* Dendritic Cell and
Cancer Characterisation Laboratories, Mater Medical Research Institute, Mater Misericordiae Hospitals, South Brisbane, Australia;
Austin Research Institute, Heidelberg, Australia; and
Mayo Clinic, Scottsdale, Arizona
Correspondence: Dr. Michael McGuckin, Senior Research Fellow, Mater Medical Research Institute, Level 3, Aubigny Place, Mater Misericordiae Hospitals, South Brisbane, 4101, Brisbane, Australia. E-mail: mmcguckin{at}mmri.mater.org.au
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: T cell • activation • glycoprotein • sialomucin
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In addition to its broad expression on epithelial tissues, it is now evident that MUC1 is expressed by many hemopoietic cell types, including T cells [6 ], B cells [7 ], and bone marrow mononuclear cells [8 ]. Based on these descriptions, some of the findings presented in this manuscript, and trials of the BC2 and BC3 MUC1-reactive antibodies [9 ] by the Non-lineage Workshop Panel, MUC1 was assigned CD227 at the 7th Workshop on Human Leucocyte Differentiation Antigens [10 ]. The function of CD227 on hemopoietic cells remains to be elucidated.
There is an increasing interest in using MUC1 as a target antigen for cancer immunotherapy for carcinomas and multiple myeloma [11 12 13 ]. The data concerning expression and function of CD227 on hemopoietic cells are likely to impact on the design and outcome of immunotherapy directed against MUC1 in patients with cancer.
We report that CD227 is expressed on human- and murine-activated dendritic cells (DC). We also show for the first time that the CD227 cytoplasmic tail is tyrosine-phosphorylated in activated DC and T cells and that CD227 is associated with other phosphorylated proteins in these cells, indicative of the involvement of CD227 in signal transduction. We discuss these findings in relation to the role of MUC1 in immunobiology and its candidacy as a cancer vaccine antigen.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The BC2 (IgG1) and BC3 (IgM) mouse mAb reactive with a peptide epitope in the tandem repeat of human MUC1 [9
] and isotype-control antibodies were produced in vitro and purified using Prosep A (IgG, Bioprocessing Ltd., Consett, UK) and polyethylene glycol (PEG) precipitation, respectively. The MFP25 and MFP32 rat IgM mAb reactive with extracellular domain peptide epitopes on murine MUC1 [14
] were produced as ascites and purified by PEG precipitation followed by gel filtration on S400HR (Amersham Pharmacia Biotech). Purified BC2 and MFP25 were labeled with biotin, FITC, and HRP using standard methods [15
]. The CT2 hamster mAb reactive with the human MUC1 cytoplasmic tail was used as culture supernatant. Secondary antibodies were obtained from Jackson Immunoresearch Laboratories (West Grove, PA), and streptavidin-Texas red was from Vector Laboratories (Burlingame, CA). Other materials included granulocyte macrophage-colony stimulating factor (GM-CSF; Novartis Pharmaceutical, Basel, Switzerland), interleukin (IL)-3 (Gibco-BRL, Grand Island, NY), IL-4, soluble CD40L (Immunex, Seattle, WA), IL-2, and tumor necrosis factor 
(TNF-; Hoffman La Roche, Basel, Switzerland). Lipopolysaccharide (LPS), sodium vanadate, saponin, FITC, DNase, collagenase, keyhole limpet hemocyanin (KLH), and phytohemagglutinin (PHA) were from Sigma Chemical Co. (St. Louis, MO). Neuraminidase was from Behring Diagnostics (Marburg, Germany), tetanus toxoid was from CSL Ltd. (Melbourne, Australia), bovine serum albumin (BSA) was from ICN Laboratories (Aurora, OH), and biotin-normal human serum and HRP were from Roche Biosciences (Castle Hill, Australia). All cell culture medium was from Sigma Chemical Co. or Gibco-BRL.
