Malignant B Lymphocyte Survival in Vivo CD22 Ligand Binding Regulates Normal

The Journal of Immunology

CD22 Ligand Binding Regulates Normal and Malignant B Lymphocyte Survival In Vivo1

Karen M. Haas, Suman Sen, Isaac G. Sanford, Ann S. Miller, Jonathan C. Poe, and Thomas F. Tedder2

The CD22 extracellular domain regulates B lymphocyte function by interacting with ␣2,6-linked sialic acid-bearing ligands. To understand how CD22 ligand interactions affect B cell function in vivo, mouse anti-mouse CD22 mAbs were generated that inhibit CD22 ligand binding to varying degrees. Remarkably, mAbs which blocked CD22 ligand binding accelerated mature B cell turnover by 2- to 4-fold in blood, spleen, and lymph nodes. CD22 ligand-blocking mAbs also inhibited the survival of adoptively transferred normal (73–88%) and malignant (90%) B cells in vivo. Moreover, mAbs that bound CD22 ligand binding domains induced signiﬁcant CD22 internalization, depleted marginal zone B cells (82–99%), and reduced mature recirculating B cell numbers by 75–85%. The CD22 mAb effects were independent of complement and FcRs, and the CD22 mAbs had minimal effects in CD22AA mice that express mutated CD22 that is not capable of ligand binding. These data demonstrate that inhibition of CD22 ligand binding can disrupt normal and malignant B cell survival in vivo and suggest a novel mechanism of action for therapeutics targeting CD22 ligand binding domains. The Journal of Immunology, 2006, 177: 3063–3073.

D22 is a B cell-specific glycoprotein of the Ig superfam- cell surface CD22, IgM, and MHC class II expression on mature B ily expressed on the surface of maturing B cells coinci- cells, whereas normal BCR signaling and Ca2ϩ mobilization are dent with IgD expression (1, 2). Following either CD22 not dependent upon ligand engagement (24). In agreement with a C Ϫ Ϫ or BCR ligation, the tyrosine-phosphorylated cytoplasmic domain negative signaling role for CD22, B cells from CD22 / mice of CD22 recruits effector molecules that regulate BCR and CD19 generate augmented calcium (Ca2ϩ) responses following BCR signaling (3–11). In addition to regulating cytoplasmic signaling, cross-linking and exhibit an IgMlowMHC class IIhigh phenotype CD22 may regulate cell-cell interactions through its seven extra- that is characteristic of stimulated B cells (25–28). Remarkably, cellular Ig-like domains (4, 12, 13). Specifically, the two mem- however, CD22Ϫ/Ϫ, CD22AA, and CD22⌬1-2 B cells each exhibit ␣ brane-distal Ig-like domains of CD22 bind to 2,6-linked sialic significantly enhanced turnover in vivo (24, 27, 28). Whether in- acid-bearing (sialoglycoconjugate) ligands expressed by hemopoi- creased turnover is due to the absence of CD22 ligand-generated etic and nonhemopoietic cells and serum proteins (13–18). Mole- signals, B cell hyperresponsiveness in CD22Ϫ/Ϫ mice, or devel- cules bearing potential CD22 ligands have been identified in vitro, Ϫ Ϫ opment/maturation defects common to CD22 / , CD22AA, and which include CD22 itself, CD45RO, IgM, members of the Ly-6 CD22⌬1-2 B cells remains unclear. family of glycoproteins, and other structurally diverse proteins and In vitro studies have also suggested roles for CD22 ligand bind- lipids (19–22). CD22 also interacts with itself in cis to form ho- ing in regulating B cell function. Interactions between CD22 and momultimeric complexes (23). Although little is known regarding ␣ the in vivo relevance of CD22 ligand binding either in cis or trans, its 2,6-linked sialoglycoconjugate ligands have been suggested to understanding this process is critical for interpreting CD22 regu- mediate innate self recognition, leading to dampened B cell auto- lation of B cell function. reactivity (29). The sialic acid binding domains of CD22 have also In addition to regulating signal transduction through its cyto- been proposed to negatively regulate BCR signaling (30, 31). The plasmic domain, CD22 regulates B cell development and function Cy34 mAb reactive with the Lyb-8.2 allele of Cd22 does not mea- through ligand-generated signals. Studies with CD22AA and surably impair B cell function in vitro (2, 32–34). However, treat- CD22⌬1-2 mice expressing gene-targeted CD22 molecules that do ment of mice with polyclonal rabbit anti-CD22 IgG reduces the not interact with sialoglycoconjugate ligands suggest that CD22 number of mature recirculating bone marrow B cells, with no other engagement influences the development or maintenance of the effects observed (35). Abs that block human CD22 ligand engage- marginal zone B cell population, BCR-induced proliferation, and ment trigger potent BCR-independent signals in vitro that can lead to apoptosis (36–38). In vivo, a humanized CD22 mAb has also been used clinically in the treatment of patients with B cell ma- Department of Immunology, Duke University Medical Center, Durham, NC 27710 lignancies and autoimmune disease (39, 40) and is currently in Received for publication February 28, 2006. Accepted for publication June 15, 2006. phase-III clinical trials for the treatment of systemic lupus ery- The costs of publication of this article were defrayed in part by the payment of page thematosus (41). This therapeutic mAb binds CD22 outside of its charges. This article must therefore be hereby marked advertisement in accordance ligand binding domains but based on in vitro assays may induce B with 18 U.S.C. Section 1734 solely to indicate this fact. cell depletion in vivo by Ab-dependent cellular cytotoxicity, in- 1 This work was supported by grants from the National Institutes of Health duced CD22 phosphorylation/signal transduction, CD22 internal- (CA96547, CA81776, and AI56363) and a Biomedical Science Grant from the Ar- thritis Foundation. K.M.H. is supported by a Career Development Fellowship Award ization, and/or other pathways (40, 42). Despite all of the above from the Leukemia and Lymphoma Society. studies, the effects of selectively blocking CD22 ligand binding on 2 Address correspondence and reprint requests to Dr. Thomas F. Tedder, Box normal human or mouse B cell survival and function in vivo have 3010, Department of Immunology, Room 353 Jones Building, Research Drive, Duke University Medical Center, Durham, NC 27710. E-mail address: not been assessed, particularly under conditions in which B cell [email protected] development proceeds normally.

