Disruption of Lymphocyte Function and Signaling in CD45–associated Protein–null Mice

A targeting DNA construct was prepared by ligating two segments, of 3 kb and 1.6 kb, of the genomic clone (10) and an HSV-tk gene cassette (obtained from H. Westphal, National Institute of Child Health and Human Development, Bethesda, MD) into pMC1neo poly(A) vector (Stratagene, La Jolla, CA). The 625-bp KpnI-NsiI genomic fragment that contains all the coding region of the unspliced form of CD45-AP (9) and all but the first five amino acid sequences of the spliced form of CD45-AP (16) was replaced by the neo (neomycin resistance) gene cassette. RW4 embryonic stem (ES) cells (Genome Systems, St. Louis, MO) were transfected by electroporation with the targeting construct linearized by SalI, and colonies resistant to G418 and gancyclovir were screened for homologous recombination by genomic Southern hybridization with a probe that is immediately 3′ to the 1.6-kb genomic sequence. NsiI digestion of genomic DNA obtained from normal mice or from ES cells without homologous recombination results in a 6.1-kb fragment on Southern hybridization. 4 ES colonies out of 216 exhibited an altered Southern hybridization pattern as a result of homologous recombination. These clones were injected into blastocysts of C57BL/6 mice and chimeric mice were produced. Two out of five male chimeras had germline transmission of the mutant CD45-AP allele and mice heterozygous for the CD45-AP deletion were identified among their offspring. Homozygous mice were then bred by crossing the heterozygotes and identified by Southern hybridization of genomic DNA.

CD45 was isolated from equal numbers of cells obtained from thymus or spleen of 7–9-wk-old CD45-AP–null mice or age-matched wild-type counterparts by immunoprecipitation. CD45 was eluted by briefly subjecting the immunoprecipitates to high pH in 50 mM of diethanolamine (8). A portion of the CD45 preparation was used for PTP activity assay with Raytide (Oncogene Research Products, Cambridge, MA) as a substrate and another portion was analyzed by immunoblotting with anti-CD45 antibody (provided by J. Marth, University of California San Diego, La Jolla, CA). The substrate for PTP assay was prepared by labeling Raytide with γ-[³²P]ATP and Src PTK (Oncogene Research Products). PTP activity was determined by measuring the release of ³²PO₄ from the labeled Raytide as previously described (8). The amount of CD45 in the samples was determined by densitometric analysis of the immunoblots.

Splenic T and B cells of 7–9-wk-old CD45-AP–null mice or wild-type littermates were purified by a combination of one round of nylon-wool fiber column followed by negative selection with antibody-mediated affinity columns (Biotex Laboratories, Edmonton, Alberta, Canada). Typical T and B cell populations thus obtained were 90 and 95% pure, respectively, by flow cytometric analysis of surface markers. T cells were stimulated by plate-bound (at 5 μg/ml) anti–TCR-α/β Ab (PharMingen, San Diego, CA; reference 17), Con A (2 μg/ml), or a combination of PMA (10 ng/ml) and ionomycin (400 ng/ml). B cells were stimulated by F(ab′)₂ fragment of anti-IgM Ab (10 μg/ml; Jackson Immunoresearch Labs., West Grove, PA), anti-CD40 Ab (5 μg/ml; Serotec, Washington DC) (18), LPS (4 μg/ml), or a combination of PMA (10 ng/ml) and ionomycin (400 ng/ml). For some assays mitomycin C-treated spleen cells of wild-type mice were added to purified T cells as a source of accessory cells. Cell proliferation was determined after 72 h by measuring [³H]thymidine incorporation (pulsed at 1 μCi/well overnight) into cells incubated at 10⁵ cells per well in 200 μl DMEM with 10% FCS. For MLR, splenic T cells (MHC H-2^b haplotype) (4 × 10⁶ cells/ml) were cultured with mitomycin C–treated spleen cells (4 × 10⁶ cells/ml) derived from MHC-incompatible BALB/c (MHC H-2^d haplotype) or MHC-compatible C57BL/6 mice (MHC H-2^b haplotype) for 4 d before overnight [³H]thymidine pulse. CTLs were generated from spleen cells (MHC H-2^b haplotype) (5 × 10⁶ cells/ml) by coculturing with mitomycin C–treated spleen cells (3 × 10⁶ cells/ ml) derived from MHC-incompatible BALB/c mice (MHC H-2^d haplotype) in the presence of conditioned medium Rat T-STIM (Collaborative Biomedical Products, Bedford, MA) for 5 d. Nonadherent CTLs were then harvested and used as effector cells. Cells of H-2^d haplotype (T cell line S49.1) or H-2^b haplotype (spleen cells of C57BL/6 mice) were labeled with ⁵¹Cr and used as targets for CTL in a radioactive chromium release assay.

