Distinct isoforms of the CD45 protein-tyrosine phosphatase differentially regulate interleukin 2 secretion and activation signal pathways involving Vav in T cells.

The CD45 family of transmembrane protein-tyrosine phosphatases plays a crucial role in the regulation of lymphocyte activation by coupling activation signals from antigen receptors to the signal transduction apparatus. Multiple CD45 isoforms, generated through regulated alternative mRNA splicing, differ only in the length and glycosylation of their extracellular domains. Differential distribution of these isoforms defines subsets of T cells having distinct functions and activation requirements. While the requirement for the intracellular protein-tyrosine phosphatase domains has been documented, the physiological role of the extracellular domains remains elusive. Here we report the generation of CD45-antisense transfected Jurkat T cell clones that lack CD45 or have been reconstituted to uniquely express either the smallest, CD45(0), or the largest, CD45(ABC), isoform. These cells exhibited marked isoform-dependent differences in IL-2 production and tyrosine phosphorylation of cellular proteins, including Vav after anti-CD3 stimulation. These results demonstrate that the distinct CD45 extracellular domains differentially regulate T cell receptor-mediated signaling pathways. Furthermore, these findings suggest that alterations in CD45 isoform expression by individual T cells during thymic ontogeny and after antigen exposure in the periphery directly affects the signaling pathways utilized.

The CD45 family of transmembrane protein-tyrosine phosphatases plays a crucial role in the regulation of lymphocyte activation by coupling activation signals from antigen receptors to the signal transduction apparatus. Multiple CD45 isoforms, generated through regulated alternative mRNA splicing, differ only in the length and glycosylation of their extracellular domains. Differential distribution of these isoforms defines subsets of T cells having distinct functions and activation requirements. While the requirement for the intracellular protein-tyrosine phosphatase domains has been documented, the physiological role of the extracellular domains remains elusive. Here we report the generation of CD45-antisense transfected Jurkat T cell clones that lack CD45 or have been reconstituted to uniquely express either the smallest, CD45(0), or the largest, CD45(ABC), isoform. These cells exhibited marked isoform-dependent differences in IL-2 production and tyrosine phosphorylation of cellular proteins, including Vav after anti-CD3 stimulation. These results demonstrate that the distinct CD45 extracellular domains differentially regulate T cell receptor-mediated signaling pathways. Furthermore, these findings suggest that alterations in CD45 isoform expression by individual T cells during thymic ontogeny and after antigen exposure in the periphery directly affects the signaling pathways utilized.
Activation of resting T lymphocytes through the T cell receptor (TCR) 1 requires expression of the CD45 family of transmembrane protein-tyrosine phosphatases (PTPases) (1,2). CD45 has been shown to regulate the basal activity of the Fyn and Lck protein-tyrosine kinases (PTKs) by dephosphorylation of their respective regulatory carboxyl-terminal tyrosine residues (3)(4)(5)(6)(7). However, it is not clear that these are CD45's sole functions. For example, new evidence suggests that CD45 can also dephosphorylate certain PTK substrates, such as the TCR chain (8) and the 32-kDa CD45-associated phosphoprotein, LPAP (9). Thus, the precise functions of the CD45 phosphatase in signal transduction are incompletely understood.
While the requirement for the intracellular PTPase domains has been documented (10 -13), the function of the CD45 extracellular domain in lymphocyte signal transduction remains a major unresolved issue. In humans, five CD45 isoforms, ranging in size from 180 -220 kDa, are generated by the regulated alternative mRNA splicing of three exons, encoded by a single gene (14 -16). The alternatively spliced exons, commonly referred to as A, B, and C, are located near the 5Ј end of the gene and give rise to isoforms that differ only in their extracellular regions. Individual lymphocytes simultaneously express more than one CD45 isoform (17,18). However, the expression of certain isoforms is highly regulated, resulting in their differential expression on lymphocytes of different lineage (e.g. T versus B cells), as well as on distinct functional subsets of T cells (19 -22). Furthermore, individual T cells alter their isoform expression in a highly regulated manner during thymic selection and upon antigen exposure in the periphery (18,(23)(24)(25)(26). The tight regulation of CD45 isoform expression by lymphocytes having distinct functions argues that these differences are likely to be of biologic importance. However, attempts to study the role of individual CD45 isoforms in signaling have been severely hampered by great difficulty in re-expressing different single intact CD45 isoforms into the same cellular background.
