Carcinoembryonic Ag-related cell adhesion molecule 1 (CEACAM1), the primordial carcinoembryonic Ag gene family member, is a transmembrane cell adhesion molecule expressed in leukocytes, epithelia, and blood vessel endothelia in humans and rodents. As a result of differential splicing, CEACAM1 occurs as several isoforms, the two major ones being CEACAM1-L and CEACAM1-S, that have long (L) or short (S) cytoplasmic domains, respectively. The L:S expression ratios vary in different cells and tissues. In addition to CEACAM1, human but not rodent cells express GPI-linked CEACAM members (CEACAM5–CEACAM8). We compared the expression patterns of CEACAM1-L, CEACAM1-S, CEACAM6, and CEACAM8 in purified populations of neutrophilic granulocytes, B lymphocytes, and T lymphocytes from rats, mice, and humans. Human granulocytes expressed CEACAM1, CEACAM6, and CEACAM8, whereas human B lymphocytes and T lymphocytes expressed only CEACAM1 and CEACAM6. Whereas granulocytes, B cells, and T cells from mice and rats expressed both CEACAM1-L and CEACAM1-S in ratios of 2.2–2.9:1, CEACAM1-S expression was totally lacking in human granulocytes, B cells, and T cells. Human leukocytes only expressed the L isoforms of CEACAM1. This suggests that the GPI-linked CEACAM members have functionally replaced CEACAM1-S in human leukocytes. Support for the replacement hypothesis was obtained from experiments in which the extracellular signal-regulated kinases (Erk)1/2 were activated by anti-CEACAM Abs. Thus, Abs against CEACAM1 activated Erk1/2 in rat granulocytes, but not in human granulocytes. Erk1/2 in human granulocytes could, however, be activated by Abs against CEACAM8. We demonstrated that CEACAM1 and CEACAM8 are physically associated in human granulocytes. The CEACAM1/CEACAM8 complex in human cells might accordingly play a similar role as CEACAM1-L/CEACAM1-S dimers known to occur in rat cells.
The functional activities of leukocytes are governed by a diversity of cell surface receptors, that in turn are regulated by costimulatory and coinhibitory receptors (1, 2). A large number of both costimulatory and coinhibitory receptors exist, many of which belong to the Ig superfamily (IgSF).3 One IgSF member that is abundantly expressed in leukocytes, as well as in epithelial and endothelial cells, is the carcinoembryonic Ag (CEA)-related cell adhesion molecule CEACAM1 (3, 4). CEACAM1 is a transmembrane signal-regulating glycoprotein that is the primordial and most well conserved member of the CEA family. In rats, CEACAM1 is the only cell surface-associated member of the CEA family, whereas mice in addition have a related gene that codes for CEACAM2, which in comparison with CEACAM1 makes up a minor component in a limited number of tissues (5, 6). CEACAM2 is not expressed in leukocytes, and the CEACAM2 gene does not exist in the human genome (5, 6). In addition to CEACAM1, the human genome contains six CEA-related genes coding for transmembrane (CEACAM3, CEACAM4) or GPI-linked (CEACAM5, CEACAM6, CEACAM7, CEACAM8) cell surface proteins (3). No GPI-linked proteins belonging to the CEA family occur in rodents (3, 4), a fact that has contributed to the difficulties in revealing the functions of the GPI-linked CEA family members.
