The Related Adaptors, Adaptor in Lymphocytes of Unknown Function X and Rlk/Itk-Binding Protein, Have Nonredundant Functions in Lymphocytes¹

Lymphocyte activation and development depend on the transmission of intracellular signals downstream of receptor engagement by adaptors (1). Adaptor molecules lack intrinsic enzymatic activity and function by bringing together two or more proteins in a complex. Adaptors can be grouped by sequence homology into families, and related adaptors may have redundant or divergent functions. In its initial description, we noted the similarity between the adaptor in lymphocytes of unknown function X (ALX;³ also known as HSH2) and the adaptor Rlk/Itk-binding protein (RIBP; also known as TSAd, VRAP, or Lad) (1, 2, 3, 4, 5, 6, 7). Each molecule contains a single, N-terminal Src homology 2 (SH2) domain, with several polyproline regions. The two proteins exhibit 36% overall homology, and 57% when only the SH2 domains are compared (6). SH2 domain mutations in either RIBP or ALX abrogate function (8, 9). Analysis of RIBP-deficient mice demonstrated that RIBP is a positive regulator of peripheral T cell activation. A moderate decrease in proliferation and IL-2 production upon TCR or TCR/CD28 stimulation was observed in RIBP-deficient mice (3). Given their similarity, we originally hypothesized that ALX and RIBP may have redundant functions during T cell development or activation. However, when ALX-deficient mice were analyzed, we demonstrated that ALX is a negative regulator of T cell activation (10). ALX-deficient T cells exhibited enhanced IL-2 production, CD25 expression, and proliferation in response to TCR/CD28 stimulation. Although this analysis suggested that ALX and RIBP have opposing functions in T cell activation, it remained possible that that functional redundancy existed that would only be revealed when double-deficient mice were examined.

Because cells may coexpress multiple related adaptors, the generation and examination of double-knockout mice are required to clarify the relationship between them. For example, mice deficient in signal transducing adaptor molecule 1 and signal transducing adaptor molecule 2 exhibited significant defects in T cell development and peripheral T cell numbers, although no effect was observed in mice deficient in only one of these adaptors (11). Double-deficient mice may also reveal new functions for adaptors, which were not apparent from initial characterization of the individual knockout mice. For example, the transmembrane adaptors linker for activated T cells (LAT) and linker for activated B cells (LAB) (also known as NTAL) are coexpressed in mast cells, and regulate signaling downstream of FcεRI (12). Mast cells deficient in LAT are impaired in their ability to activate downstream signaling pathways such as calcium mobilization and Erk activation in response to FcεRI cross-linking (13), whereas mice deficient in LAB are hyperresponsive to FcεRI engagement (14, 15). However, when mast cells doubly deficient in LAT and LAB were analyzed, the impairment of FcεRI activation was more severe than mast cells deficient in LAT alone, demonstrating that LAB also plays a positive role in regulating FcεRI signaling, which was only uncovered in the context of LAT deficiency (14, 15). Therefore, additional functions of related adaptors, as well as potential functional redundancy, might only be identified by generating and examining double-deficient mice.

In this study, we extend our prior studies to analyze the function of RIBP and ALX in lymphocytes, as well as to determine whether ALX and RIBP are functionally redundant during lymphocyte activation or development. However, no synergy was observed in lymphocyte development or activation in the ALX/RIBP-deficient mice. Therefore, despite similarities in sequence and structure that would predict similarities in function, ALX and RIBP function independently in lymphocytes.

RIBP-deficient mice were the gift from J. Bluestone (University of California, San Francisco, CA). PCR genotyping of tail DNA was used to identify the genotype of progeny, as previously described (3, 10). Mice were housed in a specific pathogen-free facility at the University of Pennsylvania and were used in accordance with the regulations of the University’s Institutional Animal Care and Use Committee. These experiments were performed with mice that had been backcrossed either three generations (RIBP deficient, ALX/RIBP deficient) or five generations (wild type and ALX deficient) to C57BL/6, except that H-Y TCR transgenic animals were backcrossed at least seven times to C57BL/6.

