Small changes in lymphocyte development and activation in mice through tissue-specific alteration of heparan sulphate

Abstract

We have examined the role of heparan sulphate in lymphocyte development and activation in mice by conditionally deleting the genes encoding the heparan sulphate biosynthetic enzymes N-deacetylase/N-sulphotransferase-1 and -2 (Ndst1 and Ndst2) and glucuronic acid/N-acetylglucosamine co-polymerase-1 (Ext1) in T cells and B cells, respectively. Ndst1 and Ndst2 are the only Ndst isoforms in T cells. In T-cell Ndst-deficient mice there were normal ratios of CD4⁺/CD8⁺ cells in the blood, spleen and thymus, indicating no dramatic effect on development. However, Ndst-deficient T cells were hyperresponsive to low-level activation, suggesting that cell surface heparan sulphate plays a role in T-cell proliferation. The hyperresponsive state correlated with a decrease in cell surface heparan sulphate that occurs in response to activation in wild-type cells. There was a slight change in the number of developing B cells in B-cell Ext1-deficient mice, but the alteration did not cause a change in antibody production. These findings demonstrate that cell surface heparan sulphate may not play a crucial role in lymphocyte development, but can modulate the sensitivity of T cells to activation.

Introduction

Lymphocytes, i.e. T and B cells, are involved in the adaptive arm of the immune system. Lymphocyte development and activation upon antigen challenge proceed through a series of well-defined intermediates characterized by expression of specific cell surface molecules. One class of cell surface proteins is comprised of proteoglycans containing heparan sulphate chains. Cell surface heparan sulphate proteoglycans bind growth factors and cytokines, and act as coreceptors for signalling complexes with conventional signalling receptors and/or as a cell surface reservoir for these factors. They also facilitate adhesion of cells to extracellular matrix proteins.

Heparan sulphate is a type of glycosaminoglycan characterized by alternating uronic acid and d-glucosamine units. After the formation of the linkage region with a serine residue in a core protein, polymerization of the alternating glucuronic acid and N-acetylglucosamine residues occurs through the action of an enzyme complex composed of glucuronic acid/N-acetylglucosamine co-polymerase-1 (Ext1) and Ext2. The chains undergo a series of enzymatically catalysed modifications that include N-deacetylation and N-sulphation of clusters of N-acetylglucosamine residues, epimerization of uronic acids and variable O-sulphation at multiple positions.¹ Four N-deacetylase/N-sulphotransferase isozymes (Ndst1–4) exist that vary in the relative ratios of N-deacetylase and N-sulphotransferase activity and expression across different tissues. The composition of the heparan sulphate chain varies among different tissues. The arrangement of the sulphated sugars and uronic acid epimers creates binding sites for various ligands such as cytokines, chemokines, morphogens, and matrix proteins.

The importance of heparan sulphate proteoglycans on lymphocytes has been only partly elucidated, and their expression on T cells is a matter of debate. Reverse transcriptase–polymerase chain reaction (RT-PCR) analysis showed that T cells express Ndst1 and Ndst2, the proteoglycan syndecan-1 (CD138)^2,3 that mediates cell adhesion,⁴ and heparan sulphate on the cell surface. However, other studies failed to demonstrate heparan sulphate and found only very low levels of syndecan or heparan sulphate.^5,6 Although chondroitin sulphate is the more abundant proteoglycan in B cells,⁷ syndecan-1 is found on pro-B cells and plasma cells,⁸ syndecan-4 is expressed on B-cell lines representing most developmental stages,⁹ and isoform v3 of CD44 (phagocytic glycoprotein-1, CD44-HS) is expressed on activated B cells where it enhances integrin-mediated adhesion.¹⁰ The function of heparan sulphate proteoglycans on lymphocytes is unknown but may include roles in cell differentiation, binding and sequestration of cytokines, such as interleukin (IL)-21,¹¹ and cell activation by antigens.