Recent studies have clearly demonstrated that TCR-mediated signaling can be reconstituted in CD45 Ϫ mutants by transfection of chimeric molecules containing the conserved PTPase domains, but lacking the CD45 transmembrane or extracellular regions (11)(12)(13). However, these results do not exclude a potentially important role for the CD45 extracellular domain, since the various extracellular domains could superimpose distinct regulatory constraints upon the cytoplasmic domain. Our present findings strongly support this hypothesis. Utilizing a unique model system, we now demonstrate that the expression of different individual CD45 isoforms is associated with differences in IL-2 production, as well as differences in the activation-related phosphorylation of cellular proteins including Vav. These findings demonstrate the preferential utilization of different signaling pathways by distinct CD45 isoforms.

MATERIALS AND METHODS
DNA Constructs-A 270-base pair segment (from the P1 transcription initiation site to the initiation codon) was amplified by polymerase chain reaction from genomic CD45 DNA (clone LCA.512, from Dr. H. Saito, Dana-Farber Cancer Institute, Boston, MA (16)). Polymerase chain reaction primers incorporating unique restriction sites allowed ligation into the RcSR␣ plasmid vector (27) in an antisense orientation, generating the AS-CD45 plasmid vector. To generate CD45 cDNA con-structs in the sense orientation, the pSPSR␣ LCA.1 and LCA.6 constructs (28), encoding the smallest, CD45R(0), and largest, CD45R(ABC), isoforms, respectively (from Dr. M. Streuli, Dana-Farber Cancer Institute), were modified by removing a 5Ј SacI-SphI segment, to reduce overlap between AS-CD45 and these CD45 cDNAs to 40 base pairs.
Cell Lines and Transfections-The Jurkat human leukemic CD4 ϩ T cell line was maintained in RPMI 1640 media containing 10% fetal calf serum, 4 mM L-glutamine, and 50 g/ml gentamycin at 37°C in humidified atmosphere with 5% CO 2 . Cells were transfected by electroporation, and G418-resistant colonies were screened for loss of CD45 by immunofluorescence. Two of the CD4 ϩ CD45 Ϫ clones (J-AS-1 and J-AS-2) were selected for further study. J-AS-1 was co-transfected at a 10:1 ratio with cDNA constructs for either the CD45R(0) or CD45R(ABC) isoform plus the pPGKhyg plasmid (29) encoding Hygromycin B resistance. Resistant colonies (G418 and Hygromycin) were screened by immunofluorescence for CD45 expression as well as for expression of appropriate cell surface markers described below. Clones were sorted (fluorescence activated cell sorting) as necessary to obtain similar CD45 and CD4 expression, as described below.
Antibodies, Immunofluorescence Phenotyping, and Cell Sorting-Immunofluorescence analysis was performed as described previously (26). Mouse anti-human mAbs reactive with CD2, CD3, CD4, CD28, CD45, and CD45RA were obtained as ascites (generously provided by Dr. Chikao Morimoto, Dana-Farber Cancer Institute) or as phycoerythrin conjugates (from Coulter Immunology, Hialeah, FL), anti-CD45RO (obtained with the kind permission of Dr. Peter Beverly, University College Hospital, London), and goat anti-mouse IgG-FITC (from Southern Biotechnology, Birmingham, AL). Cell phenotype was routinely monitored for these markers using a BD FACSTAR IV (10,000 cells/sample). Cell sorting was performed using a BD FACSTAR IV, after staining by direct or indirect immunofluorescence, as described (26). Fluorochrome conjugates were dialyzed to remove sodium azide prior to use.