CEACAM1, also known as CD66a, is an important multipotent signaling molecule (4) that regulates a variety of cellular activities such as cell proliferation (7), tumor growth (8, 9), apoptosis (10), angiogenesis (11), T cell cytotoxicity (12), dendritic cell function (13), granulocyte activation (14), and epithelial cell polarization (10). It has been demonstrated that CEACAM1 can act as a homophilic cell-cell adhesion molecule, binding to itself (4). The homophilic interaction represents the only species-independent physiological binding activity of the extracellular domain that has been unambiguously demonstrated thus far. However, CEACAM1 also functions as a microbial receptor. Thus, in human tissues CEACAM1 together with CEACAM3, CEA (CEACAM5), and CEACAM6 serves as the receptor for Opa protein-expressing gonococci and meningococci (15), and murine CEACAM1 is the receptor for mouse hepatitis virus (16). None of these ligands, or any monoclonal or polyclonal Abs analyzed thus far, cross-reacts with CEACAM1 from other species. In both rodents and humans, CEACAM1 is subject to differential splicing. The two major splice isoforms have four extracellular Ig domains but differ in their cytoplasmic domain and are denoted CEACAM1-L and CEACAM1-S, respectively, in which L denotes long and S denotes short (3, 4). The L cytoplasmic domain, which consists of 71- to 73-aa residues, have 2 phosphorylatable tyrosine residues that can recruit and activate SH2 domain-containing tyrosine kinases (17) and tyrosine phosphatases (18). The interactions with the tyrosine kinases and tyrosine phosphatases are believed to be important for the signaling activities of CEACAM1. The S cytoplasmic domain is 10–12 aa long and lacks phosphorylatable tyrosine residues. Soluble isoforms of CEACAM1 have also been found. Thus, PC12 cells express and secrete a fully glycosylated 4-Ig domain CEACAM1 lacking the transmembrane portion, due to an outsplicing of exons 6 and 7 (19).
In many CEACAM1-expressing cell types, which have been analyzed in depth, it has been found that CEACAM1-L and CEACAM1-S are coexpressed, although at varying ratios (7, 20). It has also been demonstrated that both the L and the S isoform can form dimers (21), which is believed to play an important role in the signal-regulating activities of CEACAM1 (4, 19). However, it is not known whether all CEACAM1-expressing cells express both the L and the S isoforms simultaneously, and under all functional states. With this in mind and because of the lack of expression of GPI-linked CEA family proteins in rats and mice, we initiated a comparative investigation of the expression of the L and S isoforms of CEACAM1 in leukocytes isolated from peripheral blood of rats, mice, and humans. Here we present data that show significant and striking differences in the CEACAM1 expression patterns in rodent and human leukocytes. Whereas rodent leukocytes express both the L and the S isoforms of CEACAM1, human leukocytes express only CEACAM1–4L and CEACAM1–3L, i.e., the long cytoplasmic domain isoforms having four or three extracellular Ig domains, respectively. The human leukocytes also express GPI-linked CEACAM8 and/or CEACAM6, which suggests that the GPI-linked CEA-related molecules may have functionally replaced the short cytoplasmic domains, CEACAM1-S, in these cells. Furthermore, we demonstrate that CEACAM1 and CEACAM8 are physically associated in human granulocytes and that CEACAM1 and CEACAM8 regulate the activation of extracellular signal-regulated kinases (Erk) in rodent and human granulocytes, respectively.
Materials and Methods
Abs and reagents