Single-cell suspensions were generated, as previously described (10). The following FACS reagents were used: B220-allophycocyanin-Cy7 (eBioscience), AA4.1-allophycocyanin (eBioscience), CD43-FITC (BD Pharmingen), IgM-biotin (Southern Biotechnology Associates), streptavidin-PerCP-Cy5.5 (BD Pharmingen), CD4-PE (Caltag Laboratories), CD8-allophycocyanin (Caltag Laboratories), CD4-biotin (BD Pharmingen), CD25-allophycocyanin (BD Pharmingen), anti-HY TCR T3.70 (eBioscience), and CD69-FITC (BD Pharmingen). Samples were analyzed on a FACSCalibur (BD Biosciences) or on a LSR (BD Biosciences). Data were analyzed using FlowJo (Tree Star).

T cells were isolated by negative selection from whole splenocytes using the SpinSep murine T cell isolation kit, per the manufacturer’s instructions (catalogue 17051; StemCell Technologies). In brief, splenocytes were incubated with a mixture of Abs to label non-T cells. Cells were then incubated with dense particles that bound to the Ab-labeled cells. The mixture was spun over density medium, after which the undesired cells pelleted and the T cells remained at the interface and were collected. This preparation resulted in ∼95% pure T cells. Similarly, purified B cells were isolated using either SpinSep murine B cell isolation kit (catalogue 17034; StemCell Technologies) or MACS B cell isolation kit (catalogue 130-090-862; Miltenyi Biotec), per the manufacturer’s instructions.

Purified B cells were CFSE labeled and incubated for 3 days with the following stimuli: F(ab′)₂ goat anti-mouse IgM (Jackson ImmunoResearch Laboratories; final concentration 10 μg/ml), alone or with anti-CD40 (BD Pharmingen; clone HM40-3, 10 μg/ml), B lymphocyte stimulator (BLyS; 200 ng/ml; PeproTech), LPS (InvivoGen; K12; 10 μg/ml), or CpG (InvivoGen; ODN-1826 with a phosphothiolate backbone; 1 mM).

Single-cell splenocyte suspensions were plated at a concentration of 1 × 10⁶ cells/ml in 96-well plates precoated with anti-CD3ε (clone 145-2C11; BioLegend) alone or with soluble anti-CD28 added at the time of plating (clone 37.51; BioLegend). Supernatants were collected after 48 h, and concentrations of IL-2 in the supernatants were determined by ELISA using a Duoset kit, according to the manufacturer’s instructions (DY402; R&D Systems). The data were normalized to the amount of IL-2 produced by ALX-deficient mice stimulated with 5 μg/ml plate-bound CD3 and 1 μg/ml CD28. Data are expressed as the average of four independent experiments with SE from the mean.

Serum Ig levels from wild-type and ALX-deficient mice were analyzed by ELISA using the SBA Clonotyping System/HRP (5300-05) from Southern Biotechnology Associates, according to the manufacturer’s instructions.

Single-cell suspensions of purified T or B cells were resuspended in 5 ml of PBS, to which was added 5 ml of PBS containing 0.6 mM CFSE (Molecular Probes). Tubes were inverted for 2–5 min and subsequently quenched with 4 ml of FBS. The labeled lymphocytes were cultured in RMPI 1640 medium containing 10% FBS, l-glutamine, and Pen/Strep (all from Invitrogen Life Technologies), and 55 μM 2-ME for 3 days before analysis, either unstimulated or stimulated, as described in the figure legend. To exclude dead cells, a final concentration of 100 nM TOPRO-3 (Molecular Probes) was added to samples 10 min before FACS analysis. A fixed number of 6-μm polystyrene microspheres (Polysciences) was added to each sample, and a known fraction of the microspheres was collected by FACS (along with a varying number of cells), permitting the calculation of absolute cell numbers recovered after stimulation.

Wild-type and ALX-deficient littermates were injected i.p. with 50 μg of SEB (Sigma-Aldrich) in 200 μl of PBS. Mice were bled via retro-orbital puncture on days 0, 2, 4, 8, and 11. Blood lymphocytes were stained with FITC anti-Vβ6 (BD Pharmingen), PE anti-Vβ8 (BD Pharmingen), and allophycocyanin anti-CD4 (Caltag Laboratories) to track the percentage of CD4⁺Vβ8⁺ cells over time. CD4⁺Vβ6⁺ cells, which do not respond to SEB, were also examined as a control.