DC, monocyte-derived DC (Mo-DC), and T cell preparations
Buffy coats from normal healthy donors were obtained from the Australian Red Cross Service (Brisbane, Australia). Peripheral blood mononuclear cells (PBMC) were isolated from buffy coats by standard density gradient centrifugation over Ficoll-Paque Plus (Amersham Pharmacia Biotech). For DC purification, the low-density, interface cells were washed in cold phosphate-buffered saline (PBS) and incubated in a cocktail of lineage mAb (anti-CD3, CD11b, CD14, CD16, CD19) for 20 min on ice, followed by a 15-min incubation with Biomag goat anti-mouse Ig-coated magnetic beads (Polysciences, Warrington, PA). Labeled cells were depleted by first preclearing with a MCP-1 magnet (Dynal, Oslo, Norway) and then passing through a magnetic cell sorter-CS column using a Variomacs magnet (Miltenyi Biotech, Gladbach, Germany). Negatively selected PBMC were labeled with FITC-goat anti-mouse (Becton Dickinson) and lineage-negative cells, further purified by sorting on a fluorescein-activated cell sorter (FACS) Vantage (Becton Dickinson). For activation, sort-purified, lineage-negative cells were incubated overnight at a concentration of 1 x 106 cells/ml in RPMI 1640 supplemented with 10% fetal calf serum (FCS) and 10 ng/mL IL-3 and 200 U/mL GM-CSF.
For T cell purification and the generation of Mo-DC, PBMC were fractionated by incubation with neuraminidase-treated sheep red blood cells followed by separation of the rosetting (ER+) and nonrosetting (ER-) populations on ficoll density gradients. For Mo-DC generation, the percentage of monocytes within the ER- fraction was determined by labeling with anti-CD14 followed by flow cytometric analysis on a FACS Calibur (Becton Dickinson). ER- cells were then cultured at 0.3 x 106 monocytes/ml in RPMI 1640 containing 10% FCS, 200 U/mL recombinant human (rh)GM-CSF, and 50 U/mL rhIL-4 for 7 days. Maturation of Mo-DC was achieved by the addition of 20 ng/mL rh-TNF-, 50 ng/mL rh-soluble trimeric CD40L, or 1 µg/mL LPS to the cultures on day 5 and harvesting after 48 h or as indicated in the results.
After lysis of the ER+ fraction with 1 M ammonium chloride, pure populations of responder T cells were prepared by magnetic immunodepletion with mAb to CD11b, CD14, CD16, CD19, and HLA-DR. By flow cytometry using an anti-CD3 mAb, 96–100% of the resulting cells expressed CD3.
Isolation, characterization, and culture of murine splenic DC
Murine splenic DC were isolated using a negative-selection technique as previously described [16
]. Briefly, spleens from 6- to 8-week-old C57/BL6 mice were digested with collagenase and DNase and then incubated in ethylenediaminetetraacetate in Hanks’ buffered salt solution without Ca2+ and Mg2+. The red blood cells were lysed, and cells were incubated in a cocktail containing at least 10 µg/mL purifed mAb or a concentrated tissue culture supernatant of B220, KT3, 3D6, FDC-M1, H3704, and Thy1.2 for 1 h at 4°C. Cells, which bound antibody, were depleted by rosetting with sheep red blood cells coated with sheep anti-rat Ig and anti-mouse IgG, IgA, and IgM and layering over a Histopaque gradient (Sigma Chemical Co.). Any remaining, contaminating cells were further depleted using anti-rat Ig-coated magnetic beads (Dynal, Skoyen, Norway). Isolated cells were characterized by flow cytometry using antibodies specific for major histocompatibility complex (MHC) class II, CD4, CD8, B220, CD3, CD11c, and CD11b and were assessed fresh or after culturing overnight in RPMI 1640/Iscove’s modified Dulbecco’s medium containing 10% heat-inactivated FCS with and without LPS at 20 µg/mL.
Staining of cell preparations for flow cytometric analysis
Cells were incubated with various mAb as per the manufacturer’s instructions or at 5–20 µg/mL in 1% BSA in PBS for 30–60 min at 4°C. Following all incubations, cells were washed 3x in 1% BSA in PBS. Unconjugated antibodies were detected with the appropriate conjugated secondary antibodies diluted in 1% BSA in PBS. After staining, all cells were fixed with 1% paraformaldehyde prior to analysis on a FACS Calibur cytometer (Becton Dickinson). For detection of intracellular antigens, following fixation cells were permeabilized with 0.5% saponin and all subsequent mAb incubations, and washes were performed in the presence of 0.5% saponin. Cell surface desialyation prior to flow cytometry was achieved by incubating cells in 50 mU/mL neuraminidase in PBS containing Ca2+ and Mg2+ for 30 min at 37°C.