Splenic T cells from 7–9-wk-old CD45-AP–null mice or age-matched wild-type counterparts were purified by two successive rounds of nylon-wool fiber columns. The typical T cell population thus obtained was 85% pure by flow cytometric analysis of surface markers. T cells, 4 × 10⁷ cells/ml in IMDM, were incubated first with anti–TCR-α/β Ab (10 μg/ml; PharMingen; reference 17) on ice for 3 min followed by the addition of anti–hamster IgG (12.5 μg/ ml) and further incubation at 37°C for 0, 5, 10, or 60 min for stimulation. After two washes in cold IMDM, the cells were lysed at 5 × 10⁷ cells/ml in 0.8% polyoxyethylene 10 oleyl ether (BRIJ 96) containing 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM MgCl₂, 0.1 mM EGTA, 2.5 mM thioglycolic acid, 1 mM PMSF, 10 mM sodium fluoride, 1 mM sodium orthovanadate, and 10 mM pyrophosphate. Postnuclear supernatants of the lysates were first precleared with Protein G–Sepharose 4 beads (Pharmacia Biotech, Piscataway, NJ) and were then immunoprecipitated either with anti-CD45 Ab (19) cross-linked to Protein G–Sepharose 4 beads with dimethyl pimelimidate (20) or with unconjugated Protein G–Sepharose 4 beads. Material bound to the beads was analyzed by immunoblotting with Abs against CD45 (provided by J. Marth) and Lck (Zymed Labs., South San Francisco, CA) followed by horseradish peroxidase–conjugated Protein A or anti–mouse Ig and the Supersignal substrate system (Pierce, Rockford, IL).

There are two forms of CD45-AP mRNA with a minor difference at the NH₂ terminus as a result of splicing (9, 10, 16). The two resulting CD45-AP proteins are identical in their capacity for specific binding to CD45, but use different mechanisms for endoplasmic reticulum membrane translocation (10). The genomic DNA segment that encodes the entire coding region of the first form of CD45-AP and all but the first five amino acids of the second form was replaced with the neo gene cassette by homologous recombination in ES cells (Fig. 1,A). NsiI digestion of genomic DNA obtained from normal mice or from ES cells without homologous recombination results in a 6.1-kb fragment on Southern hybridization. ES colonies that exhibited an altered Southern hybridization pattern as a result of homologous recombination were injected into blastocysts of C57BL/6 mice. Mice heterozygous and homozygous for the CD45-AP deletion were identified by Southern blot analysis of genomic DNA (Fig. 1,B). No CD45-AP protein was detected in cell lysates of spleen cells and thymocytes obtained from the homozygous mutants by immunoblot analysis with antiserum against CD45-AP (11), confirming that these homozygous mice are indeed CD45-AP–null (Fig. 1 C).