A hybridoma secreting the rat anti-mouse CEACAM1 mAb AgB10 was obtained from Drs. N. I. Kuprina and T. D. Rudinskaya (Cancer Research Center, Moscow, Russia), and the mouse anti-mouse CEACAM1 mAb CC1 was kindly provided by Dr. K. Holmes (University of Colorado, Denver, CO). The hybridoma secreting the mouse anti-rat CEACAM1 mAb 5.4 was generously provided by Dr. D. Hixson (Rhode Island Hospital, Brown University, Providence, RI). The mouse mAb 5F4 against human CEACAM1 was a gift from Dr. R. S. Blumberg (Brigham and Women’s Hospital, Harvard Medical School, Boston, MA). The following mouse mAbs (Sixth Leukocyte Typing Workshop, Osaka, Japan) were obtained from the laboratory of Dr. F. Grunert (University of Freiburg, Freiburg, Germany): 4/3/17 (specific for CEACAM1/CEA); 12/140/4 (specific for CEACAM1/CEA); 9A6 (specific for CEACAM6); 47 (specific for CEACAM8); BEAR1 (specific for CD11b); and MEM48 (specific for CD18). The CEACAM8-specific mAb 80H3 was from Coulter International (Miami, FL). Human leukocytes do not express CEA (CEACAM5); therefore the Abs 4/3/17 and 12/140/4 will specifically detect CEACAM1 when leukocytes are analyzed. Abs were purified from hybridoma supernatants by affinity chromatography on fast flow protein G-Sepharose (Amersham Pharmacia Biotech, Uppsala, Sweden). A fraction of the purified Abs was coupled to biotin using a biotinylation kit according to the manufacturer’s protocol (Sigma-Aldrich, St. Louis, MO). FITC-labeled mouse anti-human CD3, rat anti-mouse CD3, hamster anti-rat CD3, mouse anti-human CD19, rat anti-mouse CD45R/B220, and hamster anti-rat CD45R/B220 Abs were obtained from BD PharMingen (San Diego, CA). Streptavidin-PE was obtained from DAKO (Copenhagen, Denmark). Abs against phosphorylated, activated Erk1/2 were obtained from New England Biolabs (NEB). [α-32P]dCTP was from Amersham. As cDNA templates for control PCR, we used pBlueScript plasmids (Stratagene, La Jolla, CA) containing the complete coding sequences for rat CEACAM1-L (22), rat CEACAM1-S (22), mouse CEACAM1-L (23), mouse CEACAM1-S (23), and pcDNA/Neo plasmids (Invitrogen, San Diego, CA) containing the complete coding sequences for human CEACAM1–4L (24), human CEACAM1–3L (24), human CEACAM1–4S (24), and human CEACAM1–3S (24). The mouse and human CEACAM1 plasmids were kindly provided by Dr. N. Beauchemin (McGill University, Montreal, Canada) and by Dr. T. Barnett (Bayer Pharmaceutical Division, Berkeley, CA), respectively.
Isolation of leukocytes from peripheral blood
Leukocytes were prepared from heparinized (5 U heparin/ml) peripheral blood from healthy donors, from BALB/c mice, and from Lewis rats (obtained from Charles River, Uppsala, Sweden) by sedimentation through Plasmasteril (Fresenius, Bad Homburg, Germany). PBMC and neutrophilic granulocytes (polymorphonuclear neutrophils (PMN)) were separated by centrifugation in Ficoll-Paque (Amersham Pharmacia Biotech). The erythrocytes in the pellet were lysed by repeated suspension in cold 0.2% NaCl for 20 s followed by washing with PBS. More than 95% of the remaining cells were PMN as judged from morphological criteria, with a viability of >97% as measured by trypan blue exclusion. The PBMC recovered from the top of the Ficoll-Paque cushion were separated into lymphocyte subpopulations by single-cell sorting in a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA). The cells were labeled with anti-CD3-FITC for T cell sorting and either anti-CD19-FITC (human) or anti-CD45R/B220-FITC (rat and mouse) for B cell sorting. The purity of the isolated lymphocyte T and B subpopulations were 87–96% as determined with the FACSCalibur instrument using CellQuest software (BD Biosciences).
Detection of CEACAM1 by flow cytometry
To analyze the surface expression of CEACAM1 in the different leukocyte subpopulations, freshly isolated cells were stained with biotinylated anti-CEACAM1 Abs (50 μg/ml) in PBS containing 3% FCS for 1 h on ice, followed by washing with ice-cold PBS and incubation with streptavidin-PE at a dilution of 1/40. Background fluorescence was determined using isotype-matched immunoglobulins instead of specific primary Abs. Purified granulocytes were measured directly, whereas a two-color flow cytometric staining of the lymphocytes in the unfractionated PBMCs was performed. After the CEACAM1 staining, the PBMC were incubated with anti-CD3-FITC or anti-CD19-FITC or anti-CD45R/B220-FITC according to the recommendations of the manufacturer. The samples were measured in a FACSCalibur instrument, and the data were analyzed using CellQuest software. Dead cells were identified by propidium iodide staining (Sigma-Aldrich) and were excluded from the determinations.