Immunofluorescence and confocal microscopy
For dual-label immunofluorescence studies, spleens from naïve and immunized mice were frozen in OCT, and frozen sections were prepared, fixed in acetone, dried, and stored at -20°C prior to staining. Sections were blocked in 2.5% BSA in PBS and were single- or double-stained using direct or indirect staining techniques or a combination. Briefly, sections were incubated with primary antibodies at 1–20 µg/mL in 2.5% BSA in PBS for 60 min at room temperature. Sections were washed following all incubations for 3 x 5 min in PBS and were incubated with anti-rat Ig-FITC for nonlabeled, primary antibodies (1/100 in 2.5% BSA in PBS) or streptavidin-Texas red for biotinylated primary antibodies (1/200) for 40 min at room temperature. Rat serum was included in all steps following incubation with the anti-rat-FITC conjugate. Appropriate controls with irrelevant primary antibodies were included. The MFP25 MUC1-reactive antibody gave equivalent patterns using indirect detection with anti-rat fluorochrome conjugates or when biotin or FITC-labeled MFP25 was used. Stained sections were mounted with Vectashield (Vector Laboratories), examined on an Olympus BX60 fluorescent microscope (Olympus, Tokyo, Japan), and photographed by consecutive exposure using green and red filters. Isolated murine splenic DC were stained for confocal microscopy as previously described [17
, 18
] and examined on a Zeiss LSM510 scanning confocal microscope (Zeiss, Jena, Germany).
Induction of T cell responses
For assessing the contribution of MUC1 to allogeneic or recall T cell responses, Mo-DC (day 7; ±48 h, 1 µg/mL LPS) were incubated with freshly purified T cells in the presence or absence of 10 µg/ml anti-CD227 mAb BC2 or isotype control. Allogeneic mixed leukocyte reactions (MLRs) were established using various numbers of Mo-DC incubated in triplicate in round-bottom, 96-well tissue-culture plates with 1 x 105 freshly isolated, allogeneic T cells at 37°C for 5 days. For recall assays, MoDC were incubated with 105 autologous T cells in the presence of Tetanus toxoid (2 µg/mL). T cell proliferation was measured by the uptake of [3H]-thymidine (1 µCi/well; 6.7 Ci/mM; Amersham Pharmacia Biotech), which was added during the final 18 h of culture. Cells were harvested onto glass fiber filter paper with an automated 96-well harvester (Tomtec, Hamden, CT), and [3H]-thymidine incorporation was measured using a Wallac Microbeta Trilux liquid scintillation counter (Perkin Elmer Life Sciences, Turku, Finland).
Reverse transcriptase-polymerase chain reaction (RT-PCR)
RNA was isolated from human Mo-DC, human breast cancer cell lines, and murine spleens using Trizol reagent according to the manufacturer’s instructions (Life Technologies, Gaithersburg, MD). cDNA was prepared from 1 µg RNA using Superscript RNase H RT (Life Technologies). Human MUC1 transcripts were amplified using primers designed to cross two introns to differentiate any contaminating genomic DNA and give a 490-bp product from RNA (5'-TCTACTCTGGTGCACAACGG-3' and 5'-TTATATCGAGAGGCTGCTTCC-3') for 40 cycles (94°C 30 s, 57°C 30 s, 72°C 30 s) with 0.5 U Amplitaq Gold Taq DNA polymerase (Perkin Elmer, Norwalk, CT). Murine MUC1 transcripts were amplified using primers designed to cross two introns in genomic DNA and give a 318-bp product (5'-GCTCCGTGGTGGTAGAATCG-3' and 5'-TGGTAGGTGTCCTGGGTTGG-3') for 35 cycles (94°C 30 s, 59°C 30 s, 72°C 30 s) with 0.4 U Red Hot Taq DNA polymerase (Abgene, Epsom, UK). PCR products were subjected to electrophoresis in 1.2% agarose gels in Tris-buffered EDTA, stained with ethidium bromide, and photographed.
Western blotting
Human Mo-DC were lysed in 50 mM Tris, 150 mM NaCl, 0.5% Brij96, and 1 mM aprotinin, pH 7.6, and MUC1 was immunoprecipitated with the BC2 or an isotype-control antibody and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting as we previously described [19
]. For detection of phosphorylation, 1 mM sodium vanadate was included during the final 2 min of the cultures and in the lysis buffer. MUC1 was detected on Western blots using HRP-conjugated BC2, and tyrosine-phosphorylated proteins were detected using RC20-HRP as previously described [19
].