Systemic knockouts of heparan sulphate biosynthetic enzymes in mice have not been informative about the role of heparan sulphate on lymphocytes. Ext1 deficiency results in embryonic lethality at E7 as a result of defects in mesodermal differentiation.¹² Systemic inactivation of Ndst1 produces reduced N-acetylglucosamine N-deacetylation and N-sulphation of heparan sulphate in most tissues and animals succumb perinatally as a result of lung failure or cerebral hypoplasia.¹³ In contrast, Ndst2 null mice are viable and exhibit selective defects in connective tissue-type mast cells, with no reported effects on B cells or T cells.¹⁴ CD44-deficient mice are normal with respect to B-cell subsets and in vitro responses to B-cell stimulation,¹⁵ but they show reduced T-cell apoptosis in a model of hepatitis.¹⁶ Syndecan-1 deficient mice show defects in wound healing but have been reported to be otherwise normal.¹⁴

In order to study the effects of disrupting heparan sulphate biosynthesis in lymphocytes without the complications described above, we used the Cre-loxP system to inactivate Ndst1 and Ndst2 (the only expressed Ndst isoforms) in T cells and Ext1 in B cells. This approach permitted investigation of lymphocyte heparan sulphate functions in an otherwise normal environment. The mutations caused dramatic changes in heparan sulphate, but had relatively mild effects on the development and response of the immune system.

Materials and methods

Mice

Wild-type C57BL/6 animals were ordered from Jackson Laboratories (Bar Habor, ME). C57BL/6 mice expressing lckCre (J. D. Marth, University of California, San Diego, CA), mice bearing a loxP-flanked allele of Ndst1 (Ndst1^f/f),¹⁷ and Ndst2^−/− mice (L. Kjellen, University of Uppsala, Uppsala, Sweden)¹⁴ were crossed to obtain Ndst1^f/f lckCre⁺; Ndst2^−/−lckCre⁺ and lckCre⁻ littermates were used in most experiments. C57BL/6 mice expressing CD19Cre were crossed with Ext1^f/f mice¹⁸ to generate Ext1^f/f CD19Cre⁺ and Ext1^f/f CD19Cre⁻ littermates. All mice were obtained in the expected Mendelian ratios and grew normally, showing no obvious abnormalities. They were killed by cervical traction while under isoflurane anaesthesia. The mice were housed under barrier conditions in approved vivaria. They were weaned at 3 weeks, maintained on a 12-hr light–dark cycle, and fed water and standard rodent chow ad libitum. Mouse handling was in accordance with protocols for the humane treatment of animals approved by the Institutional Animal Care and Use Committee (IACUC) and Animal Subjects Committee at the University of California, San Diego.

Genotyping

Genotyping of mice was performed by PCR. The primers were: Cre (forward: 5′-ACGTTCACCGGCATCAACGT-3′; reverse: 5′-CTGCATTACCGGTCGATGCA-3′), Ndst1 (forward: 5′-CCAGGGCGTCAGGGCCTCCTG-3′; reverse: 5′-CATCCTCTGAGGTGACCGC-3′) and Ext1 (forward: 5′-GGAGTGTGGATGAGTTGAAG-3′; reverse: 5′-CAACACTTTCAGCTCCAGTC-3′).

T-cell isolation

The spleen and thymus were removed and placed in phosphate-buffered saline (PBS) with 1% fetal calf serum (FCS; Atlanta Biologicals, Lawrenceville, GA) and 10 mm HEPES. The organs were mashed between two sterile slides and resuspended in ACK lysis buffer [ammonium chloride, potassium bicarbonate and ethylenediaminetetraacetic acid (EDTA); Fisher Scientific, Pittsburgh, PA] for 2 min to lyse red blood cells. The sample was filtered through a 40-μm cell strainer (BD Falcon, Franklin Lakes, NJ) and T cells in the filtrate were isolated using anti-CD90 antibody-coated paramagnetic beads as described by the manufacturer [magnetic-activated cell sorting (MACS); Miltenyi Biotec, Auburn, CA]. T cells were isolated from whole blood collected by tail vein incision as described above.

B-cell isolation

B cells were isolated from peritoneal lavage fluid (10 ml of sterile PBS) or from peripheral blood obtained by cardiac puncture using EDTA as an anticoagulant. In some experiments, cells were isolated from spleen or from bone marrow that was obtained by flushing the femurs, tibiae and humeri with sterile buffer [Hank’s balanced salt solution (HBSS) with 1% FCS and 10 mm HEPES]. B cells were purified using anti-B220 antibody-coated paramagnetic (MACS) beads.