RT-PCR for different CEACAM molecules
Total RNA was isolated from granulocytes, sorted CD3+ T cells, and human CD19+ and rodent CD45R/B220+ B cells by guanidinium thiocyanate extraction, using the Qiagen RNAeasy minikit (Qiagen, Valencia, CA). Samples of RNA in a final volume of 20 μl were reverse transcribed by Moloney murine leukemia virus reverse transcriptase (MBI Fermentas, Hanover, MD) according to the manufacturer’s recommendation, using the following oligonucleotide primers that specifically hybridize to the 3′ regions of the various CEACAMs: rat CEACAM1, 5′-GGCATTGAAGTTCAG-3′; mouse CEACAM1, 5′-ACAGTGTATGCGACG-3′; human CEACAM1, 5′-GTTGTTTCTGTCCC-3′; CEACAM6, 5′-CCAGTGGCTGAGTT-3′; CEACAM8, CCAGTGGCTGAGTT-3′. The PCR were performed in a total volume of 30 μl containing 5 μl of first-strand cDNA solution (or plasmid cDNA for human, mouse, or rat CEACAM1-L or CEACAM1-S), 0.2 mM dNTPs, 3 U of Taq DNA polymerase (Amersham Pharmacia Biotech), 3 μl of 10 × PCR buffer, and 0.6 μM concentrations of each of the PCR primers. The reactions were initiated by heating the samples to 94°C for 60 s, followed by 30 cycles at 94°C for 45 s, 64°C for 45 s, and 72°C for 60 s and an extension at 72°C for 10 min. The products were analyzed on 2.7% agarose gels in Tris-borate-EDTA buffer and visualized by ethidium bromide staining. The relative amounts of the PCR products were analyzed by scanning the gels and determining the intensities in the ethidium bromide-stained bands with Scan analyses software (Biosoft, Milltown, NJ). The PCRs were performed with the primer combinations shown in Table I. For CEACAM6 and CEACAM8 amplifications, a single primer set was used for each CEACAM, respectively. For CEACAM1 amplifications, a common sense primer that recognized both CEACAM1 splice variants equally well and two antisense primers that were specific for the two spliced isoforms were used (Table I and Fig. 4 A). The antisense primers for the L isoforms recognized the alternatively spliced exon 7 present only in CEACAM1-L; the antisense primer for the S isoform was constructed to anneal across the splice junction between exon 6 and exon 8. CEACAM1 PCRs were performed both in a conventional way with one sense and one antisense primer and as a triple-primer PCR in which one sense primer was used together with both antisense primers selectively recognizing the L and the S isoforms, respectively. To test the ability of the triple-primer PCR to return correct values of the input ratios, some PCR were run with cDNA plasmids for rat CEACAM1-L and CEACAM1-S as templates, and with the addition of [α-32P]dCTP in the deoxynucleotide mixture. The PCR products were quantified by cutting out the ethidium bromide-stained bands and counting in a β-scintillation counter or by analysis of the gels in a PhosphoImager (Molecular Dynamics, Sunnyvale, CA).
|Transcript .||Primer .||Sequence (5′–3′) .||Position from ATG .||Fragment Length (bp) .||Ref. .|
|Transcript .||Primer .||Sequence (5′–3′) .||Position from ATG .||Fragment Length (bp) .||Ref. .|
Immunoblot analysis for CEACAM and activated MAP kinase
Human granulocytes and IL-2-stimulated lymphocytes were solubilized, immunoprecipitated with mAb 4/3/17, and analyzed for CEACAM1 expression by Western blotting as described in Ref. 25 . For immunoblot analysis of CEACAM1 in rat and mouse cells, 2 × 106 freshly isolated granulocytes or lymphocytes were solubilized in 100 μl of lysis buffer containing 1% Triton X-100, 0.1% SDS, 50 mM Tris-HCl (pH 7.5), 5 mM sodium pyrophosphate, 1 mM EDTA, 1 mM EGTA, 50 mM sodium fluoride, 1 mM PMSF, 10 μg/ml leupeptin, 10 μg/ml chymostatin, 10 μg/ml pepstatin A, and 1000 kIU/ml Trasylol (aprotinin). After centrifugation for 30 min at 15 000 × g, the supernatants were incubated overnight at 4°C with 10 μg/ml CEACAM1-specific mAbs (mAb 5.4, mAb CC1). BSA-saturated protein G-Sepharose was then used to collect the immune complexes. After thorough washing, the protein G-Sepharose beads were boiled for 5 min in 2× SDS sample buffer (250 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 0.01% bromphenol blue, 50 mM dl-DTT). The samples were electrophoresed on 8% SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membrane (Millipore, Bedford, MA). After blocking of nonspecific binding with 1% skim milk powder in TBS, the membranes were incubated with primary anti-CEACAM1 Abs (10 μg/ml) and washed twice with TBS containing 0.1% Tween 20 (Merck, Rahway, NJ). HRP-coupled secondary Abs (DAKO) were added, and the filters were developed by ECL and documented using the Fuji gel documentation system.
