Induction of T Helper Type 2 Immunity by a Point Mutation in the LAT Adaptor

See allHide authors and affiliations

Science  14 Jun 2002:
Vol. 296, Issue 5575, pp. 2036-2040
DOI: 10.1126/science.1069057


The transmembrane protein LAT (linker for activation of T cells) couples the T cell receptor (TCR) to downstream signaling effectors. Mice homozygous for a mutation of a single LAT tyrosine residue showed impeded T cell development. However, later they accumulated polyclonal helper T (TH) cells that chronically produced type 2 cytokines in large amounts. This exaggerated TH2 differentiation caused tissue eosinophilia and massive maturation of plasma cells secreting to immunoglobulins of the E and G1 isotypes. This paradoxical phenotype establishes an unanticipated inhibitory function for LAT that is critical for the differentiation and homeostasis of TH cells.

The TCR recognizes peptides bound to major histocompatibility complex (MHC) molecules and relays this information to the T cell through adaptor proteins. The adaptor LAT coordinates the assembly of signaling complexes through multiple tyrosine residues within its intracytoplasmic segment (1). Upon TCR-induced phosphorylation, each of these tyrosines manifests some specialization in the signaling proteins it recruits. For instance, mutation of tyrosine 136 (Y136) selectively eliminates binding of phospholipase C–γ1 (PLC-γ1), whereas the simultaneous mutation of Y175 and Y195 results in loss of binding of the Gads adaptor (2–4). Mice deficient in LAT (LAT–/–) or having a mutation of the four COOH-terminal tyrosine residues revealed that LAT is essential for the function of the pre-TCR, a molecular sensor that controls early T cell development and shares common signaling complexes with the TCR (5, 6).

To address the importance of LAT Y136 in vivo and to analyze the consequence of restricting LAT to only a subset of its docking function, we generated knock-in mice with a mutation that replaced tyrosine 136 with phenylalanine (Y136F). Mice homozygous for this mutation, LATY136F mice, were born at expected Mendelian frequencies, and T cell expression of LAT was comparable to wild-type T cells [fig. S1 (7)]. Although LATY136F mice displayed normal peripheral lymphoid organs at birth, their spleen and lymph nodes started to enlarge, so that by 7 weeks of age, spleen cellularity was about 5 times that of wild-type mice (Fig. 1). Despite prominent lymphocytic infiltrations in the lung, liver, and kidney, homozygotes lived to at least 17 weeks of age, and no chronic intestinal inflammation or tumor formation was observed (8). The effects of the LAT Y136F mutation were only detectable after breeding mice to homozygosity or to offspring carrying a null allele of the LAT gene.

Figure 1

Enlarged secondary lymphoid organs in LATY136F mice. (A) Spleen (right) and inguinal and mesenteric lymph nodes (left) from 7-week-old wild-type and LATY136F mice. (B) CD4/CD8 profiles of spleen cells from wild-type and from LATY136F mice. (C) Light scatter analysis and CD4/CD8 staining profiles of lymph node cells from wild-type and from LATY136F mice at 6 weeks of age. The cells with an intermediate forward scatter (FSC) and a high side scatter (SSC) correspond to eosinophils (see fig. S3). (D) Comparison of the levels of CD3, CD69, CD62L, CD44, and CD95 on CD4 T cells from 6-week-old mice expressing wild-type LAT or LATY136F molecules. CD4 T cells from LATY136Fmice express lower surface levels of CD95 relative to wild-type CD4 T cells.

The thymuses of LATY136F mice had about one-tenth the cells of wild-type thymuses and reduced numbers of CD4+CD8+ double-positive (DP) thymocytes (Fig. 2). After reaching a peak in mutant newborn mice, DP cell numbers decreased and were almost undetectable in mice older than 7 weeks (Fig. 2A and fig. S2A). Coincident with this progressive DP cell erosion, discrete populations of CD4 and CD8 single-positive (SP) thymocytes started to dominate the thymus and showed a CD4/CD8 ratio skewed toward CD4 cells (Fig. 2B). TCR expression in LATY136F thymocytes was one-half (DP) or one-sixth (SP) that observed in wild-type mice (Fig. 2C). The DP thymocytes found in young mutant mice lacked CD5 molecules at their surface (fig. S2B). CD5 is a negative regulator of TCR signaling, and its expression increases during T cell development in a manner proportional to the intensity of pre-TCR and TCR signaling (9–11). From these data and the presence of a normal complement of CD4CD8 double-negative (DN) cells in LATY136F thymuses (8), we conclude that the LAT Y136F mutation resulted in a severe but partial impairment of αβ T cell development. Although development of T cells bearing TCR made of γδ chains (γδ T cells) is ablated in LAT–/– mice (5), the LAT Y136F mutation did not affect γδ T cell development detectably (fig. S2C).