The biology of CD22 ligand binding is not only important for Flow cytometry analysis understanding normal B cell regulation, but is critical for eluci- Blood leukocyte numbers were quantified by hemocytometer following red dating the mechanisms through which CD22-directed therapies cell lysis, with cell frequencies determined by immunofluorescence stain- function. To address these issues, a panel of mouse anti-mouse ing with flow cytometry analysis. Single-cell leukocyte suspensions from CD22 mAbs that do not inhibit, or inhibit CD22 ligand binding to spleen, bone marrow (bilateral femur), peritoneal lavage, and peripheral varying degrees, were generated and assessed for their in vivo lymph node (bilateral inguinal) were isolated, and the erythrocytes were lysed in Tris-buffered 100 mM ammonium chloride solution. Leukocytes effects. mAbs that bind CD22 ligand binding domains caused strik- were then stained at 4°C using predetermined optimal concentrations of ing reductions in the blood and marginal zone B cell compart- Abs for 30 min. For whole blood, erythrocytes were lysed after staining ments. Even more remarkable was the finding that CD22 mAbs using FACS Lysing Solution (BD Biosciences). Ab binding was analyzed that blocked CD22 ligand binding significantly inhibited normal on a FACScan flow cytometer (BD Biosciences) by gating on cells with the forward and side light scatter properties of lymphocytes. Nonreactive, iso- and malignant B cell survival within tissues in vivo. Thus, CD22 type-matched Abs (Caltag Laboratories/BD Pharmingen) were used as ligand binding is required for the maintenance of mature B cells in controls for background staining. the periphery. Adhesion and CD22-Fc binding assays Materials and Methods For cellular adhesion blocking experiments, L cells stably transfected with mouse CD22 cDNA (mCD22-L) were preincubated with purified mAbs at Mice 4oC for 30 min, washed twice with DMEM, and incubated with mouse o CD22Ϫ/Ϫ, CD22⌬1-2, and CD22AA mice on a mixed B6/129 background splenocytes for 30 min at 4 C as described previously (14, 18). Cells were were generated as described previously (24). C57BL/6, Fc␥RIIϪ/Ϫ then washed three times with DMEM, and the remaining splenocytes Ϫ/Ϫ tm1Mom bound to CD22-transfected L cells were counted under a microscope. Pa- (B6.129S-Fcgr2tm1Rav), and Rag1 (B6.129S7-Rag1 /J) mice Ϫ were obtained from The Jackson Laboratory. FcR common ␥-chain rental (CD22 ) L cells were used as a negative control. The CD22-Fc Ϫ/Ϫ tm1 (FcR␥)-deficient mice (FcR␥ , B6.129P2-Fcerg1 ) were obtained fusion protein was constructed by fusing the first three NH2-terminal ex- from Taconic Farms. C3Ϫ/Ϫ mice were obtained from Dr. M. Carroll (Cen- tracellular Ig-like domains of murine CD22 to the Fc portion of human ter for Blood Research, Boston, MA) (43). Experiments were performed on IgG1. CD22-Fc was cloned into pIRES-2EGFP (Clontech) and expressed 2- to 3-mo-old mice housed under specific pathogen-free conditions. All in 293T cells. CD22-Fc (0.2 ␮g) was preincubated with CD22 mAbs (2 studies and procedures were approved by the Duke University Animal Care ␮g) for 15 min at room temperature. CD22-Fc-CD22 mAb complexes were 5 Ϫ/Ϫ and Use Committee. then incubated with 2.5 ϫ 10 CD22 splenocytes in 100 ml of DMEM containing 10% FCS for 30 min on ice. CD22-Fc protein binding was Ј ␥ CD22 mAb generation assessed using PE-conjugated F(ab )2 goat anti-human IgG Abs (Fc frag- ment-specific; Jackson ImmunoResearch Laboratories) with flow cytom- The MB22-8, -9, -10, and -11 mAbs were generated by the fusion of NS-1 etry analysis. myeloma cells with spleen cells from CD22Ϫ/Ϫ mice immunized with mouse CD22 cDNA-transfected baby hamster kidney cells. Hybridomas B cell turnover analysis (BrdU uptake) producing mAbs specifically reactive with CD22-transfected mouse L Mice were given drinking water containing 0.8 mg/ml BrdU for 7 days and cells, but not with untransfected cells, were subcloned twice with isotypes sacrificed. Tissues were harvested, and single leukocyte populations were determined using a Mouse mAb Isotyping kit (Amersham). CD22 mAbs analyzed for BrdU incorporation. Briefly, cells were fixed in 0.5% para- were purified using protein G Hi-Trap columns according to the manufac- formaldehyde, permeabilized with 3 N HCl containing 0.5% Tween 20, turer’s instructions (Amersham). MB22-8 and -9 mAbs were IgG1 Abs. neutralized with 0.1 M disodium tetraborate, stained with a FITC-conju- MB22-10 and -11 were originally isotyped as IgG2a mAbs; however, gated anti-BrdU Ab (BD Biosciences) and a PE-conjugated anti-mouse based on a study by Martin et al. (44), MB22-10 and -11 were reisotyped B220 mAb (Caltag Laboratories), and analyzed by flow cytometry. as IgG2c mAbs due to their C57BL/6 origin. Nonetheless, IgG2a- and IgG2c-specific reagents both show significant reactivity with MB22-10 and Ca2ϩ responses -11 mAbs. Because IgG2c-specific reagents were only recently made avail- able, mouse IgG2a isotype controls and IgG2a-specific reagents were used Splenocytes were isolated at room temperature and suspended (1 ϫ 107/ in experiments. Reactivity of mAbs with membrane-proximal CD22 Ig ml) in RPMI 1640 (Sigma-Aldrich) containing 5% FCS and 10 mM domains was determined using CD22AA and CD22⌬1-2 mice as well as HEPES (Invitrogen Life Technologies). The cells were loaded with 1 ␮M cDNA-transfected COS cells and immunofluorescence staining. MB22-11 indo-1 AM ester (Molecular Probes) for 30 min at 37°C, labeled with Ј F(ab )2 were prepared using immobilized pepsin (Pierce) in 20 mM sodium FITC-conjugated anti-B220 mAb for an additional 20 min, washed, and acetate buffer (pH 4.1), according to manufacturer’s instructions. Complete suspended in warm medium at 2 ϫ 106/ml for flow cytometry. Baseline digestion was confirmed by reducing and nonreducing SDS-PAGE. Sterile indo-1 fluorescence emission intensities at 405 nm and 525 nm were col- Ј ␮ protein G-purified CD22 and isotype control mAbs (10–250 ␮g of mAb/ lected for 1 min before the addition of F(ab )2 anti-IgM Abs (40 g/ml), ϩ 200 ␮l of PBS) were injected i.v. with fluorescence ratios (405 nm/525 nm) plotted for B220 -gated cells at 20-s intervals. Some samples were preincubated for 2 min with MB22-10 ϩ Abs and immunohistochemistry (20 ␮g/ml) before beginning sample collection. Increases in Ca2 mobilization are depicted by increased indo-1 fluorescence intensity ratios after Abs used included FITC-conjugated anti-mouse CD22 N terminus Ab treatment relative to the fluorescence intensity of untreated cells. (Cy34.1, TIB163; American Type Culture Collection); FITC-conjugated I-A/I-E MHC class II (clone M5/114.15.2; American Type Culture Col- B cell proliferation