A total of 100–200 μl of blood was obtained via retro-orbital puncture and was added to an equal volume of heparin solution (20 U/ml in PBS; Sigma-Aldrich). Blood was lysed in 10 ml of ACK lysis buffer (BioSource International) on ice for 10 min and then quenched with 10 ml of medium. Cells were spun at 1200 rpm, 4°C, for 8 min, resuspended in 5 ml of PBS, and filtered through mesh. After spinning, cells were resuspended in FACS buffer.

Splenocytes from adult mice of the four genotypes were enriched for CD4⁺ T cells using CD4⁺ negative selection MACS. This resulted in a residual CD8⁺ T cell contamination of <1% at the start of the culture. Cells were cultured for 7 days, as follows: Th1 polarizing (stimulation with anti-CD3, anti-CD28, IL-2, IL-12, and anti-IL-4 neutralizing Ab) or Th2 polarizing (stimulation with anti-CD3, anti-CD28, IL-2, IL-4, and neutralizing Abs against IL-12 and IFN-γ). At day 7, cells were washed and restimulated for 4 h with plate-bound anti-CD3, and brefeldin A was added for the last 2 h of the 4-h period. Cells were harvested and stained for CD4, and were then fixed, permeabilized, stained for intracellular IL-4 and IFN-γ, and analyzed by flow cytometry. Alternatively, at day 7, cells were restimulated for 24 h with plate-bound anti-CD3, and the tissue culture supernatants were examined by ELISA for IFN-γ (BioLegend) or IL-4 (R&D Systems).

Splenocytes were prepared and stimulated, as described previously (10). Lysates were analyzed by electrophoresis and Western blotting using anti-phospho-p38 (Cell Signaling Technology; no. 9216) and anti-p38 (Cell Signaling Technology; no. 9212).

Progression through checkpoints in thymocyte development requires signals of appropriate intensity, first via the pre-TCR and then via the TCR. Although both ALX and RIBP are highly expressed in the thymus, deficiency in either gene alone did not alter T cell development. One explanation for the failure of either ALX or RIBP deficiency to alter T cell development may be that these molecules are not expressed during T cell development until after both positive and negative selection has occurred. Initial work demonstrated that ALX and RIBP are expressed in thymocytes by Northern analysis, without examination of thymic subsets. To examine the expression of these adaptors during T cell development, we performed quantitative real-time PCR (QPCR) on sorted thymocyte subsets (Fig. 1 A), from early T lineage progenitors (ETP) to CD4/CD8 single positive. ETP, double-negative (DN)2, DN3, and DN4 subsets were sorted based on the lack of expression of a mixture of lineage markers, and expression of c-kit⁺CD25⁻, c-kit⁺CD25⁺, c-kit⁻CD25⁺, and c-kit⁻CD25⁻, respectively. The expression of ALX is highest in the earliest thymic T cell ETP population, and decreases with continued maturation until the double-positive stage, before it increases in CD4 and CD8 single-positive (SP) cells. In contrast, the levels of RIBP do not change appreciably throughout T cell development. Therefore, ALX and RIBP are expressed throughout T cell development and could potentially modulate pre-TCR or TCR signaling.

To determine whether there was functional redundancy between these two related adaptors in thymic development, ALX/RIBP-deficient mice were generated. ALX/RIBP-deficient mice are viable, fertile, and are born at the expected Mendelian ratios. ALX/RIBP-deficient thymus and spleen were analyzed by FACS for various T cell populations, and compared with either wild-type mice, or mice deficient in either ALX or RIBP (Fig. 1,B). No significant differences were observed in the double-negative, double-positive, or CD4 or CD8 SP thymocyte populations, in either percentages (Fig. 1,B) or absolute numbers (data not shown). Additionally, both CD4⁺ and CD8⁺ populations were found in the expected numbers and proportions in peripheral lymphoid organs such as the spleen (Fig. 1 B) in ALX/RIBP-deficient mice. Therefore, T cell development and homeostasis do not require either ALX or RIBP. However, it is possible that effects of ALX and RIBP deficiency on T cell development may not have been evident within the context of a normal TCR repertoire. Consequently, ALX-, RIBP-, and ALX/RIBP-deficient mice were crossed onto the H-Y TCR transgenic mice to test this explicitly.

