Altered T-dependent antigen responses and development of autoimmune symptoms in mice lacking E2A in T lymphocytes

Abstract

E2A has been shown to be an important transcription factor downstream of the T-cell receptor (TCR) signal during T-cell development. The TCR signal is known to elicit different cellular responses at different stages of T-cell development. Whether E2A is still required for normal TCR signalling in mature T cells is unknown. Here we examined T-cell function after disruption of the E2A gene exclusively in the T-cell lineage. The conditional E2A-deficient mice show enhanced humoral immunity to a T-dependent antigen. We further show that E2A is involved in regulating TCR-induced T-cell proliferation events. However, E2A seems to play opposite roles in naïve and effector T cells. In the absence of E2A, TCR-induced proliferation is increased in naïve T cells and decreased in effector T cells. At older ages, these mice frequently develop antinuclear antibodies and proteinuria. Our studies suggest that E2A regulates T-cell function and the loss of E2A may promote age-dependent autoimmune diseases.

Introduction

T lymphocytes are a major cellular component of the adaptive immune system. The specificity of T-cell-mediated immune responses is determined by the T-cell receptor (TCR) on CD4 helper or CD8 cytotoxic T cells, which recognize a peptide bound to major histocompatibility complex (MHC) class II or class I molecules on target cells, respectively. CD4 T cells control the course of immune reactions by regulating the functions of other immune cells, such as B cells, whereas CD8 T cells directly participate in cytotoxic killing of target cells. In both cases, TCR-mediated events are critical in initiating and regulating T-cell-mediated immune function and defects in TCR signalling often impair T-cell development and/or alter T-cell function.¹

E2A is a member of the basic-helix-loop-helix (bHLH) transcription factor family, which has been shown to regulate cell differentiation and proliferation in many cell types including lymphocytes.² The bHLH domain of E2A mediates protein dimerization and DNA binding to canonical E-box DNA sequences (CANNTG) found in the enhancers of tissue-specific genes.³ In the lymphoid system, E2A regulates the transcription of many lineage-specific genes, including the immunoglobulin genes.⁴ Studies of E2A knockout mice have revealed a unique and essential role for E2A in B-cell development.^5,6 In the absence of E2A, B-cell development is completely arrested prior to immunoglobulin gene rearrangement, although the downstream targets of E2A directly responsible for this developmental block are not known.

The importance of E2A in T-cell development was also defined by the analysis of E2A knockout mice, which revealed three distinct roles for E2A in thymocyte differentiation. First, an accumulation of the most immature double-negative (DN1) thymocytes was observed in E2A knockout mice. This phenotype is probably the result of a partial developmental block in T-cell lineage commitment.⁷ Second, genetic crosses between E2A knockout mice and RAG2 knockout mice demonstrated a role for E2A in pre-TCR selection, a checkpoint to ensure proper rearrangement and expression of the TCR β gene during early T-cell development.⁸ E2A appears to be involved in the apoptotic pathway responsible for eliminating T cells lacking a functional pre-TCR.⁹ Third, E2A has been implicated in regulating the TCR signal during positive and negative selection of double positive (DP) cells.¹⁰ The role of E2A in T-cell selection was mapped downstream of the TCR signal and within the mitogen activated protein (MAP) kinase pathway.¹¹

The function of E2A in T-cell-mediated immunity has not been investigated in E2A knockout mice because of defects in early lymphocyte development. To circumvent these developmental problems and to be able to study E2A function in mature lymphocytes, we have recently generated a conditional E2A knockout mouse model in which E2A is specifically disrupted in the developing T cells, leaving other cell-type functions unaffected. We have shown that this system allows the complete and efficient disruption of the E2A gene in developing T cells before they reach maturity.¹² This genetic manipulation does not affect normal T-cell development, presumably because E2A proteins may not be immediately eliminated after gene disruption. These residual E2A proteins may be sufficient to support the completion of T-cell development in the thymus. In this mouse model, E2A-negative T cells are phenotypically indistinguishable from T cells of control littermates.¹² However, evaluation of T-cell mediated immunity in mice harbouring E2A-deficient T cells reveals that these mice develop altered humoral immune responses to T-dependent antigens. Furthermore, a high frequency of autoantibody production is observed in aged mice deficient for E2A in T cells. Together, these observations demonstrate an important role for E2A in regulating T-cell function and T-cell-mediated autoimmunity.