lection; TIB120); FITC-conjugated rat anti-mouse CD21/35 (clone 7G6), Splenic B cells were enriched by Thy1.2-magnetic bead depletion (Dynal PE-conjugated rat anti-mouse CD1d (clone 1B1), PE-conjugated rat anti- Biotech) and labeled with CFSE according to the manufacturer’s instruc- mouse CD19 (clone 1D3), and cychrome-conjugated rat anti-mouse B220 tions at a final concentration of 0.5 ␮m (Molecular Probes). A total of 1 ϫ (clone RA3-6B2), all obtained from BD Pharmingen; FITC-conjugated ϩ 106 B cells (ϳ90% B220 ) was cultured in 0.5 ml of complete RPMI 1640 polyclonal goat anti-mouse IgM, PE-conjugated rat anti-mouse IgM (clone containing 10% FCS in 48-well flat-bottom plates. B cells were cultured for 1B4B1), PE-conjugated rat anti-mouse IgD (clone 11-26), PE-conjugated 4 days in medium alone, or in the presence of F(abЈ) anti-mouse IgM Abs polyclonal goat anti-mouse IgG1, and IgG2a and FITC-conjugated goat 2 (Cappel), LPS (Escherichia coli 0111:B4; Sigma-Aldrich), or hamster anti- anti-mouse IgM, all obtained from Southern Biotechnology Associates. For mouse CD40 mAb (HM40-3; BD Pharmingen). Proliferation was assessed immunohistochemistry, 5-␮m-thick spleen sections were fixed in acetone, by the cellular loss of CFSE fluorescence intensity over successive rounds blocked with 5% rabbit serum, and stained with FITC-conjugated anti- of division as assessed by flow cytometry. Dead cells were excluded by B220 mAb and PE-conjugated anti-CD1d mAb or AP-conjugated goat propidium iodide staining. Fifty thousand total events were collected per anti-mouse IgM (Southern Biotechnology Associates) and biotinylated anti- culture condition. mouse CD23 (BD Pharmingen), followed by streptavidin-HRP (Southern Biotechnology Associates) with detection using 3,3Ј-diaminobenzidine Adoptive transfer experiments substrate and Vector Blue AP substrate kits (Vector Laboratories). Fluo- rescent digital images were merged to indicate marginal zone B cell A total of 5 ϫ 107 CFSE-labeled wild-type splenocytes was i.v. injected location. into wild-type C57BL/6 or Rag1Ϫ/Ϫ mice. On days 2 and 9, mice received The Journal of Immunology 3065

250 ␮g of MB22-10, MB22-8, IgG2a, or IgG1 control mAbs i.v. Tissues These CD22 mAbs were also assessed for their ability to bind were harvested on day 14, with single-cell suspensions isolated and splenocytes from wild-type, CD22Ϫ/Ϫ, and CD22⌬1-2 (lacking counted. CFSEϩ cells were analyzed for B220, CD19, IgM, IgD, CD22, 7 two amino-terminal Ig-like domains) mice. The MB22-8 mAb CD4, and CD8 expression by flow cytometry. For A20 studies, 1 ϫ 10 Ϫ Ϫ Ϫ Ϫ ⌬ / A20 cells were injected i.v. into Rag1 / mice. Mice received 250 ␮gof bound wild-type and CD22 1-2 splenocytes but not CD22 MB22-10 or IgG2a control mAb i.v. on day 2. Mice were sacrificed on day splenocytes, whereas the Cy34.1, MB22-9, -10, and -11 mAbs 7, with splenocytes counted and CD19ϩB220ϩ cells analyzed by flow bound wild-type but not CD22⌬1-2 or CD22Ϫ/Ϫ splenocytes (Fig. cytometry. 1E, data not shown). The MB22-9, -10, and -11 mAbs bind a Statistical analysis similar region of CD22 because MB22-11 mAb binding was completely blocked by prior MB22-10 mAb binding and partially Ϯ All data are shown as means SEM. The Student’s t test was used to blocked by MB22-9 mAb binding (data not shown). By contrast, determine the significance of differences between population means. none of the MB22 mAbs inhibited Cy34.1 mAb binding. Thus, Results among the MB22 mAbs binding to the two amino-terminal ligand Generation of anti-mouse CD22 mAbs that block ligand binding binding domains of CD22, the MB22-10 and MB22-11 mAbs exhibited the highest levels of blocking activity in vitro. A panel of mouse anti-mouse CD22 mAbs was generated using B cells from CD22Ϫ/Ϫ mice and assessed for their ability to inhibit heterotypic CD22-mediated intercellular adhesion in vitro. Five CD22 ligand binding domain mAbs deplete blood and marginal different CD22 mAbs were selected that inhibited CD22 ligand zone B cells in vivo binding to different degrees. The MB22-8 and Cy34.1 (IgG1) The effects of CD22 mAb treatment on B cell numbers were as- mAbs had little, if any, effect on CD22-mediated adhesion (Ͻ10%) sessed in vivo. Although MB22-8 mAb treatment did not signifi- when used at 1 or 10 ␮g/ml concentrations (Fig. 1A, data not cantly affect blood B cell numbers, the Cy34.1, MB22-9, -10, and shown). The MB22-9 (IgG1), MB22-10 (IgG2c), and MB22-11 -11 mAbs reduced B cell numbers by ϳ75–85% within 3–10 days (IgG2c) mAbs inhibited CD22-mediated adhesion by ϳ35, 60, and (Fig. 2, A–C). T cell numbers were not affected (data not shown). 90%, respectively, at 1–10 ␮g/ml mAb concentrations (Fig. 1, A The degree and duration of peripheral blood B cell depletion was and B, data not shown). The ability of each CD22 mAb to block dependent on mAb dose, as shown for the MB22-10 mAb (Fig. mouse CD22-Fc fusion protein binding to CD22Ϫ/Ϫ mouse T and 2B). Reductions were comparable for 100- to 250-␮g doses over a B splenocytes was also assessed in vitro. The MB22-8 mAb did 2-wk period, with B cell numbers recovering earlier with 100-␮g not bind the CD22-Fc fusion protein and therefore had no effect on doses. A 25-␮g dose of MB22-10 mAb gave moderate depletion, its binding to B or T cells. By contrast, the Cy34.1 and MB22-9 whereas 10 ␮g had little effect. The frequency of mature recircu- mAbs enhanced the binding of CD22-Fc fusion protein (5- to 13- lating B220highIgDϩ B cells in the bone marrow was significantly fold) to both T (Fig. 1C) and B (Fig. 1D) cells, presumably due to reduced by all CD22 mAbs, including the MB22-8 mAb (Fig. 2D). CD22-Fc multimerization following mAb binding. The MB22-10 However, no changes were observed in the frequencies of pre/ mAb enhanced CD22-Fc binding to B cells (3-fold) but had no pro-B or immature B cells (data not shown), which do not express effect on T cell binding, suggesting that it partially inhibited or express low levels of cell surface CD22, respectively. Spleen CD22-Fc binding in comparison with the Cy34.1 and MB22-9 and peritoneal cavity B cell numbers were unchanged 7 days after mAbs. The MB22-11 mAb completely blocked CD22-Fc binding CD22 mAb treatment (Fig. 2E, data not shown). Lymph node B to both T and B cells. Collectively, these results indicate that the cell numbers were reduced slightly following MB22-9 and -10 MB22-9, -10, and -11 mAbs block CD22 ligand binding to in- treatments, with only the MB22-11 mAb causing significant re- creasing degrees, respectively, whereas the Cy34.1 and MB22-8 ductions (Fig. 2F). Thus, CD22 mAb treatment predominantly re- mAbs do not inhibit CD22 ligand binding. duced mature bone marrow and blood B cell numbers.