Gross morphological and histological analysis of all organ systems of CD45-AP–null mice did not show any abnormality, nor did flow cytometric analysis of cells obtained from thymus, spleen, lymph nodes, bone marrow, and peritoneal cavity. Therefore, leukocytes appeared to differentiate normally in these mice unlike in CD45-deficient (14) or -null mice (15). However, the surface expression of CD45 measured by flow cytometry was reduced in the mutant mice compared with their wild-type littermates (Fig. 2,A). The average fluorescence intensities of thymocytes and spleen cells obtained from CD45-AP–null mice were 86.2% (± 6.7% SD; 4 experiments) and 73.9% (± 2.4% SD; 6 experiments), respectively, of wild-type littermates. Expression of all isotypes of CD45 appeared reduced since the average fluorescence intensities of CD45-AP–null spleen cells stained with Abs against CD45RA (clone 14.8, PharMingen), CD45RB (clone 16A, PharMingen), and CD45RC (clone DNL-1.9, PharMingen) were 71.8% (± 5.3% SD; 4 experiments), 75.6% (± 2.5% SD; 4 experiments), and 79.7% (± 14.6% SD; 5 experiments), respectively, of wild-type cells. Interestingly, a subset of thymocytes that expresses high levels of CD45, seen as a second peak in the thymic CD45 profile of +/+ mice (Fig. 2 A), was absent in CD45-AP–null mice. Thymocyte subpopulations known to express high levels of CD45 are mature CD4⁺ or CD8⁺ single-positive cells and large, cycling, positively selected CD4⁺CD8⁺ cells (21). Purified splenic T and B cells of CD45-AP–null mice exhibited equally reduced levels of CD45 expression, i.e., 73.5% (± 2.5% SD; 5 experiments) and 73.2% (± 1.8% SD; 4 experiments) average fluorescence intensities, respectively, of wild-type controls.

To determine whether the reduction in surface CD45 correlates with a reduction in total cellular CD45, the amount of total cellular CD45 in thymocytes and spleen cells was measured by immunoblotting of total cell lysates with anti-CD45 antisera (Fig. 2 B) followed by densitometric analysis. Total CD45 in thymus and spleen was reduced to 56.7% (± 12.5% SD; 5 experiments) and 69.8% (± 9.3% SD; 4 experiments), respectively, of wild-type counterparts. Thus, the amount of total CD45 in thymocytes is disproportionately small compared with the surface level detected by flow cytometry, whereas the total amount in spleen corresponds to the surface expression level. These results indicate that the thymocytes of CD45-AP–null mice contain a much smaller amount of intracellular CD45 compared with wild-type thymocytes and suggest the intriguing possibility that CD45-AP affects the subcellular localization of CD45 differentially in thymocytes and spleen cells.

The reduced expression of CD45 in CD45-AP–null cells suggests that CD45 is unstable in the absence of its binding partner, CD45-AP. Likewise, it has been reported that CD45-AP protein levels were much lower in three out of four CD45-negative T cell lines compared with their CD45-positive wild-type parental lymphocytes, suggesting that CD45-AP is unstable in the absence of CD45 (10).

Amino acid sequence analysis and protease susceptibility analysis of CD45-AP indicate that only a short segment at the NH₂ terminus of CD45-AP is located extracellularly, with the bulk of the protein being intracellular (11). CD45-AP and CD45 form a specific association through their respective transmembrane segments (11–13), whereas the two PTP domains of CD45 reside intracellularly (2). Therefore, it seems unlikely that binding of CD45-AP directly affects CD45 PTP activity since the binding does not involve the PTP domains. In fact, the PTP activity of CD45 isolated from a T cell line, YAC-1, was not altered by binding to recombinant CD45-AP in vitro (Takeda, A., unpublished observations). However, the possibility remains that CD45-AP coordinates the interaction between CD45 and other regulatory molecules and thus affects its PTP activity indirectly. Therefore, PTP activities of CD45 isolated from thymus and spleen of CD45-AP–null mice were compared with their wild-type counterparts by measuring the release of ³²PO₄ from ³²P-labeled Raytide. The amount of CD45 was estimated by immunoblotting of the CD45 samples with anti-CD45 antisera followed by densitometric analysis of the immunoblots. Specific activities of CD45 PTP were obtained by dividing CD45 PTP activity by the amount of CD45. As shown in Table 1, the deletion of CD45-AP did not affect CD45 PTP activity in either thymus or spleen.

