Coreceptor Signal Strength Regulates Positive Selection but Does Not Determine CD4/CD8 Lineage Choice in a Physiologic In Vivo Model This information is current as of September 27, 2021. Batu Erman, Amala S. Alag, Oyvind Dahle, François van Laethem, Sophia D. Sarafova, Terry I. Guinter, Susan O. Sharrow, Alexander Grinberg, Paul E. Love and Alfred Singer

J Immunol 2006; 177:6613-6625; ; Downloaded from doi: 10.4049/jimmunol.177.10.6613 http://www.jimmunol.org/content/177/10/6613 http://www.jimmunol.org/ References This article cites 50 articles, 18 of which you can access for free at: http://www.jimmunol.org/content/177/10/6613.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 27, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Coreceptor Signal Strength Regulates Positive Selection but Does Not Determine CD4/CD8 Lineage Choice in a Physiologic In Vivo Model1

Batu Erman,*† Amala S. Alag,* Oyvind Dahle,* Franc¸ois van Laethem,* Sophia D. Sarafova,* Terry I. Guinter,* Susan O. Sharrow,* Alexander Grinberg,‡ Paul E. Love,‡ and Alfred Singer2*

TCR signals drive development, but it remains controversial what impact, if any, the intensity of those signals have on differentiation in the . In this study, we assess the impact of CD8 coreceptor signal strength on positive selection and CD4/CD8 lineage choice using novel gene knockin mice in which the endogenous CD8␣ gene has been re-engineered to encode the stronger signaling cytoplasmic tail of CD4, with the re-engineered CD8␣ gene referred to as CD8.4. We found that stronger signaling CD8.4 coreceptors specifically improved the efficiency of CD8-dependent positive selection and quantitatively increased Downloaded from -the number of MHC class I (MHC-I)-specific signaled to differentiate into CD8؉ T cells, even for thymocytes ex pressing a single, transgenic TCR. Importantly, however, stronger signaling CD8.4 coreceptors did not alter the CD8 lineage choice of any MHC-I-specific thymocytes, even MHC-I-specific thymocytes expressing the high-affinity F5 transgenic TCR. This study documents in a physiologic in vivo model that coreceptor signal strength alters TCR-signaling thresholds for positive selection and so is a major determinant of the CD4:CD8 ratio, but it does not influence CD4/CD8 lineage choice. The Journal of

Immunology, 2006, 177: 6613–6625. http://www.jimmunol.org/

cell immunocompetence requires CD4ϩ T cells to express intracellular Lck with TCR components, initiating TCR signal trans- Ag receptors specific for MHC class II (MHC-II)3 deter- duction (12, 13). However, the CD4 cytosolic tail associates with ϩ T minants and requires CD8 T cells to express Ag recep- intracellular Lck more efficiently than does the CD8 cytosolic tail, tors specific for MHC class I (MHC-I) determinants (1–4). Such with the result that CD4 coreceptor signaling intensity is greater than concordance between TCR MHC specificity and either CD4 or that of CD8 (14–17). The difference in coreceptor signaling intensity CD8 coreceptor expression is established during differentiation in between CD4 and CD8 is thought to have an important impact on the thymus (4), but the mechanism by which TCR MHC specificity positive selection, although there is significant disagreement as to the by guest on September 27, 2021 determines lineage fate remains highly controversial despite sig- nature of the impact that coreceptor signaling intensity has on positive nificant progress in understanding CD4 and CD8 gene regulation selection. The original “strength-of-signal” hypothesis remains highly (5–10). In addition, it is not known why the thymus normally gen- ϩ ϩ popular (18) and proposes that the intensity of coreceptor signals qual- erates more CD4 than CD8 T cells with the result that the itatively affects positive selection by dictating CD4/CD8 lineage Ͼ CD4:CD8 ratio is 1 in nearly all mammalian species. choice, with strong coreceptor signals from CD4 directing thymocytes Double-positive (DP) thymocytes expressing TCR with appropri- into the CD4 lineage and weak coreceptor signals from CD8 directing ate affinity for intrathymic self-ligands are signaled to undergo posi- thymocytes into the CD8 lineage (19–23). A more recent modifica- tive selection and to further differentiate into single-positive (SP) T tion of the strength-of-signal model proposes that strong coreceptor cells (4, 11). As both CD4 and CD8 coreceptors are associated with signals from CD4 result in long duration TCR signals that direct thy- the nonreceptor protein tyrosine kinase Lck, ligand-induced coen- mocytes into the CD4 lineage, whereas weak coreceptor signals from gagement of either CD4 or CD8 with TCR physically approximates CD8 result in short duration TCR signals that direct thymocytes into the CD8 lineage (23). In contrast, we have proposed that signal strength does not affect CD4/CD8 lineage choice but only has a quan- *Experimental Immunology Branch, National Cancer Institute, Bethesda, MD 20892; †Biological Sciences and Bioengineering Program, Faculty of Engineering and Nat- titative effect on positive selection in that it influences the number of ural Sciences, Sabanci University, Istanbul, Turkey; and ‡Laboratory of Mammalian DP thymocytes signaled to undergo positive selection, with stronger Genes and Development, National Institute of Child Health and Human Development, Bethesda, MD 20892 CD4 signals stimulating more DP thymocytes to undergo positive selection than weaker CD8 signals (24), a perspective supported by Received for publication June 15, 2006. Accepted for publication August 25, 2006. findings from two laboratories (25, 26). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance This study was undertaken to rigorously determine the impact of with 18 U.S.C. Section 1734 solely to indicate this fact. increased signal strength on positive selection by creating a new 1 This work was supported by the Intramural Research Program of the National In- gene knockin model in which the endogenous CD8␣ gene was stitutes of Health, National Cancer Institute, Center for Cancer Research and the Intramural Research Program of the National Institutes of Health, National Institute re-engineered to encode a chimeric CD8/CD4 protein consisting of for Child Health and Human Development. the extracellular and transmembrane domains of CD8␣ and the 2 Address correspondence and reprint requests to Dr. Alfred Singer, Experimental cytosolic tail of CD4 (“CD8.4”). Because expression of the CD8.4 Immunology Branch, National Cancer Institute, Building 10 Room 4B36, Bethesda, knockin allele was regulated by endogenous CD8␣ transcriptional MD 20892. E-mail address: [email protected] control elements, the timing and expression pattern of chimeric 3 Abbreviations used in this paper: MHC-II, MHC class II; MHC-I, MHC class I; DP, ␣ double positive; SP, single positive; WT, wild type; TK, thymidine kinase; NEO, CD8.4 proteins were identical with those of wild-type (WT) CD8 neomycin; ES, embryonic stem; LN, lymph node; Tk, T-killer. proteins. Consequently, CD8.4 mice provide a physiologic model

Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$02.00 6614 POSITIVE SELECTION AND LINEAGE CHOICE in which to examine coreceptor signal strength without the poten- TCR-V␣3 (RR3-16), TCR-V␣8 (B21.14), TCR V␣11 (RR8-1), CD4 tial artifacts induced by aberrant CD8 coreceptor expression that (GK1.5), and Qa2 (1-1-2) were obtained from BD Pharmingen; mAbs spe- ␣ ␣ plagued CD8 coreceptor transgenic mice. Indeed, the results ob- cific for CD8 (CT-CD8 ) were obtained from Caltag Laboratories; and mAbs specific for granzyme B (clone 16G6) and CD154 (CD40L) were tained with CD8.4 knockin mice are remarkably unequivocal: in- obtained from eBioscience. creased CD8 coreceptor signal strength significantly increased the efficiency of MHC-I-specific positive selection and the number of Cell lysis and immunoprecipitations CD8ϩ T cells generated in the thymus, but it had no effect on DP thymocytes were lysed for 30 min at 4°C in buffer containing 1% CD4/CD8 lineage choice. Thus, coreceptor signaling intensity af- Nonidet P-40 (Pierce), 10 mM Tris-HCl (pH 7.2), 140 mM NaCl, 2 mM fects TCR signaling thresholds for positive selection and so is a EDTA, 1 mM NaF, 1 mM orthovanadate, and protease inhibitors. Insoluble material was removed by centrifugation for 10 min at 4°C. Lysates were key determinant of the CD4:CD8 ratio, but it does not influence incubated overnight with anti-CD8␣ mAbs coupled to protein G-Sepharose CD4/CD8 lineage choice. beads (Amersham Biosciences). Proteins were resolved by reducing SDS- PAGE and transferred to polyvinylidene difluoride membranes. Blots con- Materials and Methods taining anti-CD8 immunoprecipitates and total cell lysates were incubated Mice with anti-Lck (3A5 mAb; Santa Cruz Biotechnology) followed by HRP- conjugated protein A, with reactivity revealed by ECL (Pierce). C57BL/6 (B6) were purchased from The Jackson Laboratory; F5 TCR transgenic mice (27), CD8␣o mice (28), CD8␤o mice (29), MHC-IIo mice Intracellular calcium mobilization (30), and EIIa-Cre transgenic mice (31) were bred in our own animal col- Thymocytes were loaded with the calcium dye Indo-1 (Molecular Probes; ony. All mice used in this study were cared for in accordance with National 1.8 mM) at 31°C for 30 min. Cells were stained for 40 min on ice with the Institutes of Health guidelines. ␤

indicated concentrations of the following biotinylated mAbs: anti-TCR Downloaded from ␣ Generation of CD8.4 knockin mice (H57-597); anti-CD4 (GK1.5) and anti-CD8 (53-6.72). DP thymocytes were detected using anti-CD4 and anti-CD8 mAbs that do not cross-react The CD8.4 knockin locus was constructed as depicted in Fig. 1. A targeting with the biotinylated Abs. For stimulation, cells were warmed for 3 min construct containing 5Ј and 3Ј flanking regions homologous to the endog- before being applied to the flow cytometer. Calcium mobilization was an- enous CD8␣ locus was constructed in a pKO917 plasmid (Stratagene) alyzed following avidin cross-linking (4 ␮g/ml) and acquisition was re- backbone that contains a 2-kb thymidine kinase (TK) selection cassette corded for 4 min. Calcium concentrations were calculated using FlowJo derived from pKO-Select-TK (Stratagene) and a 1.65 kb floxed neomycin software (Tree Star). http://www.jimmunol.org/ (NEO) selection cassette from the pGKneoLoxP plasmid (32). A 3.9-kb BrdU staining region of the CD8␣ locus that includes exons 1–3 was PCR amplified to generate the 5Ј flank. This fragment was digested with BglII and XhoI and For in vivo BrdU labeling, mice received daily i.p. injections of 100 ␮lof ligated into the corresponding sites of pKO917, between the TK and NEO a 10 mg/ml solution of BrdU in PBS in addition to having BrdU in the selection cassettes. One part of the 3Ј flank was generated by PCR ampli- drinking water at a final concentration of 0.8 mg/ml. Single-cell suspen- fying CD4 cytoplasmic tail sequences from a plasmid containing a murine sions of thymocytes were prepared on each time point and stained for CD4 cDNA using a hybrid forward primer containing bases from CD8␣ TCR␤, CD4, and CD8␣. Following surface staining, cells were fixed, intron 3, exon 4 and CD4 cDNA and a reverse primer homologous to CD4 stained for intracellular BrdU by using a FITC-BrdU Flow kit (BD Pharm- cDNA. This product was cloned into the pCRII-Topo TA plasmid (Invitro- ingen), and analyzed by flow cytometry. gen Life Technologies) excised with EcoRI (blunted with Klenow) and