Materials and methods

Fluorescence-activated cell sorter(FACS) analysis of lymphocytes

Expression of E2A-green fluroscent protein (GFP) in T cells was measured using FACSCaliber (Becton Dickinson, Mountain View, CA) after stimulation with anti-CD3 (2 μg/ml), alone or in T helper type 1 (Th1) or Th2 priming conditions as described previously. Cell viability was evaluated by 7AAD staining. Expression of CD3, CD4, CD5, CD8, CD44, CD69, B220, immunoglobulin M (IgM) and GL-7 was visualized using either fluorescein isothiocyanate-, phycoerythrin-, or antigen-presenting cell-conjugated antibodies (Pharmingen, San Diego, CA).

Proliferation assay, cell cycle analysis, cytokine detection and ribonulease protection assay

Cells from peripheral, inguinal and axillary lymph nodes and spleen were depleted of red blood cells using ammonium chloride. T cells were then purified using T-cell-enrichment columns (R&D systems, Minneapolis, MN). Proliferation assay of purified T cells was carried out with either phorbol 12-myristate 13-acetate (PMA; 20 ng/ml) plus ionomycin (1 μg/ml) or plate-bound anti-CD3 (2 μg/ml). Cells were cultured for 54 hr, with [³H]thymidine added during the last 16 hr. To identify cell-cycle state, cells stimulated by plate-bound anti-CD3 (2 μg/ml) were stained with Hoechst stain and subjected to FACS analysis. T-cell differentiation assay was performed according to Kaplan et al.¹³ Enzyme-linked immunosorbent assay (ELISA) was used to determined the amount of interleukin-4 (IL-4) and interferon-γ (IFN-γ) secretion from the Th2 and Th1 culture conditions, respectively. A ribonuclease protection assay (RPA) was performed according to the procedure listed in Pharmingen's RPA kit.

Immunization, enzyme-linked immunospot (ELISPOT), and proliferation assay

Mice between 2 and 3 months old were immunized with a single intraperitoneal injection of 50 μg alum-precipitated 4-hydroxy-3-nitrophenyl acetyl (NP; Cambridge Research Biomedicals, Cambridge, UK) conjugated to chicken gamma-globulin (CGG; Accurate Chemical & Scientific Corp., Westburg, NY).¹⁴ Mice were killed and analysed 15 days after immunization. The frequency of NP-specific antibody-forming cells (AFCs) was determined by ELISPOT analysis as described.¹⁴ Briefly, nitrocellulose filters were first coated with 50 μg/ml NP5–bovine serum albumin (BSA), NP25–BSA, or BSA in phosphate-buffered saline (PBS). The coated filters were incubated with splenocytes (10⁵ cells/well) or bone marrow cells (10⁶ cells/well) in 96-well plates at 37° for 2 hr. Filters were washed and stained with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG1 antibodies (Southern Biotechnologies, Birmingham, AL); 3-amino-ethyl-carbozole (Sigma, St Louis, MO) was used as the substrate to visualize the HRP activity. The frequency of high affinity and total AFCs was determined from NP5-BSA-coated and NP25-BSA-coated filters, respectively, after subtraction of background on BSA-coated filter. For the proliferation assay, mice were immunized with 75 μg CGG subcutaneously at the tail base. Lymphocytes were harvested from inguinal lymph nodes 7 days post-immunization. Isolated lymphocytes were cultured in the presence of the indicated amounts of either NP-CCG, concanavalin A, or pigeon cytochrome C. Cells were cultured for 54 hr and [³H]thymidine was added during the last 16 hr.

Antinuclear antibody detection

Slides containing Hep-2 substrates (Sigma) were rehydrated with PBS, pH 7·4 for 20 min and blocked with PBS containing 10% fetal calf serum and 0·1% Tween-20 for 1 hr at room temperature. Slides were then washed with PBS, 1% BSA and 0·1% Tween-20 at room temperature for 5 min. Serum samples were diluted 1 : 10 in wash buffer and incubated on the slides for 1·5 h at room temperature. Slides were washed with Tris buffered saline (TBS) wash buffer and incubated at room temperature for 1 hr with TRITC-conjugated goat anti-mouse IgG (Sigma) at a 1 : 400 dilution in wash buffer. Following three successive washes with PBS, slides were mounted using a glycerol-based mounting medium.