FIGURE 1. In vitro properties of CD22 mAbs. A–D, CD22 ligand blocking by MB22 mAbs. A and B, CD22 mAbs were analyzed for their ability to block splenocyte binding to mCD22-L cells. A, Blocking efficiency of MB22 mAbs relative to isotype control mAbs at 10 ␮g/ ml. B, Blocking efficiency of MB22-10 and -11 mAbs at 0–10 ␮g/ml. Significant differences in the mean (ϮSEM) number of splenocytes bound per field of mCD22-L cells following preincubation with each CD22 mAb compared with isotype control mAbs are p Ͻ 0.05. In all experiments, nonspecific ,ء ;indicated background splenocyte binding to parental L cells was subtracted (ϳ10% or less of maximum splenocyte binding). Results represent those obtained in 3–5 separate experiments. C and D, CD22 mAb effects on the binding of mCD22-Fc to CD22Ϫ/Ϫ spleen T (C)andB(D) cells. E, Reactivity of MB22 mAbs (10 ␮g/ml) with spleen B cells from wild-type (thin line) and CD22⌬1-2 mice (thick line). Dotted lines represent isotype control staining, except for MB22-8 mAb staining, where the dotted line represents reactivity of MB22-8 mAb with CD22Ϫ/Ϫ spleen cells. 3066 CD22 LIGAND BINDING REGULATES B CELL SURVIVAL

FIGURE 2. Depletion of blood, mature recirculating bone marrow, and marginal zone B cells by CD22 mAbs in vivo. A and B, Blood B cell frequencies (A; day 6, 250-␮g dose) and numbers (ϮSEM) over an 18-day time course (B) following one i.v. dose (10–250 ␮g) of MB22-10 or IgG2a isotype control mAb as determined by immunofluorescence staining with flow cytometric analysis. C–F, Blood B cell numbers per milliliter (C), mature IgDϩB220high bone marrow frequencies (D), spleen (E), and lymph node (F) B cell numbers 7 days following treatment with CD22 or isotype-control mAbs. G–I, Depletion of marginal zone B cells by mAbs that bind CD22 ligand binding domains. G and H, Spleno- cytes from mAb-treated mice (day 7) were stained with CD21, CD1d, and B220 mAbs and analyzed by flow cytometry. G, The mean percentages of gated lymphocytes expressing a B220ϩCD21highCD1dhigh marginal zone phenotype following MB22-10 or isotype control mAb treatment are indicated. H, Numbers of B220ϩCD1dhighCD21high cells 7 days following CD22 mAb treatment. I, Immunohistochemical analysis of spleen sections 7 days following MB22-10 or control mAb treatments. Sections were stained for CD23 (brown) and IgM (blue) or for CD1d (red) and B220 (green), with CD1dhighB220ϩ cells characterized by yel- low fluorescence in overlaid digital microscopy images. Original magnification, ϫ200. Significant differences between mean (ϮSEM) values for CD22 and isotype .p Ͻ 0.05 ,ء ;control mAb-treated mice are indicated Data are representative of results obtained from analysis of Ն3 mice per mAb treatment.

The effect of CD22 mAbs on the marginal zone B cell population within in these tissues increased or stayed the same. Similarly, was assessed in wild-type mice. The Cy34.1, MB22-9, -10, and -11 CD22AA and CD22Ϫ/Ϫ mice also had 2- to 3-fold increased frequen- mAbs depleted significant numbers of CD1dhighCD21highB220ϩ mar- cies of BrdUϩ B cells in the bone marrow, blood, and spleen com- ginal zone B cells as determined by flow cytometry (Fig. 2, G and H), pared with littermates with identical genetic backgrounds (Fig. 3, C whereas the MB22-8 mAb did not have a significant effect (Fig. 2H). and D), as described previously (24, 27, 28). These results demon- Notably, marginal zone B cell numbers had begun to replenish 3 wks strate that inhibiting CD22 ligand binding alters normal B cell turn- following a single 250-␮g MB22-10 mAb injection (2.4 ϫ 106/ over in the periphery. mouse; n ϭ 2) compared with control mAb-treated littermates (3.8 ϫ 106/mouse; n ϭ 2). In vivo IgMbrightCD23Ϫ or B220ϩCD1dbright CD22 ligand-blocking mAbs inhibit normal and malignant B marginal zone B cell depletion was confirmed by immunohistochem- cell survival in vivo istry following MB22-10 (Fig. 2I) and MB22-11 (data not shown) To directly address whether blocking CD22 ligand binding inter- mAb treatments. Thus, mAb binding to the CD22 ligand binding do- fered with peripheral B cell survival, CFSE-labeled splenocytes mains inhibited circulating and marginal zone B cell survival or were transferred into wild-type littermates and monitored for their localization. ability to survive in vivo following CD22 mAb treatment. After 2 wk, the frequency and number of adoptively transferred (CFSEϩ) B cells CD22 ligand-blocking mAbs increase B cell turnover in vivo recovered from the blood, bone marrow, lymph nodes, and spleens of Although CD22 mAb treatment did not significantly alter total B cell MB22-10 mAb-treated mice were decreased by 4- to 15-fold com- numbers in peripheral tissues, striking alterations in B cell turnover pared with control mAb-treated mice (Fig. 4, A and B). No significant rates were observed 7 days following mAb treatment. No significant differences were found in CFSEϩ T cell numbers (Fig. 4B). More- differences in B cell turnover were found among MB22-8, Cy34.1, over, the ratios of CFSEϩ B cells:CFSEϩ T cells within these tissues and MB22-9 mAb treatments (Fig. 3A). However, blood, spleen, and were decreased by 4- to 19-fold following MB22-10 mAb-treatment lymph node B cell turnover was significantly higher in mice treated (Fig. 4C). Similar results were observed 1 wk following MB22-10 with the CD22 ligand-blocking mAbs, MB22-10 and MB22-11, in mAb treatment (data not shown). By contrast, MB22-8 mAb treat- comparison to nonblocking mAbs. The relative frequencies of BrdUϩ ment had no effect on the ratio of CFSEϩ B cells:CFSEϩ T cells in (cycling) blood, spleen, and lymph node B220ϩ cells were increased lymph nodes or spleen and only decreased this ratio by 1.6-fold in by as much as 4-fold in MB22-10 and MB22-11 mAb-treated mice blood and bone marrow (Fig. 4D). Thus, blocking CD22 ligand bind- compared with control mAb-treated mice 7 days following BrdU ad- ing significantly reduced mature B cell survival in vivo. ministration. Both the MB22-10 and MB22-11 mAbs had significant To assess whether CD22 similarly regulates the survival of ma- effects on the numbers of BrdU-labeled and unlabeled blood, lymph lignant B cells, A20 lymphoma cells were transferred into node, and spleen B cells (Fig. 3B). The total numbers of BrdUϪ (non- Rag1Ϫ/Ϫ mice and examined for their ability to survive following cycling) spleen and lymph node B cells were significantly decreased CD22 mAb treatment. The frequency and number of A20 lym- by MB22-10 and -11 treatment, whereas numbers of BrdUϩ B cells phoma cells recovered from spleens of MB22-10 mAb-treated The Journal of Immunology 3067