Alterations in T cell development in either CD5- or Cbl-deficient mice were revealed only when the T cell repertoire was fixed by the use of a TCR transgene (16, 17). Therefore, we interbred mice of each genotype to the H-Y TCR transgenic line, a well-characterized model used to examine both positive and negative selection (18). One caveat with TCR transgenic models is that the TCR transgene may be expressed earlier in development than TCR chains generated through endogenous rearrangement (19), and thus, before the gene being examined normally functions during T cell development. However, both ALX and RIBP are expressed throughout T cell development starting at the ETP population. Therefore, early expression of TCR transgenes, including H-Y, during development will not preclude analysis of potentially altered selection within ALX-, RIBP-, and ALX/RIBP-deficient mice. As shown in Fig. 1,C, all populations were first gated on the H-Y TCR transgene using the clonotypic-specific mAb T3.70 (data not shown). The H-Y transgenic TCR recognizes a male Ag, resulting in a strong negative selection signal in male mice and the deletion of most thymocytes before the double-positive stage. Efficient negative selection of the H-Y TCR transgene occurred similarly between wild-type, ALX-, RIBP-, and double-deficient mice, resulting in failure to produce CD8 SP H-Y-expressing T cells in the thymus. We have noted in the RIBP-deficient and ALX/RIBP-deficient, and to a lesser extent, in the ALX-deficient mice, what appears to be a second population that is CD4^lowCD8^low as compared with the wild-type H-Y TCR transgenic T cells. The significance of this population is unclear. It does not appreciably alter thymocyte number, nor the failure to differentiate into CD8⁺ H-Y-expressing thymocytes or T cells in the periphery. Therefore, we conclude that negative selection is essentially unaffected in ALX-deficient, RIBP-deficient, and ALX/RIBP-deficient mice. In female mice, the H-Y TCR transgene is efficiently positively selected to produce large numbers of CD8 SP T cells, as shown in Fig. 1,C. No significant differences in positive selection were observed between wild-type, ALX-, RIBP-, and double-deficient mice, either in proportion (Fig. 1,C) or absolute number of H-Y-expressing CD8 SP T cells (Fig. 1 D). Therefore, T cell development, including positive and negative selection, occurs normally in the absence of both ALX and RIBP.

Similarly, one explanation for the lack of a phenotype during B cell development in ALX-deficient mice may be a lack of expression. To examine this, cells from various stages of B cell development were isolated from wild-type mice. From the bone marrow, the pro-B + pre-B cell fraction (B220⁺AA4.1⁺IgM⁻), and immature cells (B220⁺AA4.1⁺CD43⁻IgM⁺) were isolated, as well as transitional T1 (B220⁺AA4.1⁺IgM^highCD23⁻), T2 (B220⁺AA4.1⁺IgM^highCD23⁺), and mature B cells (B220⁺AA4.1⁻) from the spleen. Expression of ALX was examined by QPCR (Fig. 2 A). ALX is expressed early in B cell development, but increases sharply as B cells develop into immature and subsequently into T1 transitional B cells. Previous studies had not examined RIBP expression in B cells. Examining these same samples for RIBP expression by QPCR revealed that RIBP is coexpressed with ALX in B cells. Although ALX expression is up-regulated in immature B cells, RIBP expression increases significantly at the transition into mature B cells. Therefore, RIBP may also have a role in B cells.