Lymphocyte subpopulations

Total lymphocyte counts in anticoagulated peripheral blood were determined using an automated haematology instrument calibrated for mouse samples. The T- and B-cell percentages of total lymphocytes were determined by flow cytometry using anti-CD3-phycoerythrin (PE) and anti-B220-fluorescein isothiocyanate (FITC) monoclonal antibodies (Pharmingen, BD Biosciences, San Diego, CA). T cells from blood, spleen and thymus were stained for flow cytometric analysis with anti-CD4-FITC and anti-CD8-PE monoclonal antibodies (Invitrogen, Carlsbad, CA). B cells were stained for flow cytometric analysis with various combinations of the following monoclonal antibodies: anti-CD43-FITC, anti-IgD-FITC, anti-IgM-PE, and anti-B220-FITC or -peridinin chlorophyll protein (PerCP) (Pharmingen). Binding to Fc receptors was blocked by pre-incubation with unlabelled anti-CD16/anti-CD32 antibody (Pharmingen). All antibodies were titrated to determine working concentrations. Flow cytometry was performed on a FACSCalibur flow cytometer (Becton Dickinson, San Diego, CA) and the data were analysed using the cellquest software (Becton Dickinson).

Staining of cell surface heparan sulphate

The relative expression of heparan sulphate on the surface of cells was determined by binding of recombinant basic fibroblast growth factor-2 (FGF-2) using flow cytometry. Isolated T cells and B cells were incubated with biotinylated FGF-2 in Hams F12 medium (Invitrogen) with 0·5% bovine serum albumin (BSA; Sigma, St Louis, MO) for 1 hr with shaking at 4°. Cells were washed twice in PBS and incubated for 20 min with shaking at 4° in PBS containing streptavidin–allophycocyanin (APC). Samples where FGF-2 was omitted were included as negative controls. To show that FGF-2 binding was related to heparan sulphate expression, controls were also performed after pretreatment of T and B cells with heparin lyases I, II and III (Seikagaku, East Falmouth, MA) for 4 hr at 37°. Cell populations were gated using cell lineage-specific antibody staining.

T-cell proliferation after activation

Twenty-four-well tissue culture plates (BD Falcon, Franklin Lakes, NJ) were coated with goat anti-hamster immunoglobulin G (IgG) (Sigma) at 37° for 2 hr and washed with PBS. Hamster anti-mouse CD3 antibody (Pharmingen) was added to the wells at increasing concentrations (25–800 ng/ml), and the plates were incubated at 37° for 1 hr. To measure cell proliferation, isolated T cells (10⁷ cells/ml) were incubated at 37° for 10 min in prewarmed PBS containing 2·5 mm carboxy-fluorescein diacetate succinimidyl ester (CFSE) (Invitrogen). The cells were washed in PBS and transferred to the anti-CD3 coated plates, to activate and induce proliferation of the cells. CFSE is a cytoplasmic dye that becomes more diffuse with each cell division. After 2 days of incubation at 37°, the cells were harvested and flow cytometry was performed. Supernatants from activated T cells were tested for IL-2 production using a commercial enzyme-linked immunosorbent (ELISA) kit (eBioscience, San Diego, CA).

B-cell syndecan-1 expression

The expression of syndecan-1 on B-cell subpopulations was determined using flow cytometry. The cells were incubated with a rat anti-syndecan-1 antibody (Research Diagnostics, Flanders, NJ) for 20 min at 37°. Binding was detected in a two-step procedure, using a biotinylated rabbit anti-rat antibody (Pharmingen) followed by streptavidin–APC. The B-cell subpopulations were gated using antibody staining as indicated above. Samples where anti-syndecan-1 antibody was omitted were used as controls for non-specific binding.

B-cell responses

Littermate pairs of Ext1^f/f CD19Cre⁺ and Ext1^f/f CD19Cre⁻ mice were immunized with dinitrophenol conjugated either to keyhole limpet haemocyanin (DNP–KLH; Biosearch Technologies, Novato, CA) to test T-dependent B-cell responses or to Ficoll (DNP–Ficoll; Biosearch Technologies) to test T-independent B-cell responses. The mice were given metoxyflurane anaesthesia (Medical Developments International, Springvale, Victoria, Australia) and all blood samples were obtained from tail veins. After a pre-immune blood sample had been obtained, 10 μg of DNP–KLH in 100 μl of Freund’s complete adjuvant (Sigma) was given subcutaneously in the first series of experiments (n = 7 littermate pairs). Samples were drawn after 7, 14 and 28 days, and a booster dose of 10 μg of DNP–KLH in 100 μl of Freund’s incomplete adjuvant (Sigma) was given immediately after the 28-day sample had been taken. A final sample was drawn on day 35, after which the mice were killed. In a second series of experiments (n = 5 littermate pairs), immunizations were given intraperitoneally, samples were drawn on days 0, 14, 21 and 28, and the booster dose was given on day 21. To measure T-cell-independent responses, mice were immunized with 10 μg of DNP–Ficoll in 100 μl of sterile PBS subcutaneously in the first series of experiments (n = 7 littermate pairs). Samples were drawn after 7 and 14 days, and the mice were killed. The second series was performed similarly (n = 5 littermate pairs), except that the immunizations were given intraperitoneally. Serum was kept at −20° and analysed in one batch.