Culture of breast cancer cell lines
The MA11, MCF7, MDA-MB-231, MDA-MB-435, T47D, and ZR-75-1 breast cancer cell lines were sourced and cultured as previously described [20
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (29K): [in a new window] |
Figure 1. Expression of CD227 by C11c- and CD11c+ human blood DC following in vitro activation. DC were isolated from human blood by a negative-selection technique (see Materials and Methods). Lineage-negative cells (PB Lin-) were stained for cell surface CD227, CD11c, and HLA-DR before (fresh) and after activation by overnight culture. Histograms presented are derived from the HLA-DRhi CD11c- and CD11c+ cells from the gates shown in the scatterplot.
|
, or CD40-L, there was a substantial up-regulation of CD227 on the cell surface of Mo-DC. LPS induced higher expression of CD227 than TNF- or CD40L, although the level of induction varied in different Mo-DC preparations (from a five- to 50-fold increase in median fluorescence intensity in six differing preparations). This differed slightly compared with the DC activation marker, CMRF44 [21
], particularly with respect to the relative response to CD40-L, which strongly up-regulated CMRF44 but only weakly up-regulated CD227 (Fig. 2B)
. Induction of cell surface CD227 expression on LPS-activated Mo-DC followed a slightly later time course than induction of CD83 and CD86, although the expression of all three proteins continued to increase over the 48 h of the experiment (Fig. 2C)
.
 View larger version (31K): [in a new window] |
Figure 2. Expression of CD227 by human monocytes and activated Mo-DC. (A) CD14+ PBMC were stained for CD227 (dark line; isotype-control, shaded) before and after overnight culture with and without LPS. (B) Mo-DC were generated in vitro from monocytes by culture for 7 days with GM-CSF and IL-4. Mo-DC were activated during the final 48 h with LPS, TNF-, or CD40L, assessed for cell surface expression of CD227 by flow cytometry, and were compared with another activation marker, CMRF44. (C) Mo-DC were generated in vitro from monocytes by culture for 7 days with GM-CSF and IL-4. Mo-DC were activated with LPS during the final 1, 2, 6, 18, 24, 30, and 48 h of culture and assessed for cell surface expression of CD227, CD83, and CD86 by flow cytometry. Results are presented as a percentage of the maximum peak channel ratio for each marker (occurring with 48 h exposure in each case).
|
than by LPS. Western blotting also demonstrated the presence of two high molecular weight BC2-reactive bands consistent with polymorphic protein CD227 glycoprotein in mature Mo-DC (Fig. 3B)
.
 View larger version (69K): [in a new window] |
Figure 3. Confirmation of CD227 expression in activated Mo-DC by RT-PCR and Western blotting. (A) Mo-DC were generated in vitro from monocytes by culture for 7 days with GM-CSF and IL-4. Mo-DC were activated with LPS during the last 24 or 48 h or with TNF- during the last 48 h of culture. RNA was extracted, and RT-PCR was used to amplify CD227 transcripts. RNA from three breast cancer cell lines was similarly amplified. PCR products were subjected to electrophoresis in 1.2% agarose gels, stained with ethidium bromide, and photographed. (B) Nonactivated and activated (48 h with LPS) Mo-DC were lysed in a Brij97 detergent buffer, and proteins were subjected to electrophoresis in a 3–15% SDS-PAGE gel and transferred to a polyvinylidene difluoride (PVDF) membrane probed with the BC2 antibody. MUC1 purified from conditioned medium of the COLO316 ovarian cancer cell line was included as a strong positive control. The two high molecular weight bands characteristic of the polymorphic CD227 mucin glycoproteins in LPS-activated Mo-DC are arrowed. The position of the 220-kDa molecular weight (Mr) marker and the origin of the resolving gel (r) are shown at the left.
|
 View larger version (70K): [in a new window] |
Figure 4. Comparison of CD227 expression in breast cancer cell lines (A–F, M–R, Y, and Z), activated T cells (G, H, S, and T), and immature (I, K, U, W, and AA) and LPS-activated Mo-DC (J, L, V, X, and AB), assessed using the extracellular tandem repeat-reactive antibody BC2 with and without prior desialyation of cells with 100 U/mL neuraminidase for 60 min at 37°C (A–L and Y–AB) and with the CT2 antibody reactive with a glycosylation-independent epitope on the cytoplasmic tail (M–X). Mo-DC 460, 477, and 478 are Mo-DC preparations (see Materials and Methods) from three different individuals, and Act-T cell 474 and 644 are CD3+ T cells from PHA/IL-2-stimulated PBMC cultures (see Materials and Methods) from two different individuals. A–X represent cells stained in a single-staining run as do Y–AB. Shaded histograms represent isotype-control antibodies; for A–L and Y–AB, light gray lines represent neuraminidase-treated cells stained with isotype-control antibodies, black lines represent nontreated cells stained with BC2, and dark gray lines represent neuraminidase-treated cells stained with BC2.
