Published online before print October 18, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
,
,1
* Clinical Research Division and
Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, Washington, USA; and Departments of
Pathology and
|| Pediatrics and
Division of Hematology, Department of Medicine, University of Washington, Seattle, Washington, USA
1 Correspondence: Clinical Research Division, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave. N., D2-373, Seattle, WA 98109-1024, USA. E-mail: rwalter{at}fhcrc.org
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: neutrophils • monocytes/macrophages • leukocyte differentiation antigen • antibodies • host defense • inhibitory immunoreceptor
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Structurally, CD33-related Siglecs are characterized by one N-terminal V-set Ig domain mediating sialic acid binding, followed by variable numbers of C2-set Ig domains, ranging from 4 (in the case of Siglec-10 and -11) to 1 (in the case of CD33) [1 , 3 , 4 ]. All members, except Siglec-14, have a conserved, intracellular proximal ITIM (with its tyrosine at position 340 in the case of CD33) and a distal ITIM-like motif (with its tyrosine at position 358 in the case of CD33) [1 , 3 4 5 ]. Generally, when ITIM-containing receptors are engaged, they can become tyrosine-phosphorylated and then transmit inhibitory signals by binding and activating Src homology-2 domain (SH2)-containing tyrosine phosphatases (Shp1 and Shp2) and/or the SH2-containing inositol polyphosphate 5'-phosphatase (SHIP) [6 ]. In the case of CD33, several studies have demonstrated that Shp1 and Shp2, but not SHIP, are recruited and activated once the ITIMs are phosphorylated [7 8 9 ].
Although it is assumed that CD33-related Siglecs have important roles in modulating leukocyte behavior, including inhibition of proliferation or cellular activation, modulation of cytokine secretion, and induction of apoptosis, their precise signaling functions remain to be determined [1
, 3
, 4
]. Nevertheless, recent studies have demonstrated their interactions with various sialylated pathogens and suggested their potential importance in host defense and pathogenicity. For example, several pathogenic microorganisms, including Neisseria meningitides, Group B Streptococci, and Campylobacter jejuni, have been shown to bind to CD33-related Siglecs [10
11
12
]. Furthermore, there is increasing evidence highlighting the endocytic capacities of human CD33-related Siglecs [13
14
15
16
], and it has been speculated that this might be of physiological relevance for clearance of sialylated antigens and modulation of antigen presentation [1
]. Moreover, the early recognition that CD33 undergoes endocytosis led to the development of gemtuzumab ozogamicin (GO; MylotargTM), an immunoconjugate consisting of a humanized IgG4 anti-CD33 mAb (hP67.6) joined to a toxic calicheamicin-
1 derivative [17
, 18
]. Endocytosis of CD33 and bound GO results in cellular uptake of the drug, which is then cleaved intracellularly to release the toxic moiety [18
, 19
]. CD33 is present on tumor cells of 85–90% of adult and pediatric patients with acute myeloid leukemia (AML), and the therapeutic success of GO depends largely on its CD33-dependent uptake. More recent data suggest the potential clinical exploitation of other CD33-related Siglecs, such as Siglec-8 for targeting eosinophils in allergic inflammation or other disease states, or Siglec-9 for treatment of AML or inflammatory diseases [15
, 16
, 20
, 21
].