Quantification of DNP-specific antibodies was performed in enzyme immunoassays. The plates were coated with 200 μg/ml DNP–BSA (Biosearch Technologies) and blocked with 10% FCS in PBS. Serum was diluted 1/200 for measurement of IgM (all samples), 1/1000 for measurement of IgG1 (DNP–KLH-immunized mice), and 1/800 for measurement of IgG3 (DNP–Ficoll-immunized mice). Detection was performed using alkaline phosphatase-conjugated goat anti-mouse isotype-specific antibodies (IgM, IgG1 or IgG3; Southern Biotechnology Associates, Birmingham, AL). The substrate was p-nitrophenylphosphate (Southern Biotechnology Associates), and the reactions were stopped after 20 min using 2 m NaOH. Optical density at 405 nm was read spectrophotometrically, and the results are given as mean optical density of duplicate wells.

Splenic histology

Histological analyses were performed by the Cancer Center Histology Core at the University of California, San Diego. Frozen sections were cut and stained for T cells with anti-CD3 (Abcam, Cambridge, MA), for macrophages with anti-F4/80 (Abcam), or for B cells with anti-B220 (R&D Systems, Minneapolis, MN). Measurement of pixel density was performed on spleen sections from three wild-type and three Ndst1^f/f lckCre⁺; Ndst2^−/− mice, calculating the follicle areas as percentage of the total area (three different areas in each specimen).

Statistics

Because of small data sets and non-normal distribution of variables, non-parametric statistics were used. Data are given as the median with 95% non-parametric confidence interval (six or more observations) or range (less than six observations) in parentheses. Differences between littermate pairs were assessed using the Mann–Whitney U-test. Differences in spleen histology and differences in activation responses in isolated T cells were analysed with the Kruskal–Wallis test. The antibody responses after immunization were analysed using two-way analysis of variance for repeated measures, employing logarithmic or rank transformation if necessary to obtain an appropriate model fit.

Results

Conditional inactivation of biosynthetic enzymes in heparan sulphate biosynthesis

Cells make heparan sulphate via a copolymerase encoded by the genes Ext1 and Ext2. As the chains polymerize they undergo various modifications by sulphation and epimerization. These modifications depend on the action of Ndst isozymes, which creates the substrate for further modifications. To study heparan sulphate in T cells, we crossed Ndst1^f/f; Ndst2^−/− mice with lckCre mice, in which Cre is activated early in T-cell development. We chose to delete Ext1 in B cells by cross-breeding Ext1^f/f mice¹⁸ with mice bearing the Cre recombinase under the control of the CD19 promoter, which results in activation of the recombinase in late pro-B cells.^19,20

To establish the extent of inactivation of Ndst1 in T cells and Ext1 in B cells, we analysed the binding of FGF-2 to plasma membrane heparan sulphate. Binding depends on the number of chains as well as their degree of sulphation.²¹ Both cell types bound FGF-2. Binding was abolished in wild-type T-cell controls and greatly reduced in wild-type B-cell controls after treatment with heparin lyases I, II and III, confirming that the assay indirectly measures heparan sulphate expression (Fig. 1a,c). Cre expression in Ndst1^f/f; Ndst2^−/− T cells (Fig. 1b) and Ext1^f/f B cells (Fig. 1d) greatly diminished FGF-2 binding, reducing it to a level seen in control cells incubated in the absence of FGF-2 but with FITC-streptavidin. The strongly reduced binding of FGF-2 indicated that the extent of gene inactivation was sufficient to cause a dramatic decrease in heparan sulphate expression.

Figure 1.