|
 View larger version (27K): [in a new window] |
Figure 5. Expression of CD227 by murine splenic CD8+, CD4-CD8-, and CD4+ DC subsets following overnight in vitro culture. DC were isolated from naïve, murine spleens using a negative-selection technique (see Materials and Methods). Freshly isolated DC and DC cultured overnight with and without LPS were stained for CD4, CD8, and CD227 on the cell surface and following permeabilization. Stained DC were assessed by three-color flow cytometry; the scatterplot and histogram at the top show the gating for this analysis and the CD3 and MHC class II expression of this DC preparation, respectively. Reactivity with an isotype-control antibody is shaded.
|
 View larger version (124K): [in a new window] |
Figure 6. Subcellular localization of CD227 in isolated murine splenic DC and in situ localization of CD227-positive cells in murine splenic tissue and skin. (A and B) DC were isolated from naïve, murine spleens using a negative-selection technique, cultured overnight, and stained for MHC class II (FITC) and CD227 (Texas red) following permeabilization. Stained DC were assessed using confocal microscopy. (C–K) Murine splenic tissue from naïve mice (C–J) and mice 7 days following intravenous immunization with 200 µg KLH (K) stained for CD227 and other markers. (C and D) CD227 green fluorescence; (E) red pulp CD227 green, CD3 red; (F) CD227 green, CD11c red; (G) CD227 red, CD80 green; (H) red pulp CD227 green, MHC class II red; (I–K) CD227 green, sialoadhesin red; (L) murine skin from transverse section of ear, CD227 green, MHC class II red (CD227+ class II- cells are arrowed). Magnifications: A and B, 1000x; C, 60x; D and I, 120x; E–H and J–L, 240x.
|
MHC class II-positive Langerhan’s cells of the skin did not typically express CD227, although unidentified CD227+ MHC class II- cells were occasionally observed in the skin (Fig. 6L) .
CD227 expression by human T, B, and natural killer (NK) cells
We also confirmed that CD227 is expressed by human peripheral blood-derived T cells activated by PHA and IL-2 (not shown), as previously described [6
]. We have also induced CD227 expression in human tonsil-derived T cells activated by PHA and IL-2 or by MLRs (not shown). Flow cytometric comparison with six breast cancer cell lines using the BC2 and CT2 antibodies demonstrated CD227 levels in activated T cells similar to those seen in many breast cancer cell lines (Fig. 4G
4H
and 4S
4T)
. Only a small proportion of peripheral blood CD19+ B cells express low levels of CD227, and we could not demonstrate CD227 on the surface of circulating CD16+CD56+ NK cells, even following in vitro activation (not shown).
The CD227 cytoplasmic tail is phosphorylated in activated human T cells and DC
The CD227 cytoplasmic tail has been shown to be involved in signal transduction in epithelial cells. Using immunoprecipitation and Western blotting, we have been able to demonstrate tyrosine phosphorylation of the CD227 cytoplasmic tail in cultures of activated human T cells and activated human Mo-DC (Fig. 7A
). In cells incubated in the presence of the tyrosine phosphatase inhibitor sodium vanadate for 2 min prior to lysis, a strong RC20-reactive band at 23–25 kDa was seen in T cells and Mo-DC. This is the expected molecular weight of the CD227 cytoplasmic tail-containing subunit, which separates from the extracellular subunit in the presence of SDS [4
]. In an attempt to mimic potential roles of CD227 signaling in DC-T cell interactions, we set up MLRs involving in vitro, LPS-matured, CD227-expressing, human Mo-DC and naïve allogeneic human PBMC-derived T cells. The level of CD227 cytoplasmic tail tyrosine phosphorylation did not alter substantially in the first 60 min of these cultures (Fig. 7A)
. In addition to detection of the 25-kDa CD227 cytoplasmic tail band, additional tyrosine-phosphorylated proteins were coprecipitated with CD227 in T cells and Mo-DC. A prominent 110-kDa phosphoprotein and a weaker 42-kDa phosphoprotein coprecipitated with CD227 in the Jurkat T cell leukemia cell line (Fig. 7B)
, whereas in Mo-DC, the 42-kDa phosphoprotein coprecipitated with CD227, but the 110-kDa band was absent (Fig. 7C)
.