For determination of activated Erk1/2 kinases, 107 rat or human granulocytes in 100 μl were incubated with mAbs (50 μg/ml in PBS) against CEACAM1 (mAb 5.4 for rat CEACAM1 and mAb 4/3/17 or mAb 5F4 for human CEACAM1), CEACAM6 (mAb 9A6), and CEACAM8 (mAb 80H3), respectively, for 5 min at room temperature. Isotype-matched Igs and PBS were used as controls. After centrifugation, the cell pellets were lysed in 50 μl of ice-cold 2× SDS sample buffer (see above), sonicated for 10 s, and boiled for 5 min. Twenty microliters of each sample were electrophoresed on 10% SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membranes (Millipore). After blocking with 5% skim milk powder in TBS overnight at 4°C, the membranes were incubated with Abs specific for the phosphorylated form of activated Erk1/2 according to the manufacturer’s protocol (NEB). After two washings with TBS containing 0.05% Tween 20, the membranes were incubated with HRP-coupled goat anti-rabbit Abs (NEB), developed by ECL, and documented by the Fuji gel documentation system. Equal loading of samples in the electrophoresis step was checked by Amido Black staining of the membranes.
Microtiter plates (Nunc, Wiesbaden, Germany) were coated overnight at 4°C with 100 μl of solutions of various mAbs (10 μg/ml in PBS) of defined specificities. After being washed and blocked with 300 μl of 2% BSA in PBS, lysates of human granulocytes or HeLa-Neo cells (HeLa cells transfected with the neomycin resistance gene, kindly provided by Dr. F. Grunert, Department of Immunology, Freiburg University, Freiburg, Germany) were added, and the plates were incubated for 4 h at 4°C and washed thoroughly. The plates were then incubated with a second set of peroxidase- or biotin-conjugated Abs of defined specificities and were washed again. For detection of biotinylated Abs, the wells were then incubated with streptavidin-peroxidase (Pierce, Rockford, IL) and washed. Finally, peroxidase activity was analyzed using tetramethylbenzidine (Fluka, Buchs, Switzerland) as substrate. The reaction was stopped with 1 M H2SO4, and the OD was measured in an ELISA reader (SLT-Spectra, Salzburg, Austria) at 450 nm.
Leukocyte expression of CEACAM1
Pure populations of PMN and PBMC were isolated from healthy human donors, rats, and mice. Analysis by flow cytometry showed that granulocytes, CD3+ T cells, human CD19+ B cells, and rodent CD45R/B220+ B cells expressed significant amounts of CEACAM1 on their surfaces (Fig. 1). The CEACAM1 expression in rat, mouse, and human leukocytes was confirmed by Western blotting, which showed that granulocytes and lymphocytes of both mouse and rat origin expressed CEACAM1 molecules with identical size, corresponding to an apparent molecular mass of 140 kDa (Fig. 2). In contrast, human granulocytes and lymphocytes expressed CEACAM1 with different apparent molecular masses, 160 and 140 kDa, respectively, due to differences in glycosylation (Fig. 2).
Discrimination between L and S isoform expression
The expression ratios of the two isoforms, CEACAM1-L and CEACAM1-S, were determined by RT-PCR using primers that could discriminate between the splice variants giving rise to the long and short cytoplasmic domains, respectively. The specificity of the primers was tested using full length cDNA for the l and S isoforms of CEACAM1 from the respective species (Fig. 3). Complete specificity was shown under the conditions used for the PCR, yielding products of the expected nucleotide lengths. Because human cells express splice variants of CEACAM1 that have either three or four Ig domains, the PCR primers amplified L and S splice isoforms of both the three Ig domain isoforms (CEACAM1–3L and CEACAM1–3S), and the four Ig domain isoforms (CEACAM1–4L and CEACAM1–4S) (Fig. 3,A). Mice and rats do not express the three Ig domain splice variants that lack Ig-domain 4, and therefore only one band for the L isoform and S isoform, respectively, was amplified by the PCR primers that were used (Fig. 3, B and C).