Disruption of heparan sulphate biosynthesis in T cells and B cells. Splenic lymphocytes of different genotypes were incubated with biotinylated fibroblast growth factor-2 (FGF-2), stained with phycoerythrin (PE)-Cy5-streptavidin and analysed by flow cytometry. Cre expression dramatically reduced binding of FGF-2 to both T cells and B cells. Data are representative of three similar experiments. (a) Wild-type T cells. Grey shaded curve, without FGF-2; black line, with FGF-2; grey line, with FGF-2 in cells pretreated with heparin lyases I, II and III. (b) T cells. Grey shaded curve, Ndst1^f/f lckCre⁻; Ndst2^−/− T cells without FGF-2; black line, Ndst1^f/f lckCre⁻; Ndst2^−/− T cells with FGF-2; grey line, Ndst1^f/f lckCre⁺; Ndst2^−/− T cells with FGF-2. (c) Wild-type B cells. Grey shaded curve, without FGF-2; black line, with FGF-2; grey line, with FGF-2 in cells pretreated with heparin lyases I, II and III. (d) B cells. Grey shaded curve, Ext1^f/f CD19Cre⁻ B cells without FGF-2; black line, Ext1^f/f CD19Cre⁻ B cells with FGF-2; grey line curve, Ext1^f/f CD19Cre⁺ B cells with FGF-2.

Alterations in T-cell development

Analysis of T cells in thymus, peripheral blood, and the spleen did not reveal significant differences in CD4⁺ and CD8⁺ subsets in littermate pairs of Ndst1^f/f lckCre⁻; Ndst2^−/− and Ndst1^f/f lckCre⁺; Ndst2^−/− mice (Table 1). Examination of spleen histology for structural abnormalities (n = 3) showed that wild-type spleens had defined lymphoid follicles with B cells (Fig. 2a) and periarteriolar lymphoid sheath (PALS) containing T cells (Fig. 2b) surrounded by areas dense with macrophages (Fig. 2c). The Ndst1^f/f lckCre⁺; Ndst2^−/− spleens showed a reduction in follicular size (Fig. 2d), with displacement of T cells into the B-cell areas (Fig. 2e). An increased density of macrophages was also noted in the intrafollicular spaces (Fig. 2f). Using densitometry, the follicle areas were found to cover 54·8% (46·9–65·0%) of the spleen sections in the wild-type spleens and 37·9% (23·3–46·6%) in the Ndst1^f/f lckCre⁺; Ndst2^−/− spleens (P < 0·05).

Table 1.

T-cell counts in Ndst1^f/f lckCre⁺; Ndst2^−/− and Ndst1^f/flckCre⁻; Ndst2^−/− mice¹

	Ndst1f/flckCre+; Ndst2−/−	Ndst1f/flckCre−; Ndst2−/−	P-value
Thymus
CD4–/CD8– cells (%)	9·7 (6·1–13·3)	9·6 (5·6–13·7)	1·00
CD4+/CD8+ cells (%)	76·9 (73·3–80·5)	76·1 (73·6–78·7)	1·00
CD4+ cells (%)2	9·7 (9·4–10·1)	10·0 (9·5–10·4)	0·44
CD8+ cells (%)2	3·7 (3·4–4·0)	4·3 (3·3–5·3)	1·00
CD4+/CD8+ ratio2	2·65 (2·32–2·97)	2·42 (1·96–2·89)	1·00
Peripheral blood
CD4+ cells (%)	12·5 (8·0–14·3)	12·4 (9·5–15·1)	1·00
CD8+ cells (%)	6·8 (4·2–8·8)	7·1 (4·7–8·2)	0·83
CD4+/CD8+ ratio	1·81 (1·54–2·16)	1·89 (1·55–2·06)	0·83
Spleen
CD4+ cells (%)	10·6 (4·5–16·7)	14·9 (9·6–20·1)	0·44
CD8+ cells (%)	5·5 (3·2–7·9)	7·2 (5·3–9·1)	0·44
CD4+/CD8+ ratio	1·76 (1·41–2·12)	2·01 (1·81–2·21)	1·00

¹n = 10 littermate pairs.

²Single positive cells.

Figure 2.

Splenic follicle architecture in Ndst1^f/flckCre⁺; Ndst2^−/− mice. Spleen sections from wild-type and Ndst1 Ndst2-deficient mice. (a, d) white arrows, B220 stain for B cells. (b, e) white arrows, CD3 stain for T cells. (c, f) white arrows, F480 macrophage stain; black arrows, splenic follicle. Original magnification, ×20. Data are representative of three similar experiments.