Figure 2

Impeded T cell development in LATY136F mice. (A) Absolute numbers of total thymocytes, DP cells, TCRαβhi cells, and T cells bearing γδ TCR (TCRγδ+) found in LATY136F (black bars) and wild-type (white bars) thymuses at different ages of embryonic life and at 1 and 2 weeks of age. (B) Light scatter analysis and CD4/CD8 staining profiles of total thymocytes from wild-type and from LATY136F mice at 5 and 7 weeks of age. The cells with an intermediate forward scatter and a high side scatter correspond to eosinophils (see fig. S3). The percentage of cells within each gate is indicated. Also shown is the total number of thymocytes (averaged from six experiments). (C) Comparison of the levels of CD3-ɛ, a TCR transduction subunit, on DP, CD4 SP, and CD8 SP thymocytes from 6-week-old mice expressing wild-type LAT or LAT Y136F molecules.

Given the scarcity of SP thymocytes found in newborn LATY136F mice (Fig. 2A and fig. S2A), one would expect very few SP cells in secondary lymphoid organs. However, SP cells appeared in the spleen and lymph nodes of LATY136F mice with the same kinetics as in wild-type mice. However, they showed a strong bias for CD4 cells and expanded over time (Fig. 1, B and C). These CD4 cells had a CD25, CD44hi, CD62Llo, CD69+ phenotype closely resembling activated and memory T cells (Fig. 1D). They also expressed low levels of TCR on their surfaces, an attribute that may in part account for their inability to proliferate in response to TCR stimulation in vitro (8). Despite the absence of CD5 on LATY136F DP thymocytes, CD5 molecules were detected on their direct CD4 and CD8 SP progeny (fig. S2B). Analysis of DNA content by propidium iodide staining showed that the CD4 populations from wild-type and LATY136F mice contained 3.3 and 7.2% of cells, respectively, in the G2-S-M phases of the cell cycle (8). Moreover, when cultured in medium alone, CD4 T cells purified from LATY136F mice showed a lower rate of spontaneous apoptosis than wild-type CD4 T cells (8). Therefore, the progressive accumulation of CD4 T cells in the periphery of LATY136Fmice is probably due to both their extended survival and increased proliferation.

When freshly isolated from LATY136F mice, CD4 T cells expressed sufficient interleukin-4 (IL-4) and IL-10 transcripts to allow their detection even without ex vivo restimulation (Fig. 3A). Upon activation by phorbol 12-myristate 13-acetate (PMA)–ionomycin IL-5, IL-13, and interferon-γ (IFN-γ) transcripts were also detected (Fig. 3B), and close to 80% of the CD4 T cells expressed very high levels of intracytoplasmic IL-4 (Fig. 3C). In contrast, wild-type CD4 T cells showed only the IL-2 and IFN-γ transcripts expected after activation of primary T cells. Moreover, despite their activated phenotype, CD8 T cells from LATY136F mice did not produce any IL-2, IL-4, IL-5, and IFN-γ when stimulated under similar conditions (12). Consistent with the notion that the CD4 T cells from LATY136F mice were refractory to TCR stimuli, none of them scored as IL-4+ in response to antibody-mediated TCR cross-linking (8). Thus, over the first weeks of their life, LATY136F mice spontaneously developed a T helper type 2 (TH2) lymphoproliferative disorder. In the case of wild-type CD4 T cells, a TH2 polarization of such magnitude is only achieved after prolonged antigenic stimulation in the presence of IL-4 (13).