Antinuclear Western blot

Nitrocellulose strips containing combined nuclear extracts from the human cell lines RAJI, 293T, BOSC23 and 3T3 were blocked overnight in blocking buffer (TBST and 2% BSA). Strips were washed with TBST and incubated with serum at a 1 : 100 dilution in diluent (TBST and 0·1% BSA) for 2 hr at room temperature. The strips were washed with TBST and an HRP-conjugated anti-mouse IgG antibody (1 : 10 000 in diluent) was added for 1 hr at room temperature. The strips were then washed with TBST and exposed to film using a chemiluminescent substrate.

Measurement of protein concentration in urine

Urine was collected from 23 E2A^loxp/loxp mice and 22 E2A^loxp/loxp Cre^tg mice. Protein levels were measured using the QuanTtest Red Total Protein Assay System (Quantimetrix Corp., Redondo Beach, CA). Protein levels were determined by comparing the optical density values of 20 μl samples at 600 nm with a standard curve.

Results

E2A up-regulation upon T-cell activation

It has been reported that E2A protein expression is low in resting B cells and increases after B-cell activation.¹⁵ However, little is known about E2A expression in mature T cells. Here, E2A expression was evaluated in mature T cells using the E2A^GFP mice, which produce functional E2A-GFP fusion proteins.¹² Purified T cells from spleen and lymph nodes were cultured for 1–3 days in either the presence of anti-CD3 alone, in the Th1-priming condition, or in the Th2-priming condition. The GFP signal was low but detectable in mature T cells at the resting stage. A significant increase in GFP signal was observed after 3 days in culture for all three activation conditions (Fig. 1). These results indicate that E2A protein levels are increased in mature T cells upon TCR activation, regardless of cytokine environment.

Figure 1.

E2A protein expression during T-cell activation. Purified T cells from E2A^GFP/GFP mice and wild-type mice were cultured in the unbiased condition (2 μg/ml anti-CD3), the Th1-priming condition, or the Th2-priming condition. The GFP signal was determined at the indicated times by FACS analysis (shaded area, wild-type T cells; line, E2A^GFP/GFP T cells).

Phenotypic analysis after E2A disruption in the T-cell lineage

A T-cell-specific E2A knockout mouse model was previously established by crossing the Lck-Cre transgene into mice carrying the E2A^loxp conditional knockout allele.¹² It has been shown that E2A is specifically disrupted in developing thymocytes and that T-cell development is not affected in this mouse model.¹² Before testing T-cell function in these mice, peripheral lymphocytes were further examined for any possible developmental defects. Spleenocytes from 3-month-old mice were analysed by lineage- and stage-specific markers (Fig. 2A). Populations of CD4 and CD8 T cells and B220 B cells are not changed in either cell numbers or relative percentage. Staining with the memory T-cell marker CD44 shows normal distribution of naïve vs. memory T cells. Analysis of 15-month-old mice also did not show any significant effect of T-cell-specific E2A disruption on either B- or T-cell populations (Fig. 2B). Neither MHC class II and IgM expression on B cells nor TCR expression on T cells are altered by E2A disruption in T cells. This result, combined with the previous studies¹² indicates that disruption of E2A in developing T cells does not affect subsequent T-cell development and that there is no inadvertent effect of Lck Cre on B-cell development.

Figure 2.

Four-colour FACS analysis of splenic lymphocytes from a littermate of 3-month-old (a) and a littermate of 15-month-old (b) E2A^loxp/loxp (−Cre) and E2A^loxp/loxp Cre^tg (+Cre) mice. Events shown are size- and 7AAD-gated for live lymphocytes. Numbers in the plots indicate the relative percentage of the subpopulation in the quadrants.

The effect of E2A disruption on cytokine production and T-cell proliferation in vitro

Next, the function of E2A-deficient T cells was evaluated for their ability to produce appropriate cytokines under conditions which promote Th1 or Th2 differentiation.¹³ Splenic T cells were purified from 2–3-month-old mice and immediately used in the differentiation assay. Detection of Th1 cytokine production under the Th1 priming condition indicated that there was not a substantial difference in IFN-γ production between the E2A^loxp/loxp mice (31 ng/ml) and the E2A^loxp/loxp Cre^tg mice (22·5 ng/ml) (Fig. 3A). However, detection of Th2 cytokine production under the Th2 priming condition revealed that there was a significant difference in IL-4 production between the two groups. Specifically, the E2A^loxp/loxp Cre^tg T cells produced more than three times the amount of IL-4 (79 ng/ml) as compared to the E2A^loxp/loxp T cells (15 ng/ml) after Th2 differentiation. The RPA revealed that the increase in IL-4 production by the E2A^loxp/loxp Cre^tg T cells was not the result of increased transcription (Fig. 3B). Furthermore, disruption of E2A did not cause a significant change in the transcription of several other Th2-type cytokines including IL-5, Il-6, IL-10, and IL-13 under the Th2 priming condition. RPA analysis of Th1-type cytokines from Th1-primed cells also failed to show any significant difference between E2A^loxp/loxp Cre^tg and E2A^loxp/loxp control samples (data not shown).