HindIII and ligated into the SmaI and HindIII sites of the plasmid pBlue- Results by guest on September 27, 2021 Ј script II (Stratagene). The remainder of the 3 flank was amplified from CD8.4 mice expressing endogenously encoded chimeric CD8␣ genomic DNA (using primers with the appropriate restriction sites and inserted into the ClaI and SalI sites of pBluescript II containing the CD8/CD4 coreceptor proteins CD8/4 hybrid sequence). A SacII-SalI fragment of the resulting plasmid To increase the signaling intensity of endogenously encoded CD8 Ј comprised the full 3 flank and was inserted into the corresponding sites of CD8␣ pKO917 already containing the 5Ј flank. All PCR were performed with 129 coreceptors, we re-engineered the endogenous gene locus in R1 embryonic stem (ES) cell DNA unless otherwise noted. two ways: we replaced the exon encoding the CD8␣ cytoplasmic The resulting targeting construct encodes a chimeric gene with CD8␣ tail with a chimeric exon containing the CD4 cytoplasmic tail, and exons 1–3 provided by the 5Ј flank, and a chimeric exon 4 provided by the we removed the alternative splice site in the CD8␣ gene that gives Ј 3 flank. The amino acid sequence at the modified junction is as follows rise to signaling-deficient tailless CD8␣Ј proteins (33). The mo- (single-letter code, a slash indicates the putative boundary between the transmembrane and intracellular domains, underlined sequences are from lecular approach we used is schematized in Fig. 1A. Our targeting CD8␣ and bold sequences are from CD4): IITLICYH/RCRHQQ.... construct consisted of five components: a TK selection cassette, Therefore, the predicted protein sequence is identical with the protein en- exons 1–3 of CD8␣ encoding the CD8␣ extracellular domain, a coded by the ␣␣4 transgene studied previously (24), except for a single floxed NEO selection cassette, chimeric exon 4 encoding the amino acid that was mutated in the Bosselut study (24) to an alanine, and ␣ is a histidine (identical with WT CD8␣) in the current study. CD8 transmembrane domain ligated to the CD4 cytoplasmic tail, The targeting construct was linearized with a unique SalI site and trans- and a truncated version of noncoding CD8␣ exon 5. The targeting fected into 129 R1 ES cells by electroporation. G418 and ganciclovir dou- construct was electroporated into ES cells so that homologous re- ble-resistant clones were screened for homologous recombination by combination with the endogenous CD8␣ locus would generate the Southern blotting of genomic DNA digested with SphI and probed with a CD8␣T locus (Fig. 1A), and an ES cell clone containing the CD8␣T 5Ј CD8␣ genomic fragment that was outside of the targeting construct. ␣T ␣T This probe identifies a 7-kb band from the CD8␣ locus that is shortened to locus was used to generate CD8 knockin mice. CD8 knockin a 5.5-kb band in the targeted CD8␣T locus. Positive clones were recon- male mice were bred to EIIa-Cre transgenic female mice whose firmed for homologous recombination by Southern blotting (5Ј end) and by offspring deleted the floxed NEO selection cassette in their germ- PCR (3Ј end). Genomic DNA from ES cell clones and tail DNA from line (31), giving rise to mice with the CD8.4 gene (Fig. 1A). We knockin mice were also screened by PCR using a forward oligo that hy- Ϫ Ϫ then bred the CD8.4 gene into CD8␣ / mice to generate CD8.4/ bridized to the CD4 sequences in the chimeric exon and a reverse oligo that Ϫ hybridized to CD8␣ genomic sequence outside (3Ј) of the targeting con- CD8␣ heterozygous mice expressing only chimeric CD8.4 struct. Positive clones were injected into B6 blastocysts. Chimeric mice coreceptor proteins, and these mice subsequently gave rise to were mated to EIIa-Cre transgenic mice to induce germline excision of the CD8.4/CD8.4 homozygous mice. Both heterozygous and homozy- NEO-resistance cassette. The deletion of the NEO cassette was confirmed gous CD8.4 mice were used in this study. by PCR on tail DNA. Southern blotting of ES cell DNA confirmed that the selected Antibodies ES cell clone contained the targeted CD8␣T allele (Fig. 1B, left mAbs with the following specificities were used for immunofluorescence panel), and PCR analyses of tail DNA from CD8.4 mice confirmed and flow cytometry: CD5 (53-7.3), TCR␤ (H57-597), TCR-V␣2 (B20.1), the presence of the CD8.4 allele (Fig. 1B, right panels). Only tail The Journal of Immunology 6615 Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 1. Re-engineering the endogenous CD8␣ locus. A, The targeting construct consisted of a TK cassette, CD8␣ exons 1–3 which encode the CD8␣ extracellular domain (E1–E3), a NEO-resistance cassette flanked by directional LoxP sites (triangles), a chimeric exon encoding the transmembrane domain of CD8␣ and the cytoplasmic domain of CD4 (E4), and a truncated noncoding CD8␣ exon 5 (E5). The targeting construct was electroporated into murine ES cells where it was integrated by homologous recombination to generate the CD8␣T locus. Homologous integration of the CD8␣T locus was confirmed by digesting ES cell DNA with SphI (digestion sites are indicated by S) and blotting the DNA with the probe indicated by the filled bar (f). CD8␣T knockin male mice were bred to EIIa-Cre transgenic female mice to induce germline deletion of the floxed neo selection cassette, giving rise to the CD8.4 gene. Oligonucleotide probes that were used in PCR to confirm homologous integration of the targeting vector (3 ϩ 4) and deletion of the NEO cassette (1 ϩ 2) are indicated (Œ). The expected protein products of the WT CD8␣ locus and chimeric CD8.4 locus are schematically indicated. Mice expressing the CD8.4 gene were bred with CD8␣Ϫ/Ϫ mice to generate heterozygous CD8.4/CD8␣Ϫ mice which were interbred to generate homozygous CD8.4/CD8.4 mice. B, Characterization of CD8␣T and CD8.4 genes. Southern blotting of SphI-digested genomic DNA from NEO-selected ES cells was performed to detect homologous integration of the targeting construct (left panels). The probe indicated by the filled bar (f)inA detects a 7.0-kb fragment from the endogenous CD8␣ locus and a 5.5-kb fragment from the CD8␣T locus (left panels). PCR with oligonucleotides 3 ϩ 4 amplified a 1670-bp band only from CD8.4 tail DNA but not from wild-type CD8␣ tail DNA (top right panel). PCR with oligonucleotides 1 ϩ 2 amplified a 550-bp CD8␣ band from WT CD8␣ tail DNA and a 592-bp band from CD8.4 tail DNA, a size difference of ϳ50 bp which is indicative of a single residual loxP site remaining after germline deletion of the NEO-resistance cassette (which was ϳ2000 bp) to generate the CD8.4 gene locus (bottom right panel). C, Detection of surface CD8.4 proteins. Surface biotinylated CD8 coreceptors on thymocytes from B6, CD8␤o, or CD8.4 mice were solubilized in Nonidet P-40 detergent, immunoprecipitated (Immpt) with anti-CD8␣ mAb, and resolved on reducing SDS-PAGE. CD8␣, tailless CD8␣Ј, and CD8␤ proteins are indicated. CD8.4 proteins are detectably larger than WT CD8␣ proteins because the CD4 cytosolic tail is longer than the CD8␣ cytosolic tail. Note that CD8␤ proteins are present in anti-CD8␣ immunoprecipitates from both B6 and CD8.4 thymocytes, revealing that CD8␤ proteins assemble with both WT CD8␣ and chimeric CD8.4 proteins.