FIGURE 3. Blocking CD22 ligand binding enhances B cell turnover. A and B, B cell turnover in wild-type mice treated i.v. with 250 ␮g of CD22 or control mAbs. C and D, B cell turnover in wild-type (mixed B6/129), CD22AA (C), and CD22Ϫ/Ϫ (D) mice. A–D, Mice were fed BrdU for 7 days. The Means significantly ,ء .mean (ϮSEM) percentages (A, C, and D) and numbers (B) of BrdUϩB220high B cells were determined by flow cytometric analysis different between mAb-treated mice (A and B) or between wild-type and CD22AA (C) or CD22Ϫ/Ϫ (D) mice; p Ͻ 0.05. N. S. indicates where no significant differences were found between mAb-treated mice. mice was decreased by ϳ90% compared with control mAb-treated mAb treatment only resulted in a 30 Ϯ 3% reduction in Cy34.1 littermates (Fig. 4E). Thus, CD22 ligand binding regulates both mAb staining (Fig. 5C), and residual cell surface bound mature and malignant B cell survival in the periphery. MB22-8 mAb was often detected 7 days postinjection (Fig. 5D). Similarly, Cy34.1 mAb treatment resulted in the partial loss (78 Ϯ CD22 mAb effects on B cell phenotypes in vitro and in vivo 2%) of CD22 expression with residual Cy34.1 mAb also detected Because blocking CD22 ligand binding increased B turnover in on the cell surface. Unlike the MB22-9 and MB22-10 mAbs, re- vivo (Fig. 3, A and B), the effects of CD22 mAbs on B cell phe- sidual cell surface MB22-11 mAb was also sometimes observed on notypes were assessed. All CD22 mAbs significantly reduced tissue B cells 7 days after treatment. Importantly, restoration of CD22 expression on spleen B cells during 20-h in vitro cultures circulating B cell numbers coincided with the appearance of ϩ (Fig. 5A: MB22-8, 45 Ϯ 8%; Cy34.1, 69–70%; MB22-9, 82 Ϯ CD22 B cells (data not shown). 3%; MB22-10, 64 Ϯ 6%; and MB22-11, 54 Ϯ 8%, compared with In vivo CD22 mAb treatment also induced significant reductions in CD22 mean fluorescence intensities (MFIs)3 for control cultures; cell surface IgM (ϳ40–50%) and CD19 (20–60%) expression by n Ն 3 experiments for MB22 mAbs, n ϭ 2 for Cy34.1). Compa- spleen, blood, lymph node, and peritoneal B cells (Fig. 5, E and F, rable results were obtained for shorter (1–2 h) incubation periods data not shown). CD21 expression was also reduced by 30–60%. In (data not shown). CD22 mAb binding to B cells from CD22AA comparison, IgD expression was only slightly decreased (ϳ5–30%), mice also induced similar reductions in cell surface CD22 expres- whereas MHC class II expression was only slightly enhanced (ϳ20%) sion (Fig. 5B: MB22-8, 51 Ϯ 10%; Cy34.1, 60–66%; MB22-9, on tissue B cells following MB22-10 mAb treatment (data not 70 Ϯ 7%; MB22-10, 55 Ϯ 14%; and MB22-11, 67 Ϯ 6%; n Ն 3 shown). CD1d and B220 expression were unaffected by CD22 mAb experiments for MB22 mAbs, n ϭ 2 for Cy34.1). Similar results treatment (Fig. 5E and data not shown). Similar to the B cells remain- were also obtained using B cells from Fc␥RIIBϪ/Ϫ mice (data not ing in wild-type mice after CD22 mAb treatment, small numbers of ϩ shown). In vitro CD22 mAb treatment had minimal effects on cell adoptively transferred CFSE B cells remained in mouse tissues fol- surface IgM, IgD, or CD19 expression (data not shown). Thus, all lowing MB22-10 mAb treatment (Fig. 4, A–C). The residual B cells ϩ CD22 mAbs induced significant receptor internalization in vitro were B220highIgD , expressed significantly lower levels of CD19 and regardless of Fc␥RIIB expression or in the case of CD22AA B IgM, and were negative for CD22 expression (Fig. 5, G and H). Thus, cells, prior CD22 ligand engagement. the treatment of mice with mAbs that bind CD22 ligand binding do- In vivo, MB22-9, -10, and -11 mAb treatments significantly mains either enriched for residual B cell populations expressing a reduced cell surface CD22 expression on B cells to nearly unde- unique phenotype or induced significant reductions in the expression tectable levels in the spleen, blood, bone marrow, lymph node, and of multiple regulatory surface molecules, including CD22, IgM, peritoneum as determined by Cy34.1 mAb staining 7 days after CD19, and CD21. treatment (Fig. 5C, data not shown). CD22 expression was similarly reduced on adoptively transferred A20 cells remaining after CD22 mAb effects in CD22AA mice MB22-10 treatment (Fig. 4E, data not shown). Reduced CD22 Whether CD22 mAbs exhibit their biological effects by inhibiting expression was likely due to CD22 internalization because the ligand-induced signals or by removing CD22 from the cell surface MB22-9 and -10 mAbs were not detected on the cell surface by was assessed using CD22AA mice that express CD22 receptors isotype-specific secondary Abs (Fig. 5D, data not shown). MB22-8 incapable of ligand binding. In vivo, MB22-10 mAb treatment down-modulated B cell CD22 expression to undetectable levels in 3 Abbreviation used in this paper: MFI, mean fluorescence intensity. both wild-type and CD22AA mice (Fig. 6A). However, in contrast 3068 CD22 LIGAND BINDING REGULATES B CELL SURVIVAL

sponses (Fig. 6D, left panel), as demonstrated previously (24–28). Ca2ϩ responses of B cells from MB22-10 mAb-treated wild-type mice were augmented and delayed following IgM cross-linking compared with control mAb-treated mice, perhaps due to decreased CD22 and CD19 expression (Fig. 6D, middle panel). B cells from CD22AA mice treated in vivo with MB22-10 mAb also generated augmented Ca2ϩ responses (Fig. 6D, right panel), likely due to the loss of cell surface CD22 (Fig. 6A). Wild-type B cells incubated with MB22-10 mAb in vitro for 2 min before BCR stimulation generated augmented Ca2ϩ responses (Fig. 6D, middle panel). In contrast, in vitro MB22-10 mAb pretreatment did not alter Ca2ϩ responses in CD22AA B cells (Fig. 6D, right panel). Finally, B cells isolated from MB22-8 mAb-treated mice had normal IgM-induced Ca2ϩ responses, whereas B cell preincubation with MB22-8 mAb before BCR stimulation resulted in augmented Ca2ϩ responses (data not shown). Thus, mAb engagement of CD22 may inhibit CD22 negative regulation independent of interfering with CD22 ligand binding. CD22 mAb treatment had no effect on B cell survival during in vitro cultures and did not induce B cell proliferation or significantly affect anti-IgM- or LPS-induced proliferation (Fig. 6E, data not shown). However, the MB22-11 mAb synergized with CD40 mAbs to induce robust B cell proliferation. The MB22-8 mAb did not significantly augment proliferation, whereas the MB22-9 and MB22-10 mAbs had intermediate effects on CD40-induced proliferation. B cells isolated from MB22-10 and isotype control mAb- treated mice exhibited identical proliferative responses in vitro (Fig. 6F), unlike B cells from CD22Ϫ/Ϫ mice with impaired prolif- FIGURE 4. CD22 mAb treatment impairs normal and malignant B cell eration in response to IgM cross-linking and augmented responses to survival in vivo. A–D, Fifty million CFSE-labeled wild-type splenocytes CD40 stimulation (25, 45). CD22 expression levels on these B cells were injected i.v. into wild-type mice on day 0. On days 2 and 9, mice received 250 ␮g of CD22 (n ϭ 4) or control mAb (n ϭ 4). Tissues were approached normal levels 24 h after culturing either due to B cell harvested on day 14, and CFSEϩ cells were analyzed for B220, CD19, maturation in vitro or recovery from CD22 mAb-induced internaliza- IgM, and CD22 expression by flow cytometry. A, Flow cytometric analysis tion (data not shown). Nonetheless, in vivo CD22 mAb treatment did ϩ Ϫ Ϫ of B220ϩCFSEϩ cells remaining after MB22-10 or IgG2a control mAb not induce CFSE B cell proliferation in either wild-type or Rag1 / treatment. B, Mean CFSEϩ B cell (B220ϩ) and T cell (CD4ϩ ϩ CD8ϩ) mice (Fig. 4A, data not shown). Thus, although CD22 mAb treatment ϩ numbers remaining following mAb treatment. C and D, Ratios of CFSE resulted in augmented BCR-induced Ca2ϩ responses in vitro, anti- ϩ B cells to CFSE T cells recovered from individual mice treated with IgM Ab-induced B cell proliferation remained normal. MB22-10 (C) or MB22-8 (D) mAbs. E, CD22 mAb treatment impairs malignant B cell survival in vivo. Ten million A20 cells were injected i.v. Ϫ/Ϫ ␮ ϭ into Rag1 mice. Mice received 250 g of MB22-10 (n 3) or IgG2a CD22 mAbs deplete B cells independent of FcRs and control mAb (n ϭ 3) i.v. on day 2. CD19ϩB220ϩ spleen cells were ana- complement lyzed by flow cytometry on day 7. A–E, Results are representative of three similar experiments. Values shown represent mean (ϮSEM) cell numbers To further verify that CD22 mAbs altered B cell survival in vivo Means significantly differ from by interfering with CD22 ligand binding, the effects of MB22-10 ,ء .(B and E) or B:T cell ratios (C and D) Ͻ Ϫ Ϫ isotype control mAb-treated mice; p 0.05. mAb treatment were assessed in C3 / mice deficient in complement-dependent cytotoxicity, and in Fc␥RIIBϪ/Ϫ and FcR␥Ϫ/Ϫ mice with enhanced or blocked Ab-dependent cellular cytotoxic- to wild-type mice, MB22-10 mAb treatment did not alter B cell ity, respectively. Fc␥RI, Fc␥RIII, and Fc␥RIV ligand binding CD19, CD21, or IgM expression in CD22AA mice. Likewise, ␣-chains associate with a common stimulatory ␥ chain (FcR␥) that MB22-10 mAb treatment did not reduce the number of mature is required for Fc␥R assembly and triggering of effector functions, bone marrow (IgDϩB220high), blood, or marginal zone B cells in including phagocytosis by macrophages and cytotoxicity by NK CD22AA mice, but significantly reduced these populations in cells (46–48). MB22-10 mAb administration significantly reduced wild-type mice (Fig. 6B). B cell turnover in CD22AA mice was blood and marginal zone B cells, and mature recirculating bone not altered by MB22-10 mAb treatment, but was significantly in- marrow B cells after 7 days in C3Ϫ/Ϫ, FcR␥Ϫ/Ϫ, and Fc␥RIIBϪ/Ϫ creased in wild-type mice (Fig. 6C). Thus, CD22 mAb treatment mice compared with control mAb-treated littermates (Fig. 7, A–C, down-modulated CD22 expression in CD22AA mice but did not data not shown). MB22-10 mAb treatment also significantly re- alter B cell phenotypes, numbers, or turnover beyond the changes duced B cell CD22, CD19, and IgM expression in C3Ϫ/Ϫ, induced by the CD22AA mutation. FcR␥Ϫ/Ϫ, and Fc␥RIIBϪ/Ϫ mice (Fig. 7, A–C, data not shown). Blood and marginal zone B cell numbers were also significantly 2ϩ CD22 mAb effects on Ca responses and proliferation reduced following treatments of wild-type mice with MB22-11 Ј Whether CD22 mAb treatment in vivo affected BCR-induced mAb F(ab )2 (Fig. 7D). Treatment of mice with intact MB22-11 Ca2ϩ mobilization or proliferation was assessed in vitro. CD22AA mAb yielded greater reductions, especially 6 days following mAb B cells generated normal Ca2ϩ responses following IgM cross- treatment, but this difference is likely due to the short in vivo Ϫ/Ϫ 2ϩ Ј linking, whereas CD22 B cells generated augmented Ca re- half-life of F(ab )2. CD22, CD19, and IgM expression levels were The Journal of Immunology 3069