The mechanisms underlying the endocytosis of CD33-related Siglecs remain elusive. Given the high sequence identity within this subset of Siglecs, some common principles seem likely. Indeed, studies of CD33, Siglec-5, Siglec-F, and Siglec-9 have shown that they all endocytose slowly when bound to antibody. Also, CD33, Siglec-F, and Siglec-9 use their ITIMs for endocytosis [14 , 16 , 22 ]. However, it remains unclear whether the ITIMs stimulate endocytosis when they are phosphorylated or when they are not phosphorylated. AP-2 is a potential endocytic adaptor protein for the nonphosphorylated ITIMs, and other adaptors may bind to phosphorylated or nonphosphorylated tyrosine motifs [23 24 25 ]. For example, the YXXØ motif (where Ø denotes an amino acid residue with a bulky hydrophobic side-chain) is also found in FcRs, where it mediates internalization in a largely AP-2-independent manner when the tyrosine is phosphorylated [26 ]. As a further complication, endocytosis of CD22 was shown recently to be increased when glycosylation is inhibited [27 ]. The primary objective of the present study was therefore to investigate the endocytosis of CD33-related Siglecs in more detail, using CD33 as paradigm. Specifically, we asked the question of whether the slow rate of endocytosis is the result of CD33 extracellular or intracellular domains and whether phosphorylation or dephosphorylation favors uptake of antibody-bound CD33. We identified proteins that bind to CD33 in an ITIM-dependent manner and assessed their importance for CD33 internalization in human myeloid cells by specific silencing of target genes through expression of small interfering (si)RNA.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Antibodies
Primary antibodies used were anti-CD33 (kindly provided by Wyeth-Ayerst Research, Radnor, PA, USA), anti-hemagglutinin (HA; Roche Diagnostics, Indianapolis, IN, USA; Covance, Berkeley, CA, USA), antiphosphotyrosine and antiphospholipase C (anti-PLC)-1 (Upstate Biotechnology, Charlottesville, VA, USA), anti-T7 (Novagen, Madison, WI, USA), anti-MAPK [29
], anti-Shp1 (BD Biosciences PharMingen, San Diego, CA, USA; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-Shp2, anti-spleen tyrosine kinase (anti-Syk), and anti-Crk-Like protein (anti-CrkL) (Santa Cruz Biotechnology), anti-glutathione-S-transferase (anti-GST) (clone 38.3), anti-AP-2 (ABR-Affinity BioReagents, Golden, CO, USA), and antimurine Thy 1.1 (used as murine IgG1 isotype control antibody; clone 31A [30
]). F(ab) and F(ab)2 fragments of the murine IgG1 anti-CD33 antibody (mP67.6) were prepared by digestion with papain and pepsin, respectively (both from Sigma Chemical Co., St. Louis, MO, USA). Secondary antibodies used were HRP-conjugated sheep anti-mouse Ig, goat anti-rat IgG, and donkey anti-rabbit Ig (GE Healthcare, Amersham Biosciences, Buckinghamshire, UK), as well as biotin-conjugated mouse anti-human IgG4 and rat anti-mouse Ig 
light chain or IgG1 (BD Biosciences PharMingen).
Cellular transfection with cDNA constructs
A C-terminal T7-tag was introduced into pcDNA3 vectors containing CD33WT, CD33Y340F, CD33Y358F, and CD33Y340F/Y358F (kindly provided by Dr. Daniel W. McVicar, National Cancer Institute, Frederick, MD, USA [9
]) by site-directed mutagenesis (see Table 1
for primer sequences); all constructs were verified by sequencing. A pME vector encoding wild-type (WT) human Fyn was kindly provided by Drs. Toru Tezuka and Tadashi Yamamoto, University of Tokyo (Japan) [31
]. Constructs encoding WT and catalytically inactive Shp1 (Shp1WT and Shp1C453S) were kindly provided by Dr. Benjamin G. Neel, Beth Israel Deaconess Medical Center (Boston, MA, USA). HEK293T and COS-7 cells, grown in six-well plates (Costar, Corning, NY, USA), were transiently transfected with polyethylenimine (Polysciences, Warrington, PA, USA) or lipofectamine 2000 (Gibco-Invitrogen) and harvested 48 h after transfection.
|
View this table: [in a new window] |
Table 1. List of Mutagenesis Primers
|
F; amino acid single letter code), 66 (TG), 69 (VL), 73 (SA), and 204 (TY). Target sequences were selected using siDESIGNTM Center (Dharmacon, Lafayette, CO, USA) and/or Block-iTTM RNAi Designer (Gibco-Invitrogen) against the following genes (target sites referred to by sense strand with base position indicated relative to translation start nucleotide): Shp1 (NM_002831), GCAAGAACCGCTACAAGAA, bp 989 (construct A), and TGACACAACCGAATACAAA, bp 1329 (construct B); Shp2 (NM_002834), GCAATGACGGCAAGTCTAA, bp 859 (construct A), ATATGGCGGTCCAGCATTA, bp 1924 (construct B), and ACACTGGTGATTACTATGA, bp 553 (construct C); and Syk, GCACTATCGCATCGAACAAA (NM_003177, bp 789). Vesicular stomatitis virus G glycoprotein-pseudotype lentiviral vectors were prepared as described [14
]. Hematopoietic cell lines were infected in fibronectin-coated wells in the presence of 8 µg/mL protamine sulfate at a multiplicity of infection (MOI) of 2.5–100. EGFP and/or YFP-positive cells were sorted by flow cytometry and recultured for analysis.