DNA from CD8.4 mice amplified a PCR product from primers Notably, the PCR product obtained with primers 1 and 2 from designated “3” and “4” (see Fig. 1A), which were specific for se- CD8.4 tail DNA was ϳ50 bp larger than that from CD8␣ tail quences in the targeting construct and for downstream sequences DNA, the size difference expected from a residual loxP site re- in the CD8␣ gene locus (Fig. 1B, top right panel). Tail DNA from maining in the CD8.4 locus after Cre-mediated deletion of the both CD8.4 and WT CD8␣ mice amplified PCR products from NEO cassette (which was nearly 2000 bp in length) from the primers designated “1” and “2” (see Fig. 1A) that were specific for CD8␣T allele (Fig. 1B, bottom right panel). sequences present in the CD8␣ gene and that flanked the NEO Biochemical characterization of proteins encoded by the CD8.4 cassette in the targeting construct (Fig. 1B, bottom right panel). gene locus confirmed that they had the structural characteristics 6616 POSITIVE SELECTION AND LINEAGE CHOICE Downloaded from http://www.jimmunol.org/

FIGURE 2. Increased Lck binding and increased signaling intensity by CD8.4 coreceptors. A, More Lck is associated with CD8.4 than WT CD8␣ proteins in DP thymocytes. CD8 coreceptors on DP thymocytes from B6 and CD8.4 mice were solubilized in Nonidet P-40 detergent, immunoprecipitated (Immpt) with anti-CD8␣ mAb, resolved by reducing SDS-PAGE, and blotted for Lck. The relative intensities of the Lck bands were determined by densitometry and compared with B6, which was set at 1.0. B, CD8.4 coreceptors increase in vivo signaling intensity in DP thymocytes. DP thymocytes from B6 and CD8.4 mice were analyzed by multicolor flow cytometry. Single parameter histograms of CD5 expression on electronically gated CD4ϩ8ϩ DP thymocytes from B6 (dashed line) and CD8.4 (solid line) mice are shown. CD5 surface expression on both cell populations was quantified into linear

fluorescence units, with CD5 expression on B6 DP thymocytes set equal to 100. C, CD8.4 coreceptors increase in vitro signaling intensity in DP thymocytes. by guest on September 27, 2021 Thymocytes from WT and CD8.4 mice were loaded with the calcium-sensitive dye Indo-1 (1.8 ␮M) and coated with biotinylated mAbs specific for TCR␤ (5 ␮g/ml), CD4 (1 ␮g/ml), or CD8␣ (1 ␮g/ml). Signaling was induced by addition of streptavidin (1 ␮g) (indicated by the arrow) that cross-linked the stimulatory biotinylated mAbs. DP thymocytes were identified using anti-CD4 and anti-CD8 mAbs which do not cross-react with the biotinylated Abs used for stimulation. Intracellular calcium concentrations were determined by the ratio of Indo-1 fluorescence at 405 vs 510 nm. expected of a chimeric CD8/CD4 protein (20, 24, 34). Because the tent with previous observations that TCR signaling in DP thymo- CD4 cytosolic tail is longer than the CD8␣ cytosolic tail, anti- cytes is highly coreceptor dependent (14, 15). Indeed, coengage- CD8␣ immunoprecipitation of surface-labeled thymocyte proteins ment of TCR with CD4 surface coreceptors induced significant confirmed that CD8.4 proteins were larger than WT CD8␣ pro- and equivalent calcium flux in DP thymocytes from both CD8.4 teins, lacked tailless CD8␣Ј forms, and assembled with CD8␤ and WT B6 mice (Fig. 2C, center panel). Most importantly, co- chains to form surface CD8 coreceptor complexes (Fig. 1C). Chi- engagement of TCR with CD8 surface coreceptors also induced meric CD8.4 proteins also bound significantly more intracellular significant calcium flux in both DP thymocyte populations, but the Lck than WT CD8␣ proteins, as anti-CD8␣ immunoprecipitates of calcium flux induced in CD8.4 thymocytes was significantly purified DP thymocytes from CD8.4 mice contained twice the Lck greater than that induced in B6 thymocytes, demonstrating that of anti-CD8␣ immunoprecipitates from WT mice, even though CD8 signaling by CD8.4 coreceptors was stronger than that of WT total intracellular amounts of Lck were equivalent (Fig. 2A). CD8 coreceptors (Fig. 2C, left panel). In fact, the calcium flux To document that the in vivo signaling intensity of CD8.4 co- induced in CD8.4 thymocytes by TCR coengagement with CD8.4 receptors on DP thymocytes was increased over that of WT CD8␣ coreceptors was equivalent to that induced by TCR coengagement coreceptors, we compared CD5 surface levels on fresh DP thymo- with CD4 coreceptors (Fig. 2C, compare left and center panels), cytes, as CD5 surface levels on DP thymocytes directly reflect the indicating that the signaling strengths of CD8.4 and CD4 corecep- intensity of in vivo TCR signaling (24, 35, 36). We found that tors were equivalent. We conclude that expression of the CD4 surface CD5 expression on CD8.4 DP thymocytes was 1.7ϫ cytosolic tail significantly enhanced CD8.4 coreceptor signaling so higher than that on WT B6 DP thymocytes, revealing that CD8.4 that it was quantitatively equivalent to CD4 coreceptor signaling. coreceptors increased the intensity of in vivo signaling in DP thy- mocytes (Fig. 2B). To more directly compare the signaling inten- Quantitative effect of enhanced CD8 coreceptor signaling on sities of CD8.4 and WT CD8 coreceptors on DP thymocytes, we ϩ used intracellular calcium flux to assess TCR signaling. In the positive selection of CD8 T cells absence of coreceptor coengagement, Ab-mediated TCR engage- To assess the impact of enhanced CD8.4 coreceptor signaling on pos- ment failed to stimulate significant calcium flux in DP thymocytes itive selection of CD4ϩ and CD8ϩ T cells, we compared thymocytes from either CD8.4 or WT B6 mice (Fig. 2C, right panel), consis- and peripheral T cells from CD8.4 mice with WT littermate mice, The Journal of Immunology 6617 Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 3. The CD4 tail on CD8.4 coreceptor molecules specifically increases CD8ϩ T cell number and reverses the CD4:CD8 ratio. A, Thymocytes from B6, WT littermate (WT LM), and CD8.4 mice were analyzed by multicolor flow cytometry. CD4 vs CD8 contour plots on various cell populations from these mice are shown (left panels). Within the contour plots, the numbers inside each box indicates the percentage of cells falling within that box. Numbers below the contour plots indicate the total number of cells obtained. The CD4:CD8 ratios (ϮSEM) in WT littermate (LM) and CD8.4 mice are shown (right panel). A total of seven mice from each strain was analyzed. SEM are indicated by the error bars. Values of p were obtained by the two-tailed Student t test. B, Effect of CD8.4 coreceptors on T cell subpopulations. The frequency (ϮSEM) and absolute number (ϮSEM) of CD4ϩ and CD8ϩ T cells in various lymphoid populations from WT LM (Ⅺ) and CD8.4 (f) mice are shown. A total of seven mice from each strain were analyzed. 6618 POSITIVE SELECTION AND LINEAGE CHOICE Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 4. CD8.4 coreceptors increase the efficiency of positive selection of CD8ϩ T cells, but do not affect CD8 T cell function. A, B6 and CD8.4 mice were continuously labeled with BrdU for up to 7 days and thymocytes were analyzed at different time points for expression of CD4, CD8␣, TCR␤, and intracellular BrdU. Appearance of BrdUϩ cells among DP thymocytes is displayed for days 1–4 (top panel), whereas appearance of BrdUϩ cells among TCR␤highCD4SP and TCR␤highCD8SP thymocytes is displayed for days 1–7 (middle and lower panels). Each data point represents the mean BrdU incorporation of up to three individual mice in each group. Slopes (ϮSE) of the linear regression lines are indicated. The rates of generation of thymocytes in B6 mice (Ⅺ, dashed lines) and CD8.4 mice (f, solid lines) were compared using grouped linear regression and p values are indicated. B, Efficiency of positive selection. The efficiency by which immature DP thymocytes converted into mature TCR␤highCD4SP cells, and TCR␤highCD8SP cells was calculated by dividing the slopes (ϮSE) of the SP regression lines by the slope of the DP regression line, as previously described (40). The conversion efficiencies of DP into TCR␤highCD4SP and TCR␤highCD8SP thymocytes were compared between B6 (Ⅺ) and CD8.4 (f) mice. Values of p were determined by the two-tailed Student t test. C, Function of CD8.4ϩ T cells. Purified LN T cells from B6 and CD8.4ϩ/ϩ mice were assessed ex vivo without stimulation for granzyme B expression by intracellular staining (lower panels), and were assessed for TCR induced up-regulation of CD40L (CD154) and CD69 expression (upper and middle panels). After overnight anti-TCR stimulation, only CD4 T cells up-regulated CD40L, although both CD4 and CD8 T cells up-regulated CD69 expression.