FIGURE 5. Effects of CD22 mAb treatments on B cell phenotypes. A and B, CD22 internalization following incubation with CD22 mAbs in vitro. Fluorescence intensity of CD22 on purified B cells from wild-type (A) or CD22AA (B) mice incubated in medium (thick solid line) or with CD22 mAbs (20 ␮g/ml; thin solid line) for 20 h. Cells were stained with CD22 mAbs or isotype- matched control (dashed line) mAbs (10 ␮g/ml). In all cases, mAb staining was visualized using PE-conjugated isotype-specific secondary Ab with flow cytometry analysis. MB22 mAb results represent those obtained in Ն3 experiments. C–F, Splenocytes were assessed for CD22 (C and D) and IgM, IgD, CD19, CD21, and CD1d (E and F) expression 7 days following mAb injection by immunofluorescence staining with flow cytometry analysis. C, CD22 expression in mice treated with MB22 mAbs (thick lines) or isotype control (thin lines) mAbs assessed by Cy34.1 mAb staining. FITC-conjugated MB22-8 was used to assess CD22 expression in Cy34.1-treated mice. D, Cell surface CD22 mAb binding in mice treated with CD22 (thick lines) or isotype control (thin lines) mAbs assessed using isotype-specific secondary Abs. Results represent those obtained in Ն3 experiments. E, Cell surface molecule expression in mice treated with MB22-10 (thick lines) or isotype control (thin lines) mAbs. F, Mean IgM, IgD, CD19, and CD21 expression levels relative to wild-type controls (ϮSEM) are indicated for mice treated with CD22 or isotype control mAbs (n Ն 3 mice per treat- Mean MFI values were significantly different ,ء .(ment from isotype control mAb-treated mice; p Ͻ 0.05. G and H, Fifty million CFSE-labeled wild-type splenocytes were i.v. injected into wild-type mice. On days 2 and 9, mice received 250 ␮g of MB22-10 (n ϭ 4) or IgG2a control mAb (n ϭ 4). Tissues were harvested on day 14, and CFSEϩ cells were analyzed for B220, CD19, IgM, and CD22 expression by flow cytometry. G, CD19 expression by CFSEϩ spleen cells, and IgM and CD22 expression by CFSEϩB220ϩ-gated spleen cells following mAb treatment. H, Mean CD19 and IgM MFIs (ϮSEM) for CFSEϩCD19ϩ and CFSEϩIgMϩ spleen and lymph node cells.

Ј also significantly reduced in blood 2 days following F(ab )2 treat- blood, spleen, and lymph node was significantly attenuated (73– ment but returned to normal levels by day 6, in contrast to treat- 88%) by mAb blockade of CD22 ligand binding (Fig. 4, A–C), ment with intact MB22-11 mAb. Thus, CD22 mAbs alter B cell whereas a nonblocking mAb only reduced mature recirculating B survival through mechanisms that do not require C3, Fc␥Rs (I, II, cells by 40% and had no effect on B cell survival within spleen or III, or IV), or the mAb Fc region. lymph nodes (Fig. 4D). Malignant B cell survival was also significantly reduced (90%) by a CD22 mAb that blocks CD22 ligand Discussion binding (Fig. 4E). Inhibiting CD22-ligand interactions through The current study is the first to describe the generation and in vivo CD22 deletion or mutation of its amino-terminal ligand binding assessment of mouse anti-mouse CD22 mAbs that bind to distinct domains is also known to enhance B cell turnover in vivo (24–28). domains of CD22 and elicit ligand-blocking effects (Fig. 1). Using Thereby, CD22 represents an important survival signal for mature these reagents, this study demonstrates that CD22 ligand binding follicular B cells, with mAb blockade of ligand engagement re- regulates peripheral B cell survival in vivo, with mAb blockade of sulting in significant and rapid B cell turnover in the periphery. CD22 ligand binding leading to significant 2.5- to 4-fold increases The MB22-10 and -11 mAbs are likely to bind at or near the in B cell turnover in the blood, spleen, and lymph nodes (Fig. 3A). CD22 ligand binding site because they blocked trans interactions In contrast, differences in B cell turnover were not observed among between CD22 and ligand-bearing splenocytes (Fig. 1, A and B), nonblocking mAbs, regardless of whether they interacted with and their binding required expression of the first two Ig-like do- CD22 ligand binding domains. Consistent with this result, the sur- mains of CD22 (Fig. 1E). The MB22-11 mAb also completely vival of adoptively transferred B cells found in bone marrow, blocked CD22-Fc fusion protein binding to ligands expressed by T 3070 CD22 LIGAND BINDING REGULATES B CELL SURVIVAL

FIGURE 6. Effects of CD22 mAbs in CD22AA mice and on BCR-induced Ca2ϩ responses and proliferation. A–C, CD22 ligand binding domain mAbs down-modulate CD22 expression in CD22AA mice, but have no effect on CD22AA B cell depletion, turnover, or IgM, CD19, and CD21 expression levels in vivo. CD22AA and wild-type (WT) mice were given 250 ␮g of MB22-10 or isotype control mAb i.v. (n Ն 3 mice/group). Splenocytes from wild-type and CD22AA mice were stained and analyzed by flow cytometry 7 days following mAb injection to assess alterations in CD22, IgM, CD19, and CD21 expression levels (A), blood, mature recirculating IgDϩB220high bone marrow, and marginal zone B cell numbers (B), and B cell turnover as determined by BrdU staining (C). Values represent mean (ϮSEM) numbers or percentages of B cells (A and B) or BrdUϩB220high B cells (C) within the indicated Means significantly different from isotype control mAb-treated mice; p Ͻ 0.05. D, Effect of MB22-10 mAb treatment on BCR-induced Ca2ϩ ,ء .tissues responses. BCR-induced Ca2ϩ responses in splenocytes from CD22AA, CD22Ϫ/Ϫ, or wild-type mice (left panel). BCR-induced Ca2ϩ responses for splenocytes from wild-type (middle panel) or CD22AA (right panel) mice treated for 7 days with MB22-10 or control mAbs in vivo and for splenocytes from untreated mice preincubated for 2 min with MB22-10 mAb (20 ␮g/ml) in vitro before data collection. Relative Ca2ϩ concentrations were assessed ϩ Ј by flow cytometry for B220 -gated cells (presented as the fluorescence ratio of 405:525 nm). F(ab )2 anti-IgM Ab was added to the cells after 1 min (depicted by the arrow). Ca2ϩ flux responses represent the results obtained in three experiments (left and middle panels) or one experiment (right panel). E and F, CD22 mAb effects on B cell proliferation in vitro. E, Purified splenic B cells from wild-type mice were CFSE-labeled and incubated in the presence ␮ Ј ␮ of CD22 or control mAbs, and cultured in medium or with 40 g/ml F(ab )2 anti-IgM, or 0.1–1 g/ml CD40 mAb. F, Wild-type mice were treated with 250-␮g MB22-10 (thick line) or isotype control mAb (thin line) i.v. (n ϭ 2 mice/group). On day 7, splenic B cells were purified, labeled with CFSE, and ␮ Ј ␮ ␮ cultured in medium alone or the presence of 40 g/ml F(ab )2 anti-IgM, 1 g/ml CD40 mAb, or 2.5 g/ml LPS. E and F, Fifty thousand events were collected, and cells excluding propidium iodide were analyzed for CFSE fluorescence by flow cytometry. and B cells, whereas the MB22-10 mAb appeared to partially says, or may enhance CD22 internalization or other functional block binding in this assay system (Fig. 1, C and D). Thereby, the changes in CD22. Select CD22 ligands that modulate physiologic MB22-10 and -11 mAbs may also inhibit CD22-CD22 cis-inter- B cell function in vivo or glycosylation sites on CD22 that mediate actions, which may represent functionally significant signaling cis-interactions have not been identified. Nonetheless, mAbs such units in vivo and regulate CD22 microdomain localization (23, 49, as MB22-8 that bind CD22 outside of its ligand binding domains 50). Most important, however, the MB22-10 and MB22-11 mAbs appear to have limited biological effects in vivo. Thus, CD22 induced significant B cell turnover in vivo and other biological mAbs that block CD22-ligand interactions were the most potent effects beyond those evoked by other CD22 mAbs. CD22 mAbs inhibitors of CD22-mediated survival signals. that bind within the two amino-terminal CD22 ligand binding do- CD22 mAbs that bind epitopes within CD22 ligand binding domains may nonetheless represent a second class of CD22 mAbs mains significantly reduced mature bone marrow frequencies (40– that also induce biological effects in vivo. Thus, mAbs that engage 63%) and circulating (75–85%) and marginal zone (82–99%) B the two amino-terminal CD22 ligand binding domains may inhibit cell numbers within 3–7 days (Fig. 2). It remains unclear whether physiologically important CD22-ligand interactions in vivo CD22 mAbs characterized as “nonblocking” induced these through steric hindrance that are not reflected during in vitro as- changes by affecting interactions between CD22 and physiologic The Journal of Immunology 3071

