Regulation of NF-κB-Dependent Lymphocyte Activation and Development by Paracaspase

See allHide authors and affiliations

Science  28 Nov 2003:
Vol. 302, Issue 5650, pp. 1581-1584
DOI: 10.1126/science.1090769


Paracaspase (MALT1), a member of an evolutionarily conserved superfamily of caspase-like proteins, has been shown to bind and colocalize with the protein Bcl10 in vitro and, because of this association, has been suggested to be involved in the CARMA1-Bcl10 pathway of antigen-induced nuclear factor κB (NF-κB) activation. We demonstrate that primary T and B lymphocytes from paracaspase-deficient mice are defective in antigen-receptor–induced NF-κB activation, cytokine production, and proliferation. Paracaspase acts downstream of Bcl10 to induce NF-κB activation and is required for the normal development of B cells, indicating that paracaspase provides the missing link between Bcl10 and activation of the IκB kinase complex.

Activation of the nuclear factor κB (NF-κB) family of transcription factors is essential to the expression of genes required for lymphocyte activation, cytokine production, the generation of immune responses (1), and lymphocyte development (2). The inhibitor of NF-κ (IκB) kinase (IKK) complex (composed of two kinase subunits, IKKα and IKKβ, and a noncatalytic subunit, NEMO/IKKγ) is responsible for the phosphorylation of inhibitory molecules known as the IκBs that retain NF-κB in an inactive state in the cytoplasm (3). The association of two caspase-recruitment domain (CARD)–containing proteins, CARMA1 (CARD11) and Bcl10, has been shown to be essential to signal transduction from the T cell receptor (TCR) to the IKK complex (4). Mice deficient for either Bcl10 or CARMA1 thus display profound defects in T and B cell proliferation and cytokine production, because of a lack of NF-κB activation (59); however, the mechanism by which the CARMA1/Bcl10 complex activates IKK remains elusive (4).

Paracaspase contains an N-terminal death domain followed by two immunoglobulin (Ig)–like domains and a C-terminal caspase-like domain (10). Substantial interest in the function of paracaspase has stemmed from its role in mucosa-associated lymphatic tissue (MALT) lymphoma (11). Bcl10 and paracaspase are involved in two recurrent chromosomal translocations in MALT lymphoma, t(1;14)(p22;q32) and t(11;18)(q21;q21) (11). The first translocation results in overexpression of Bcl10, whereas the second translocation leads to production of a fusion protein that contains the N-terminal portion of inhibitor of apoptosis protein 2 (c-IAP2) linked to the caspase-like domain of paracaspase (11). The consequence of overexpression of either Bcl10 or this chimeric protein is potent NF-κB activation, which is thought to contribute to the increased B cell proliferation and survival responsible for MALT lymphoma (12, 13). Given the role of Bcl10 as a positive regulator of lymphocyte activation and proliferation and its ability to interact with paracaspase, we investigated whether paracaspase might link CARMA1 or Bcl10 to the IKK complex.

To examine the in vivo function of paracaspase, we disrupted the murine paracaspase gene using deletion in embryonic stem cells (fig. S1A) (14). Paracaspase–/– mice were born at the expected Mendelian frequency, were fertile, and appeared to be healthy.

The lymphoid compartments of paracaspase–/– mice were first examined to determine the effect of paracaspase deficiency on T and B cell development. Paracaspase–/– mice displayed normal numbers and differentiation of B cells in the bone marrow (15). However, analysis of CD19+ splenocytes showed a decrease in IgMhiIgDlo B cells (Fig. 1A), suggesting a defect in the immature or T1 subset and/or the marginal zone (MZ) and B1 B cells. Further examination of CD19+IgM+ splenocytes demonstrated a complete absence of the CD21hiCD23lo subset of MZ B cells in paracaspase–/– mice (Fig. 1B). We further confirmed the lack of MZ B cells by immunofluorescence staining of spleens from wild-type and paracaspase–/– mice with antibodies to IgM and to MOMA-1, a marker specific for the metallophilic macrophages that form a border between follicular and MZ B cells (2). Consistent with these findings, the width of the MZ B cell area in histological sections was barely detectable in spleens from paracaspase–/– mice (fig. S2). Analysis of peritoneal cells from paracaspase–/– mice revealed a complete absence of CD19+IgM+CD5+ B cells (Fig. 1C).

Fig. 1.

Lymphocyte populations in paracaspase–/– mice. Flow cytometric analysis of B cell development is shown for (A and B) splenic B cells, (C) peritoneal CD5+ B cells, and (D to F) thymocytes. Cells from 8-week-old mice were stained with the indicated antibodies. Percentages of positive cells within each quadrant are shown. Results are representative of four different experiments. Vα2TCR, TCRα.

