymph nodes were isolated from two mice (one male and one female), and the analysis was independently performed three times using two different transgenic lines. The numbers indicate the ranges of positive cells in percentages.

Activated T cells from Lck-mC2GnT Transgenic Mice Inefficiently Induce B Cell Proliferation Compared with Control-To further examine whether overexpression of core 2 O-glycans
affects T-B interaction, we assayed T cell-dependent B cell proliferation in vitro. This assay is critical because CD40-CD40L interaction is too weak to be detected by adhesion (41). T cells were activated for 6 h with anti-mouse CD3 and a mouse ICAM-1-IgG fusion protein shown in Fig. 3A. B cells from the transgenic mice (closed symbols) were cocultured with activated T cells from control mice (shown as C) or the transgenic mice (shown as Lck). B cell proliferation was induced more by activated T cells from wild-type mice than by those from the transgenic mice (shown as Lck). The proliferative responses of B cells from the transgenic mice were enhanced by addition of IL-4. However, T cells derived from the transgenic mice yielded a reduced proliferative response even in the presence of IL-4 (Fig. 4A).
Similarly, B cells derived from control mice were cocultured with T cells from the transgenic and control mice (open symbols in Fig. 4A). T cells derived from the transgenic mice produced a reduced proliferative response of control B cells both in the absence and presence of IL-4 (Fig. 4A). These results indicate that T cells derived from the transgenic mice induced an impaired T-B interaction, regardless of whether the B cells were derived from wild-type or the transgenic mice.
In contrast, B cells from the control and transgenic mice showed comparable proliferative responses to lipopolysaccha-ride, to lipopolysaccharide plus IL-4, or to anti-CD40 treatment (Fig. 4B), where T-B cell interaction was not required. These results indicate that overexpression of core 2 O-glycans on T cell surface glycoproteins inhibits T-B interaction.
Isotype Switching Is Impaired in Lck-mC2GnT Transgenic Mice-To investigate how overexpression of core 2 O-glycans in T cells influences antibody production, we tested primary and secondary antibody responses of the transgenic mice. IgM antibody is produced in primary immune responses, and production switches to other immunoglobulin subtypes, such as IgG, in secondary responses. As shown in Fig. 5, the production of anti-KLH IgG1, IgG2a, and IgG2b responses was significantly lower in Lck-mC2GnT transgenic mice (closed symbols) than in control mice (open symbols). In addition, the amount of IgM product was higher in transgenic mice than in the control mice

FIG. 2. Western blot analysis of whole-cell extracts of purified T cells (A) and transfected CHO cells (B).
A, monoclonal antibodies specific to the CD43-tetrasaccharide form (CD43-tetra, S7, 2 g/ml), CD43-hexasaccharide form (CD43-hexa, activation-associated form, 1B11, 2 g/ml), CD44 (IM7, culture supernatant, 1:40 dilution), CD45RB (23G2, 10 g/ml), and CD2 (RM2-5, 0.5 g/ml) were used. The results were obtained on lysates (100 g of protein) of T cells derived from four spleens. B, the parent CHO cells (lanes 3 and 5) and CHO-leu⅐C2 (lanes 4 and 6) were transfected with pcDNAI-CD45 ABC . Fortyeight h after transfection, cell lysates were subjected to Western blotting with 23G2 (lanes 5 and 6) and 30-F11 (lanes 3 and 4) antibodies. Open histograms indicate the negative staining control. In both A and B, cells isolated from spleens of three mice (one male and two females) were pooled and analyzed. In A, T cell activation and flow cytometric analysis were independently performed three times using offspring from Lck-mC2GnT-1 and Lck-mC2GnT-2. The numbers indicate the ranges of the percentages of the CD40L ϩ cells.
due to impaired switching to other immunoglobulin isotypes (Fig. 5). The production of anti-KLH IgG3 and IgA in the transgenic mice was similar to that of control mice. These results, as a whole, indicate that isotype switching was impaired in Lck-mC2GnT transgenic mice.