To determine whether RIBP deficiency altered B cell development or homeostasis, different B cell populations were analyzed via flow cytometry in wild-type, ALX-, RIBP-, and double-deficient mice. In the bone marrow, no differences in either the absolute number (data not shown) or proportions of pro-B cells (B220⁺AA4.1⁺IgM⁻CD43⁺), pre-B cells (B220⁺AA4.1⁺IgM⁻CD43⁻), or immature B cells (B220⁺AA4.1⁺IgM⁺CD43⁻) were observed (Fig. 2,B). Similarly, in the spleen, neither were differences observed in absolute number (data not shown) or proportion of T1 transitional B cells (B220⁺AA4.1⁺IgM^highIgD^low) or T2/T3 (B220⁺AA4.1⁺IgM⁺IgD^high) as well as mature follicular B cells (B220⁺AA4.1⁻IgM⁺CD21/35⁺) or marginal zone B cells (B220⁺AA4.1⁻IgM^highCD21/35^high), as shown in Fig. 2,B. In addition, examination of serum Ig concentration in a cohort of 4-mo mice did not reveal any differences between wild-type, ALX-, RIBP-, or ALX/RIBP-deficient mice (Fig. 2 C). Therefore, neither ALX nor RIBP is essential for B cell development, homeostasis, or class-switching.

It has been reported that ALX expression is up-regulated in B cells in response to activation/survival signals, including anti-CD40, LPS, CpG, or BLyS/B cell-activating factor, whereas expression was reduced upon addition of IL-21, which induces apoptosis (20). This implied ALX may participate in B cell survival. In fact, ALX overexpression in the WEHI-231 cell line inhibited BCR-mediated apoptosis (21). However, when survival and proliferation of purified, CFSE-labeled B cells in response to anti-IgM, anti-CD40, BLyS/B cell-activating factor, LPS, or CpG were examined, no differences were observed between wild-type and ALX-deficient B cells (Fig. 3) (10). One possible explanation of these results is that the presence of RIBP in B cells may compensate for the lack of ALX. However, no differences in B cell responses were observed in RIBP- or ALX/RIBP-deficient B cells (Fig. 3). Therefore, it appears that ALX and RIBP are dispensable for the survival and proliferation of B cells in vitro to a variety of stimuli.

Peripheral T cell responses are altered in both ALX-deficient and RIBP-deficient mice. We found that the absence of ALX resulted in augmented IL-2 production in vitro in response to stimulation with anti-CD3 and anti-CD28 (10). In contrast, RIBP-deficient mice were found to have a mild defect in IL-2 production (3, 22, 23). T cell function in ALX/RIBP-deficient mice was examined to clarify the relationship between these two proteins. One possibility was that the T cell responses in the doubly-deficient mice would lie between the responses of the singly-deficient mice. Alternatively, if ALX and RIBP function at similar points, or if one is downstream of the other on the same pathway, then the ALX/RIBP-deficient T cells would mirror the phenotype of either ALX- or RIBP-deficient T cells. Purified T cells were stimulated with anti-CD3, in the absence or presence of anti-CD28 (Fig. 4). As expected, purified T cells in the absence of costimulation did not produce appreciable amounts of IL-2. Under suboptimal conditions, ALX-deficient T cells exhibit increased IL-2 production compared with wild type, as previously reported (10). However, with increasing stimulation, the difference between ALX-deficient and wild type is decreased. Therefore, ALX deficiency appears to alter the responsiveness of T cells, enhancing IL-2 production when stimulation is limiting. In contrast, the diminished IL-2 production resulting from RIBP deficiency was most apparent under conditions when wild-type T cells responded robustly. Under conditions in which the greatest effect of RIBP deficiency was observed (10 μg/ml CD3 and 0.3 μg/ml CD28), deficiency in ALX completely masked the effect of RIBP deficiency. In fact, under no conditions did the doubly-deficient T cells respond differently than the ALX-deficient T cells, implying that ALX and RIBP may function separately. Similar to that previously described for ALX-deficient T cells, ALX/RIBP-deficient T cells had enhanced CD25 expression as compared with wild type (Fig. 4,B). In addition, when proliferation was measured by CFSE dilution, ALX/RIBP-deficient T cells also had enhanced proliferation, resulting in a greater proportion of cells proliferating, as compared with wild-type T cells, similar to that observed in ALX-deficient T cells (10) (Fig. 4,C). Previously, we analyzed the signaling pathways downstream of TCR/CD28 to identify the molecular basis for the hyperresponsiveness of ALX-deficient T cells, and found that ALX deficiency leads to constitutive p38 MAPK activation, whereas other signaling pathways are unaffected (10). Because T cells deficient in both ALX and RIBP have increased IL-2 production, CD25 expression, and proliferative responses in response to TCR/CD28 stimulation, we thought it possible that p38 MAPK would also be constitutively activated in ALX/RIBP-deficient mice. Constitutive p38 MAPK activation was observed in ALX/RIBP-deficient splenocytes with no further activation in response to either anti-TCR/CD28 or PMA stimulation (Fig. 4 D), as was observed in our previous analysis of ALX-deficient mice (10). Therefore, in examining T cell responses in vitro, no synergistic effects of ALX and RIBP deficiency were observed.