T-cell activation and proliferation

To determine whether altering heparan sulphate might affect T-cell activation, T cells from wild-type, Ndst1^f/flckCre⁻; Ndst2^–/– and Ndst1^f/f lckCre⁺; Ndst2^−/− animals were activated by anti-CD3 monoclonal antibody and proliferation was measured by CFSE staining. At 0·8 and 0·1 mg/ml anti-CD3, T cells from mutant mice proliferated to the same extent as the wild type (Fig. 3). At lower antibody doses, however, the Ndst-deficient T cells were more sensitive than the wild-type cells, as shown by the increase in proliferation. IL-2 production during activation was not affected (data not shown).

Figure 3.

T-cell activation by anti-CD3 in Ndst1^f/f lckCre⁺; Ndst2^−/− mice. Isolated splenic T cells from wild-type (Ndst1^f/f lckCre⁻; Ndst2^+/+), Ndst1^f/f lckCre⁻; Ndst2^−/− and Ndst1^f/f lckCre⁺; Ndst2^−/− animals were stimulated with varying concentrations of anti-CD3 for 3 days, and proliferation was measured as a decrease in carboxy-fluorescein diacetate succinimidyl (CFSE) staining. The data represent proliferating cells as a percentage of the total cell population. Statistical comparisons (Kruskal–Wallis test): 0·8 μg/ml anti-CD3: P = 0·60; 0·1 μg/ml anti-CD3: P = 0·19; 0·05 μg/ml anti-CD3: wild-type versus Ndst1^f/f lckCre⁺; Ndst2^−/−: P < 0·01; 0·025 μg/ml anti-CD3: P = 0·096.

To examine the effect of T-cell heparan sulphate on normal T-cell proliferation, we incubated cells with CFSE. Unstimulated T cells took up CFSE, but did not divide (Fig. 4). After activation by anti-CD3, multiple rounds of proliferation occurred, resulting in decreased amounts of CFSE. The population of activated T cells showed a reduced ability to bind to FGF-2. Analysis of the five individual subsets of cells showed that heparan sulphate expression decreased as the cells proliferated (Fig. 4).

Figure 4.

T-cell heparan sulphate expression during in vitro activation. Isolated splenic T cells were stimulated with anti-CD3, and stained with carboxy-fluorescein diacetate succinimidyl (CFSE) and fibroblast growth factor-2 (FGF-2) to measure proliferation and cell surface heparan sulphate. Upper left panel: Black line, isolated T cells stained with CFSE showing five different proliferative subsets; grey shaded area, CFSE stain in non-activated control T cells. Remaining panels: FGF-2 binding in individual subsets of activated T cells as indicated by CFSE staining in upper left panel. Grey shaded curves, Ndst1^f/f lckCre⁻; Ndst2^−/− T cells without FGF-2; black line curves, individual populations of activated Ndst1^f/f lckCre⁻; Ndst2^−/− T cells with FGF-2. Data are representative of three similar experiments.

Alterations in B-cell development

All subpopulations of B cells expressed heparan sulphate based on FGF-2 binding, and expression of Cre resulted in 50–85% loss of expression except in pro B cells (Fig. 5a). This was not surprising as CD19 is expressed as cells transition from pro-B to pre-B cells and therefore the extent of recombination would be lower in the pro-B cells. The percentage of bone marrow pro-B cells was statistically higher in the mutant by ∼25% (P = 0·018, Table 2). In spite of the dramatic change in B-cell heparan sulphate at later stages of development, the distribution of B-cell subsets was only mildly affected in Ext1^f/f CD19Cre⁺ animals, with the percentage of splenic marginal and transitional B cells was somewhat lower (Table 2). These changes in FGF-2 binding were not related to changes in expression of one of the major core proteins of heparan sulphate-bearing proteoglycans found in B cells, syndecan-1 (Fig. 5b).

Figure 5.

B-cell fibroblast growth factor-2 (FGF-2) binding and syndecan-1 expression in Ext1^f/f CD19Cre⁺ mice. (a) FGF-2 binding to B cells (pro-B, pre-B, immature B and mature B cells), peripheral blood, and spleen [transitional (T) and marginal (M) B cells, and follicular (F) B cells] in Ext1^f/f CD19Cre⁺ and Ext1^f/f CD19Cre⁻ mice (n = 3 littermate pairs). All subsets were affected with the exception of pro-B cells. (b) Syndecan-1-expressing B cells as percentages of total B cells at a similar developmental stage for bone marrow, peripheral blood, and spleen. Grey bars, Ext1^f/f CD19Cre⁺ mice; white bars, Ext1^f/f CD19Cre⁻ mice (n = 2 littermate pairs).

