CEACAM1 as a mediator of B-cell receptor signaling in mantle cell lymphoma

Abstract

B-cell receptor (BCR) signaling plays an important role in the pathogenesis of mantle cell lymphoma (MCL), but the detailed mechanisms are not fully understood. In this study, through a genome-wide loss-of-function screen, we identify carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) as an essential factor in a subset of MCL tumors. Our signal transduction studies reveal that CEACAM1 plays a critical role in BCR activation through involvement in two dynamic processes. First, following BCR engagement, CEACAM1 co-localizes to the membrane microdomains (lipid rafts) by anchoring to the F-actin cytoskeleton through the adaptor protein filamin A. Second, CEACAM1 recruits and increases the abundance of SYK in the BCR complex leading to BCR activation. These activities of CEACAM1 require its cytoplasmic tail and the N-terminal ectodomain. Considering that previous studies have extensively characterized CEACAM1 as an ITIM-bearing inhibitory receptor, our findings regarding its activating role are both surprising and context-dependent, which may have implications for BCR-targeting therapies.

Fig. 1. CRISPR library screen identifies CEACAM1 as essential in MCL.

a Left panel. Log₂ fold-change of sgRNA counts between days 14 and 0 (P < 0.01; two-sided permutation test) are plotted against the Z ratio of mRNA expression between MCL and naïve-B cells (P < 0.05; two-sided t-test). Right panel. Overlap of selected genes depleted at least twofold and overexpressed in MCL with pan-dependent genes and genes associated with poor prognosis. Numbers in green boxes indicate filtering strategy (see text). b Shown are mean values of selected normalized sgRNA counts on days 0 and 14 from two biological replicates. c Immunoblots show CEACAM1 knockout in JEKO-1 cells transduced with control (gNTC) or CEACAM1 gRNAs (gCC1) followed by anti-IgM antibody stimulation (2 μg/ml, 5 min). d JEKO-1 cells were transduced with indicated sgRNAs, and viable, GFP⁺-transduced cells were monitored over time by FACS. Shown are the means of GFP+ fractions compared to day-2 samples from three biological replicates. Error bars, SD. *P < 0.05 by a two-sided, paired t-test. e Immunoblots show CEACAM1 knockdown by shRNA in JEKO-1 cells. f Indicated cells were transduced with CEACAM1 shRNA and viable, propidium iodide (PI)-negative cells were assessed by FACS over time. Shown are the means of PI-negative fractions compared to day-2 samples from three biological replicates. Error bars, SD. *P < 0.05, **P < 0.01 by a two-sided, paired t-test. g CEACAM1 is required for MCL survival in vivo. JEKO-1 cells were transduced with control or CEACAM1 shRNA and intravenously transplanted into NSG mice. Shown are weekly bioluminescence images of six mice/group. h Line graphs show the means of bioluminescence signals measuring tumor growth in mice described in (g). Error bars, SEM. *P < 0.05 by a one-sided t-test. i Kaplan–Meier survival analysis of mice shown in (g). P value, log-rank test. j Top panel, Generation of double SOX11/CCND1 transgenic (DT) and CEACAM1-deficient mice. Middle panel, Representative FACS plots show MCL-like population (CD19 + CD5 + CD23-). Bottom panel, bar graphs show means of %MCL-like cells from the peripheral blood of indicated mice and sample sizes. Error bars, SD. P values, two-sided unpaired t-test. Source data are provided as a Source Data file.

Fig. 2. CEACAM1 expression in lymphoma cells.

a–c Box plots show CEACAM1 mRNA expression levels in MCL and other cell types from the datasets GSE2350 (a, b) and GSE132929 (c). The boxes extend from the 25th to 75th percentiles with the center line as median and whiskers drawn from 10th to 90th percentiles. ****P < 0.0001 by a two-sided Mann–Whitney U- test. GC germinal center, Mem memory, Fo follicular, BL Burkitt lymphoma, DLBCL diffuse large B-cell lymphoma, FL follicular lymphoma, PEL primary effusion lymphoma. d Representative immunoblot analysis of CEACAM1 expression in indicated cells from at least three independent experiments. Numbers below bands represent densitometric values of CEACAM1 signals normalized over GAPDH loading controls. e Flow cytometry analysis of surface CEACAM1 expression in MCL PDXs compared to CEACAM1-negative RAMOS cells from at least two independent experiments. Live cell gating in this experiment and throughout the study is described in Supplementary Fig. 13. f Representative immunohistochemistry images showing varying CEACAM1 staining levels for indicated formalin-fixed paraffin-embedded tissue microarrays from three independent experiments. In cHL tissue, background plasma cells are also stained positive. cHL classic Hodgkin lymphoma, CLL chronic lymphocytic leukemia. See Table 1 for a summary of CEACAM1 positivity in specific diseases. Source data are provided as a Source Data file.