Impaired Germinal Center Formation in Lck-mC2GnT Transgenic Mice-Germinal center formation is dependent on the interaction between CD40L and CD40 (30,31,42). We thus tested if germinal center formation in Lck-mC2GnT transgenic mice is reduced, since CD40L-CD40 interaction is impaired in transgenic mice as shown in Fig. 4A. Mice were immunized with KLH on day 0. Spleens from immunized mice were taken on day 7, the peak time of germinal center formation. Germinal centers from control mice detected by either anti-IgM antibody or peanut agglutinin were larger and denser than those from the transgenic mice (Fig. 6). Moreover, significantly fewer germinal centers were found in the transgenic mice than in controls (Table II). There are no obvious differences in numbers and ratios of T and B cells between the transgenic and control mice (Fig. 1). These observations are consistent with the phenotypes of CD40L-deficient or CD40-deficient mice. Mice carrying null mutations in either gene do not display abnormalities in the number and ratios of T and B cells, although germinal center formation is defective in both mice (30, 31, 42).

DISCUSSION
This study demonstrated that B cells derived from the transgenic mice overexpressing core 2 oligosaccharides in T cells exhibit reduced humoral immune responses to thymus-dependent antigen, which depend on T-B interaction. B cells derived from the transgenic mice exhibited delayed immunoglobulin isotype switching (Fig. 5) and impaired germinal center formation (Fig. 6). Moreover, T cells derived from the transgenic mice inefficiently stimulated B cell proliferation, regardless of whether B cells were derived from wild-type or the transgenic mice (Fig. 4A). Overexpression of core 2 O-glycans on various surface glycoproteins of T cells apparently renders T-B interaction inefficient, resulting in reduced humoral immune responses in B cells.
B cell proliferation is induced by T-B interaction, which is mainly mediated by CD40L and CD40 (29 -31, 38 -40). Binding of CD40L, a member of the tumor necrosis factor gene family, to CD40, a protein related to tumor necrosis factor receptor, is thought to trigger signal transduction in B cells (43). A soluble CD40-IgG fusion protein was shown to reduce T cell-dependent B cell proliferation by inhibiting the interaction between T and B cells (39). No significant difference was detected in the ex-pression levels of CD40L and CD40 induced in vitro on CD4 ϩ T cells derived from the transgenic and control mice (Fig. 3). This observation was supported by the fact that anti-CD40 antibody treatment induced almost identical proliferation responses in B cells from the transgenic and control mice (Fig. 4B). CD40L expression is very rapid and transient. CD40L expression reaches maximal levels 6 -8 h after activation and then returns to resting levels in 24 h (40). In contrast, B cell proliferation stimulated by T cells from the transgenic mice, assayed after a 72-h incubation, was decreased compared with that stimulated by T cells from control mice (Fig. 4A). CD40L contains few if any O-glycans (44). Core 2 O-glycans overexpressed on various glycoproteins on the T cell surface interfere with the interaction between T and B cells most likely in later stages, resulting in reduced T cell-dependent B cell proliferation (Fig. 4). These results, taken together, suggest that T cell activation in early stages is not impaired, but T-B interaction in later stages, which may be mediated by other than CD40-CD40L interaction, is impaired in the transgenic mice.
Previously, CD40L-deficient mice were shown to lack completely the ability to undergo immunoglobulin isotype switching in response to KLH (31). Similarly, B cells from CD40deficient mice were unable to undergo isotype switching in response to thymus-dependent antigens (30,42), and no germinal centers were found in CD40L or CD40-deficient mice (30,31,42). In CD40L-and CD40-deficient mice, the impairment in immune responses such as isotype switching or germinal center formation is complete. In contrast, delayed immunoglobulin isotype switching and reduced formation of germinal centers were observed in the transgenic mice overexpressing core 2 O-glycans (Figs. 5 and 6). The results strongly suggest that the overexpression of core 2 O-glycans interferes with the interaction between T and B cells and CD40L-CD40 interaction is reduced but not abolished in the transgenic mice.
IL-4 and IFN-␥ produced in activated T cells also play critical roles in B cell differentiation and immunoglobulin isotype switching (41). In Lck-mC2GnT transgenic mice, both in vivo and in vitro T cell activation was reduced, resulting in reduction of IL-4 and IFN-␥ production (20). However, the difference in T cell-dependent B cell proliferation between the transgenic and control mice was observed even after addition of IL-4 ( Fig.  4A). We conclude that overexpression of core 2 O-glycans on T cell surface glycoproteins results in impaired T-B cell interaction and reduced production of cytokines, leading to impaired humoral immune responses in Lck-mC2GnT transgenic mice.