A triple-primer PCR procedure was used to produce quantitative information on the expression ratios of the L and S splice isoforms. In this procedure, the sense (forward (FP)) primer is common for the two isoforms, whereas the antisense (backward (BP)) primers are selective for the L or the S isoform, respectively (Fig. 4,A). The cDNAs of the two isoforms accordingly compete for the sense primer, which results in a ratio of the products that corresponds to the input ratios of the templates. By using specified ratios of CEACAM1-L and CEACAM1-S cDNA as templates and radioactive nucleotide precursors, we could demonstrate that the procedure worked and indeed gave the same ratio of the products as the ratio of the templates in the input mixture (Fig. 4,B). Furthermore, determining the ratio of the products by quantifying the fluorescence signals of the ethidium bromide-stained electrophoresis bands gave values almost identical with the ones determined from the radioactivity of the separated bands. By applying the triple-primer PCR assay to reverse transcribed RNA preparations, we found that granulocytes, B cells, and T cells of rats and mice expressed both the L and the S isoforms of CEACAM1 (Fig. 4,C). In all of these cell types the L isoform was expressed to a higher extent than the S isoform, with a L:S ratio of 2.2–2.9. In striking contrast, however, granulocytes, B lymphocytes, and T lymphocytes from human peripheral blood did not express any trace of the S isoforms as analyzed by RT-triple primer-PCR (Fig. 5,A). This was confirmed when the PCR was repeated with specific primers for either the L or the S isoforms (Fig. 5 B). Thus, human leukocytes only express the L isoforms (CEACAM1–3L and CEACAM1–4L) of CEACAM1.
Expression of GPI-linked CEACAMs
It is well documented that human granulocytes express two GPI-linked CEACAM molecules, CEACAM6 and CEACAM8 (14), but the expression of GPI-linked CEACAM molecules has not been reported in normal lymphocytes. We therefore analyzed purified populations of human granulocytes, B lymphocytes, and T lymphocytes for expression of CEACAM6 and CEACAM8 by RT-PCR. We could confirm the expression of CEACAM6 and CEACAM8 in granulocytes (Fig. 6). In addition, we found that both B lymphocytes and T lymphocytes expressed CEACAM6, whereas no signals were found for CEACAM8 expression in these cells (Fig. 6).
Signaling activities of CEACAM1 in human and rat granulocytes
CEACAM1 is a signal-regulating cell surface molecule. One of the pathways that are regulated by CEACAM1 in epithelial cells is the Erk/MAP kinase pathway (I. Scheffrahn, B. B. Singer, and B. Öbrink, unpublished observations). We therefore wanted to know whether the Erk pathway could be influenced by CEACAM molecules also in leukocytes. To that end, we investigated whether Erk1 and Erk2 could be activated by CEACAM Abs added to human and rat granulocytes, respectively. The mAb 5.4 against rat CEACAM1 had a strong activation effect on Erk1/2 in rat granulocytes (Fig. 7), whereas neither mAb 4/3/17 (Fig. 7) nor mAb 5F4 (data not shown) against human CEACAM1 activated Erk1/2 in human granulocytes. Erk1/2 in human granulocytes could, however, be significantly activated by the mAb 80H3 against CEACAM8 (Fig. 7). The mAb 9A6 against CEACAM6, in contrast, did not affect Erk1/2 activation differently than the Abs against CEACAM1 (Fig. 7).
CEACAM1 and CEACAM8 are physically associated in human granulocytes
The finding that Ab perturbation of the GPI-linked CEACAM8 caused activation of Erk1/2 prompted us to investigate whether CEACAM8 might be associated with CEACAM1, because this would offer an explanation for the signaling activity of CEACAM8. To that end, we designed a sandwich ELISA that could detect bimolecular complexes of membrane-bound proteins. In this method, a capturing Ab was used to bind one molecular species to a microtiter plate. A second Ab with a different specificity was then used to detect molecules that were physically associated with the captured protein. Using Abs against different epitopes in the same molecule, we showed that this sandwich ELISA could capture and detect both CEACAM1 and CEACAM8 (Table II). Its ability to detect bimolecular complexes was demonstrated for the β2 integrin CD11b/CD18, where anti-CD11b was used as capturing Ab, and anti-CD18 was used as detecting Ab (Table II). The specificity in detecting bimolecular complexes was demonstrated using anti-CD11b as capturing Ab and anti-CEACAM8 as detecting Ab, which did not give any signal either in granulocytes or in HeLa-Neo cells. We could then unambiguously demonstrate that CEACAM8 solubilized from human granulocytes was physically associated with CEACAM1, using anti-CEACAM1 as capturing Ab and anti-CEACAM8 as detecting Ab (Table II).