 View larger version (64K): [in a new window] |
Figure 7. CD227 is tyrosine-phosphorylated in activated T cells and DC. (A) Mo-DC were generated in vitro from monocytes by culture for 7 days with GM-CSF and IL-4 and were activated with 1 µg/mL LPS during the last 48 h of culture. Allogeneic T cells were derived from two donor PBMCs by resetting and lineage-depletion (see Materials and Methods). T cells (TC) and activated Mo-DC (DC) were cultured alone or together at a 5:1 T cell:DC ratio for 0, 5, 30, and 60 min. Cells were lysed in Brij97 buffer and immunoprecipitated with BC2 for CD227 (M) or a control antibody (C). Immunoprecipitates were subjected to electrophoresis in 3–20% SDS-PAGE gels, transferred to PVDF, and probed with the RC20-HRP conjugate to detect tyrosine-phosphorylated proteins. Each lane represents 4 x 106 T cells (where present) and 8 x 105 DC (where present). CD227 tyrosine phosphorylation was detected in T cells and DC, and there were no dramatic changes in phosphorylation during the first 60 min of allogeneic MLR. (B) Jurkat T cells were lysed in Brij97 buffer and immunoprecipitated with BC2 for CD227 (M) or a control antibody (C). Immunoprecipitates were subjected to electrophoresis (each lane represents 1x106 cells) in 3–20% SDS-PAGE gels, transferred to PVDF, and probed with the RC20-HRP conjugate to detect tyrosine-phosphorylated proteins. In addition to the 23-kDa band representing the CD227 cytoplasmic tail containing a subunit, a major coprecipitating tyrosine-phosphorylated protein was seen at 110 kDa and a minor band at 42 kDa. The kDa of molecular weight markers and the origin of the resolving gel (r) are shown at the left. (C) Mo-DC were activated in the presence of LPS for 48 h, lysed in Brij97 buffer, and immunoprecipitated (each lane represents 1x106 cells) and probed with RC20-HRP as in B. In addition to the 23-kDa band representing the CD227 cytoplasmic tail-containing subunit, a coprecipitating, 42-kDa phosphoprotein was seen.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Until recently, MUC1 (CD227) was considered to be an epithelial mucin. It is now clear that this molecule is expressed by a wide variety of hemopoietic cell types from early differentiating bone marrow mononuclear cells [8 ] through to mature cell types. Additionally, it is now apparent that other "epithelial" cell surface mucins, such as MUC11 (M. McGuckin, unpublished observations) and MUC13 [2 ], are also expressed by hemopoietic cells. The wide expression of this family of molecules in a variety of cell types, often in an activation-dependent manner, suggests they play an important, functional role in hemopoietic cells, and these glycoproteins can no longer be regarded as restricted to epithelial tissues.
Cell surface expression of CD227 was activation-dependent in murine splenic DC and in human blood DC and Mo-DC, as it was in human T cells. Our findings that only a small proportion of peripheral blood T cells express CD227 using the BC2 antibody agree with the initial report of CD227 in T cells using B27.29 [6 ]. However, using the DF3 antibody, another group has recently claimed that CD227 is expressed on all peripheral T cells and that DF3 coated on plates blocks CD3-stimulated T cell activation [25 ]. BC2 coated on plates does not block T cell activation in our hands (data not shown). These discrepancies remain to be resolved. One possibility is that the differential antibody reactivity is a result of differences in CD227 glycosylation following T cell activation. However, the post-activation increase in CD227 expression that we have observed using BC2 is mirrored by CT2 reactivity (glycosylation-independent antibody) and induction of mRNA expression (not shown).
CD227 can be shed from the cell surface of epithelial cells, and directly secreted isoforms of CD227 generated by alternative splicing have also been described [26 ]. Consequently, cytoplasmic CD227 may represent stored transmembrane molecules packaged for transport to the cell surface or the secreted isoform stored prior to direct secretion from the cell. Activated T cells have previously been shown to secrete or shed CD227 [6 ]. However, detection with the CT2 antibody, which will not react with the secreted isoform, suggests that an intracellular pool of the transmembrane form of CD227 exists in immature DC. Activation then results in transmembrane CD227 being transported to the cell surface. This does not exclude coproduction of the secreted isoform following activation. The activation-dependent expression of CD227 by T cells and DC contrasts with the CD43 sialomucin, which is highly expressed by both cell types regardless of activation state [27 , 28 ].