Figure 3

Type-2 cytokine production in CD4 T cells freshly isolated from LATY136F peripheral lymphoid organs. (A). Analysis of the cytokine transcripts expressed in ex vivo CD4 T cells isolated from the spleens of wild-type mice (lane 3) and LATY136F mice (lane 4). Total RNA was analyzed by a multiprobe ribonuclease protection assay using a MCK1 RiboQuant mouse template set. The autoradiogram also shows the MCK1-probe set not treated with RNase (lane 1) and a control sample provided by the supplier (lane 2). The identity of the various protected bands is indicated on the right. (B) Analysis of the cytokine transcripts expressed in CD4 T cells isolated from wild-type (lane 2) and LATY136F (lane 3) mice after stimulation by PMA/ionomycin for 15 hours. Lane 1 corresponds to wild-type CD4 T cells that were grown under TH2 polarizing conditions. Samples were processed as described in (A). (C) IL-2, IL-4, IL-5, and IFN-γ production analyzed in single cells. Ex vivo CD4 T cells purified from wild-type and from LATY136F lymph nodes were cultured for 4 hours in the presence of monensin to trap cytokine in the endoplasmic reticulum. During the culture period, cells were stimulated with PMA/ionomycin. At the end of the culture, cells were processed for intracellular staining. Numbers indicate percentages of cells in the respective gates.

Analysis of thymic and lymph node cells from LATY136F mice older than 4 weeks showed high levels of eosinophils (fig. S3). LAT transcripts were undetectable in these eosinophils, suggesting that the observed eosinophilia resulted from the production of IL-5 by the abnormal CD4 cells present in these mutant mice. Most of the CD4 thymocytes found in LATY136F mice older than 4 weeks had a phenotype (CD44hi, CD62Llo, CD69+, and HSA) distinct from that expected for genuine CD4 SP thymocytes, but closely resembled that of the abnormal peripheral CD4 cells. Provided that the latter cells effectively recirculated to the thymus, the IL-5 and IL-13 cytokines they produced in situ are likely to be primarily responsible for the progressive disappearance of DP thymocytes and for the thymic eosinophilia (14,15).

Secondary lymphoid organs of 6-week-old LATY136F mice contained 7 to 10 times as many B cells as their wild-type counterparts. Most of the mature B cells found in 6-week-old wild-type littermates had a resting phenotype (B220high, MHC class II+, IgM+, IgD+; Fig. 4A). In contrast, among the B cells found in the enlarged secondary lymphoid organs of age-matched LATY136F littermates, 25% showed a hyperactivated phenotype (B220high, MHC class II high, IgD; Fig. 4A), 50% expressed a phenotype typical of antibody-producing cells (B220low, MHC class II+, IgD, IgM, CD44high), and only 25% had a resting phenotype. Serum IgG1 and IgE concentrations were elevated about 200 times and up to 10,000 times, respectively, compared with wild-type mice (Fig. 4B). In contrast, levels of other immunoglobulin isotypes did not differ from those of wild-type serum. Increased concentrations of κ and λ light chains in the serum of LATY136F mice (Fig. 4B), also indicated polyclonal hypergammaglobulinemia E and G1. IgE and IgG1 antibody concentrations peaked at 5 weeks of age (Fig. 4 C), the values of which exceeded those reported for mice deprived of NFATc2 and NFATc3 transcription factors (16). Given that B cells do not express LAT proteins (1) and considering that isotype switching to IgE and IgG1 depends on IL-4 and IL-13 (14), overproduction of IgE and IgG1 noted in LATY136F mice is probably secondary to the presence of an abnormally high frequency of TH2 effectors.

Figure 4

Hyperactivated B lymphocytes and serum levels of IgG1 and IgE antibodies in unimmunized LATY136Fmice. (A) Dot plots show the B220 versus MHC class II profiles of wild-type and LATY136F B cells. Boxes define B220low and B220high B cell populations. Data are representative of four mice aged 6 to 7 weeks. (B) Serum samples from 6-week-old wild-type and LATY136F mice were subjected to a serial dilution three times, and the titers of the specified immunoglobulin isotypes were determined by ELISA. Data are representative of six mice. (C) Early appearance of IgG1 and IgE in the serum of LATY136F mice. The concentrations of IgG1 and IgE found in individual mice were plotted on a logarithmic scale.