Figure 3.

(a) ELISA of cytokine production from T-cell culture. T cells were purified from E2A^loxp/loxp and E2A^loxp/loxp Cre^tg mice and cultured under either the Th1 (right) or Th2 (left) -priming condition. The concentration of IFN-γ and IL4 in the supernatant during anti-CD3 restimulation was determined by cytokine ELISA. The result is shown as means ± standard error of triplicate culture and is representative of three independent experiments. (b) Cytokine gene expression in Th2-primed splenocytes. red blood cell-depleted splenocytes from E2A^loxp/loxp and E2A^loxp/loxp Cre^tg mice were cultured in Th2 priming condition. Total RNAs were prepared and cytokine production was determined using an RNase protection assay. Lanes 1–4 represent RNA samples collected on days 0, 2, 4·5 and 7 in the primary culture. Lanes 5–7 represent samples from 2, 14 and 24 hr after restimulation with anti-CD3. The identity of each band is indicated on the right. L32 and GAPDH are controls for mRNA quantity and quality.

It has been reported that E2A proteins negatively regulate G1 progression in the cell cycle.¹⁶ This raises the possibility that the increase in IL-4 protein levels could be the result of abnormal regulation of T-cell proliferation in the absence of E2A. Proliferation assays were therefore performed using T cells from E2A^loxp/loxp and E2A^loxp/loxp Cre^tg mice. Interestingly, when cells were stimulated with anti-CD3 E2A^loxp/loxp Cre^tg T cells proliferated at a significantly greater rate than E2A^loxp/loxp T cells (Fig. 4A). However, when stimulated with PMA and ionomycin, both groups of T cells proliferated at a similar rate. Using Hoechst staining, cell-cycle analysis demonstrated that an increased percentage of E2A^loxp/loxp Cre^tg T cells were entering the S-G2/M phase after CD3 stimulation at either a suboptimal concentration (0·5 μg/ml, data not shown) or an optimal concentration (2 μg/ml) (Fig. 4B). Most T cells from either E2A^loxp/loxp or E2A^loxp/loxp Cre^tg mice were in G1 phase at the resting stage (day 0). Two days after stimulation, 21 ± 4% of E2A^loxp/loxp Cre^tg T cells were found in S-G2/M phase versus 11 ± 3% for the control T cells. The difference persisted to day 3 (48 ± 5% versus 35 ± 5%). Analysis of apoptotic cells by annexin V staining did not show any significant difference between E2A^loxp/loxp Cre^tg T cells and E2A^loxp/loxp T cells (data not shown). These data suggest that E2A plays a negative role in naïve T-cell proliferation induced by TCR cross linking.

Figure 4.

(a) T-cell proliferation responses induced by PMA (20 ng/ml) plus ionomycin (1 μg/ml) or plate-bound anti-CD3 (2 μg/ml). Purified splenic T cells were cultured for 54 hr with [³H]thymidine added during the last 16 hr. Results are shown as mean ± standard error of triplicate culture and are representative of three independent experiments. (b) Hoechst staining of cells stimulated by plate-bound anti-CD3 (2 μg/ml) at the indicated time. Percentages of cells in G1 and S-G2/M phases are indicated in each FACS plot. Data shown represent one mouse from each genotype. Similar results were obtained in three independent experiments.