The function of CD227 on the surface of activated DC and T cells remains to be elucidated. In fact, the precise role of the entire family of cell surface mucins remains obscure, even in glandular epithelial cells on which they are very highly expressed. The large, extracellular tandem repeat mucin domains of these molecules are predicted to form rigid, elongated structures, which together with their highly glycosylated nature, suggest that these molecules form a dominating presence above other molecules present in the glycocalyx [11 ]. Mucin carbohydrates, although potentially providing specific ligands for lectins, are generally thought to be repulsive in nature. Therefore, it has been suggested that like secreted mucins, the cell surface mucins play a protective role when highly expressed on the apical surface of glandular epithelial cells. However, all members of this family identified to date contain highly conserved, cytoplasmic domains likely to be involved in signal transduction [2 , 4 , 29 30 31 ]. Additionally, several cell surface mucins, including CD227, can be endocytosed into clathrin-mediated pits [32 ]. Their function is likely to be more complex than providing a physical barrier on the cell surface.
Our studies showing tyrosine phosphorylation of CD227 in T cells and Mo-DC, together with coprecipitation of other tyrosine-phosphorylated proteins, suggest a signaling role for CD227 in these cells. There were no substantial changes in tyrosine phosphorylation during the early phases of MLRs involving activated Mo-DC and allogeneic T cells. However, it appears that CD227 may also signal through serine phosphorylation [33 ], which was not assessed in our current study. Like CD227, CD43 has a somewhat obscure function on T cells and DC. It is generally accepted that CD43 is antiadhesive but may also have specific roles in cell adhesion and motility; for a review, see ref. [28 ]. Antibody cross-linking of CD43 in neutrophils causes capping and triggers polarization to uropodia and cell locomotion [34 ]. Initial studies in CD43 null mice suggested that CD43-/- T cells were hyper-responsive to activation stimuli [35 ], although a recent report questions these observations [36 ]. However, it is evident that the influence of CD43 on T cell activation and adhesion requires its cytoplasmic domain [37 ], which like CD227, appears to be involved in signal transduction [38 , 39 ].
Cell surface CD227 may be an important modulator of intracellular adhesion and/or motility in activated DC and T cells. High cell surface expression of CD227 has been shown to interfere with cell-cell and cell-extracellular matrix adhesion [40 , 41 ]. The presence of CD227 in activated DC and T cells may indicate a role for CD227 during migration of these cells, possibly acting to restrict unwanted, adhesive events. It is interesting that the CD227 cytoplasmic domain has been shown to interact with the actin cytoskeleton [42 ] and to colocalize with ezrin in filopodial protrusions in epithelial cells [43 ]. These interactions may provide a very important link between the external glycocalyx and the cytoskeleton during cell migration. In another similarity with CD227, the CD43 cytoplasmic domain has been shown to interact with the cytoskeleton via ezrin and moesin [44 ]. Although CD227 has similarities to CD43, their expression is differentially regulated, CD43 and CD227 knockout mice have vastly differing immunological phenotypes (S. J. Gendler, unpublished observations), and it is likely that CD227 plays a distinct role in activated hemopoietic cells.
Appropriate localization of CD227 may be important in T cells and DC during cellular interactions and perhaps during formation of supramolecular activation clusters. Our inability to interfere with antigen-specific or alloreactive, DC-stimulated T cell proliferation using CD227 antibodies does not rule out an important role for CD227 in DC-T cell interactions, as these antibodies may be incapable of interfering with CD227 function. CD227 may act as a specific ligand for cell surface molecules on other cell types. Our desialyation studies in Mo-DC demonstrate that CD227 is sialyated in DC. Our demonstration of contact between CD227+ cells in the outer MZ and sialoadhesin-expressing MZ metallophillic macrophages (MZMM) places CD227 and a proposed ligand, the sialic acid-binding lectin, sialoadhesin [24 ], together in vivo. This observation suggests a hereto-unrecognized interaction between migrating DC and MZMM, which warrants further investigation.