Immunoprecipitation and Western blotting
Because of the poor performance of anti-CD33 antibodies, cells expressing T7- or HA-tagged CD33 were used for immunoprecipitations. Cells were washed twice in ice-cold Dulbecco’s PBS (Gibco-Invitrogen) and then lysed in 25 mM HEPES (pH 7.4), 100 mM NaCl, 50 mM NaF, 1% Nonidet P-40 (Igepal CA-630, Sigma Chemical Co.), 10% glycerol, 1 mM PMSF, and protease inhibitors (Roche Diagnostics, Mannheim, Germany); in some experiments, cells were treated with 100 µM pervanadate for 15–30 min at 37°C prior to harvest, and 1 mM sodium orthovanadate was added to the lysis buffer. After incubation on ice for 45 min, lysates were precleared by centrifugation. For immunoprecipitations, precleared lysates were incubated with 1 µg antibody overnight at 4°C, and then Protein A Sepharose CL-4B or Protein G Sepharose 4 Fast Flow beads (both GE Healthcare) added for 2 h. Beads were then washed four times in ice-cold lysis buffer and boiled in 1x reducing sample buffer {2% SDS, 10% glycerol, 20% β-ME, 105 mM Tris [tris(hydroxymethyl)aminomethane]/HCl, 0.03% bromophenol blue}. In a subset of experiments, washed beads were treated with N-glycosidase F (PNGase F; New England Biolabs, Ipswich, MA, USA), according to the manufacturer’s instructions, before boiling in reducing sample buffer. For Western blotting, samples were separated on 7.5% or 10% SDS-PAGE, transferred to nitrocellulose, blocked with 5% nonfat dry milk in TBS [50 mM Tris (pH 7.5), 150 mM NaCl] containing 0.1% Tween 20 (TBST) for 30 min at room temperature, and then incubated with primary antibodies overnight at 4°C. For assessment of tyrosine phosphorylation, blots were blocked with TBS containing 5% BSA (Sigma Chemical Co.), 0.05% Tween 20, 5 mM NaF, and 0.005% NaN3 for 24–48 h prior to incubation with primary antibody. After washing, blots were incubated for 60 min with HRP-conjugated secondary antibody in TBST/BSA and washed, and immunoreactive signals were visualized with ECL (PerkinElmer, Boston, MA, USA) and CL-XPosure film (Pierce, Rockford, IL, USA).
GST fusion proteins and pull-down assays
To generate CD33/GST fusion proteins, cytoplasmic tails of WT or mutant CD33 (residues 289–364) were subcloned into pGEX2T. Fusion proteins were expressed after isopropyl-β-D-thiogalactoside (IPTG) induction in Escherichia coli TKB1 (Stratagene, La Jolla, CA, USA) as tyrosine-phosphorylated proteins, respectively. The TKB1 strain contains an Elk1 tyrosine kinase that can be induced with indoleacrylic acid to phosphorylate bacterial fusion proteins expressed after IPTG induction in the same cells. Bacteria were lysed in ice-cold PBS containing 1% Triton X-100 (Sigma Chemical Co.), 1 mM phenylarsine oxide, 0.1% β-ME, 1 mM PMSF, and 1 mM EDTA. After sonication, lysates were cleared by centrifugation and stored at –20°C. Phosphorylation of the CD33 tails was verified by Western blot. For pull-down assays, GST fusion proteins were incubated with a 50% slurry of Glutathione Sepharose 4B beads (GE Healthcare) in lysis buffer and rotated for 1 h at 4°C. Beads were isolated by centrifugation, washed four times with ice-cold lysis buffer, and then incubated with cell lysates from human myeloid cell lines, which were or were not stimulated with pervanadate prior to lysis to promote tyrosine phosphorylation. After rotating for 2 h at 4°C, beads were isolated again by centrifugation, washed four times with ice-cold lysis buffer, boiled in 1x reducing sample buffer, and analyzed by Western blotting.