along with similar reductions in marginal zone B cells observed in CD22⌬1-2, CD22AA, CD22Ϫ/Ϫ, and ST6Gal-IϪ/Ϫ mice, which lack CD22 ligands, clearly demonstrates that this unique B cell subpopulation requires CD22 ligand-induced signals for survival in vivo (24, 25, 50, 54). Thus, CD22-regulated homeostasis de- pends on ligands ubiquitously expressed by multiple cell types in vivo (1). Indeed, marginal zone B cells are reported to express higher levels of CD22 in an unmasked form, which would poten- tially allow for enhanced ligand binding (55). Although a higher proportion of CD22 is similarly reported to be unmasked on B1 cells of the peritoneum (55), inhibition of ligand binding function did not appear to affect this population (data not shown; Ref. 24). Thus, marginal zone B cells represent a unique B cell subpopulation with enhanced susceptibility to CD22 mutations or mAbs that interfere with CD22 ligand binding. Although blocking CD22 ligand binding significantly reduced recirculating and marginal zone B cell numbers following mAb treatment, B cell turnover was significantly enhanced in all other tissues (Figs. 2 and 3). CD22AA mice also have significantly higher B cell turnover rates in all compartments, yet normal lymph node and spleen B cell numbers in addition to reduced recirculating and marginal zone B cells numbers (Fig. 6, B and C, and Ref. 24). These differences therefore appear to represent an intrinsic consequence of CD22-ligand interactions within different lymphoid compartments rather than a lack of effective CD22 mAb penetration into tissues. In fact, CD22 mAbs effectively penetrated most tissues within minutes of treatment, with CD22 mAb found at saturating levels on the surface of spleen and lymph node B cells within 1–4 h (K. M. Haas and T. F. Tedder, unpublished obser- vations). In contrast to recirculating blood and marginal zone B cell populations that appear more dependent on CD22-ligand interactions for survival, the spleen and lymph node environments may provide additional survival factors that rescue B cells from depletion in the absence of CD22 ligand binding. For example, CD40-generated signals rescue CD22Ϫ/Ϫ, CD22⌬1-2, and CD22AA B cell survival following BCR cross-linking in vitro (24, 45). The differences in B cell survival between tissues may also reflect intrinsic differences between the lymphocyte subsets within FIGURE 7. B cell depletion by CD22 mAbs is independent of C3, each tissue as well as their individual turnover rates. Thus, CD22 Fc␥Rs, and the Fc portion of mAb. A–C, Blood and marginal zone (MZ) is one of multiple cell surface receptors that significantly influ- Ϯ Ϫ/Ϫ B cell numbers ( SEM), and CD22 expression in spleens of C3 (A), ences B cell survival. ␥ Ϫ/Ϫ ␥Ϫ/Ϫ Fc RII (B), and FcR (C) mice were assessed 7 days after Although CD22 adhesive function was originally envisioned to MB22-10 or isotype-control mAb (250 ␮g) treatment. In CD22 expression mediate B cell migration or physical interactions within tissues histograms, mice treated with MB22-10 (thick line) or IgG2a isotype control mAb (thin line) are indicated. D, Blood and marginal zone B cell (12), CD22 engagement of its ligand(s) may primarily promote B numbers and CD22 expression (day 2 blood and day 6 spleen) in normal cell survival through unknown pathways. In the current study, ma- ␮ Ј ture recirculating blood and bone marrow B cell numbers were mice after one 250 g of i.v. dose of whole MB22-11, F(ab )2 MB22-11, or isotype control mAb. Significant differences between means for MB22 significantly reduced following CD22 mAb treatment (Figs. 2, p Ͻ 0.05. A–D,3B, and 4, A–C). The treatment of mice with either a ,ء ;or isotype-control mAb-treated mice are indicated CD22-Fc fusion protein or polyclonal Abs against CD22 also reduces the mature recirculating IgDϩ B cell population in the bone ligands in vivo or inducing receptor internalization (Fig. 5) thereby marrow with no change in spleen B cell frequencies (35). Although rendering B cells unable to interact with ligands due to reduced these results were originally interpreted to suggest that CD22 me- cell surface CD22 expression. Normally, CD22 is constitutively diated B cell migration to the bone marrow, B cells from internalized and subsequently degraded in an acidic intracellular CD22⌬1-2 and CD22AA mice migrate normally to bone marrow compartment without recycling back to the surface (51). By con- and lymphoid tissues (24). Thereby, accelerated B cell turnover trast, mAb binding to CD22 leads to a significant increase in CD22 due to CD22 blockade is likely to reduce the size of the mature internalization without increasing its intracellular degradation (51, recirculating B cell pool rather than blocking B cell migration to 52), making CD22 an effective target for the treatment of B cell the bone marrow. tumors by immunotoxins (51, 53). Recirculating and marginal The mechanisms through which CD22 mAbs depleted B cells zone B cell populations may thereby represent B cell subsets that were distinct from those induced by CD20- and CD19-targeted are exquisitely sensitive to blockade of CD22 ligand binding or immunotherapies. Effective CD19 and CD20 mAbs deplete B cells mAb-induced reductions in CD22 expression. Indeed, marginal in vivo via Fc␥R-dependent, complement-independent pathways zone B cells were extremely sensitive to treatment with mAbs that that require mononuclear phagocytes (56–59). CD22 mAb-medi- bind CD22 ligand binding domains (Fig. 2, G–I). This finding, ated effects did not require the presence of FcRs (Fc␥RI, -II, -III, 3072 CD22 LIGAND BINDING REGULATES B CELL SURVIVAL or -IV) or complement component C3 (Fig. 7, A–C). In fact, intact following CD40 stimulation, suggesting that CD22 ligand binding Ј and F(ab )2 CD22 mAb fragments elicited similar effects, demon- negatively regulates CD40 signaling. Thus, CD22 regulates phys- strating that the Fc portion of the mAb was not required for B cell iologically relevant ligand-dependent and -independent functions depletion (Fig. 7D). The rapid internalization of CD22 mAbs may that have distinct effects on B cell biology (24, 45). be a partial explanation for the lack of dependence on FcRs and The results of this study and the recent characterization of complement. However, the MB22-8 mAb did not induce signifi- CD22⌬1-2 and CD22AA mice expressing CD22 molecules unable cant CD22 internalization and persisted on the cell surface, but did to bind ligands (24) collectively demonstrate that CD22 ligand not significantly affect B cell survival. Nonetheless, FcRs and com- binding significantly influences B cell turnover and is required for plement may contribute to the moderate reductions in recirculating normal B cell survival in the periphery. Similarly, mAb blockade B cell numbers caused by the MB22-8 mAb. Although mAb bind- of CD22 ligand binding inhibits human B cell survival in vitro, and ing to CD22 may also induce B cell Fc␥RIIb cross-linking, this either induces or inhibits B cell signal transduction (36–38, 63). does not appear to contribute significantly to the results described These anti-human CD22 mAbs are also tumorcidal in xenotrans- in this study. Specifically, B cell depletion is similar in plantation models of human lymphomas (38). In the current stud- Fc␥RIIbϪ/Ϫ and wild-type mice (Fig. 7B). On CD22AA B cells, ies, mAb blockade of mouse CD22 ligand binding also signifi- CD22 and Fc␥RIIb may be cross-linked by CD22 mAb, yet this cantly reduced the survival of A20 tumor cells in vivo (Fig. 4E), does not affect the phenotype or survival of these cells in vivo or suggesting the ability of CD22 to regulate both normal and ma- in vitro (Fig. 6, A–D). Moreover, with the exception of inducing lignant B cell survival. This raises the possibility that mAb antag- CD22 internalization, CD22 ligand blocking mAbs had no mea- onists that interfere with CD22 ligand binding may demonstrate surable effects in CD22AA mice on mature recirculating bone mar- therapeutic effectiveness for both B cell malignancies and some row, blood, or marginal zone B cell numbers, B cell turnover, or autoimmune diseases. Moreover, the current studies clearly dem- the expression of other cell surface markers as occurred in wild- onstrate that CD22 mAbs have variable effects on B cells in vivo type mice (Fig. 6, A–C). These data suggest that CD22 mAbs most depending on the region of CD22 bound. Thereby, mAbs that likely elicit their effects by interfering with the ability of CD22 to block human CD22 ligand binding may have effects on B cells that bind its ligands rather than indirect effects. are not induced by mAbs that target nonligand binding domains of Wild-type mice treated with CD22 mAbs share similar pheno- CD22. Indeed, an anti-human CD22 ligand-blocking mAb, types with CD22Ϫ/Ϫ, CD22AA, and CD22⌬1-2 mice, including HB22-7, exhibits increased cytotoxic effects on human CD22ϩ significantly reduced numbers of circulating and marginal zone B tumor cells in mice compared with nonblocking mAbs (37, 38). cells (24). Like B220highIgDϩ B cells from mice treated with Interestingly, the therapeutic anti-human CD22 mAb, LL2 (epratu- CD22 ligand-blocking mAbs, mature B cells from CD22Ϫ/Ϫ, zumab), binds to a region outside of the two amino-terminal ligand CD22⌬1-2, and CD22AA mice express low levels of IgM and binding domains of CD22 (64). Thus, a further understanding of CD21, with normal to slightly reduced IgD levels (24, 25, 60). the factors that influence CD22 ligand interactions may lead to However, CD22 mAb-treated mice (MB22-10) only had slightly novel therapies that inhibit the survival of pathogenic B cells. enhanced levels of MHC class II Ag expression, while B cells from CD22Ϫ/Ϫ, CD22⌬1-2, and CD22AA mice have significantly ele- Disclosures vated MHC class II levels (24, 25, 60). Furthermore, CD22 mAb T. F. Tedder is a consultant and equity holder in Angelica Therapeutics Inc. treatment resulted in a dramatic decrease in CD19 expression by and also a consultant for MedImmune Inc. the remaining B cells (Fig. 5, E–H). Whether this phenotypic change directly results from the appearance of a minor B cell sub- References 1. Tedder, T. F., J. Tuscano, S. Sato, and J. H. Kehrl. 1997. CD22, a B lymphocyte- population resistant to CD22 mAb-mediated depletion, CD22 en- specific adhesion molecule that regulates antigen receptor signaling. Annu. Rev. gagement and its effects on B cell signal transduction, or CD22 Immunol. 15: 481–504. internalization in vivo is currently unknown. CD22 has been sug- 2. Erickson, L. D., L. T. Tygrett, S. K. Bhatia, K. H. Grabstein, and T. J. Waldschmidt. 1996. Differential expression of CD22 (Lyb8) on murine B gested to associate with multiple proteins, including CD19 and cells. Int. Immunol. 8: 1121–1129. IgM (19, 20, 61, 62). Thus, phenotypic changes may be due to 3. Schulte, R. J., M. A. Campbell, W. H. Fischer, and B. M. Sefton. 1992. Tyrosine internalization of these proteins along with CD22. Interestingly, B phosphorylation of CD22 during B cell activation. Science 258: 1001–1004. 4. Wilson, G. L., C. H. Fox, A. S. Fauci, and J. H. Kehrl. 1991. cDNA cloning of cells from CD22AA mice demonstrate no additional alterations in the B cell membrane protein CD22: a mediator of B-B cell interactions. J. Exp. cell surface phenotype following CD22 mAb treatment despite sig- Med. 173: 137–146. 2ϩ 5. Poe, J. C., M. Fujimoto, P. J. Jansen, A. S. Miller, and T. F. Tedder. 2000. CD22 nificant CD22 internalization (Fig. 6A). The Ca responses ob- forms a quaternary complex with SHIP, Grb2 and Shc: a pathway for regulation tained with CD22AA B cells following CD22 mAb treatment in of B lymphocyte antigen receptor-induced calcium flux. J. Biol. Chem. 275: vivo also support this possibility. CD22AA and wild-type mice 17420–17427. 2ϩ 6. Doody, G. M., L. B. Justement, C. C. Delibrias, R. J. Mathews, J. Lin, treated in vivo with MB22-10 mAb generated augmented Ca M. L. Thomas, and D. T. Fearon. 1995. A role in B cell activation for CD22 and responses (Fig. 6D), likely due to the loss of cell surface CD22 the protein tyrosine phosphatase SHP. Science 269: 242–244. (Fig. 5, C and D). However, MB22-10 mAb pretreatment in vitro 7. Lankester, A. C., G. M. van Schijndel, and R. A. van Lier. 1995. Hematopoietic cell phosphatase is recruited to CD22 following B cell antigen receptor ligation. 2ϩ did not alter Ca responses in CD22AA B cells, but did in wild- J. Biol. Chem. 270: 20305–20308. type B cells. This may reflect the inability of CD22AA molecules 8. Campbell, M. A., and N. R. Klinman. 1995. Phosphotyrosine-dependent association between CD22 and protein tyrosine phosphatase 1C. Eur. J. Immunol. 25: to associate with cis-ligands and/or specialized microdomains, 1573–1579. such that mAb manipulation had no effect on altering the ability of 9. Law, C.-L., S. P. Sidorenko, K. A. Chandran, Z. Zhao, S.-H. Shen, E. H. Fischer, 2ϩ and E. A. Clark. 1996. CD22 associates with protein tyrosine phosphatase 1C, CD22AA to regulate BCR-mediated Ca responses, while acute ␥ 2ϩ Syk, and phospholipase C- 1 upon B cell activation. J. Exp. Med. 183: 547–560. mAb engagement of wild-type CD22 augmented Ca responses. 10. Blasioli, J., S. Paust, and M. L. Thomas. 1999. Definition of the sites of inter- Alternatively, phenotypic changes in CD22 mAb-treated mice may action between the protein tyrosine phosphatase SHP-1 and CD22. J. Biol. Chem. be due to the emergence of a minor B cell population that does not 274: 2303–2307. 11. Yohannan, J., J. Wienands, K. M. Coggeshall, and L. B. Justement. 1999. Anal- depend on CD22 ligand binding for survival. In addition to alter- ysis of tyrosine phosphorylation-dependent interactions between stimulatory ef- ations in cell surface phenotype, B cells from wild-type mice fector proteins and the B cell co-receptor CD22. J. Biol. Chem. 274: treated with CD22 ligand-blocking mAbs in vitro and B cells from 18769–18776. Ϫ Ϫ 12. Stamenkovic, I., and B. Seed. 1990. The B cell antigen CD22 mediates monocyte CD22⌬1-2, CD22AA, and CD22 / mice are hyperproliferative and erythrocyte adhesion. Nature 344: 74–77. The Journal of Immunology 3073