Fig. 3. CEACAM1 is required for BCR signaling.

a–e Indicated cell lines were transduced with control or CEACAM1 shRNA, followed by stimulation with anti-IgM (1 μg/mL). a Shown are normalized Ca²⁺ signals. ****P < 0.0001 by two-way ANOVA. Data were representative of at least three independent experiments. b, c Immunoblots show effects of CEACAM1 knockdown on BCR signaling components. Data were representative of three independent experiments. d, e Immunoblots show CEACAM1 knockdown (top panels) and its effect on Ca²⁺ signals (bottom panels). f Top panel, immunoblots of splenocytes from wild type (+/+) or Ceacam1-deficient (−/−) mouse with indicated antibodies. Bottom panel, Ca²⁺ signals of indicated splenic B220⁺ B cells. Data from (d–f) are representative of at least two independent experiments. g CEACAM1 mRNA expression levels in CD19+ sorted MCL cells from peripheral blood (PB) or lymph nodes (LN) analyzed from the GSE70910 dataset. Closed circles or squares represent individual patient samples. Horizontal bars from each group indicate mean mRNA expression. P value is from a two-sided unpaired t-test with Welch’s correction. h CEACAM1 expression is correlated with ibrutinib response. Left panels, FACS plots showing surface CEACAM1 expression on indicated MCL and MZL cell lines. Isotype, negative isotype antibody on Z-138 cells. Right panels, Indicated cell lines were treated with indicated doses of ibrutinib for 4 days, and viable propidium iodide (PI)-negative cells were assessed by flow cytometry. Line graphs showing means of normalized PI-negative fractions from three independent experiments. Error bars, SD. Fifty-percent inhibition concentration (IC₅₀) values were calculated by GraphPad Prism v8. i, j ITIM tyrosine residues are required for BCR signaling. JEKO-1 cells transduced with control (gNTC) or CEACAM1 gRNA (gCC1), followed by reintroducing WT CEACAM1 (4 L), CC1-4L-Y493F/Y520F mutant (YY/FF), or short cytoplasmic tail (4S). Controls or reconstituted cell lines were verified for CEACAM1 expression by FACS using B1.1 antibody (i) and stimulated with 1 μg/mL of anti-IgM for 5 min, followed by immunoblotting with indicated antibodies, including the CEACAM1 E1 antibody (j). Data from (i, j) are representative of at least two independent experiments. Source data are provided as a Source Data file.

Fig. 4. CEACAM1 co-localizes to lipid rafts and promotes F-actin reorganization during BCR activation.

a Diagram of sucrose-density gradient fractionation of lipid rafts. b Left panels, Immunoblots of fractions indicated in (a) from control or CEACAM1-knockout JEKO-1 cells stimulated with control or anti-IgM antibody (1 μg/ml) for 2 min. Right panels, immunoblot signal quantification of fraction I (red dashed-line box) after normalization to Flotillin-1 and unstimulated controls. Shown are the means of fold changes from three independent experiments. Error bars, SD. **P < 0.01, *P < 0.05 by a two-sided, paired t-test. ns not significant. c Left panels, Control (gNTC) or CEACAM1 knockout (gCC1) JEKO-1 cells were stimulated with 2 μg/ml anti-IgM antibody for 2 min. Shown are representative cells from confocal immunofluorescence images in Supplementary Fig.8 of control and IgM-stimulated cells co-stained with anti-FLNA (green) and anti-LYN (red) antibodies, followed by nuclear staining with DAPI (blue). Scale bar, 2 μm. Right panels, quantified fluorescent signals for each cell from the samples described. Shown are the sums of intensities in signal-positive areas per cell in arbitrary units from three independent experiments. Horizontal red bars indicate the mean. Approximately 200 cells from each sample were analyzed. ****P < 0.0001 by a two-sided unpaired t-test. ns not significant. d CEACAM1 is required for lipid-raft assembly. Top panels, Control, CEACAM1 knockout (CC1 KO), or CC1 KO + CEACAM1-4L (4 L) JEKO-1 cells were stimulated with 2 μg/ml anti-IgM antibody for 5 min. Shown are representative confocal immunofluorescence images of control and IgM-stimulated cells stained with anti-CEACAM1 antibody (green), GM1 via cholera toxin B (red), p-SRC^Y416 antibody (green), or F-actin via Actin-Stain 555 Phalloidin (red) followed by nuclear staining with DAPI (blue). Scale bar, 10 μm. Bottom panels, quantified fluorescent signals from the samples described in the top panels. Shown are the sums of intensities in signal-positive areas per cell in arbitrary units from three independent experiments. Horizontal red bars indicate the mean. Approximately 200 cells from each sample were analyzed. ****P < 0.0001 by a two-sided unpaired t-test. Source data are provided as a Source Data file.