Flow cytometry assays for determination of cell surface receptor expression and internalization
Cell surface expression and internalization of antibody-bound CD33 or HA-tagged chimeric receptor constructs were measured using flow cytometry-based assays as described [14
]. Briefly, to measure internalization of antibody-bound receptors, cells (typically 1–1.5x106) were transferred into 5 mL polystyrene round-bottom tubes (BD FalconTM, BD Biosciences PharMingen) and incubated for at least 30 min with IMDM containing 2.5 µg/mL unconjugated, unlabeled anti-HA antibody, anti-CD33 antibody, or anti-CD33 antibody fragments in ice water to prevent internalization during the staining procedure. Cells were then washed in ice-cold PBS to remove unbound antibody or antibody fragments; experiments in CD33-negative 32D cells indicated that this washing step removed all unbound and nonspecifically bound CD33 antibody (data not shown), thus allowing us to measure receptor-mediated uptake of CD33 antibody or antibody fragments separate from possible nonspecific antibody uptake in our assay [37
]. Subsequently, cells were resuspended in IMDM without antibody or antibody fragment, split into several tubes, and incubated at 37°C (in 5% CO2 and air) for various periods of time, after which cells were chilled and incubated with appropriate biotin-conjugated secondary mAb (used at 5 µg/mL in PBS/2% FBS), followed by incubation with streptavidin-PE conjugate (used at 5 µg/mL in PBS/2% FBS, BD Biosciences PharMingen) to detect remaining antibody on the cell surface. One sample that was kept in ice water was used to determine the starting level of antibody bound to the cell surface and to estimate cell surface expression of CD33/constructs. To identify nonviable cells, all samples were stained with propidium iodide (PI; Sigma Chemical Co.). At least 10,000 events were acquired, and PI– cells were analyzed on a FACScan flow cytometer using Cellquest software (BD Biosciences PharMingen). Linear fluorescence values were used to calculate the percentage of antibody internalization. In some experiments, cells were incubated with pervanadate and/or the Src family kinase inhibitor PP2 (Calbiochem, EMD Biosciences, La Jolla, CA, USA) or DMSO vehicle during antibody internalization. Antibody titration experiments performed with hP67.6 labeling (0.1–5 µg/mL) of ML-1 cells indicated that 2.5 µg/mL were saturating and that the percentage of antibody internalization was affected little by the degree of saturation of binding sites, i.e., whether saturating or nonsaturating antibodies were used for labeling of binding sites (data not shown). Additional antibody titration experiments performed with anti-HA labeling (0.1–5 µg/mL) of 32D cells transduced with CD33/LDLR indicated that 2.5 µg/mL was subsaturating but that the percentage of antibody internalization was not affected by the degree of saturation of binding sites (data not shown).
Statistical analysis
Results from CD33 expression and internalization studies are presented as means ± SEM from at least three independent experiments. Parametric statistical analysis was performed using repeated measures of ANOVA with Tukey-Kramer multiple comparisons test (InStat 3.05, GraphPad, San Diego, CA, USA); P < 0.05 was considered significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (32K): [in a new window] |
Figure 1. Endocytosis of chimeric receptor proteins. Human Jurkat T cells (A), murine myeloid 32D (B), human NB4 cells (C), and human TF-1 cells (D) were virally transduced with HA-tagged chimeric receptors at a MOI of 5 as indicated. To assess endocytosis of chimeric receptors (left panel), cells were labeled with anti-HA antibody (HA.11, clone 16B12, 2.5 µg/mL) in ice water before washing and incubation in 37°C in antibody-free medium to allow internalization for up to 2 h as indicated. Subsequently, remaining cell surface-associated antibody was detected with biotin-conjugated anti-mouse IgG1 secondary antibody and a streptavidin-PE conjugate. The percentage of internalized antibody/antibody fragment is expressed relative to cells kept at 0°C. The arbitrary fluorescence units (AFU) of cells kept at 0°C were used as a measure of the relative receptor expression levels (right panels). Results are shown as mean ± SEM from three to four independent experiments.
|
 View larger version (28K): [in a new window] |
Figure 2. ITIM-dependent internalization of F(ab) fragments of anti-CD33 antibody in transduced 32D cells. Sublines of mouse myeloid 32D cells transduced with WT CD33 or CD33 mutants were labeled with unconjugated F(ab) fragments of mP67.6 (2.5 µg/mL) on ice water before the cells were incubated at 37°C in antibody-free medium to allow internalization of antibody-bound CD33 for up to 1 h. Subsequently, remaining cell surface-associated F(ab) fragments were detected with biotin-conjugated secondary antibody and a streptavidin-PE conjugate. Results are shown as mean ± SEM from three to five independent experiments. The cell surface levels of the CD33 constructs, determined as AFU, varied less than 1.8-fold (WT: 1482.8±205.2; Y340F: 2594.6±76.6; Y358F: 2699.3±230.2; and Y340/358F: 2455.0±499.0).