Table 2.

B-cell counts in Ext1^f/f CD19Cre⁻ and Ext1^f/f CD19Cre⁺ mice¹

	Ext1f/fCD19Cre+	Ext1f/fCD19Cre−	P-value
Bone marrow
Pro-B cells (%)	10·4 (8·9–15·3)	8·4 (6·4–10·2)	0·018
Pre-B cells (%)	25·2 (16·4–32·5)	26·6 (19·3–31·8)	0·50
Immature B cells (%)	18·0 (14·5–21·4)	19·6 (14·6–24·4)	0·27
Mature B cells (%)	7·8 (6·0–9·7)	7·5 (5·2–10·1)	0·67
Peripheral blood
B cells (× 109/l)	4·6 (3·8–5·2)	5·2 (4·6–6·2)	0·06
T cells (× 109/l)	4·0 (3·1–4·8)	4·1 (3·6–4·9)	0·51
Spleen
Marginal and transitional B cells (%)	6·1 (4·6–8·3)	8·4 (6·0–11·3)	0·043
Follicular B cells (%)	21·7 (17·1–26·3)	19·7 (17·3–23·1)	0·24

¹n= 11 littermate pairs.

Significant P-values are given in bold.

B-cell activation

After subcutaneous injection of the T-dependent antigen DNP–KLH there was a significant increase in DNP-specific IgM and IgG1 (P < 0·001) (Fig. 6a,b). However, no significant differences were observed between Ext1-deficient and wild-type mice (IgM: P = 0·51; IgG: P = 0·37). Similar results were also obtained after intraperitoneal injection of antigen (Fig. 6c,d; IgM: P = 0·27; IgG: P = 0·28). Subcutaneous immunization with the T-independent antigen DNP–Ficoll produced very small responses (data not shown), but after intraperitoneal immunization there were significant increases in DNP-specific IgM and IgG3 (P < 0·001). Again, no differences between mutant and wild-type groups were observed (Fig. 6e,f; IgM: P = 0·72; IgG3: P = 0·60).

Figure 6.

Antibody responses in Ext1^f/f CD19Cre⁺ mice. (a, b) Dinitrophenol (DNP)-specific immunoglobulin M (IgM) (a) and IgG1 (b) antibody response after subcutaneous immunization with the T-dependent antigen DNP–keyhole limpet haemocyanin (KLH) (n = 7 littermate pairs, medians with 95% confidence intervals). (c, d) DNP-specific IgM (c) and IgG1 (d) antibody response after intraperitoneal immunization with DNP–KLH (n = 5 littermate pairs, medians and ranges). (e, f) DNP-specific IgM (e) and IgG3 (f) antibody response after intraperitoneal immunization with the T-independent antigen DNP–Ficoll (n = 5 littermate pairs, medians and ranges). Open circles, Ext1^f/f CD19Cre⁺ mice; filled squares, Ext1^f/f CD19Cre⁻ mice.

Discussion

T-cell biology

T-cell development was normal in the Ndst1^f/f lckCre⁺; Ndst2^−/− animals, but the distribution of cells in the spleen was altered, with T cells invested into B-cell areas of splenic lymphoid follicles in the mutant mice. Thus, T-cell heparan sulphate may play a role in defining cellular compartments in splenic follicles, perhaps by organizing growth factors and extracellular matrix proteins to form trafficking regions. Cell surface integrins have been proposed to regulate the spatial organization of T cells within lymphoid organs,²² and integrins are known to interact with cell surface heparan sulphate.²³ Another possible role is in the interaction between heparan sulphate and chemokines. CCL19 and CCL21 are involved in attracting naïve T cells into peripheral lymph node T-cell areas and retaining them there,²⁴ and splenic follicles may use a similar mechanism. Almost all known chemokines bind heparin or heparan sulphate in vitro.²⁵

Genetically altered T cells showed a hyperresponsive proliferative response to low levels of activation. This could be a result of the interaction between cell surface heparan sulphate and IL-2. Cell surface heparan sulphate may bind IL-2 and sequester this cytokine away from the IL-2 receptor as part of normal regulation of the immune response. Wild-type T cells down-regulated cell surface heparan sulphate in response to proliferative signals, which would facilitate rapid clonal expansion. Removal of cell surface heparan sulphate in the genetically altered T cells would make them more sensitive to IL-2, as reduced heparan sulphate expression essentially recapitulates the cell response to activation and creates a ‘feed-forward’ effect. This proposed mechanism of heparan sulphate sequestration of a ligand from a receptor has been demonstrated for the growth factor Wnt in fibroblasts.²⁶