E2A^loxp/loxp Cre^tg mice display an altered adaptive immune response after immunization with a T-dependent antigen

T-dependent humoral immunity was evaluated by immunizing E2A^loxp/loxp Cre^tg mice with 4-hydroxy-3-nitrophenyl acetyl conjugated to chicken gamma-globulin (NP-CGG). Flow cytometry analysis of T and B cells in the spleen at 15 days post-immunization did not reveal any significant difference between E2A^loxp/loxp Cre^tg and E2A^loxp/loxp mice. Specifically, the number of activated T cells (determined by CD69 staining) and germinal centre B cells (determined by GL-7 staining) were similar in both genotype groups (Fig. 5A). In addition, frozen spleen sections from either unimmunized or immunized mice stained with peanut agglutinin (PNA) and antibodies against IgM, IgD, or B220 revealed normal distribution of B and T cells and normal germinal centre formation (data not shown). The number of NP-specific AFCs in the bone marrow and spleen was determined 15 days post-immunization. The number of NP-specific IgG1 AFCs in the bone marrow of E2A^loxp/loxp Cre^tg mice (NP5, 93 ± 3 per 10⁶ cells; NP25, 122 ± 11 per 10⁶ cells) were approximately twice those in the E2A^loxp/loxp mice (NP5, 44 ± 5 per 10⁶ cells; NP25, 62 ± 7 per 10⁶ cells; P < 0·05) (Fig. 5B). These increases were observed in the production of both high-affinity (NP5) and total (NP25) AFCs. However, similar numbers of NP-specific IgG1 AFCs were present in the spleen of E2A^loxp/loxp Cre^tg and E2A^loxp/loxp mice, suggesting a minimal perturbation on the kinetics of AFC in the spleen at 15-day post-immunization.¹⁷

Figure 5.

(a) FACS analysis of splenic lymphocytes from NP-CGG immunized E2A^loxp/loxp Cre^tg (left) and E2A^loxp/loxp (right) mice. Events shown are size- and 7AAD-gated for live lymphocytes. Numbers in the plots indicate the relative percentage of the subpopulation in the quadrants. (b) Frequencies of NP-specific IgG1 AFCs in the spleen and bone marrow of E2A^loxp/loxp and E2A^loxp/loxp Cre^tg mice 15 days after immunization. The number of NP-specific IgG1 AFCs were determined by ELISPOT on NP25-BSA for low-affinity AFCs and NP5-BSA for high-affinity AFCs. Results are presented as mean number ± standard error of three mice for each genotype. Asterisks denote values that are significantly different from the corresponding controls (Student's t-test, P < 0·05). This result has been reproduced with another group of six mice with three for each genotype.

Following this, the antigen-induced proliferation of effector T cells was examined. CGG was added in culture to stimulate the proliferation of effector T cells isolated 7 days post-immunization with CGG. Lymphocytes from E2A^loxp/loxp Cre^tg mice show a reduced dose-dependent proliferation response to CGG as compared to E2A^loxp/loxp cultures. Specifically, stimulation of cultures containing E2A^loxp/loxp lymphocytes in the presence of 100, 200 and 400 μg/ml of CGG incorporated [³H]thymidine counts of 9670 ± 385, 16 605 ± 2400, and 28 004 ± 4048, respectively, in comparison to counts of 3211 ± 473, 6374 ± 358 and 9948 ± 2432, respectively, for E2A^loxp/loxp Cre^tg T cells (Fig. 6A). To determine that proliferation in the presence of CCG is the result of an antigen-specific response, lymphocytes from both mouse groups were cultured in the presence of pigeon cytochrome C. Results showed that both groups of lymphocytes had minimal incorporation of tritiated-thymidine at several doses of pigeon cytochrome C (Fig. 6B). In contrast, cells from both genotype groups proliferated equally well after stimulation with concanavalin A (Fig. 6C). These data suggest that E2A may play a positive role in regulating the proliferation of antigen-experienced T cells during their re-exposure to the same antigen.

Figure 6.

Antigen-specific T-cell proliferation response. Purified lymph node lymphocytes from NP-CGG-immunized mice were cultured in the presence of NP-CGG (a), pigeon cytochrome C (b), or concanavalin A (c) for 54 hr with [³H]thymidine added during the last 16 hr. Results are shown as mean ± standard error of three animals for each genotype group.