Having found evidence of altered T cell activation in vitro, in vivo activation-induced cell death was examined, using the SEB-deletion model. SEB provides a potent activating signal to CD4⁺ T cells bearing Vβ8⁺ TCR subunits, but does not affect cells bearing other Vβ subunits such as Vβ6, which was examined as a negative control. To examine SEB-induced deletion, mice of all four genotypes were immunized with SEB. As expected, CD4⁺Vβ8⁺ T cells disappear from blood as a result of SEB-induced deletion within days postinjection (Fig. 5), whereas CD4⁺Vβ6⁺ T cells remain unaffected. In these experiments, we failed to observe a significant difference in the percentage of Vβ8⁺ cells among CD4⁺ T cells at any time point in either ALX-, RIBP-, or ALX/RIBP-deficient mice. Therefore, SEB-induced deletion does not depend on either ALX or RIBP.

Because previous reports demonstrated that T cells from RIBP-deficient mice had a mild decrease in the production of IFN-γ after 24 or 48 h, it seemed possible that the differentiation of naive T cells into Th1 or Th2 effector cells would be compromised in the absence of RIBP and/or ALX. To investigate this possibility, we cultured CD8-depleted splenocytes under Th1- or Th2-polarizing conditions for 7 days. Polarized cells were restimulated with anti-CD3, and their ability to produce IFN-γ or IL-4 was examined by intracellular FACS. As shown in Fig. 6,A, equivalent differentiation into Th1 and Th2 effector cells in wild-type, ALX-, RIBP-, and ALX/RIBP-deficient T cells was observed. Similar results were obtained when the polarized cultures were restimulated with PMA/ionomycin (data not shown). In addition, IFN-γ and IL-4 secretion from polarized T cells in response to anti-CD3 restimulation was equivalent between all genotypes (Fig. 6 B). Therefore, T cell differentiation into Th1 or Th2 effector cells is unaffected by deficiency in ALX and/or RIBP.

Adaptors related by structure and sequence homology may also share function, which by redundancy could mask effects of deficiency in any one family member. ALX and RIBP are related by structure and have 36% overall sequence homology. In addition, a comparative analysis of all human SH2 domains demonstrated that those in ALX and RIBP are more closely related to each other than to all other SH2 domains (24). Initial characterization of ALX demonstrated that overexpression in Jurkat T cells inhibited activation of the IL-2 promoter (6), similar to what had been previously reported for RIBP/TSAd overexpression (25). Analyses of mice deficient in either ALX or RIBP, however, demonstrated that they were negative and positive regulators of T cell activation, respectively (3, 10, 23). Although ALX and RIBP were initially described to have different, opposing roles in regulating T cell activation, this did not preclude the possibility that there were areas of functional redundancy between these two adaptors, as was uncovered in the analysis of LAT/LAB-deficient mast cells (14, 15). However, in our analysis of T and B cell development, activation, and effector functions, we have to date not observed a synergistic phenotype in the ALX/RIBP-deficient mice, demonstrating that they function independently.