Rabbit polyclonal Ab to Vav was developed by immunization of rabbits with a synthetic peptide containing residues 575-594 of (murine) Vav. Immunoprecipitation of a 95-kDa band with the antiserum was specifically blocked by addition of the immunizing peptide to the lysate mixture (data not shown).

legend).
Cellular Activation-For whole cell lysates, 3 ϫ 10 6 cell aliquots were prewarmed to 37°C and then treated for the indicated time periods with prewarmed goat anti-mouse (GAM) alone (unstimulated control), or anti-CD3 (30 g/ml) plus GAM (7.5 g/ml) (stimulated), as described (31). For analysis of Vav tyrosine phosphorylation, 15 ϫ 10 6 cells were stimulated with anti-CD3 for 1 min at 37°C, as described above. Unstimulated controls were treated with anti-CD3 after addition of lysis buffer. At the appropriate times, an excess of ice-cold stop solution (phosphate-buffered saline containing 5 mM EDTA and phosphatase inhibitors) was added, followed immediately by brief centrifugation in a Microfuge, removal of the supernatant, and resuspension of the pellets in ice cold lysis buffer (described below).
Cell Lysis, Western Blotting, and Immunoprecipitation-Cells were lysed in 1% Nonidet P-40 lysis buffer containing 25 mM Tris-HCl (pH 8.0) with 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 10 mM iodoacetamide, 1 mM sodium vanadate, 10 mM NaF, and 10 mM sodium pyrophosphate, for 20 min at 4°C, followed by centrifugation at 14,000 rpm for 15 min, as described (31,32). For experiments using whole cell lysates, postnuclear supernatants were boiled in Laemmli sample buffer and loaded onto SDS-PAGE gels. For analysis of CD45 isoform expression, (1-2 ϫ 10 6 cell eq for Jurkat and 5 ϫ 10 6 for transfectants) were run on 6% SDS-PAGE, transferred to nitrocellulose membranes, and immunoblotted with anti-CD45 (GAP 8.3, ATCC). For analysis of tyrosine phosphorylation of cellular proteins, 3 ϫ 10 6 cell eq/lane were lysed before or after stimulation as above, and postnuclear supernatants separated on 10% SDS-PAGE (nonreducing conditions). Following transfer to nitrocellulose, membranes were blocked with 5% milk in phosphate-buffered saline and probed with anti-phosphotyrosine (anti-Tyr(P)) (mAb 4G10 generously provided by Dr. Brian Drucker, Oregon Health Sciences Center, Portland, OR). For immunoprecipitation of Vav, 15 ϫ 10 6 cells/lane were activated and lysed as described above, and postnuclear supernatants were precleared as described (31,32), twice. Protein concentration was determined using the micro-BCA protein assay kit (Pierce), and equivalent amounts of protein in precleared lysates were incubated with anti-Vav antisera, followed by immunoprecipitation with Protein A-Sepharose (Pharmacia Biotech Inc.). Immunoprecipitates were washed five times with lysis buffer, boiled in sample buffer, and separated on 8% SDS-PAGE under reducing conditions. After transfer to nitrocellulose, blots were blocked as above and then probed with anti-Tyr(P) (4G10), as described (31,33). After stripping, the same membrane was reprobed with anti-Vav (kindly provided by Dr. A. Altman, La Jolla Research Institute, La Jolla, CA, or from UBI, Lake Placid, NY). All immunoblots were developed using horseradish peroxidaseconjugated secondary reagents and developed using ECL (Amersham).