Fig. 5. CEACAM1 interactions with BCR signaling components.

a, b Top panels, Proximity ligation assay (PLA) showing interactions (visualized as red dots using Airyscan FAST 2D confocal microscope and a 40x/1.2NA water objective) between CEACAM1 (CC1) and the indicated proteins in JEKO-1 and MINO cells stimulated with 2 μg/ml of anti-IgM antibody for the indicated times. Bottom panels, Quantification of PLA signals shown in the top panels for ~200 cells on average from three independent experiments using QuPath 0.3.2 software. ****P < 0.0001, ***P < 0.001 by a two-tailed unpaired t-test. ns not significant. c, d Immunoprecipitation analysis of CEACAM1 interactions. Top panels, JEKO-1 or MINO cells were stimulated with 2 μg/ml of anti-IgM antibody for the indicated times, and CEACAM1 was immunoprecipitated with a CEACAM1-specific antibody or IgG control antibodies, followed by immunoblotting with the indicated antibodies. One percent of the total lysates was used as an input control. Bottom panels, Quantification of indicated co-IP signals shown in the top panels. Bar graphs show the means and individual densitometric values from three independent experiments for each timepoint normalized to the CEACAM1 pull-down signals. Error bars, SD. ***P < 0.001, **P < 0.01, *P < 0.05 by a two-sided unpaired t-test. ns not significant. Source data are provided as a Source Data file.

Fig. 6. CEACAM1 interactions with BCR signaling components require an intact N-domain.

a Left panel, Immunoprecipitation analysis of CEACAM1 interactions. CEACAM1-knockout JEKO-1 cells were transduced with either full-length (4 L) or N-domain truncated (ΔN) CEACAM1 constructs, stimulated with 2 μg/ml anti-IgM antibody for the indicated times, followed by CEACAM1 immunoprecipitation and immunoblotting with the indicated antibodies. One percent of the total lysates was used as an input control. p-Y, anti-phosphotyrosine antibody clone 4G10. Right panels, Quantification of indicated co-IP signals shown in the left panels. Bar graphs show the means and individual densitometric values from four independent experiments for each timepoint normalized to the CEACAM1 pull-down signals. Error bars, SD. **P < 0.01, *P < 0.05 by two-way ANOVA. b Proximity ligation assay (PLA) showing interactions (visualized as red dots using Airyscan FAST 2D confocal microscope and a 40x/1.2NA water objective) between 4 L or ΔN and the indicated proteins. c Quantitation of PLA signals shown in (b) for 100–300 cells on average from three independent experiments using QuPath 0.3.2 software. Error bars indicate means with S.D. ****P < 0.0001 by two-sided Mann–Whitney U-test. ns not significant. Source data are provided as a Source Data file.

Fig. 7. A dual role of CEACAM1 in BCR signaling.

a Left panel, CEACAM1 suppresses BCR signaling in Z-138 cells. Z-138 cells were transduced with control empty vector (EV), WT CEACAM1 (CC1-4L), or short cytoplasmic tail CEACAM1 (CC1-4S). Transduced cells were stimulated with 2 μg/mL of anti-IgM F(ab’)2 fragments for the indicated times, followed by immunoblot analysis probed with the indicated antibodies. HA, hemagglutinin, a protein tag in-framed with CEACAM1 to detect the 4S isoform. Right panel, Quantification of the immunoblot signals shown in the left panel. Bar graphs show the means of densitometric values from the indicated time points normalized to GAPDH loading controls from two independent experiments. P values, one-sided permutation test. b, c CEACAM1-knockout JEKO-1 cells (gCC1) or Z-138 cells were transduced with either empty vector control or full-length (4 L) CEACAM1 construct, stimulated with 2 μg/ml anti-IgM antibody for the indicated times, followed by CEACAM1 immunoprecipitation and immunoblotting with indicated antibodies. One percent of the total lysates was used as an input control (b). The samples shown in (b) derive from the same experiment, but different gels for CEACAM1, p-SYK^Y352, SYK, SHP-1, GAPDH, and another for SHP-2 were processed in parallel. Bar graphs show the quantification of the co-IP signals shown in (c). Shown are the means of densitometric values from two independent experiments for each timepoint normalized to the CEACAM1 pull-down signals. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05 by two-way ANOVA. d Left panels, Proximity ligation assay (PLA) showing interactions (visualized as red dots using Airyscan FAST 2D confocal microscope and a 40x/1.2NA water objective) between CEACAM1 (CC1) and the indicated proteins in Z-138 cells or Z-138 cells transduced with CEACAM1-4L followed by stimulation with 2 μg/ml of anti-IgM antibody for the indicated times. Right panels, Quantification of PLA signals shown in the right panels for ~200 cells on average from three independent experiments using QuPath 0.3.2 software. ****P < 0.0001 by a two-sided unpaired t-test with Welch’s correction. Source data are provided as a Source Data file.

Fig. 8. Proposed model for CEACAM1 function in BCR signaling.

Following antigen stimulation in B-cell lymphomas with abundant CEACAM1 expression, CEACAM1 binds to FLNA, which anchors to the actin cytoskeleton and lipid rafts, and recruits SYK to the proximity of CD79A/B to enhance BCR activity. In cells with low or no CEACAM1 expression (and potentially low SYK expression), SHP-1 and SHP-2 outcompete for CEACAM1 binding, leading to signal attenuation. Created in BioRender. Ngo, V. (2025) https://BioRender.com/dl16wxh.