The hypersensitivity of heparan sulphate-deficient T cells may implicate heparan sulphate in autoimmunity. Autoimmune diseases may arise from overactive or especially sensitive T cells. Other T-cell-specific changes in cell surface glycosylation have led to autoimmune phenotypes. The Mgat5 (an enzyme involved in N-acetylglucosamine transfer)-deficient animal developed kidney autoimmune disease, possibly as a result of lowering T-cell activation thresholds.²⁷ It is possible that these T-cell-specific, heparan sulphate-deficient animals may also show autoimmune phenotypes.

B-cell biology

Only minor changes in B-cell subpopulations and no significant changes in antibody responses to T-dependent or T-independent antigens were observed. These findings are somewhat surprising as differentiation of early B-cell stages in the mouse are dependent on stromal contact and cytokines, especially stem cell factor, IL-7, and Flt-3 ligand.^28,29 All of these factors bind to heparan sulphate, and pro-B cells lacking plasma membrane heparan sulphate as a result of heparinase treatment show a 50–75% reduction in IL-7-driven proliferation in parallel with a substantial reduction in their ability to bind biotinylated IL-7³⁰. Although the extent of removal of heparan sulphate afforded by the mutation would appear to be sufficient to reproduce the effect of heparinase treatment, apparently the dependence on heparan sulphate is not as great as predicted from enzyme digestion. Perhaps this reflects the activity of tissue heparan sulphate in the bone marrow, thymus, and spleen which is abundant in the extracellular matrix of these tissues.

It has been estimated that about 90% of the immature B cells produced in the bone marrow are lost as a result of selection against self-reactive cells.³¹ It is not known whether heparan sulphate is involved in this selection, which is strongly dependent on the B-cell receptor. This selection process may have favoured the more normal B cells in our experimental mouse, as a larger proportion of the B cells in the peripheral blood than of the immature or mature B cells in the bone marrow were able to bind FGF-2 (Fig. 5). Even so, there was a tendency towards lower B-cell counts in peripheral blood of the mice with Ext1 gene inactivation (Table 2). Because we did not measure bone marrow B-cell turnover and apoptosis in our mice, it is not known whether there may have been a compensatory enhancement in the production of B-committed precursor cells. The finding of an increased percentage of pro-B cells in the Ext-1^f/f CD19Cre⁺ mice may support this hypothesis.

Antibody responses to immunizations were similar in the mice with Ext1 gene inactivation and control littermates. There may have also been a selection of functional cells at this stage, where normally only a very limited number of activated cells are necessary to give rise to large clones of antibody-producing plasma cells. Upon antigen stimulation, human tonsillar B cells show a strong transient increase in the specific heparan sulphate-containing isoform of CD44.¹⁰ This strongly promotes the binding of hepatocyte growth factor through a heparan sulphate-dependent mechanism, which in turn may strengthen B-cell adhesion, for example to antigen-presenting cells. Mice homozygous for CD44 gene inactivation have normal B cells and show no defects in response to type II collagen immunization.³² However, their B cells show altered migration patterns during inflammation,³³ resulting in reduced responses in a model of autoimmune arthritis.³² As the Ext1 gene-inactivated mice were only challenged by immunization, we cannot exclude the possibility that their responses to infection, perhaps with selected pathogens, may be altered.

Conclusions

In this study, we found that T-cell and B-cell surface heparan sulphate does not play a major role in lymphoid development. There were slight changes in the number of developing B cells, but these did not correlate with a change in antibody production. Heparan sulphate-deficient T cells were hyperresponsive to low-level activation, suggesting that cell surface heparan sulphate plays a role in the proliferation of activated T cells.

Abbreviations

APC

allophycocyanin

CFSE

carboxy-fluorescein diacetate succinimidyl ester

DNP

dinitrophenol

EDTA

ethylenediaminetetraacetic acid

FCS

fetal calf serum

FGF-2

fibroblast growth factor-2

FITC

fluorescein isothiocyanate

HBSS

Hank’s balanced salt solution

IACUC

Institutional Animal Care and Use Committee

KLH

keyhole limpet haemocyanin

Ndst1/2

N-deacetylase/N-sulphotransferases 1 and 2

PBS

phosphate-buffered saline

PE

phycoerythrin

PerCP

peridinin chlorophyll protein