|
 View larger version (45K): [in a new window] |
Figure 3. Effect of pervanadate on CD33 tyrosine phosphorylation and uptake of antibody-bound CD33. (A) HL-60 cells infected with HA-tagged WT CD33 were left untreated or stimulated with pervanadate (100 µM) for 30 min at 37°C prior to lysis and immunoprecipitation (IP) with anti-HA antibody (HA.11, clone 16B12); parallel samples were then treated with glycosidase (PNGase F) or left untreated. The blot was probed subsequently for phosphotyrosine (p-Tyr; upper), stripped, and reprobed for HA (clone 3F10) to visualize CD33 (lower). (B and C) The indicated human CD33+ AML cell lines were labeled with (B) the unconjugated human IgG4 anti-CD33 antibody hP67.6 (2.5 µg/mL) or (C) the unconjugated F(ab)2 fragment of the murine IgG1 anti-CD33 antibody mP67.6 (2.5 µg/mL) on ice water before the cells were incubated in 37°C in antibody-free medium in the presence or absence of pervanadate (100 µM) to allow internalization for up to 1 h as indicated. Subsequently, remaining cell surface-associated hP67.6 or mP67.6 F(ab)2 was detected with biotin-conjugated secondary antibody and a streptavidin-PE conjugate. The percentage of internalized antibody/antibody fragment is expressed relative to cells kept at 0°C. Results are shown as mean ± SEM from three to seven independent experiments with the exception of TF-1 cells (n=2). (D) HEK293T cells were transfected with T7-tagged WT or mutant CD33 as indicated. After 2 days, cells were left untreated or stimulated with pervanadate (100 µM) for 15 min at 37°C prior to lysis and immunoprecipitation with anti-T7 antibody; the blot was then probed for phosphotyrosine (upper), stripped, and reprobed for T7 to visualize CD33 (lower). Presented is one representative experiment; similar findings were obtained with COS-7 cells.
|
 View larger version (21K): [in a new window] |
Figure 4. Effect of pervanadate and/or PP2 on uptake of antibody-bound CD33. (A) The human CD33– Jurkat T cell line was virally transduced with WT or mutant CD33 constructs as indicated, and internalization of the human IgG4 anti-CD33 antibody hP67.6 (incubated at 2.5 µg/mL to label CD33-binding sites) was assessed for up to 1 h in the presence or absence of pervanadate (100 µM). (B) Internalization of the unconjugated F(ab)2 fragment of the murine IgG1 anti-CD33 antibody mP67.6 (2.5 µg/mL) was assessed in the presence or absence of pervanadate (100 µM) for up to 45 and 30 min in Jurkat and murine myeloid 32D cells virally transduced with WT or mutant CD33 constructs as indicated, respectively. (C) Endocytosis assays were performed with Jurkat cells transduced with WT CD33 (labeled with hP67.6 at 2.5 µg/mL) or HL-60 cells endogenously expressing CD33 [labeled with the F(ab)2 fragment of mP67.6 at 2.5 µg/mL] for up to 1 h in the presence or absence of pervanadate (100 µM) and/or the Src family tyrosine kinase inhibitor PP2 or DMSO vehicle as indicated. The percentage of internalized antibody/antibody fragment is expressed relative to cells kept at 0°C. Results are shown as mean ± SEM from two to four independent experiments. The cell surface levels of the CD33 constructs in the transduced Jurkat sublines, determined as AFU, varied less than 1.6-fold (WT: 1135.4±131.41; Y340F: 1166.5±215.22; L343A: 1807.2±313.6; Y358F: 1795.2±325.83; and Y340/358F: 1521.0±300.91).
|
We then tested for binding of Shp1 and other SH2 domain-containing proteins, including the related tyrosine phosphatase Shp2, Syk, CrkL, and PLC-1 to GST/CD33 using Western blotting. As shown in Figure 5
, all of these proteins bound to the phosphorylated GST/CD33 protein in vitro. To determine the specific residues in CD33 that are involved, WT or mutant GST/CD33 proteins were incubated with cell extracts and bound proteins detected using Western blotting. Although Shp1, Shp2, and Syk bound to phosphorylated GST/CD33WT, they only interacted weakly with phosphorylated GST/CD33Y358F and failed to interact with GST/CD33Y340F or CD33Y340F/Y358F (Fig. 5A)
. Thus, binding of CD33 mutants to Shp1, Shp2, and Syk paralleled their internalization, revealing Shp1, Shp2, and Syk as potential candidates playing a role in anti-CD33 antibody endocytosis. In contrast, binding of CrkL and PLC-1 was more dependent on Y358 than on Y340, inconsistent with internalization. To test whether CD33 binds to Shp1, Shp2, and Syk in cells in response to pervanadate treatment, we tested whether each protein coimmunoprecipitated with CD33. As shown in Figure 5B
and 5C
, Shp1 and Shp2 bound to CD33 in pervanadate-pretreated but not control cell lysates (Fig. 5B
and 5C)
, and Syk failed to bind to CD33. Consistent with the inability of anti-CD33 antibody to elicit sustained CD33 tyrosine phosphorylation, we were unable to detect binding of Shp1 or Shp2 to CD33 following cellular stimulation with anti-CD33 antibodies (data not shown).