There is increasing interest in using MUC1 (CD227) as the antigen in cancer vaccines. Initially, this interest was based on the high expression of MUC1 by adenocarcinomas such as breast cancer and by multiple myelomas and on cellular immune responses against the particular glycoform of MUC1 produced by these cancers [12 ]. However, more recently, the approach of many groups has broadened to attempts to generate classical MHC-restricted T cell responses against VNTR and non-VNTR MUC1 epitopes. Several early clinical trials of MUC1 vaccines have been completed or are currently in progress [45 46 47 48 49 50 ]. It has been more difficult to induce strong responses against human MUC1 in mice expressing human MUC1 as a transgene than in wild-type mice. However, using DC immunization in MUC1 transgenic mice, immune responses against MUC1 have been induced that are capable of tumor rejection without inducing autoimmunity [51 ]. The expression of MUC1 on leukocytes as well as epithelial cells may influence its use as a tumor-rejection antigen; however, in the end, this will be determined by clinical trials of MUC1 vaccines using approaches capable of inducing strong MUC1-specific immune responses, such as DC therapy.
Future research needs to concentrate on the functional role of CD227 in hemopoietic cells. CD227 knockout mice are available to aid these studies; however, interruptions to hemopoietic cell development in these mice (S. J. Gendler, unpublished observations) may complicate their use in elucidating CD227 function in mature hemopoietic cells. We recommend that clinical trials using MUC1 as a cancer vaccine antigen should carefully monitor for epithelial and hemopoietic autoimmunity in patients with tumor responses.
Received August 22, 2001; revised May 5, 2002; accepted June 11, 2002.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Agrawal and B. M. Longenecker MUC1 mucin-mediated regulation of human T cells Int. Immunol., April 1, 2005; 17(4): 391 - 399. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Levitin, A. Baruch, M. Weiss, K. Stiegman, M.-l. Hartmann, M. Yoeli-Lerner, R. Ziv, S. Zrihan-Licht, S. Shina, A. Gat, et al. A Novel Protein Derived from the MUC1 Gene by Alternative Splicing and Frameshifting J. Biol. Chem., March 18, 2005; 280(11): 10655 - 10663. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Vasir, D. Avigan, Z. Wu, K. Crawford, S. Turnquist, J. Ren, and D. Kufe Dendritic Cells Induce MUC1 Expression and Polarization on Human T Cells by an IL-7-Dependent Mechanism J. Immunol., February 15, 2005; 174(4): 2376 - 2386. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Y. Tsang, C. Palena, J. Yokokawa, P. M. Arlen, J. L. Gulley, G. P. Mazzara, L. Gritz, A. Gomez Yafal, S. Ogueta, P. Greenhalgh, et al. Analyses of Recombinant Vaccinia and Fowlpox Vaccine Vectors Expressing Transgenes for Two Human Tumor Antigens and Three Human Costimulatory Molecules Clin. Cancer Res., February 15, 2005; 11(4): 1597 - 1607. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Mukherjee, T. L. Tinder, G. D. Basu, and S. J. Gendler MUC1 (CD227) interacts with lck tyrosine kinase in Jurkat lymphoma cells and normal T cells J. Leukoc. Biol., January 1, 2005; 77(1): 90 - 99. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Herbert, J. F. Grosso, M. Dorsey Jr., T. Fu, I. Keydar, M. A. Cejas, D. H. Wreschner, N. Smorodinski, and D. M. Lopez A Unique Mucin Immunoenhancing Peptide with Antitumor Properties Cancer Res., November 1, 2004; 64(21): 8077 - 8084. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Cloosen, M. Thio, A. Vanclee, E. B. M. van Leeuwen, B. L. M. G. Senden-Gijsbers, E. B. H. Oving, W. T. V. Germeraad, and G. M. J. Bos Mucin-1 is expressed on dendritic cells, both in vitro and in vivo Int. Immunol., November 1, 2004; 16(11): 1561 - 1571. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. P. Lillehoj, H. Kim, E. Y. Chun, and K. C. Kim Pseudomonas aeruginosa stimulates phosphorylation of the airway epithelial membrane glycoprotein Muc1 and activates MAP kinase Am J Physiol Lung Cell Mol Physiol, October 1, 2004; 287(4): L809 - L815. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. F. Grosso, L. M. Herbert, J. L. Owen, and D. M. Lopez MUC1/sec-Expressing Tumors Are Rejected In Vivo by a T Cell-Dependent Mechanism and Secrete High Levels of CCL2 J. Immunol., August 1, 2004; 173(3): 1721 - 1730. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