In contrast to B cells, the frequency of T cell subsets in the spleen, lymph nodes (15), and thymus was comparable between paracaspase–/– mice and their wild-type littermates (Fig. 1D). However, staining of double-negative thymocytes (CD48) with antibodies to CD44 and CD25 revealed a significant reduction in the proportion of DN3 cells (CD25+CD44lo) and an increase in the DN4 cells (CD25CD44lo) (Fig. 1E). This appears to be due to premature maturation of the double-negative (DN) thymocytes, which display elevated levels of the αβ TCR on their surface (Fig. 1F). This phenotype mimics the phenotype of Bcl10–/– thymocytes and suggests a link between Bcl10 and paracaspase in normal T cell development (5).

We next examined the proliferative response of paracaspase–/– T and B cells. Purified T cells from paracaspase–/– mice displayed defective proliferation in response to stimulation both with antibodies to CD3 and CD28 and with phorbol 12-myristate 13-acetate (PMA) and ionomycin (Fig. 2A), suggesting that paracaspase regulates T cell activation downstream of protein kinase C. Consistent with a lack of proliferation, paracaspase–/– T cells also displayed reduced production of endogenous interleukin (IL)–2 after TCR stimulation (Fig. 2B). Addition of exogenous IL-2 was unable to rescue the defect in proliferation (Fig. 2A), and surface expression of CD25 was markedly reduced in paracaspase–/– T cells (Fig. 2C, top), demonstrating a defect in the up-regulation of the IL-2 receptor. Although no significant difference in spontaneous apoptosis of resting purified T cells was observed (15), viability was consistently decreased in stimulated paracaspase–/– T cells (Fig. 2C, middle). Analysis of cell size illustrated that viable paracaspase–/– T cells failed to increase volume or form blasts (Fig. 2C, bottom). Taken together, these results indicate that the block in proliferation of paracaspase–/– T cells occurs at an early time point before the initiation of cell growth.

Fig. 2.

Paracaspase is essential for T cell activation. (A) Proliferation of purified T cells stimulated either with different concentrations of plate-bound antibody to CD3 with or without plate-bound CD28 or with various concentrations of PMA plus ionomycin (PMA/iono), in the presence or absence of IL-2 (40 ng/ml). Proliferation was determined 48 hours later by [H3] thymidine incorporation (shown in thousands). Values are mean ± SD for triplicate samples and are representative of three different experiments. (B) IL-2 production of purified T cells stimulated with different concentrations of plate-bound antibody to CD3 with or without plate-bound CD28. Values are mean ± SD for triplicate samples and are representative of three different experiments. (C) Flow cytometric analysis of CD25 expression (top), annexin V and propidium iodide (PI) staining (middle), and forward light scatter (FSC) profiles (bottom) of purified T cells, after 24 hours of incubation with 10 μg/ml plate-bound antibody to CD3 with or without 10 μg/ml plate-bound antibody to CD28, in the absence or presence of IL-2 (40 ng/ml). Percentage of CD25+ cells (top) and viable (annexin V, PI) cells (middle) is shown. Only live annexin V, PI cells were collected for analysis of CD25 expression and FSC profiles (bottom). Control FSC profiles for untreated cells are denoted by dashed lines and treated cells by bold lines. All data are representative of three separate experiments.

Paracaspase–/– B cells were completely defective in response to stimulation with either antibody to IgM or to CD40, even in the presence of exogenous IL-4 (Fig. 3A). Paracaspase–/– B cells were also defective in their response to lipopolysaccharide (LPS), which signals through toll-like receptor 4 (Fig. 3A). This defect could be due to the lack of MZ B cells, which are required for an immediate and potent response to LPS (2).

Fig. 3.

Impaired B cell activation and T and B cell immunity in paracaspase–/– mice. (A) Proliferation of purified splenic B cells stimulated with antibody to CD40 (20 μg/ml) with or without IL-4 (40 ng/ml), antibody to IgM (20 μg/ml), or LPS (20 μg/ml). Proliferation was determined 48 hours later by [H3] thymidine incorporation. Values are mean ± SD for triplicate samples and are representative of three different experiments. (B) Impaired T cell immunity in paracaspase–/– mice in response to the T cell–dependent antigen TNP-OVA. Serum antibody titers to TNP-IgM and -IgG1 were determined by ELISA. Solid symbols, paracaspase+/+; open symbols, paracaspase–/–. (C) Reduced basal serum Ig levels in paracaspase–/– mice. Serum was collected from 6-week-old paracaspase+/+ and paracaspase–/– mice, and Ig levels were determined by ELISA. Results from individual mice are shown. (D) Lack of germinal center formation in spleens from paracaspase–/– mice. Red, B220 staining; green, PNA staining.

We assessed in vivo immune responses by challenging wild-type and paracaspase–/– mice with the T cell–dependent antigen 2,4-dinitrophenol–conjugated ovalbumin (DNP-OVA). Whereas paracaspase+/+ littermates produced high levels of DNP-specific IgM and IgG1, this response was significantly reduced in paracaspase–/– mice (Fig. 3B). In addition, basal immunoglobulin serum levels were markedly reduced in paracaspase–/– mice (Fig. 3C). Immunofluorescence staining with the CD45 receptor B220 and peanut agglutinin (PNA) revealed the complete absence of germinal center formation in spleens from paracaspase–/– mice, consistent with the lack of B cell proliferation in vivo (Fig. 3D).