 View larger version (32K): [in a new window] |
Figure 5. ITIM- and phosphotyrosine-dependent interactions between CD33 and SH2 domain proteins. (A) GST proteins, fused to cytoplasmic tails of WT and mutant CD33, were expressed in E. coli and used in pull-down assays with cell lysates from human myeloid cells that were stimulated with pervanadate (100 µM) for 30 min at 37°C prior to lysis. After beads were washed, proteins were separated by SDS-PAGE, and blots were probed for Shp1, Shp2, Syk, CrkL, and PLC-1 as indicated. Blots were then stripped and reprobed with anti-GST antibody to ascertain equal amounts of fusion proteins. Blots were also probed with antiphosphotyrosine antibody to verify phosphorylation of the GST fusion proteins (not shown). Similar results were obtained with cell lysates from HL-60, NB4, and ML-1 cells. (B) NB4 cells transduced with HA-tagged CD33WT were incubated in the presence or absence of pervanadate (100 µM) for 15 min at 37°C, lysed, and immunoprecipitated with anti-HA antibody. Blots were then probed with anti-Shp1, anti-Shp2, and anti-Syk antibodies as indicated. Similar results were obtained with HL-60 and ML-1 cells. (C) NB4 cells transduced with HA-tagged CD33WT were incubated in the presence or absence of pervanadate (100 µM) for 15 min at 37°C, lysed, and immunoprecipitated with isotype control antibody (31.A), anti-Shp1, anti-Shp2, or anti-Syk as indicated. Blots were then probed with anti-HA antibody. Similar findings were obtained with transduced HL-60, U937, and ML-1 cells.
|
 View larger version (45K): [in a new window] |
Figure 6. Effect of tyrosine phosphatases on CD33 phosphorylation and internalization of anti-CD33 antibody. (A) NB4 cells were virally transduced with siRNAs targeting Shp1 and/or Shp2 as indicated. Cell lysates were prepared, and Western blots were performed to assess protein knockdown. Shown is one representative experiment. (B–E) Internalization of hP67.6 (incubated at 2.5 µg/mL to label CD33-binding sites) was assessed in TF-1 (B), ML-1 (C), HL-60 (D), and NB4 (E) cells over the indicated times and expressed as percentage of internalized antibody relative to cells kept at 0°C. Results are shown as mean ± SEM from three to four independent experiments. *, P < 0.05; **, P < 0.01, compared with corresponding, parental cells.
|
Depletion of Syk by siRNA does not affect CD33 endocytosis
Last, infection of human myeloid cells with a siRNA construct targeting Syk resulted in efficient Syk protein knockdown but unaltered levels of Shp1 and Shp2 (Fig. 7A
). However, in all cell lines tested (TF-1, ML-1, HL-60, NB4), Syk siRNA-infected sublines showed similar rates of anti-CD33 antibody internalization (Fig. 7B
and 7C)
and comparable cell surface expression levels of CD33 relative to parental cells (data not shown), indicating that Syk is dispensable for internalization of antibody-bound CD33 and providing a control for the specificity of the Shp1 and Shp2 siRNA experiments.