|Single Ag or Ag Heterocomplex to Be Detected .||OD .||.||.||Ab Combinations (capturing–detecting) .|
|.||Granulocyte lysate .||.||HeLa-Neo lysate .||.|
|.||Dilution 1/20 .||Dilution 1/10 .||Dilution 1/10 .||.|
|Single Ag or Ag Heterocomplex to Be Detected .||OD .||.||.||Ab Combinations (capturing–detecting) .|
|.||Granulocyte lysate .||.||HeLa-Neo lysate .||.|
|.||Dilution 1/20 .||Dilution 1/10 .||Dilution 1/10 .||.|
The sandwich ELISA is described in detail in Materials and Methods. Microtiter plates were coated with capturing Abs. Cell lysates were added, the dishes were washed, and captured Ags or Ag complexes were quantified by conjugated detecting Abs that were monitored by determination of the OD for the appropriate conjugate reaction. The tabulated values are means of duplicate determinations. The mAbs that were used had the following specificities: 12/140/4, CEACAM1 and CEA; 4/3/17, CEACAM1 and CEA; 47, CEACAM8; 80H3, CEACAM8; BEAR1, CD11b; MEM48, CD18. CEA is not expressed in granulocytes.
We have analyzed peripheral blood leukocytes of human, mouse, and rat origins for the expression pattern of CEACAM1 with particular emphasis on the two cytoplasmic domain isoforms, CEACAM1-L and CEACAM1-S. Granulocytes, B lymphocytes, and T lymphocytes from all three species expressed significant amounts of the signaling, long cytoplasmic domain isoforms CEACAM1-L. The most striking result of the present investigation was, however, that human granulocytes, B lymphocytes, and T lymphocytes did not express any detectable amounts of the short cytoplasmic domain isoforms, CEACAM1-S, whereas this isoform was significantly expressed in all three cell types both in mice and rats. This is in contrast to the finding of both the L and the S isoforms in human T cells of intestinal epithelial origin (26). Thus, different subpopulations of human T cells may exhibit different expression patterns with regard to the cytoplasmic domain isoforms of CEACAM1. Another interesting finding was that the expression ratio of the cytoplasmic domain isoforms, CEACAM1-L:CEACAM1-S, was 2.2–2.9 in all the rodent leukocytes, which is contrasting to the expression ratios in epithelial cells where the expression of the S isoform always dominates over the L isoform. The dominant expression of the S isoforms is also seen in human epithelial cells. Although it is well described that human granulocytes express the two GPI-linked molecules CEACAM6 and CEACAM8 (14), only one report has described the expression of GPI-linked CEA-family members in human lymphocytes, but only in lymphocytes of leukemic origin (27). Our finding is therefore the first report describing expression of CEACAM6 in normal human adult B and T cells.
CEACAM1-L has signaling or signal-regulating abilities due to the presence of phosphorylatable tyrosine residues in immunoreceptor tyrosine-based inhibitory motif sequences (4) and immunoreceptor tyrosine-based inhibitory motif-related switch sequences (28), which can recruit and activate src family kinases or SH2 domain-containing protein tyrosine phosphatases. CEACAM1-S lacks these tyrosine-containing sequences, but it can regulate the signaling activities of CEACAM1-L, probably via dimer formation with the L isoform (4, 19). This raises the question whether CEACAM1-mediated signaling or signal regulation is different in human vs rodent leukocytes. In this context, it is interesting that human cells, in contrast to rodent cells, express several CEACAM proteins that are associated with the cell surfaces via GPI anchors. No satisfactory explanation to the functions of these GPI-linked CEACAM molecular species have been given so far. However, it has been reported that both CEACAM6 and CEACAM8 can form complexes with CEACAM1 in human granulocytes (29), and we demonstrated unambiguously that CEACAM8 is physically associated with CEACAM1 in these cells. Thus, it seems plausible that these GPI-linked CEACAM proteins have functionally replaced the short cytoplasmic domain isoform CEACAM1-S in human leukocytes.