both of which contained one CD8␣Ϫ null allele and so were heterozy- actually be CD8 lineage cells in the process of differentiating into gous for either the CD8.4 or WT CD8␣ gene (Fig. 3A). Thymus CD8ϩ T cells. Consequently, to accurately determine the frequencies cellularity and thymocyte profiles were similar in these two strains, of committed CD4 lineage and CD8 lineage thymocytes in each except that CD8.4 mice contained a higher frequency of SP thymo- strain, we focused on Qa2ϩ thymocytes as these are terminally dif- cytes (both CD4ϩ and CD8ϩ) than WT littermate mice (Fig. 3A). ferentiated and functionally mature T cells (39) (Fig. 3A). Unlike WT Because positively selected DP thymocytes first appear as CD4ϩ8low littermate mice that contained a greater number of CD4SP than transitional cells whether they are differentiating into CD4ϩ or CD8ϩ CD8SP cells among Qa2ϩ thymocytes, CD8.4 mice contained the T cells (37, 38), some thymocytes that fall into the CD4SP gate may reverse (i.e., a greater number of CD8SP than CD4SP cells among The Journal of Immunology 6619

lection in CD8.4 and B6 mice. We injected young adult mice daily with BrdU and also put BrdU in their drinking water to continu- ously label proliferating pre-DP thymocytes so that we could track their subsequent differentiation into DP and SP thymocytes (Fig. 4). BrdUϩ cells appeared as DP thymocytes with equal rapidity in CD8.4 and B6 mice (Fig. 4A, top panel). Positive selection in- duced BrdUϩ DP thymocytes to convert into BrdUϩ CD4SP and BrdUϩ CD8SP thymocytes, with their conversion rates indicated by the slopes of their linear regression lines (Fig. 4A, middle and bottom panels) (40). Note that CD4SP and CD8SP thymocytes in this experiment were also gated to be TCR␤high cells. Although BrdUϩ DP thymocytes converted into BrdUϩ CD4SP (TCR␤high) thymocytes at equivalent rates in CD8.4 and B6 mice (Fig. 4A, middle panel), BrdUϩ DP thymocytes converted into BrdUϩ CD8SP (TCR␤high) thymocytes at significantly different rates in CD8.4 and B6 mice (Fig. 4A, bottom panel). As indicated by the slopes of their linear regression lines, the conversion rate of BrdUϩ DP thymocytes into BrdUϩ CD8SP thymocytes was 2.5-

fold greater in CD8.4 mice than in B6 mice (Fig. 4A, bottom Downloaded from panel). Thus, increased CD8.4 coreceptor signaling by the CD4 cytosolic tail specifically increased positive selection of CD8ϩ T cells without effecting positive selection of CD4ϩ T cells. The efficiency of positive selection is defined as the relative number of DP thymocytes signaled to undergo further differenti-