RESULTS AND DISCUSSION
Like peripheral T cells, the Jurkat human T cell leukemia line normally expresses CD45 at high levels, and individual cells express multiple isoforms simultaneously (17)(18)(19). To examine the role of CD45 and its individual isoforms, free from the potential confounding influences of unknown mutations, we directly targeted endogenous CD45 expression by stable transfection of a plasmid construct (AS-CD45) expressing an antisense RNA directed at a 270-base pair region of genomic CD45 just upstream from the coding region. (Fig. 1A). Of the several independent CD45 Ϫ colonies selected and subcloned, two, denoted J-AS-1 and J-AS-2, were selected on the basis of CD4 expression comparable to that of parental Jurkat. (Fig.  1B). Jurkat expresses high levels of total CD45 and lower levels of both the smallest CD45 isoform (CD45RO) and the largest (two) isoforms which contain exon A (CD45RA). J-AS-1 and -2 lack detectable CD45 expression either on the surface or in the cytoplasm (Fig. 1, B and C).
J-AS-1 was then stably transfected with CD45 cDNA constructs modified to minimize overlap with AS-CD45 antisense RNA and encoding either the smallest isoform, denoted CD45(0), lacking alternative exons, or the largest isoform, denoted CD45(ABC), which includes all three alternative exons. These isoforms best exemplify differential distribution on T cell subsets having distinct functions and activation preferences (19,21,23). Each of the CD45 ϩ clones arising expressed solely the transfected CD45 isoform by both immunofluorescence and by immunoblotting. (Fig. 1, B and C). Three independent CD45(ABC) transfected isolates (J[ABC]-1, -2, and -3) and two CD45(0) transfected isolates (J[0]-1 and -2) were selected for further study, based on CD45 expression and wild-type levels of CD3. The clones were then sorted to obtain stable populations expressing similar levels of CD4 and CD45. When matched for their surface expression, J[0] and J[ABC] clones expressed identical levels of CD45 by immunoblotting, indicating no inherent differences in the relative distribution of intracellular and extracellular CD45 (data not shown). Although total CD45 expression was lower in the transfectants, the expression of individual CD45(0) and CD45(ABC) isoforms by J[0] and J[ABC] cells, respectively, was similar to their expression in wild-type cells. The expression of CD3, CD2, and CD28 was nearly identical in each of the cell lines ( Fig. 1B and data  not shown).
Current evidence indicates that CD45 regulates the activity of proximal components of the signaling apparatus such as the Src family PTKs, Lck and Fyn, and, presumably, their substrates (3)(4)(5)(6). First, TCR/CD3-induced IL-2 secretion, which depends on the coordinated activation of multiple transcription factors (34), was examined as an integrated measure of such signaling events. The dose-response curve to anti-CD3 ( Fig. 2A) reveals that, in contrast to Jurkat, the CD45 Ϫ (J-AS) cell lines secreted minimal IL-2 in response to all doses of anti-CD3 tested. Furthermore, no enhancement was seen after co-stimulation by cross-linking anti-CD3 and anti-CD4 (data not shown). Reconstitution with the CD45(0) isoform resulted in wild-type levels of IL-2 secretion after stimulation with anti-CD3 (1.0 g/ml). In contrast, the CD45(ABC) transfected cell lines produced significantly less IL-2 than either Jurkat or the CD45(0) transfectants at both 0.05 and 1.0 g/ml anti-CD3, secreting at maximum, 30% of wild-type levels. Increasing the anti-CD3 dose to 5 g/ml had no additional effect on IL-2 secretion by any of the cell lines (data not shown). However, at lower doses of anti-CD3 (0.005 g/ml), transfectants expressing either individual isoform secrete much less IL-2 than Jurkat, possibly owing to the lower levels of overall CD45 expression. Stimulation with anti-CD2 gave overall results similar to those observed above (not shown).
Similar responses by J-AS-1 and each of its single isoformreconstituted derivatives after stimulation by Ab-mediated cross-linking of CD3 and CD28 (Fig. 2B), or with PMA plus ionomycin (Fig. 2C), documents similar inherent capacity of each cell line to secrete IL-2 when the proximal signaling machinery, or the requirement for CD45 (30), are bypassed, respectively.