 View larger version (29K): [in a new window] |
Figure 7. Effect of Syk siRNA on internalization of anti-CD33 antibody. (A) NB4 and HL-60 cells were virally transduced with siRNAs targeting Syk at a MOI of 5 or left untreated as indicated. Cell lysates were prepared, and Western blots were performed to verify efficient Syk protein knock-down and unaffected Shp1 and Shp2 protein levels. (B and C) Internalization of hP67.6 (incubated at 2.5 µg/mL to label CD33-binding sites) was assessed in TF-1 and ML-1 (B) as well as NB4 and HL-60 (C) cells over the indicated times and expressed as percentage of internalized antibody relative to cells kept at 0°C. Results are shown as mean ± SEM from three to four independent experiments.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1, but only Shp1, Shp2, and Syk bind with appropriate specificity to participate in CD33 internalization. Last, the tyrosine phosphatases Shp1 and Shp2 can, to a modest degree, interfere with the internalization of CD33, presumably through dephosphorylation or masking of phosphorylated CD33, whereas the tyrosine kinase Syk is not involved in CD33 endocytosis. In addition to several members of the CD33-related subset of Siglecs, the more distantly related CD22/Siglec-2, which sets a threshold for antigen-induced activation and regulates B cell survival and homeostasis, also undergoes endocytosis [1 ]. The cytoplasmic tail of CD22 contains as many as six tyrosine-based motifs, including three ITIMs [1 ], the two distal of which have been shown to bind the AP50 subunit of the AP-2 adaptor complex that mediates association with clathrin [38 ]. As a consequence, CD22 is localized predominantly in clathrin-rich microdomains, where it undergoes constitutive endocytosis [27 , 38 , 43 44 45 46 ]. Recent studies showed that sialic acid not only modulates the signaling functions of CD22 but also affects its endocytosis. Using experimental mice deficient in ST6Gal I, the sialyltransferase that produces ligands for CD22, Grewal et al. [27 ] provided evidence that protein glycosylation limits the basal rate of CD22 endocytosis and degradation. However, our experiments using chimeric receptor proteins indicate that endocytosis of CD33 is only slightly restricted by its extracellular and/or transmembrane domains. Restriction was most evident in cell lines with a low rate of CD33 internalization (Jurkat and NB4 cells, see Fig. 1 ). Additional experiments will be necessary to test whether this restraint is a result of sialic acid-dependent interactions, as suggested by the CD22 studies, or involves other types of interactions.
The major reason for slow endocytosis of CD33 is its intracellular domain, as replacement of the intracellular domain of CD33 by the cytoplasmic tail of LDLR resulted in a significant increase in the rate of receptor endocytosis (Fig. 1) . Cell surface proteins that enter the receptor-mediated endocytosis pathway can be cleared rapidly from the surface of our transduced cells if they are concentrated efficiently in clathrin-coated pits [47 ]. By comparison, endocytosis of chimeric receptors with CD33 tails occurred at significantly lower rates than those with LDLR tails, indicating that the cytoplasmic tail of CD33 is not able to recruit the endocytic machinery as efficiently as the LDLR tail. Furthermore, although all tested cell lines showed similarly high rates of endocytosis of the receptors with LDLR tails, we noted significant differences in the ability of the cells to internalize receptors with CD33 tails, ranging from barely measurable internalization (in Jurkat cells) to internalization that approached that of LDLR constructs (in 32D cells), pointing to significant, cell-type specific differences of the CD33 cytoplasmic tail, for example, with regard to phosphorylation status and/or recruitment of endocytic adaptor proteins.
Previous findings about CD33, CD22, Siglec-F, and Siglec-9 have demonstrated the importance of ITIM or ITIM-like sequences for internalization [14 , 16 , 22 ]. However, our current studies indicate significant mechanistic differences between CD33 and CD22. First, although AP-2 was identified to bind to two nonphosphorylated tyrosine motifs in the carboxyl-terminus of CD22 to mediate clathrin-dependent internalization [38 ], we were unable to demonstrate AP-2 binding to CD33 by GST fusion protein pull-down experiments or coimmunoprecipitations (data not shown). Second, although phosphorylation of CD22 inhibits CD22 internalization [38 ], consistent with the preferred binding of AP-2 to nonphosphorylated tyrosine motifs, we found that forced tyrosine phosphorylation of CD33 by pervanadate is associated with increased uptake of antibody-bound CD33 in a Src family kinase-dependent and ITIM-dependent manner. Our data would therefore be consistent with a model in which the adaptor protein(s) involved in this process favor(s) binding to tyrosine-phosphorylated rather than nonphosphorylated CD33. The low level of CD33 tyrosine phosphorylation in the absence of pervanadate may explain why CD33 endocytosis is normally very slow. Thus, the mechanism underlying endocytosis of ligated CD33 appears distinct from the mechanisms underlying CD22 endocytosis.
Recently, Tateno et al. [22 ] reported a major difference between the CD22 and the murine CD33-related Siglec, Siglec-F. Although CD22 was found to undergo clathrin-mediated endocytosis, sorting to early endosomes and recycling compartments, Siglec-F showed clathrin- and dynamin-independent but ADP ribosylation factor 6-dependent endocytosis and trafficking to lysosomes [22 ]. Although further studies are needed to investigate whether CD33 endocytosis is also clathrin-independent, it is noteworthy that CD33, similar to Siglec-F, is directed to lysosomes (data not shown).