ation and can be quantitated in BrdU-labeling experiments as con- http://www.jimmunol.org/ version efficiency (40). In B6 mice, the conversion efficiency of DP thymocytes into CD4SP and CD8SP thymocytes was 3 and 1%, respectively, indicating that 3% of B6 DP thymocytes were positively selected to differentiate into CD4ϩ T cells while only 1% of B6 DP thymocytes were positively selected to differentiate ϩ FIGURE 5. Stronger signaling CD8.4 coreceptors increase the selection into CD8 T cells (Fig. 4B). However, in CD8.4 mice, the con- of MHC-I-specific thymocytes but do not alter their lineage choice. Thy- version efficiency of DP thymocytes into CD4SP and CD8SP thy- mocytes and LN cells from B6, CD8.4(MHC-IIo), and LM(MHC-IIo) mice mocytes was ϳ3% for each T cell subset (Fig. 4B), revealing that were analyzed by multicolor flow cytometry. CD4 vs CD8 contour plots increased CD8 coreceptor signaling by the CD4 cytosolic tail by guest on September 27, 2021 are shown (upper panels), and the numbers inside each box indicate the quantitatively increased the efficiency of CD8ϩ T cell positive se- percentage of cells falling within that box. Numbers below the contour lection to the point that it was equivalent to that of CD4ϩ T cell plots indicate the total number of cells obtained. The bar graph displays the positive selection. total number (ϮSEM) of CD4ϩ and CD8ϩ LN T cells obtained from CD8.4(MHC-IIo) and LM(MHC-IIo) mice (bottom panel). Five to seven Cellular function of CD8.4ϩ T cells mice in each group were analyzed. ϩ ϩ CD8 T cells function as T-killer (Tk) cells and CD4 T cells function as Th cells in normal B6 mice. To determine whether stronger signaling CD8.4 coreceptors altered these cellular func- Qa2ϩ thymocytes) and therefore displayed a CD4:CD8 ratio of Ͻ1 tions in CD8.4 mice, we measured CD8 Tk function by assessing (Fig. 3A). Assessment of peripheral T cells in the spleen and lymph intracellular expression of granzyme B and we measured CD4 Th node (LN) of CD8.4 mice revealed precisely the same finding, namely function by assessing TCR induced up-regulation of CD40L ex- that CD8.4 mice had more CD8ϩ than CD4ϩ T cells in the periphery pression (Fig. 4C). We found that expression of stronger signaling and therefore also displayed a CD4:CD8 ratio of Ͻ1 (Fig. 3A). CD8.4 coreceptors had no effect on cellular function, as CD8 T Reversal of the normal CD4:CD8 ratio in every lymphoid com- cells from CD8.4 and B6 mice stained identically for granzyme B partment in CD8.4 mice might have resulted from reduced CD4ϩ (Fig. 4C), and anti-TCR stimulation failed to induce CD8 T cells T cell numbers, increased CD8ϩ T cell numbers, or both. Assess- in either CD8.4 or B6 mice to up-regulate CD40L (CD154) ex- ment of CD4ϩ and CD8ϩ T cell frequencies in CD8.4 mice re- pression, even though both T cell populations were stimulated by vealed that, compared with WT littermates, CD4ϩ T cell frequen- anti-TCR to up-regulate CD69 (Fig. 4C). cies were reduced and CD8ϩ T cell frequencies were increased (Fig. 3B, right). However, assessment of absolute T cell numbers Does enhanced CD8 coreceptor signaling affect lineage choice? revealed that CD8.4 coreceptors had no effect on CD4ϩ T cell Finally, we wished to determine whether strengthened signaling by numbers as absolute numbers of CD4ϩ T cells were equivalent in endogenously encoded CD8.4 coreceptors affected the lineage all lymphoid compartments in CD8.4 and WT littermate mice (Fig. choice of MHC-I-specific thymocytes. Notably, the strength of sig- 3B, left). In contrast, CD8.4 coreceptors specifically increased ab- nal hypothesis of lineage choice originated to explain why a co- solute numbers of CD8ϩ T cells in every lymphoid compartment receptor transgene encoding a chimeric CD8/CD4 coreceptor pro- of CD8.4 mice (Fig. 3B, left). tein identical with the CD8.4 protein appeared to direct the We thought it likely that the increased number of mature CD8ϩ differentiation of MHC-I-specific thymocytes into CD4ϩ T cells T cells in CD8.4 mice was due to increased positive selection of (20). To determine the effect of endogenously encoded CD8.4 co- CD8 lineage thymocytes by stronger signaling CD8.4 coreceptors. receptors on the lineage choice of MHC-I-specific thymocytes, we Consequently, we wished to quantitate and compare positive se- bred the CD8.4 gene locus into A␤Ϫ/Ϫ mice deficient in MHC-II 6620 POSITIVE SELECTION AND LINEAGE CHOICE Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 6. Effect of stronger signaling CD8.4 coreceptors on positive selection and lineage choice of thymocytes selected by transgenic F5 TCR. A, The F5 TCR transgene was bred into RAGo mice expressing either CD8␣ or CD8.4 coreceptors to generate F5(RAGo)CD8␣ and F5(RAGo)CD8.4 mice. Thymocytes (left panels) and lymph node cells (right panels) from these mice were analyzed by multicolor flow cytometry and analyzed for expression of CD4 vs CD8␣. The numbers inside each box indicate the percentage of cells within that box. The numbers under each strain name indicate the total numbers of cell obtained. B, The F5 TCR transgene (TCR-V␣4V␤11) was bred into RAGϩ mice expressing either CD8␣ or CD8.4 coreceptors to generate F5 CD8␣ and F5 CD8.4 RAGϩ mice. Thymocytes (left panels) and lymph node cells (right panels) from these mice were subjected to multicolor flow cytometry and analyzed for expression of CD4, CD8␣, TCR-V␤11, and TCR-V␣2ϩ3ϩ8ϩ11. Note that mAb to TCR-V␤11 detects the transgenic F5 TCR ␤-chain, but that the mixture of mAbs to The Journal of Immunology 6621 Downloaded from

FIGURE 7. Schematic representations of findings and the kinetic signaling model. Schematic representation that CD8.4 coreceptors improve the efficiency of positive selection of thymocytes expressing the F5 TCR transgene. All preselection DP thymocytes in F5(RAGo) mice obligatorily express the MHC-I-specific transgenic F5 TCR. Because positive selection is inherently inefficient, only a small minority of F5-expressing preselection DP http://www.jimmunol.org/ thymocytes are signaled to undergo positive selection into intermediate thymocytes and to differentiate into CD8SP thymocytes. However, the numberof F5-expressing DP thymocytes that can be signaled to undergo positive selection is influenced by the signaling intensity of CD8 coreceptors, with the effect that stronger signaling CD8.4 coreceptors increase the number of F5-expressing DP thymocytes that are signaled to undergo positive selection into intermediate thymocytes and to differentiate into CD8SP thymocytes. expression (MHC-IIo), because in these mice, only MHC-I-spe- mocytes into CD4 lineage T cells, stronger signaling CD8.4 core- cific thymocytes are signaled to undergo positive selection and to ceptors increased by Ͼ2.5-fold the number of F5 CD8ϩ T cells