Thus, after stimulation with anti-CD3 (at 0.05 to 1 g/ml), IL-2 secretion by J[0] transfectants is not significantly different from wild-type cells, despite 6 -7-fold lower CD45 expression. Nonetheless, it is possible that decreased IL-2 secretion by J[ABC] cells compared to J[0] cells is due to their somewhat lower levels of CD45 expression. To rule out this possibility, we sorted J[ABC] clones to obtain CD45 expression equal to that of J[0] cells and then compared IL-2 secretion by these cell populations after anti-CD3 stimulation (see Fig. 3). As before, clones expressing CD45(0) secrete wild-type levels of IL-2. As shown, increased CD45 expression by J[ABC] transfectants did not augment IL-2 secretion. Both sorted and unsorted J[ABC] populations averaged just 24% of the wild-type levels of IL-2.
Given differential anti-CD3-induced IL-2 secretion by these cell lines, more proximal signaling events were next examined. Comparison of Lck and Fyn activities by immune complex kinase assays failed to reveal isoform-dependent differences (data not shown). T cell activation is associated with alterations in the tyrosine phosphorylation of a number of cellular proteins. Therefore, we compared the tyrosine phosphorylation of cellular proteins in each cell line before, and at various time points after, anti-CD3 stimulation (Fig. 4). Under basal conditions, J-AS-1 consistently revealed hyperphosphorylation of a limited set of bands at ϳ70 -76 kDa and decreased tyrosine phosphorylation of several other bands (ϳ105, ϳ95, and ϳ50 -52 kDa) when compared to Jurkat (Fig. 4A).
After anti-CD3 stimulation of Jurkat, there was rapid phosphorylation (peaking at 30 s to 1 min) and subsequent dephosphorylation of a number of bands. Although many of the same bands were ultimately phosphorylated (within 5-10 min) after stimulation of J-AS-1 cells, the kinetics were significantly slowed. Furthermore, once phosphorylated, these bands did not undergo dephosphorylation, consistent with decreased action of the CD45 PTPase and perhaps of other cellular PTPases whose activities depend on regulated tyrosine phosphorylation (35,36).
Re-expression of either the CD45(0) or the CD45(ABC) isoforms generally restored basal and activation-related tyrosine phosphorylation, although the kinetics were somewhat prolonged compared to wild-type Jurkat (Fig. 4B). This may reflect the lower overall levels of CD45 expression in these cells. More importantly, direct comparison reveals clear isoform-dependent differences in the relative phosphorylation of several bands. For example, J[ABC] cells consistently exhibited relative hyperphosphorylation of a band at ϳ95 kDa when compared to J[0] cells.
This prompted a comparison in our cells of the tyrosine phosphorylation of p95 vav (Vav) which is rapidly and transiently phosphorylated on tyrosine after ligation of the TCR (33,37), CD28 (38), or upon the binding of IL-2 to its receptor (39). While the exact function of this proto-oncogene product in signal transduction is unclear, gene ablation studies document the important role of Vav in the activation and proliferation of mature lymphocytes as well as in the normal developmental expansion of lymphocyte precursors in the marrow and thymus (40,41).
Basal Vav tyrosine phosphorylation was minimal but detectable in each of our cell lines (Fig. 5A). Anti-CD3 stimulation consistently induced significantly greater tyrosine phosphorylation of Vav within 1 min, in Jurkat and particularly in all three J[ABC] transfectants compared to either of the two J[0] transfectants or the CD45 Ϫ J-AS-1 cells. Reprobing the same membrane with anti-Vav antisera confirmed similar loading of Vav protein in each lane (Fig. 5B). These differences are not secondary to altered kinetics, since the same pattern is observed 4 min after anti-CD3 stimulation, at which time phosphorylation of Vav in Jurkat and single-isoform transfectants is decreasing (data not shown and Ref. 33).