Development of antinuclear antibodies in aged E2A^loxp/loxp Cre^tg mice

Autoantibodies are often detected in mice displaying altered immune responses. We therefore investigated autoantibody production in E2A^loxp/loxp Cre^tg mice using three detection methods. First, immunofluorescent microscopy for antinuclear antibodies (ANA) showed that more than half of the E2A^loxp/loxp Cre^tg population (15 out of 27 analysed) developed ANA at age 8 months or older (Table 1). No gender bias was observed in these ANA-positive mice. In contrast, the frequency of detecting ANA among the control groups with similar genetic background was only about 15%. Production of ANA from various E2A^loxp/loxp Cre^tg mice as determined by Hep2/slide detection can be divided into three antigenic groups: rimmed, speckled and nucleolar (Fig. 7A). This was in contrast to the staining pattern from MRL^lpr/lpr mice, which invariably showed homogeneous nuclear staining. Second, ELISA analysis of the 15 ANA-positive E2A^loxp/loxp Cre^tg mice showed that eight were positive for one or combinations of antibodies against histones, single-stranded DNA, and double-stranded DNA (data not shown). ELISA analysis of 46 E2A^loxp/loxp Cre^tg mice between the ages of 2 and 7 months did not detect any significant amount of autoantibodies (data not shown). Third, Western analysis was used to evaluate a group of 2-year old mice for ANA. ANA with strong reactivity to nuclear antigens were detected in two out of five E2A^loxp/loxp Cre^tg mice but not in four control mice (Fig. 7B). In comparison to control lanes in which antigen-specific antibodies were reacted against test extract, the two mice showed autoantibody reactivity most closely resembling that against ribonuclear proteins (lane 5). In all three assays, we consistently found that the ANA present in E2A^loxp/loxp Cre^tg mice have varying specificity.

Table 1.

Summary of fluorescent microscopy ANA analysis of mouse sera

	No. of mice showing antinuclear staining*
Genotype (no. of mice)	A	B	C	D	% responders
E2Aloxp/loxp	0	5	8	2	56%
Lck Cretg (27)					(15/27)
E2Aloxp/loxp (13)	0	1	0	1	15%(2/13)
E2A+/+ Lck Cretg and	0	0	1	0	14%
E2Aloxp/+ Lck Cretg (7)					(1/7)
E2A+/+ and	0	1	0	0	17%
E2Aloxp/+ (6)					(1/6)
MRL-lpr (4)	4	0	0	0	100%(4/4)

^*The staining patterns on HEp-2 cells are classified as follows: A, homogeneous; B, rimmed; C, speckled; D, nucleolar.¹⁸

Figure 7.

Detection of ANA in E2A^loxp/loxp Cre^tg mice. (a) Immunofluorescent microscopy analysis of serum (diluted 1 : 10) applied to the human epithelial (HEp-2) tissue culture substrate (Sigma) exhibited three antinuclear antibody staining patterns that include nucleolar (a1), speckled (a2), and rimmed (a4). Serum from MRL-lpr mice exhibited a homogeneous antinuclear antibody pattern (a3). (b) Western analysis of mouse serum (diluted 1 : 100). Positive control autoantibodies are against scleroderma (lane 1), La/SS-B (lane 2), Jo-1 (lane 3), Smith antigen (lane 4), and RNP (lane 5) (Immunovision). Experimental samples are E2A^loxp/loxp Cre^tg (lanes 7, 10, 11, 12 and 13), E2A^loxp/+Cre^tg (lane 8), E2A^loxp/loxp (lanes 6 and 14), and E2A^loxp/+ (lane 9).

Development of proteinuria and lymphocyte infiltration of kidney tissue in aged E2A^loxp/loxp Cre^tg mice

A phenotype typical of several autoimmune disorders is an increase in urine protein. A population of 45 mice, ranging from 1·5 to 2 years in age, was used to test kidney function with a proteinurea test. Protein amounts from male and female E2A^loxp/loxp mice demonstrated that on average, male mice had a larger amount of protein in their urine than female mice. Within each sex group, total urine protein levels among the E2A^loxp/loxp Cre^tg mice were found to be significantly higher than the E2A^loxp/loxp controls (Fig. 8A). Histological analysis revealed the presence of perivascular lymphocyte infiltrate in the kidneys of the aged E2A^loxp/loxp Cre^tg mice (Fig. 8B). The same animals did not show significant lymphocyte infiltrates in several other organs examined, including pancreas, liver and heart.

Figure 8.

Autoantibody production affects kidney function. (a) Comparison of urine protein levels for 1·5–2-year-old E2A^loxp/loxp and E2A^loxp/loxp Cre^tg mice grouped by sex. Urine protein amounts were found to be significantly different, as determined by a Student's t-test. (b) A representative haematoxylin & eosin staining of a kidney section from an E2A^loxp/loxp Cre^tg mouse. Arrowheads indicate lymphocyte infiltrates. An arrow shows the presence of a normal glomerulus.