differentiate into mature T cells. As expected, LN T cells from positively selected in the thymus and present in peripheral LNs by guest on September 27, 2021 MHC-IIo mice expressing WT CD8␣ coreceptors were almost ex- (Fig. 6A). It might be noted that CD8.4 coreceptors also increased clusively CD8ϩ, with few CD4ϩ T cells (Fig. 5). But, in direct the frequency and number of CD4ϩ8low intermediate cells among conflict with the strength-of-signaling hypothesis, LN T cells from F5(RAGo) TCR transgenic thymocytes, which are the immediate MHC-IIo mice expressing CD8.4 coreceptors were also almost ex- precursors of CD8SP thymocytes (37, 38). clusively CD8ϩ, with few CD4ϩ T cells (Fig. 5). In fact, CD8.4 Because these results directly contradicted the original experi- coreceptors resulted in greater absolute numbers of MHC-I-spe- mental observations on which the strength-of-signal hypothesis cific CD8ϩ T cells than did WT CD8 coreceptors (Fig. 5, bottom was based (20), we repeated this experiment with RAG-replete panel). The few CD4ϩ T cells found in both CD8.4 and WT lit- F5(RAGϩ) mice that were used in the original coreceptor trans- termate MHC-IIo mice were presumably NKT cells with specific- gene experiments, with the key exception that chimeric CD8/CD4 ity for minor MHC-I determinants (41). coreceptors were now encoded by endogenous rather than trans- To assess the effect of CD8.4 coreceptors on lineage choice by genic coreceptor genes (Fig. 6B). Because these F5 thymocytes thymocytes expressing an MHC-I-specific transgenic TCR, we in- were RAGϩ, F5 transgenic mice with CD8.4 or CD8␣ coreceptors troduced the F5 TCR transgene that encodes TCR-V␣4V␤11 re- contained a few CD4ϩ T cells in the thymus and LN. However, ceptors into CD8.4RAGo mice (Fig. 6A). We specifically chose the unlike CD8ϩ T cells in these F5 transgenic mice, these few CD4ϩ relatively high-affinity F5 TCR transgene for these experiments T cells had diminished surface expression of transgenic TCR- because the strength-of-signal hypothesis was originally formu- V␤11 proteins and expressed nontransgenic TCR-V␣ proteins lated to explain the differentiation of F5 thymocytes into CD4ϩ T (Fig. 6B), indicating that they had not been selected by F5 trans- cells in coreceptor transgenic mice expressing chimeric CD8/CD4 genic TCR. Importantly, stronger signaling CD8.4 coreceptors sig- coreceptor molecules (20). Importantly, our experiments with en- nificantly increased the number of F5 CD8ϩ T cells generated in dogenously encoded CD8.4 coreceptors revealed that F5(RAGo) CD8.4 mice compared with WT CD8␣ mice (Fig. 6B). TCR transgenic thymocytes differentiated exclusively into CD8ϩ Finally, we performed BrdU-labeling experiments to quanti- T cells even in the presence of stronger signaling CD8.4 corecep- tate the effect of CD8 coreceptor signal strength on the effi- tors (Fig. 6A). In fact, rather than redirecting F5 transgenic thy- ciency of CD8ϩ T cell positive selection in F5(RAGo) TCR

TCR-V␣2ϩ3ϩ8ϩ11 detects endogenous TCR ␣- chains (not transgenic F5 TCR-V␣4 chains for which there is no available mAb). C, Efficiency of MHC-I-positive selection in F5(RAGo) TCR-transgenic mice. F5(RAGo)CD8␣ and F5(RAGo)CD8.4 mice were continuously labeled with BrdU as described in Fig. 4. One to three mice were used in each group at each time point. The efficiency by which DP thymocytes converted into CD8SP thymocytes was calculated by dividing the slope (ϮSE) of the CD8SP regression line by the slope (ϮSE) of the DP regression line, as described (40). The conversion efficiencies of DP into CD8SP thymocytes in F5(RAGo)CD8␣ mice (Ⅺ) and F5(RAGo)CD8.4 mice (f) were compared. Values of p were determined by the two-tailed Student’s t test. 6622 POSITIVE SELECTION AND LINEAGE CHOICE Downloaded from