In vitro analysis of proliferative responses of splenic T and B cells to various stimuli revealed interesting differences between CD45-AP–null mice and their wild-type littermates. Proliferative responses of CD45-AP–null T cells to TCR-mediated stimuli were significantly reduced regardless of the presence or absence of accessory cells (AC; Fig. 3,A). Strikingly, the response of CD45-AP–null T cells to a combination of PMA and ionomycin was reduced to <50% of wild-type whereas the response to Con A was comparable to that of the control. The proliferative response of mutant B cells to anti-IgM Ab was reduced to half of the wild-type, whereas their responses to anti-CD40 Ab, LPS, or a combination of PMA and ionomycin were comparable to control levels (Fig. 3 B).

No difference in fluorescence intensity was observed between CD45-AP–null and wild-type littermates when their spleen cells were stained with FITC-labeled anti–TCR-α/β (clone H57-597; PharMingen) or anti-IgM (clone R6-60.2; PharMingen) Ab and analyzed by flow cytometry (data not shown). Therefore, the reduced proliferative responses of CD45-AP–null lymphocytes to antigen receptor ligation are due to defects in signal transduction and not differences in receptor numbers. Furthermore, cells obtained by two successive rounds of nylon-wool fiber columns gave identical results in the proliferation assays to those obtained by negative selection with antibody-mediated affinity columns (data not shown). Thus, incubation of cells with antibodies during the cell purification procedures had no effect on the outcome of the cell proliferation and function assays.

B cells of both CD45-deficient and CD45-null mice differentiate normally but fail to proliferate in response to anti-IgM stimulation (14, 15). The mechanism by which the absence of CD45 leads to failed proliferation is unknown. Thus, it is possible that the reduced B cell response to anti-IgM stimulation observed in CD45-AP–null mice (Fig. 3,B) is due to the decreased expression of CD45 in CD45-AP–null B cells. On the other hand, there is a considerable amount of evidence that CD45 plays a similar role in antigen receptor–mediated signal transduction of T and B cells, namely, by dephosphorylating regulatory tyrosine residues of their respective Src family PTKs (1, 2). As described below (Fig. 4 B), the interaction of CD45 with Lck in T cells is significantly decreased in CD45-AP–null mice. Therefore, it is also possible that the reduced B cell response to anti-IgM stimulation observed in CD45-AP–null mice is a result of a reduced interaction of CD45 with B cell PTKs.

As expected from the impaired proliferative response of CD45-AP–null T cells to anti-TCR antibodies, spleen cells of CD45-AP–null mice exhibited reduced responses in MLR and CTL assays. CD45-AP–null T cells showed a markedly reduced proliferative response in MLR when they were cultured with mitomycin C–treated allogeneic lymphocytes obtained from BALB/c mice (Fig. 3,C). Syngeneic lymphocytes obtained from C57BL/6 mice did not trigger proliferation in either mutant or wild-type T cells. An ∼50% decrease in CTL activity against MHC-incompatible S49.1 cells was observed in spleen cells derived from CD45-AP–null mice compared with their wild-type littermates (Fig. 3 D). When MHC-compatible spleen cells obtained from C57BL/6 mice were used as targets no significant killing was observed with either mutant or wild-type CTLs.

These data show that T cells of CD45-AP–null mice failed to mount adequate responses not only with stimuli that trigger signaling events at the membrane surface but also with a combination of PMA and ionomycin. Stimulation with a combination of PMA and ionomycin provides a measure of cell response irrespective of upstream signaling events by directly stimulating protein kinase(s) C (PKC) and causing intracellular calcium influx. Since the response to a combination of PMA and ionomycin in CD45-AP–null B cells is unaffected, it is possible that CD45-AP has a role in signaling events that is T cell specific and downstream of PKC. PKC appears to have multifaceted and complex functions in TCR signal transduction pathways (22). PKC may function either upstream or downstream of Ras in the Ras pathway, or parallel to the Ras pathway for enhancing TCR signals (23, 24). Effectors of PKC in these pathways have not been established but candidates include the Ras GTPase-activating protein (p120-GAP), Lck, Raf-1, and Sos (guanine nucleotide exchange factor). Studies to identify the cause of PKC-mediated signaling blockage in CD45-AP–null mice are in progress.