We tested Shp1, Shp2, and other candidate ITIM-binding proteins, for their effects on CD33 endocytosis. In addition to Shp1 and Shp2, which have been recognized previously as binding partners of tyrosine-phosphorylated CD33 [7
8
9
], and the nonreceptor tyrosine kinase Syk that complexes with CD33 in some circumstances [48
, 49
], we found that two other SH2 domain proteins, CrkL and PLC-1, interacted with phosphorylated CD33 in vitro. Additional studies will be required to identify the full repertoire of proteins that bind to CD33 and similar immunoreceptors and address the functions of these interactions. All of these SH2 domain-containing proteins have been implicated to play a direct or indirect role in endocytosis, for example, by interaction with immunoreceptor tyrosine-based activation motif (ITAM)-containing FcRs [26
, 50
, 51
], by recruiting growth factor receptor-bound protein 2 and linkage to clathrin-mediated endocytosis pathways or Cbl family proteins [41
, 42
], or by acting as a guanine nucleotide exchange factor of dynamin-1 [52
]. However, CrkL and PLC-1 favored binding to the C-terminal, ITIM-like motif, which is relatively less important for endocytosis [14
], and Shp1, Shp2, and Syk bound preferentially to the proximal ITIM of CD33, which is more important for endocytosis [14
]. Using RNA interference, we found that Syk can be removed from myeloid cell lines without affecting CD33 endocytosis and that Shp1 and Shp2 can oppose, rather than promote, endocytosis.
Shp1 and/or Shp2 might potentially be involved in CD33 internalization by affecting the phosphorylation state of CD33 and/or by acting as adaptor proteins [41 , 42 ]. Indeed, in ML-1 and TF-1 cell lines, simultaneous depletion of Shp1 and Shp2 increased CD33 internalization, slightly but statistically significantly. This finding is consistent with a role of these phosphatases in the regulation of CD33 phosphorylation, and Shp1- and Shp2-catalyzed CD33 dephosphorylation resulted in reduced internalization. This finding is not consistent with a role of Shp1 and Shp2 as adaptors mediating endocytosis. It is interesting that depletion of Shp1 and Shp2 failed to affect CD33 internalization in HL-60 and NB4 cells, suggesting again important cell type-specific factors affecting the uptake of CD33. Several possible explanations might account for variation between cell lines, including differences in the baseline tyrosine phosphorylation state of CD33 or the extent of tyrosine phosphorylation induced by anti-CD33 antibody, the presence of additional, yet-unrecognized tyrosine phosphatases that keep CD33 in its dephosphorylated state despite the functional absence of Shp1 and Shp2, and differences in the ability to recruit and activate Shp1 and Shp2 despite CD33 phosphorylation. It is noteworthy that TF-1 and ML-1 cells have higher cell surface levels of CD33 compared with NB4 and HL-60 cells (data not shown), and ligation by antibody may thus provoke different levels of responses. However, we have only been able to detect antibody-induced tyrosine phosphorylation inconsistently in myeloid cell lines (data not shown), indicating that the level of phosphorylation induced by primary anti-CD33 antibody in the absence of cross-linking with a secondary antibody is relatively low and not well sustained and consistent with the rather slow rate of CD33 antibody internalization.
In summary, the present studies provide evidence that endocytosis of CD33 and, by extrapolation, other members of the CD33-related subset of Siglecs is restricted to a small degree by extracellular interactions and show the importance of tyrosine phosphorylation of the cytoplasmic domain for uptake of CD33. Thus, signaling and endocytosis appear intimately intertwined. Furthermore, the manipulation of the tyrosine phosphorylation status of CD33, for example, by activating pivotal tyrosine kinases or interfering with tyrosine phosphatases may be a novel, pharmacological means that could be exploited to enhance antibody internalization and possibly increase the effectiveness of Siglec-targeting therapies.
Received June 11, 2007; revised September 25, 2007; accepted September 26, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
receptor II (CD32) on malignant B cells influences modulation induced by anti-CD19 monoclonal antibody Blood 83,1632-1639 R J. Immunol. 169,6787-6794 receptor-mediated phagocytosis and the activation of RAC Blood 100,1852-18591 is a guanine nucleotide exchange factor for dynamin-1 and enhances dynamin-1-dependent epidermal growth factor receptor endocytosis J. Cell Sci. 117,3785-3795This article has been cited by other articles:
 |
|
 |
|
  R. B. Walter, K. M. Boyle, F. R. Appelbaum, I. D. Bernstein, and J. M. Pagel Simultaneously targeting CD45 significantly increases cytotoxicity of the anti-CD33 immunoconjugate, gemtuzumab ozogamicin, against acute myeloid leukemia (AML) cells and improves survival of mice bearing human AML xenografts Blood, May 1, 2008; 111(9): 4813 - 4816. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