To establish that CD4 cells were responsible for the disorders documented in LATY136F mice, we first used a mutation that prevents the development of all mature T cells [CD3-ɛ Δ5 (17)]. Introduction of the LAT Y136F mutation had no effect on the B cells and eosinophils present in CD3-ɛ Δ5 mice (fig. S5). Moreover, the DN cells found in CD3-ɛ Δ5 × LATY136F mice were identical to those of CD3-ɛ Δ5 mice, indicating that the effects of the LAT Y136F mutation become manifest only at the DN to DP transition. By breeding the LATY136Fmice to β2-microglobulin–deficient mice, we observed that the residual CD8 T cells found in LATY136F mice were dispensable for the development of the disorders (8). In contrast, in mice deficient for both MHC class I and class II molecules (MHC KO), the absence of CD4 T cells protected them from the pathological effects of the LAT Y136F mutation, and their small complement of DP cells remained stable over time (fig. S5). Because noxious CD4 T cells cannot develop in MHC KO mice and interfere with developing thymocytes, this established that the LAT Y136F mutation has a direct impact on the DN to DP transition and that the CD4 T cells were responsible for the erosion of the DP cell compartment observed in LATY136F mice.

The absence of CD4 T cells in LATY136F × MHC KO mice suggests that their development requires a selective process involving MHC class II molecules. Given that the LATY136F CD4 T cells do not respond to TCR cross-linking, we converted them into T cell hybridomas. All the hybridomas derived from LATY136F CD4 T cells reacted with syngeneic MHC class II molecules (table S1), whereas none of those derived from wild-type CD4 T cells showed autoreactivity (8). These two series of hybridomas expressed comparable levels of TCR at their surface and used a heterogeneous set of Vα and Vβ segments (8). This indicates that the pathogenic CD4 T cells that develop in LATY136F mice do not correspond to CD1-d–restricted T cells (18). Their unexpected reactivity against self-MHC class II molecules may result from inadequate negative selection and may account for the exaggerated help they provide to B cells in vivo. Alternatively, their TCR may have been appropriately calibrated in relation to the context of LATY136F mice, and it is only after introducing them in T hybridomas and artificially increasing both their surface density and output that they started displaying reactivity toward self-MHC class II molecules. Accordingly, the polyclonal B cell activation found in LATY136F mice may result from noncognate interactions with aberrant T cells overproducing IL-4 in a chronic manner (19,20).

Although LATY136F mice showed a markedly impeded sequence of T cell development, an early and massive accumulation of TH2 effectors occurred in their periphery. This paradoxical phenotype can be accounted for if residue Y136 activates both a positive and a negative feedback loop. The positive function of residue Y136 may dominate during T cell development, whereas in CD4 T cells, Y136 may dominantly turn off signaling pathways that are activated at the time of positive selection or that keep in check the low-level of TCR signaling that CD4 T cells require for their survival in the periphery (21). Once this negative feedback function is relieved, a chronic and low-intensity signaling probably takes place and ultimately leads, for reasons that have yet to be defined, to the selective differentiation of TH2 effector cells. This dual role is consistent with the pattern of CD5 expression found in LATY136F mice. Its lack on the few DP cells that develop in LATY136F mice probably allows the pre-TCR to adapt to the lowered signaling potential of the LATY136F molecules. Conversely, the high levels of CD5 found on CD4 T cells probably correspond to a failed attempt to desensitize persistent TCR signals. Mutation Y136F does not affect the partition of LAT into glycolipid-enriched microdomains (fig. S6). Therefore, the negative role postulated for the LAT Y136 residue can be reconciled with the recessive nature of the LAT Y136 mutation, provided that in heterozygous mice the aberrant signals delivered by LAT Y136F molecules are blunted by signals from the wild-type LAT molecules that colocalize to glycolipid-enriched microdomains. The LATY136F mice differ from other mouse strains with TH2-based diseases (22) or T cell lymphoproliferative disorders (23), but strikingly resemble mice deprived of NFATc2 and NFATc3 transcription factors (15). Therefore, by unleashing signal propagation at two distinct levels of the TCR transduction cassette, the LAT Y136F and NFATc2/c3 mutations presumably triggered a runaway feedback pathway that resulted in similar effects on TH2 polarization. Although the underlying mechanisms by which the LAT Y136F mutation impacts on T cell development and differentiation remain to be defined, our results reveal that restricting LAT to a subset of its docking functions critically affects the differentiation of TH cells. It is therefore possible that even in a physiological context, the differential phosphorylation of some of the tyrosines found in LAT also influences TH cell differentiation.

  • * On leave from Institut Méditerranéen de Recherche en Nutrition, UMR-INRA, Marseille, France.

  • To whom correspondence should be addressed. E-mail: bernardm{at}


Stay Connected to Science

Navigate This Article