Most of the CD45 population in lymphocytes exists as a complex with CD45-AP with a stoichiometry of 1:1 (8). It has been hypothesized that CD45-AP coordinates the interaction between CD45 and other molecules involved in CD45-mediated signaling (9). Therefore, we examined the interaction of CD45 with Lck, a Src family PTK associated with the TCR complex and a likely CD45 substrate in vivo (3, 4). Splenic T cells of CD45-AP–null mice and wild-type mice were stimulated by TCR ligation with anti–TCR-α/β Ab and cell lysates were prepared at various time points after stimulation. The total amount of Lck present in the cell lysates was equal between the samples obtained from CD45-AP–null and wild-type cells and also remained unchanged during the stimulation as shown in Fig. 4,A. Material immunoprecipitated with anti-CD45 Ab or with control beads was analyzed by immunoblotting with anti-CD45 and anti-Lck Abs (Fig. 4 B). In wild-type mice, the amount of Lck associated with CD45 increased after 5 and 10 min of stimulation and returned to the baseline level after 60 min of stimulation. In contrast, in CD45-AP–null mice the amount of Lck coimmunoprecipitating with CD45 was significantly diminished compared with that seen in wild-type mice and was not affected by stimulation. The amounts of Lck and CD45 in immunoprecipitates were quantitated by densitometry of the immunoblots and Lck/CD45 ratios were obtained. The Lck/CD45 ratio increased in wild-type cells after 5 and 10 min of stimulation but remained at constantly lower levels in CD45-AP–null cells. The Lck/CD45 ratios of wild-type T cells were 1.9-, 2.5-, 2.7-, and 2.7-fold more than those of CD45-AP–null T cells at stimulation times 0, 5, 10, and 60 min, respectively. The low, but detectable, levels of association between CD45 and Lck in CD45-AP–null mice indicate that CD45 is capable of associating with Lck at a lower level in the absence of CD45-AP. In fact, direct association has been observed between recombinant CD45 and Lck in vitro (25). However, when CD45 and Lck were expressed in fibroblasts that do not express CD45-AP, CD45 failed to stably dephosphorylate the kinase-active form of Lck (26), suggesting that an CD45-AP is required for optimally functioning association of CD45 with Lck.

Dephosphorylation of downregulatory tyrosine residues by CD45 is thought to be required for the activity of Src family PTKs that function early in lymphocyte signal transduction (1–4). However, the PTP activity of CD45 is potentially capable of counteracting the activity of PTKs by dephosphorylating their substrates (27, 28). Given the important role of CD45 in lymphocyte differentiation and signal transduction, there must be a mechanism in vivo that either regulates the CD45 PTP activity itself or restricts its access to substrates. Our results demonstrate that CD45-AP is essential for normal interactions of CD45 with a TCR-associated PTK, Lck, in vivo (Fig. 4). It remains to be determined whether CD45-AP directly mediates the interaction between CD45 and Lck. Combined with the functional impairments exhibited by T and B lymphocytes of CD45-AP–null mice (Fig. 3), our results strongly support the notion that CD45-AP plays an important role in coordinating the interaction between CD45 and the appropriate PTKs involved in CD45-mediated signal transduction. This function of CD45-AP may provide a means for regulation of CD45 PTP by targeting appropriate substrates.

We thank F. Gonzalez and N. Yaseen for useful comments on the paper; J. Buters for helpful technical advice; J. Marth, I. Trowbridge, and H. Westphal for generously providing reagents; N. Kouttab for assisting in flow cytometry; and C. Samek for providing superb care for the mice.

Supported in part by National Institutes of Health Grant GM-48188 to A. Takeda.

AP

associated protein

ES

embryonic stem

PKC

protein kinase C

PTK

protein tyrosine kinase

PTP

protein tyrosine phosphatase