To determine the molecular events responsible for the defect in antigen-receptor signaling in the absence of paracaspase, we analyzed pathways activated by T and B cell receptor (TCR and BCR) engagement. Paracaspase+/+ and paracaspase–/– T and B cells displayed comparable levels of total tyrosine phosphorylation after antigen-receptor stimulation, indicating that proximal signaling events directly after TCR and BCR engagement are unaffected by loss of paracaspase (fig. S2, A and B). Likewise, deletion of paracaspase had no effect on the phosphorylation of extracellular signal–regulated kinases (Erks) 1 and 2 or of c-Jun N-terminal kinase (JNK) after T and B cell stimulation, demonstrating that the activating protein 1 (AP1) and mitogen-activated protein kinase (MAPK) pathways remained intact in paracaspase–/– mice (Fig. 4). In contrast, phosphorylation and concomitant degradation of IκBα were substantially reduced in stimulated paracaspase–/– T and B cells (Fig. 4). Phosphorylation and degradation of IκBα induced by treatment with tumor necrosis factor–α (TNFα), however, was comparable between wild-type and paracaspase–/– T cells, demonstrating that the components of the IKK signaling complex downstream of paracaspase remained intact in paracaspase–/– cells and that loss of paracaspase affected NF-κB in a receptor-specific manner (Fig. 4C). Lack of NF-κB DNA binding activity in stimulated paracaspase–/– T and B cells was confirmed by enzyme-linked immunosorbent assay (ELISA), with specific oligonucleotides for the NF-κB p50 and p65 responsive elements and with antibodies directed against p50 and p65 (fig. S3, C and D).

Fig. 4.

Paracaspase regulates NF-κB activation in T and B cells. (A and B) Phosphorylation and degradation of IκBα and activation of Erk1/Erk2 and JNK in purified paracaspase+/+ and paracaspase–/– T cells activated with (A) 10 μg/ml plate-bound antibody to CD3 and antibody to CD28 or (B) 50 ng/ml each of PMA and ionomycin. (C) Phosphorylation and degradation of IκBα in purified paracaspase+/+ and paracaspase–/– T cells stimulated with TNFα (10 ng/ml). (D and E) Phosphorylation and degradation of IκBα and activation of Erk1/Erk2 and JNK in purified paracaspase+/+ and paracaspase–/– B cells activated with (D) 10 μg/ml antibody to IgM or (E) 50 ng/ml each of PMA and ionomycin. (F) Luciferase reporter activity of paracaspase+/+ and paracaspase–/– MEFs cotransfected with Bcl10 in the presence or absence of paracaspase (hPC).

To determine whether paracaspase was required for Bcl10-induced NF-κB activation, mouse embryonic fibroblasts (MEFs) from either wild-type or paracaspase–/– mice were cotransfected with Bcl10 and a luciferase reporter for NF-κB. Whereas Bcl10 induced NF-κB activation in wild-type MEFs, NF-κB activation in paracaspase–/– MEFs was significantly decreased (Fig. 4F), but was restored by transfection of exogenous paracaspase. These results indicate that paracaspase is a downstream mediator required for optimal Bcl10-induced NF-κB activation.

Our data demonstrate that paracaspase is required for the activation of the IKK complex and that the defective proliferative responses observed in vitro and in vivo are most likely due to the defect in NF-κB activation and subsequent gene transcription. Together with published evidence that paracaspase binds Bcl10 but not CARMA1 in overexpression studies (16), our data places paracaspase downstream of both signaling molecules in TCR- and BCR-induced NF-κB activation in vivo.

Paracaspase–/– mice displayed reduced serum immunoglobulin levels and lacked CD5+ peritoneal B cells. This phenotype is reminiscent of not only the recently described CARMA1 mutant mice (69) but also mice deficient for both the Rel and NF-κB1 transcription factors (17). In addition, paracaspase–/– mice displayed significantly reduced numbers of MZ B cells, similar to the phenotype recently described in Bcl10-deficient mice (18). Given the similarities in phenotypes of mice that lack CARMA1, Bcl10, or paracaspase, it appears likely that these proteins form a signaling complex required for optimal NF-κB activation, which is essential for both the development and function of B cells. The cIAP-paracaspase chimeric gene has been associated with MZ B cell MALT lymphomas (19). These data support the view that normal expression of paracaspase allows for appropriate NF-κB signaling in the maintenance of MZ B cells, but overexpression of the chimeric protein induces inappropriate NF-κB activation, resulting in increased survival and growth.

Although paracaspase contains a caspase-like domain with the highly conserved cysteine and histidine dyad required for catalysis by cysteine proteases, no protease activity has been reported to date (10). Paracaspase may either directly associate with the IKK complex to induce its activation, or it may recruit other adapter or enzymatic proteins that are required for IKK activation. Further analysis of its biochemical activity will provide more detail as to how paracaspase is connected to the IKK complex and may provide potential therapeutic targets for the treatment of lymphomas and potentially other immune disorders.

Supporting Online Material

Materials and Methods

Figs. S1 to S3

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article