Our results are the first to demonstrate that signaling pathways utilized by the TCR are differentially regulated by the extracellular domain of distinct CD45 isoforms. Stimulation of the TCR leads to the phosphorylation of a number of cellular proteins including Vav. Although the signaling pathways in-volving Vav have not yet been clarified, Vav contains an array of signaling and DNA-binding motifs, including SH2 and SH3 domains, a Dbl domain, and a helix-loop-helix, which all appear to be involved in the generation of downstream signals (33,(42)(43)(44)(45). Activation-related tyrosine phosphorylation directs SH2-mediated interactions between Vav and several other signaling molecules. Thus, Vav has been shown to associate with Shc, Grb2, ZAP-70, phosphatidylinositol 3-kinase (p85), CD19, VAP-1, and several other uncharacterized bands through SH2 and/or SH3-mediated interactions after activation of B or T lymphocytes (37, 46 -48).
How different CD45 isoforms might regulate this pathway remains speculative. Particular CD45 isoforms might directly dephosphorylate Vav or could differentially regulate the activ- ity of the PTK(s) responsible for Vav phosphorylation. However, preferential dephosphorylation of Vav by a particular isoform is difficult to envision given that Vav phosphorylation is decreased in both CD45 Ϫ cells and those expressing CD45(0), yet increased in wild-type cells (which express both CD45(0) and CD45(ABC) isoforms) and in cells expressing CD45(ABC). Activation-related phosphorylation of Vav clearly does not directly correlate with absolute levels of CD45 expression. Our results are more consistent with augmented activity of the PTK responsible for Vav phosphorylation by cells expressing the CD45(ABC) isoform. However, the PTK(s) responsible for the in vivo phosphorylation of Vav are presently unknown. Although Lck is capable of phosphorylating Vav in vitro, IL-2 mediated phosphorylation of Vav occurs in the absence of Lck (39). CD28 ligation results in Itk phosphorylation followed temporally by that of Vav, suggesting a possible link (38). Recently, it was reported that ZAP-70 can physically associate with the Vav-SH2 domain after T cell activation, although it is unknown whether Vav serves as a substrate for this PTK (47). Further analysis of these PTKs and Vav-associated molecules in our single isoform transfectants may help delineate those pathways relevant to CD45.
As mentioned above, chimeric PTPase molecules lacking the CD45 transmembrane or extracellular domains are able to restore nearly normal patterns of tyrosine phosphorylation and calcium flux (11)(12)(13). In general agreement, we showed that either the CD45(0) or CD45(ABC) isoforms are capable of reconstituting activation-related tyrosine phosphorylation. However, isoform-specific differences in IL-2 production and the tyrosine phosphorylation of cellular proteins indicate that the extracellular domain can superimpose regulatory influences on a "constitutive" cellular PTPase requirement. Thus, even though many of the signaling pathways are conserved, a subset (that includes Vav) appears subject to differential regulation by the various CD45 extracellular domains. Differences in the utilization of these pathways can lead to rather substantial isoform-dependent differences in IL-2 secretion as demonstrated in our model and in the mouse thymocyte model of Novak et al. (49).
Exactly how the CD45 extracellular domains may regulate the PTPase domains is unknown. One long-standing hypothesis supported by our findings is that the distinct extracellular domains of the various CD45 isoforms interact with different molecules on the surface of the same or other cells, thereby directing the cytoplasmic phosphatase domains toward distinct substrates. In this regard, the co-capping studies of Dianzani et al. (50) support the notion that differential interactions between CD45 isoforms and other molecules on the surface of the same cell can occur.
In conclusion, our results indicate that the regulated expression of distinct CD45 isoforms on different developmental and functional subsets of T cells may impose preferential utilization of particular TCR-mediated signaling pathways. Alterations in CD45 isoform expression by individual T cells in response to thymic selection or peripheral antigen exposure, may consequently allow that cell to respond to TCR ligation using a different subset of signals. We speculate that the delivery of these different signals to the cell nucleus might have a significant influence on cell differentiation, the expression of functional repertoire, or in allowing T cells to "fine-tune" their responsiveness. A more complete understanding of these differences is likely to have important implications for signal transduction and for the interpretation of the highly regulated expression of CD45 isoforms in lymphocytes.