The replacement hypothesis is supported by our demonstration that the activation of the Erk/MAP kinase pathway could be regulated by CEACAM proteins in granulocytes. In rat granulocytes, Erk1/2 was strongly activated by an Ab directed against the N-terminal Ig-domain of CEACAM1, which confirmed the signal-regulating activity of CEACAM1 in this cell type. However, neither of two different mAbs against the N-terminal Ig-domain of human CEACAM1 activated Erk1/2 in human granulocytes. This discrepancy might be due to recognition of different epitopes in human and rat CEACAM1 by these Abs but could also be due to participation of CEACAM1-L in different types of complexes in rat and human granulocytes. The latter possibility is supported by our finding that Abs against CEACAM8 strongly activated Erk1/2 in human granulocytes. GPI-linked cell surface proteins do not have any cytoplasmic domains and can therefore trigger transmembrane signals only by complexing with other signaling proteins. Our finding that CEACAM8 occurs in a complex with CEACAM1 suggests that the signal triggered by perturbation of CEACAM8 is mediated by CEACAM1-L. Complexes between CEACAM1-L and CEACAM8 may therefore play a similar role in transmembrane signaling in human granulocytes as complexes between CEACAM1-L and CEACAM1-S in rat or mouse granulocytes. Interestingly, Abs against CEACAM6 did not cause any significant activation of Erk1/2, although both CEACAM6 and CEACAM8 can complex with CEACAM1 in human granulocytes (28). CEACAM1/CEACAM6 and CEACAM1/CEACAM8 complexes in human granulocytes therefore seem to have different functions. This suggests that human granulocytes have a more varied repertoire in terms of signal regulation than rodent granulocytes, which lack the GPI-linked CEACAM-molecules.
The activation of Erk1/2 by CEACAM Abs clearly shows that these cell surface molecules can regulate signaling activities in both rodent and human granulocytes. However, it seems unlikely that anti-CEACAM Abs are the physiological triggering substances of such signaling. Several different CEACAM molecules can act as homophilic cell adhesion molecules (4), and therefore it is possible that CEACAM-mediated signaling activities under physiological and pathological conditions are triggered by leukocyte adhesive events. This might either be adhesion between different cells or CEACAM-mediated binding of microbes to the cell surface. However, unactivated leukocytes are nonadhesive cells that require some kind of activation to become adhesive. This has clearly been demonstrated for cell surface-exposed integrins that are inactive in unactivated leukocytes and platelets but become adhesive when the cells are exposed to various activating agents (30). A similar situation seems to prevail for the CEACAM proteins. It is a challenge for future work to unravel the mechanisms of CEACAM activation and the molecular events that trigger CEACAM-mediated cell signaling.
We thank Drs. N. I. Kuprina, T. D. Rudinskaya, D. Hixson, K. Holmes, F. Grunert, and R. S. Blumberg for their generous gifts of mAbs and Drs. N. Beauchemin and T. Barnett for CEACAM1 plasmids. We thank Marie-Louise Alun, Karin Blomgren, and Ulla Sundberg for their excellent technical assistance in obtaining blood samples from mice, rats, and humans.
This work was supported by grants from the Swedish Medical Research Council (Project 05200), the Swedish Research Council (Project 05200), the Swedish Cancer Foundation (Project 3957), and Polysackaridforskning AB. B.B.S. and I.S. were supported by fellowships from the Wenner-Gren Foundations.
Abbreviations used in this paper: IgSF, Ig superfamily; CEA, carcinoembryonic Ag; CEACAM, CEA-related cell adhesion molecule; PMN, polymorphonuclear neutrophil; Erk, extracellular signal-regulated kinase; MAP, mitogen-activated protein kinase; NEB, New England Biolabs; FP, forward primer; BP, backward primer.