Discussion

E2A regulates T-cell function

This study is the first to report a physiological role of E2A in mature T cells. Using the B-cell antibody response as a functional readout, we found that E2A^loxp/loxp Cre^tg mice displayed an altered humoral immunity when challenged by NP-CGG. Abnormal T-cell function was further exemplified by the development of autoimmune symptoms in the aged animals. Although the mechanism by which E2A regulates T-cell immunity requires further investigation, our study suggests that E2A may be involved in mediating the proliferation signals from TCR.

Analysis of naïve and antigen-experienced T cells showed that E2A may play both negative and positive roles downstream of TCR signals. We show that E2A disruption led to an enhanced T-cell proliferation in naïve T cells and reduced proliferation in antigen-experienced T cells. This result indicates that the downstream targets of E2A in naïve and antigen-experienced T cells may be different. It has been well established that naïve, effector, and memory T cells are intrinsically different in their ability to respond to TCR signals¹⁹ and they have different proliferation kinetics. Perhaps, E2A is involved in the differential response to the TCR signals in these cells. Given the reported roles played by E2A in other cell types, it is possible that E2A may control antiproliferation genes such as p21 in naïve T cells^20,21 and pro-proliferation genes such as cyclin D3 in antigen-experienced T cells.²²

The CDK inhibitor p21^CIP/WAF1 has been shown to be directly regulated by E2A in fibroblasts.²¹ A recent study of p21^CIP/WAF1 knockout mice revealed hyperproliferation of activated T cells during sustained stimulation.²³ Interestingly, p21^CIP/WAF1 knockout mice developed ANA at a similar age to E2A^loxp/loxp Cre^tg mice. At face value, the T-cell phenotype in p21^CIP/WAF1 knockout mice and T-cell-specific E2A knockout mice are very similar. However, several unique aspects of p21^CIP/WAF1 knockout mice should be noted. First, the enhanced T-cell proliferation in these mice was only observed after prolonged stimulation (3 days in concanavalin A plus 6 days in IL-2). Second, ANA were only detected among female p21^CIP/WAF1 knockout mice and were mostly positive for anti-DNA antibodies. Finally, the autoimmune disorders reported in p21^CIP/WAF1 knockout mice may not be simply the result of a loss of p21 function in T cells because the p21^CIP/WAF1 gene is inactivated in all cell types. We argue that p21^CIP/WAF1 is unlikely to be the only E2A target responsible for the phenotypes observed in E2A^loxp/loxp Cre^tg mice. Whether E2A regulates the expression of p21^CIP/WAF1 and/or other cell-cycle genes in antigen-activated T cells requires further evaluation. The fact that E2A^loxp/loxp Cre^tg mice are fully viable and fertile makes future genetic and biochemical approaches to downstream target genes feasible.

An animal model for age-dependent autoimmune disorders

Although E2A^loxp/loxp Cre^tg mice appear normal at young ages, most of them develop ANA when they are older. A higher proportion of aged E2A^loxp/loxp Cre^tg mice also showed reduced kidney function by the urine protein test. It is not entirely clear how E2A disruption in T cells may result in these autoimmune symptoms. However, several scenarios may be considered. First, certain autoreactive T cells lacking functional E2A gene products may escape negative selection during thymopoiesis, leading to the development of autoimmunity. The break of self-tolerance may also occur in peripheral rather than central lymphoid organs. Second, the autoimmune symptoms may be the result of a prolonged survival of the activated T cells. A role for E2A in promoting T-cell apoptosis has been indicated in the study of T-cell tumours.⁹ Although a change in cell survival is not observed for E2A-deficient T cells, our assay conditions may not be sensitive enough to detect any subtle changes. Third, E protein deletion may alter the TCR signalling threshold, leading to disregulated proliferation of T cells carrying an inappropriate TCR. However, the effect of E2A disruption on TCR-triggered proliferation can be either increase or decrease depending on naïve or antigen-experienced status of T cells, respectively. Finally, the development of ANA may be the result of several independent events, which explains why individual mice develop different types of ANA. These and other possibilities could be further investigated through the identification of E2A downstream target genes. The normal lymphocyte development and survival rate of E2A^loxp/loxp Cre^tg mice makes them a unique animal model for studying E2A-dependent transcription in the context of immune responses and age-dependent autoimmunity.