As part of our initial characterization of ALX-deficient mice, we examined the role of ALX in B cell development and activation. Work from Herrin and Justement (20) had demonstrated that ALX expression is up-regulated in B cells upon activation in response to anti-apoptotic stimuli. In addition, overexpression of ALX in the WEHI231 B cell line blocked induction of apoptosis in response to BCR cross-linking (21). However, no defects in B cell development or activation were observed in ALX-deficient mice. We considered that RIBP may play a role in B cells as well. Although initially reported to be expressed only in activated T cells and NK cells, Spurkland et al. (4) demonstrated strong expression of RIBP in PBLs by Northern analysis, of which activated T and NK cells comprise only a minor population. Therefore, it was highly likely that RIBP was expressed in additional cell types. In this study, we demonstrate that RIBP is coexpressed with ALX in B cells; however, neither ALX nor RIBP appears to function in B cell development, or B cell responses to a variety of stimuli. Consequently, the function of these adaptors in B cells remains to be understood, and they may perhaps function downstream of receptors other than those examined.

We have extended the analysis of RIBP and ALX in T cell development and selection beyond that reported previously. By QPCR, we demonstrate that both ALX and RIBP are expressed in all developing thymocyte populations, from ETP to SP. Similar to results from mice deficient in each adaptor, mice deficient in both ALX and RIBP did not have any alterations in T cell development. Using the H-Y TCR transgenic model, we demonstrate that there are no alterations in either positive or negative selection in either ALX-deficient, RIBP-deficient, or ALX/RIBP-deficient mice. This is similar to recently published results examining H-Y TCR transgenic, RIBP-deficient mice (26). In addition, we examined the effects of ALX and/or RIBP deficiency on Th cell differentiation, but found that Th1 and Th2 development under polarizing conditions occurred normally in the absence of ALX and RIBP.

For reasons that remain unclear, our results differ with regard to the role of RIBP in SEB-induced deletion (23). In our hands, depletion of CD4⁺Vβ8⁺ cells occurred similarly between wild-type, RIBP-deficient, and ALX/RIBP-deficient T cells. Although these differences could be due to the degree of backcrossing, this is unlikely because CD4⁺Vβ8⁺ depletion was found to be the same in mice backcrossed 3 generations (Fig. 5) or 10 generations (data not shown). These differences may also be due to the strength of superantigen stimulation, because at 4 days postinjection we observe a considerable increase in the number of CD4⁺Vβ8⁺-expressing cells in response to superantigen stimulation before deletion, whereas this was not observed previously at the same time point (23). In fact, the strength of stimuli used had significant effects on the ability of ALX and RIBP to regulate T cell activation (Fig. 4,A). Using a range of TCR/CD28 stimulation conditions, we have shown that ALX and RIBP deficiency alters T cell responsiveness. In the absence of ALX, T cells respond more readily, whereas in the absence of RIBP, T cell responses were muted. Under very weak or very strong stimulation conditions, effects due to absence of RIBP or ALX were minimal, implying that ALX and RIBP modulate T cell responsiveness under limiting conditions. Interestingly, the response of ALX/RIBP-deficient T cells does not lie between that of T cells deficient in either ALX or RIBP, but instead recapitulates the ALX-deficient phenotype. Because deficiency in ALX can overcome any effect of deficiency in RIBP, ALX most likely functions at a point downstream of RIBP. Current understanding of the biochemical pathways regulated by these adaptors fits with this interpretation, because RIBP deficiency has an effect on the earliest events in T cell activation, whereas these events are intact in ALX-deficient mice that exhibit constitutive p38 MAPK activation (10, 22). Consistent with this, constitutive p38 MAPK activation is also observed in ALX/RIBP-deficient splenocytes (Fig. 4 D). Therefore, despite considerable sequence identity and structural similarity that would predict a functional overlap, ALX and RIBP function independently in lymphocytes.

We thank Jeffrey Bluestone (University of California, San Francisco, CA) for sharing his RIBP-deficient mice with us for these studies. We thank Marielena Velez and Avinash Bhandoola for assistance in FACS sorting different thymic populations, as well as Will Quinn III and Michael Cancro for assistance in FACS sorting various B cell developmental stages. We thank Gary Koretzky and members of his laboratory for thoughtful discussions, and Jon Maltzman for careful reading of the manuscript. C.E.P. also thanks C. D. Perchonock for his support.

The authors have no financial conflict of interest.