The results obtained in this present study are consistent with those of other reports demonstrating the effect of glycosylation on cell-cell interaction. In particular, cell adhesion of A375 melanoma cells was disrupted after cells were induced to express ASGP-1 mucin glycoprotein, a protein similar to leukosialin (CD43) (45). In contrast, it was reported that T cells derived from the N-acetylglucosaminyltransferase V null mice produced hyperimmune responses to T cell mitogens (46). Since the side chain synthesized by N-acetylglucosaminyltransferase V is further modified to produce poly-N-acetyllactosamines, the results strongly suggest that bulky N-glycans are essential to weaken unnecessary hyperimmune responses. These, and the findings presented here, indicate that the complexity and amount of O-glycans and N-glycans play critical roles in cell adhesion.
Leukocytes from patients with immunodeficient syndromes such as Wiskott-Aldrich syndrome (WAS) (13,14,47) and AIDS (16 -18) show increased activity of C2GnT and levels of core 2 O-glycans. C2GnT activity of peripheral blood lymphocytes from WAS patients (0.4 nmol/h/mg of protein) is 8-fold higher than that of normal individuals (0.05 nmol/h/mg of protein) FIG. 5. Isotype switching is impaired in Lck-mC2GnT transgenic mice. Mice were immunized with KLH in complete Freund's adjuvant (100 g, intraperitoneally), boosted 14 days later, and bled at the indicated time after the first injection. Mouse anti-KLH IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA levels were determined by isotype-specific ELISA. Sera were diluted 1:10,000 to assay IgM and IgG1 and diluted 1:1000 to assay IgG2a, IgG2b, IgG3, and IgA. Open symbols, control mice; closed symbols, Lck-mC2GnT transgenic mice. Each symbol represents three different sets of experiments using three different controls and transgenic mice. The experiments were independently repeated three times using offspring derived from Lck-mC2GnT-1 (four females and three males) and Lck-mC2GnT-2 (two females), respectively. A representative result from one experiment is shown here. The differences observed in the amount of IgM at 7 days and IgG1, IgG2a, and IgG2b at 20 days obtained in three separate experiments were statistically significant (p Ͻ 0.05). (13). This difference is similar to that seen in peripheral blood lymphocytes of the Lck-mC2GnT transgenic mice (0.46 nmol/ h/mg of protein) compared with control mice (0.09 nmol/h/mg of protein) (20). Moreover, it has been shown that immunoglobulin isotype switching and IL-4 and INF-␥ production are impaired in WAS patients (48,49). These results suggest that overexpression of core 2 O-glycans on cell surface glycoproteins may be a cause of impaired B cell immune responses in WAS patients. The gene defective in WAS has been identified (50), and that gene product interacts with both the cytoskeleton and intracellular signaling molecules (51). Our preliminary studies indicate that introduction of a normal WAS gene into lymphocytes from WAS patients results in dramatically reduced levels of core 2 oligosaccharides. 2 In future studies, it will be important to determine whether the mutated WAS gene promotes constitutive activation of the C2GnT gene, resulting in aberrant expression of core 2 O-glycans in T cells of those patients.  D, F, and H). Sections seen in A and B were prepared from a different mouse than those in C and D. The sections in E and F were also prepared from different transgenic lines than those in G and H. Slides were evaluated by a Zeiss Axioplan fluorescence microscope and photographed. Bar ϭ 50 m.

TABLE II Number of germinal centers in spleens from control and
Lck-mC2GnT transgenic mice. Spleen flat sections (5 m thick), taken from mice 7 days after immunization with KLH, were stained with biotinylated peanut agglutinin, followed by PE-streptavidin. Parallel sections were stained with biotinylated anti-mouse IgM, followed by PE-streptavidin. The slides were evaluated by a Zeiss Axioplan fluorescence microscope, and the numbers of germinal centers per section were counted. (mean Ϯ S.E.; p Ͻ 0.05).