13. Law, C.-L., A. Aruffo, K. A. Chandran, R. T. Doty, and E. A. Clark. 1995. Ig ments the efficacy of 90Y-DOTA-peptide-Lym-1 in lymphoma xenografts. Blood domains 1 and 2 of murine CD22 constitute the ligand-binding domain and bind 101: 3641–3647. multiple sialylated ligands expressed on B and T cells. J. Immunol. 155: 39. Leonard, J. P., and B. K. Link. 2002. Immunotherapy of non-Hodgkin’s lym- 3368–3376. phoma with hLL2 (epratuzumab, an anti-CD22 monoclonal antibody) and 14. Engel, P., Y. Nojima, D. Rothstein, L.-J. Zhou, G. L. Wilson, J. H. Kehrl, and Hu1D10 (apolizumab). Semin. Oncol. 29: 81–86. T. F. Tedder. 1993. The same epitope on CD22 of B lymphocytes mediates the 40. Carnahan, J., P. Wang, R. Kendall, C. Chen, S. Hu, T. Boone, T. Juan, adhesion of erythrocytes, T and B lymphocytes, neutrophils and monocytes. J. Talvenheimo, S. Montestruque, J. Sun, et al. 2003. Epratuzumab, a humanized J. Immunol. 150: 4719–4732. monoclonal antibody targeting CD22: characterization of in vitro properties. Clin. 15. Kelm, S., A. Pelz, R. Schauer, M. T. Filbin, S. Tang, M.-E. de Bellard, Cancer Res. 9: 3982S–3990S. R. L. Schnaar, J. A. Mahoney, A. Hartnell, P. Bradfield, and P. R. Crocker. 1994. 41. St. Clair, W. E., and T. F. Tedder. 2006. New prospects for autoimmune disease Sialoadhesin, myelin-associated glycoprotein and CD22 define a new family of therapy: B cells on deathwatch. Arthritis Rheum. 54: 1–9. sialic acid-dependent adhesion molecules of the immunoglobulin superfamily. 42. Siegel, A. B., D. M. Goldenberg, A. Cesano, M. Coleman, and J. P. Leonard. Curr. Biol. 4: 965–972. 2003. CD22-directed monoclonal antibody therapy for lymphoma. Semin. Oncol. 16. Sgroi, D., and I. Stamenkovic. 1994. A B-cell superfamily receptor with sialic 30: 457–464. acid-binding lectin activity. The Immunologist 2: 161. 43. Wessels, M. R., P. Butko, M. Ma, H. B. Warren, A. Lage, and M. C. Carroll. 17. Kelm, S., R. Schaur, J. C. Manuguerra, H. J. Gross, and P. R. Crocker. 1994. 1995. Studies of group B streptococcal infection in mice deficient in complement Modifications of cell surface sialic acids modulate cell adhesion mediated by C3 or C4 demonstrate an essential role for complement in both innate and ac- sialoadhesin and CD22. Glycoconjugate J. 11: 576–585. quired immunity. Proc. Natl. Acad. Sci. USA 92: 11490–11494. 18. Engel, P., N. Wagner, A. Miller, and T. F. Tedder. 1995. Identification of the 44. Martin, R. M., J. L. Brady, and A. M. Lew. 1998. The need for IgG2c specific ligand binding domains of CD22, a member of the immunoglobulin superfamily antiserum when isotyping antibodies from C57BL/6 and NOD mice. J. Immunol. that uniquely binds a sialic acid-dependent ligand. J. Exp. Med. 181: 1581–1586. Methods 212: 187–192. 19. Peaker, C. J. G., and M. S. Neuberger. 1993. Association of CD22 with the B cell 45. Poe, J. C., K. M. Haas, J. Uchida, Y. Lee, M. Fujimoto, and T. F. Tedder. 2004. antigen receptor. Eur. J. Immunol. 23: 1358–1363. Severely-impaired B lymphocyte proliferation, survival and induction of the c- 20. Leprince, C., K. E. Draves, R. L. Geahlen, J. A. Ledbetter, and E. A. Clark. 1993. Myc:Cullin 1 ubiquitin ligase pathway resulting from CD22 deficiency on the CD22 associates with the human surface IgM-B cell antigen receptor complex. C57BL/6 genetic background. J. Immunol. 172: 2100–2110. ␥ Proc. Natl. Acad. Sci. USA 90: 3236–3240. 46. Takai, T., M. Li, D. Sylvestre, R. Clynes, and J. V. Ravetch. 1994. FcR chain 21. Pflugh, D. L., S. E. Maher, and A. L. M. Bothwell. 2002. Ly-6 superfamily depletion results in pleiotrophic effector cell defects. Cell 76: 519–529. members Ly-6A/E, Ly-6C, and Ly-6I recognize two potential ligands expressed 47. Ravetch, J. V., and R. A. Clynes. 1998. Divergent roles for Fc receptors and by B lymphocytes. J. Immunol. 169: 5130–5136. complement in vivo. Annu. Rev. Immunol. 16: 421–432. ␥ 22. Stamenkovic, I., D. Sgroi, A. Aruffo, M. S. Sy, and T. Anderson. 1991. The B 48. Nimmerjahn, F., P. Bruhns, K. Horiuchi, and J. V. Ravetch. 2005. Fc RIV: a lymphocyte adhesion molecule CD22 interacts with leukocyte common antigen novel FcR with distinct IgG subclass specificity. Immunity 23: 41–51. CD45RO on T cells and ␣2,6 sialyltransferase, CD75, on B cells. Cell 66: 49. Collins, B. E., O. Blixt, A. R. DeSieno, N. Bovin, J. D. Marth, and J. C. Paulson. 1133–1144. 2004. Masking of CD22 by cis ligands does not prevent redistribution of CD22 23. Han, S., B. E. Collins, P. Bengtson, and J. C. Paulson. 2005. Homomultimeric to sites of cell contact. Proc. Natl. Acad. Sci. USA 101: 6104–6109. complexes of CD22 in B cells revealed by protein-glycan cross-linking. Nat. 50. Collins, B. E., B. A. Smith, P. Bengtson, and J. C. Paulson. 2006. Ablation of Chem. Biol. 1: 93–97. CD22 in ligand-deficient mice restores B cell receptor signaling. Nat. Immunol. 24. Poe, J. C., Y. Fujimoto, M. Hasegawa, K. M. Haas, A. S. Miller, I. G. Sanford, 7: 199–206. C. B. Bock, M. Fujimoto, and T. F. Tedder. 2004. CD22 regulates B lymphocyte 51. Shan, D., and O. W. Press. 1995. Constitutive endocytosis and degradation of J. Immunol. function in vivo through both ligand-dependent and ligand-independent mecha- CD22 by human B cells. 154: 4466–4475. nisms. Nat. Immunol. 5: 1078–1087. 52. John, B., B. R. Herrin, C. Raman, Y. Wang, K. R. Bobbitt, B. A. Brody, and L. B. Justement. 2003. The B cell coreceptor CD22 associates with AP50, a 25. Sato, S., A. S. Miller, M. Inaoki, C. B. Bock, P. J. Jansen, M. L. K. Tang, and clathrin-coated pit adapter protein, via tyrosine-dependent interaction. J. Immu- T. F. Tedder. 1996. CD22 is both a positive and negative regulator of B lym- nol. 170: 3534–3543. phocyte antigen receptor signal transduction: altered signaling in CD22-deficient 53. Ghetie, M.-A., R. D. May, M. Till, J. W. Uhr, V. Ghetie, P. P. Knowles, M. Relf, mice. Immunity 5: 551–562. A. Brown, P. M. Wallace, G. Janossy, et al. 1988. Evaluation of ricin A chain- 26. O’Keefe, T. L., G. T. Williams, S. L. Davies, and M. S. Neuberger. 1996. Hy- containing immunotoxins directed against CD19 and CD22 antigens on normal perresponsive B cells in CD22-deficient mice. Science 274: 798–801. and malignant human B-cells as potential reagents for in vivo therapy. Cancer 27. Otipoby, K. L., K. B. Andersson, K. E. Draves, S. J. Klaus, A. G. Farr, Res. 48: 2610–2617. J. D. Kerner, R. M. Perlmutter, C.-L. Law, and E. A. Clark. 1996. CD22 regulates 54. Samardzic, T., D. Marinkovic, C.-P. Danzer, J. Gerlach, L. Nitschke, and thymus-independent responses and the lifespan of B cells. Nature 384: 634–637. T. Wirth. 2002. Reduction of marginal zone B cells in CD22-deficient mice. Eur. 28. Nitschke, L., R. Carsetti, B. Ocker, G. Kohler, and M. C. Lamers. 1997. CD22 is J. Immunol. 32: 561–567. a negative regulator of B-cell receptor signaling. Curr. Biol. 7: 133–143. 55. Danzer, C. P., B. E. Collins, O. Blixt, J. C. Paulson, and L. Nitschke. 2003. 29. Lanoue, A., F. D. Batista, M. Stewart, and M. S. Neuberger. 2002. Interaction of Transitional and marginal zone B cells have a high proportion of unmasked ␣ CD22 with 2,6-linked sialoglycoconjugates: innate recognition of self to CD22: implications for BCR signaling. Int. Immunol. 15: 1137–1147. dampen B cell autoreactivity? Eur. J. Immunol. 32: 348–355. 56. Uchida, J., Y. Hamaguchi, J. A. Oliver, J. V. Ravetch, J. C. Poe, K. M. Haas, and 30. Jin, L., P. A. McLean, B. G. Neel, and H. H. Wortis. 2002. Sialic acid binding T. F. Tedder. 2004. The innate mononuclear phagocyte network depletes B lym- domains of CD22 are required for negative regulation of B cell receptor signal- phocytes through Fc receptor-dependent mechanisms during anti-CD20 antibody ing. J. Exp. Med. 195: 1199–1205. immunotherapy. J. Exp. Med. 199: 1659–1669. 31. Kelm, S., J. Gerlach, R. Brossmer, C.-P. Danzer, and L. Nitschke. 2002. The 57. Hamaguchi, Y., J. Uchida, D. W. Cain, G. M. Venturi, J. C. Poe, K. M. Haas, and ligand-binding domain of CD22 is needed for inhibition of the B cell receptor T. F. Tedder. 2005. The peritoneal cavity provides a protective niche for B1 and signal, as demonstrated by a novel human CD22-specific inhibitor compound. conventional B lymphocytes during anti-CD20 immunotherapy in mice. J. Im- J. Exp. Med. 195: 1207–1213. munol. 7: 4389–4399. 32. Symington, F. W., B. Subbarao, D. E. Mosier, and J. Sprent. 1982. Lyb-8.2: a 58. Hamaguchi, Y., Y. Xiu, K. Komura, F. Nimmerjahn, and T. F. Tedder. 2006. new B cell antigen defined and characterized with a monoclonal antibody. Im- Antibody isotype-specific engagement of Fc␥ receptors regulates B lymphocyte munogenetics 16: 381–391. depletion during CD20 immunotherapy. J. Exp. Med. 203: 743–753. 33. Law, C.-L., R. M. Torres, H. A. Sundberg, R. M. E. Parkhouse, C. I. Brannan, 59. Yazawa, N., Y. Hamaguchi, J. C. Poe, and T. F. Tedder. 2005. Immunotherapy N. G. Copeland, N. A. Jenkins, and E. A. Clark. 1993. Organization of the murine using unconjugated CD19 monoclonal antibodies in animal models for B lym- Cd22 locus: mapping to chromosome 7 and characterization of two alleles. J. Im- phocyte malignancies and autoimmune disease. Proc. Natl. Acad. Sci. USA 102: munol. 151: 175–187. 15178–15183. 34. Dunussi-Joannopoulos, K., G. E. Hancock, A. Kunz, M. Hegen, X. X. Zhou, 60. Amano, H., E. Amano, T. Moll, D. Marinkovic, N. Ibnou-Zekri, B. J. Sheppard, J. Lamothe, E. Li, H. L. Ma, P. R. Hamann, et al. 2005. B-cell E. Martinez-Soria, I. Semac, T. Wirth, L. Nitschke, and S. Izui. 2003. The Yaa depletion inhibits arthritis in a collagen-induced arthritis (CIA) model, but does mutation promoting murine lupus causes defective development of marginal zone not adversely affect humoral responses in a respiratory syncytial virus (RSV) B cells. J. Immunol. 170: 2293–2301. vaccination model. Blood 106: 2235–2243. 61. Carter, R. H., G. M. Doody, J. B. Bolen, and D. T. Fearon. 1997. Membrane 35. Nitschke, L., H. Floyd, D. J. P. Ferguson, and P. R. Crocker. 1999. Identification IgM-induced tyrosine phosphorylation of CD19 requires a CD19 domain that of CD22 ligands on bone marrow sinusoidal endothelium implicated in CD22- mediates association with components of the B cell antigen receptor complex. dependent homing of recirculating B cells. J. Exp. Med. 189: 1513–1518. J. Immunol. 158: 3062–3069. 36. Tuscano, J., P. Engel, T. F. Tedder, and J. H. Kehrl. 1996. Engagement of the 62. Zhang, M., and A. Varki. 2004. Cell surface sialic acids do not affect primary adhesion receptor CD22 triggers a potent stimulatory signal for B cells and block- CD22 interactions with CD45 and surface IgM nor the rate of constitutive CD22 ing CD22/CD22L interactions impairs T-cell proliferation. Blood 87: endocytosis. Glycobiology 14: 939–949. 4723–4730. 63. Tuscano, J. M., P. Engel, T. F. Tedder, A. Agarwal, and J. H. Kehrl. 1996. 37. Tuscano, J. M., A. Riva, S. N. Toscano, T. F. Tedder, and J. H. Kehrl. 1999. Involvement of p72syk kinase, p53/56lyn kinase and phosphatidyl inositol-3 ki- CD22 cross-linking generates B-cell antigen receptor-independent signals that nase in signal transduction via the human B lymphocyte antigen CD22. Eur. activate the JNK/SAPK signaling cascade. Blood 94: 1382–1392. J. Immunol. 26: 1246–1252. 38. Tuscano, J. M., R. T. O’Donnell, L. A. Miers, L. A. Kroger, D. L. Kukis, 64. Stein, R., E. Belisle, H. J. Hansen, and D. M. Goldenberg. 1993. Epitope spec- K. R. Lamborn, T. F. Tedder, and G. L. DeNardo. 2003. The anti-CD22 ligand ificity of the anti-(B cell lymphoma) monoclonal antibody, LL2. Cancer Immu- blocking antibody, HB22.7, has independent lymphomacidal properties, and aug- nol. Immunother. 37: 293–298.