FIGURE 8. Kinetic signaling model of lineage determination. The kinetic signaling model postulates that positive selection and lineage commitment are distinct and separable developmental events. In this model, thymocytes signaled to undergo positive selection make their lineage choice by assessing the CD8 dependence of their TCR. According to the kinetic signaling model, preselection DP thymocytes are first signaled by TCR plus coreceptor coengagements to undergo positive selection which invariably terminates CD8 gene expression. Note that CD8 gene expression is terminated by positive selection signals regardless of the MHC specificity of the TCR and regardless of the coreceptor used. Termination of CD8 gene expression transcriptionally converts signaled DP thymocytes into CD4ϩ8Ϫ ϩ Ϫ ϩ Ϫ intermediate thymocytes, and it is in CD4 8 intermediate thymocytes that lineage choice occurs. If TCR-mediated positive selection signals persist in CD4 8 http://www.jimmunol.org/ intermediate thymocytes despite absent CD8 gene expression, intermediate thymocytes differentiate into CD4ϩ T cells. However, if TCR-mediated positive selection signals cease in CD4ϩ8Ϫ intermediate thymocytes, intermediate thymocytes differentiate into CD8ϩ T cells. transgenic mice (Fig. 6C). These experiments revealed two im- during positive selection is a quantitative regulator of positive portant points. First, they confirmed that MHC-I-specific posi- selection and a major determinant of the CD4:CD8 ratio, but it tive selection of DP into CD8SP thymocytes was 20-fold more does not determine CD4/CD8 lineage choice. efficient when DP thymocytes expressed the F5 TCR transgene than when they expressed a diversity of polyclonal TCR (com- Regulating positive selection and the CD4:CD8 ratio by guest on September 27, 2021 pare Figs. 6C and 4B). Second, and more importantly, they In this study, replacement of the CD8 cytosolic tail with the revealed that positive selection of DP into CD8SP thymocytes stronger signaling tail of CD4 increased the number of CD8ϩ T signaled by F5 transgenic TCR was increased more than 2-fold cells generated in the thymus as a direct result of more efficient by CD8.4 coreceptors compared with CD8␣ coreceptors (Fig. positive selection. We measured positive selection efficiency in 6C). Remarkably, Ͼ60% of F5 DP thymocytes were positively BrdU continuous labeling experiments and determined that the ϩ selected in CD8.4 mice to differentiate into CD8 T cells (Fig. efficiency of MHC-I-specific CD8ϩ T cell positive selection 6C). Thus, stronger signaling CD8.4 coreceptors increased the was improved 3-fold (3 vs 1%) by stronger signaling CD8.4 efficiency of MHC-I-specific T cell positive selection even for coreceptors compared with WT CD8␣ coreceptors. In fact, DP thymocytes expressing a single TCR specificity. CD8.4 coreceptors improved positive selection of MHC-I-spe- We conclude that stronger signaling CD8.4 coreceptors do not cific CD8ϩ T cells to the point that it occurred with equal ef- effect CD8 lineage choice by MHC-I-specific thymocytes, but they ficiency as MHC-II-specific CD4ϩ T cells. It is likely that stron- ϩ do increase the efficiency of CD8 T cell positive selection and ger signaling CD8.4 coreceptors induced thymocytes with ϩ promote the generation of greater numbers of CD8 T cells in the lower affinity TCR to undergo positive selection than were in- thymus, even for thymocytes expressing a single transgenic TCR duced by WT CD8␣ coreceptors. However, selection of thymo- specificity (schematically illustrated in Fig. 7). cytes with lower affinity TCR does not fully explain the effect of CD8.4 coreceptors we observed on positive selection, as Discussion CD8.4 coreceptors also increased positive selection of F5 TCR This study documents that the signaling intensity of CD8 co- transgenic thymocytes expressing a single, fixed TCR specific- receptors quantitatively regulates the efficiency of CD8-depen- ity. That is, CD8.4 coreceptors increased positive selection of dent positive selection and thereby determines the number of MHC-I-specific thymocytes expressing diverse TCR with a CD8ϩ T cells generated in the thymus, but it does not influence spectrum of affinities and also increased positive selection of F5 CD4/CD8 lineage fate. The present results were remarkably thymocytes expressing only a single TCR specificity. So how clear cut and definitive because they were obtained in a phys- might stronger CD8.4 coreceptors increase CD8ϩ T cell posi- iologic in vivo model using gene knockin mice whose endog- tive selection when DP thymocytes all express a single TCR enous CD8␣ gene locus had been re-engineered to encode CD8 specificity? coreceptors expressing the stronger signaling tail of CD4. The Our results indicate that stronger signaling CD8.4 corecep- in vivo expression pattern of CD8.4 coreceptors was identical tors reduce the TCR signaling threshold required for positive with that of WT CD8␣ coreceptors and therefore avoided the selection, with the result that more DP thymocytes receive pos- expression artifacts that plagued CD8 coreceptor transgenic itive selection signals—even when all thymocytes express a mice. Thus, this study reveals that coreceptor signal strength single transgenic TCR (schematized in Fig. 7). Our reasoning is The Journal of Immunology 6623 based on the fact that positive selection is inherently inefficient receptor signals transduced by the CD4 cytosolic tail do not determine (40, 42–44), because DP thymocytes express relatively few sur- CD4/CD8 lineage choice. face TCR, and these TCR transduce intracellular signals only Because the strength-of-signal model provides a widely ac- when they engage intrathymic ligands with the appropriate af- cepted explanation for CD4/CD8 lineage choice that remains finity, and because TCR signal transduction requires the protein highly popular (18), it is worth discussing two issues related to the tyrosine kinase Lck that is present in limiting amounts and is strength of signal model in greater detail (1). First, the strength- associated with CD4 and CD8 coreceptors, making TCR signal of-signal model was originally proposed to explain observations in transduction in DP thymocytes highly coreceptor dependent mice expressing a coreceptor transgene that encoded a chimeric (14, 15). As a consequence, TCR signaling is frequently unsuc- CD8/CD4 molecule in which the cytosolic tail of CD8␣ was re- cessful in DP thymocytes even when their TCR have appropri- placed by the stronger signaling tail of CD4 (20). In these mice, ate specificity for intrathymic ligands, with the result that rel- transgenic CD8/CD4 chimeric coreceptors appeared to redirect the atively few eligible DP thymocytes are successfully signaled to differentiation of MHC-I-specific thymocytes, especially those ex- undergo positive selection. Because Lck preferentially associ- pressing the F5 transgenic TCR, into CD4ϩ T cells. Indeed, the ates with the CD4 tail rather than the CD8 tail, DP thymocytes fundamental observation on which the strength of signal model are more likely to be successfully signaled to undergo positive was based was that increased signaling intensity by transgenically selection signal when their TCR are coengaged with corecep- encoded CD8/CD4 chimeric coreceptors redirected F5 TCR trans- tors bearing the CD4 tail than the CD8 tail, regardless of the genic thymocytes to differentiate into CD4ϩ T cells (20). How- diversity or specificity of their TCR (schematized in Fig. 7). ever, in contrast to these previous observations in coreceptor trans- This study fulfills predictions that we previously based on ex- genic mice, our study demonstrates that when the identical CD8/ Downloaded from periments in CD8 coreceptor transgenic mice (24), namely that the CD4 chimeric coreceptor proteins are encoded by endogenous CD4:CD8 ratio might reflect the relative signaling intensities of CD8␣ genes rather than extrinsic transgenes, F5 TCR transgenic CD4 and CD8 coreceptors. Importantly, this study significantly thymocytes differentiate exclusively into CD8ϩ T cells. The crit- extends our original hypothesis by documenting that CD8 core- ϩ ical difference between endogenously encoded CD8 coreceptors ceptor signal strength influences the number of CD8 T cells gen- and transgenically encoded CD8 coreceptors is that TCR-mediated erated even when all thymocytes express only a single TCR spec- positive selection signals are unable to terminate coreceptor trans- http://www.jimmunol.org/ ificity. Moreover, this study documents the mechanism by which gene expression, leading to persistent expression of transgenic this occurs, namely by improving the positive selection efficiency CD8 coreceptor proteins throughout positive selection that abnor- of MHC-I-specific thymocytes and thereby increasing the number mally prolongs CD8-dependent TCR signaling (discussed in more of CD8ϩ T cells generated. Because CD4 molecules normally detail below) (2). Second, the strength of signal model requires transduce stronger coreceptor signals than CD8 molecules, core- that strong signals transduced by the CD4 tail direct thymocytes to ceptor signal strength is a major reason why mammalian immune ϩ ϩ differentiate into CD4 T cells, regardless of the MHC specificity systems normally contain more MHC-II-specific CD4 T cells ϩ of their TCR. Conflicting with this requirement, CD8.4 corecep- than MHC-I-specific CD8 T cells. Of course, other factors such

tors—which signaled via the CD4 tail and which signaled with the by guest on September 27, 2021 as TCR specificity, quantitative differences in MHC-I and MHC-II same intensity as CD4 coreceptors—failed to redirect MHC-I-spe- expression, and peptide diversity may also contribute to the CD4: cific thymocytes to differentiate into CD4ϩ T cells. Thus, this CD8 ratio (45). But we think that the CD4:CD8 ratio primarily study directly contradicts the original observation that prompted reflects the relative signaling intensities of CD4 vs CD8 corecep- tors during positive selection. formulation of the strength-of-signal model and directly contra- Because both CD8.4 and CD4 coreceptors expressed the CD4 dicts a necessary requirement of the strength-of-signal model. cytosolic tail and both signaled with equal intensity, why was the The strength-of-signal model has recently been reformulated as CD4:CD8 ratio in CD8.4 mice less than 1 rather than equal to 1? a “duration-of-signal” model (23) in which signal strength is pro- DP thymocytes express 4-fold more CD8 than CD4 surface mol- posed to affect both the intensity and duration of positive selection signaling in DP thymocytes such that long/strong signals lead to ecules (14) so that the total amount of Lck associated with CD8.4 ϩ ϩ coreceptors is greater than the total amount of Lck associated with CD4 T cell differentiation and short/weak signals lead to CD8 CD4 coreceptors in each CD8.4 DP thymocyte, providing an ex- T cell differentiation. As described in this model (23), the duration/ planation for why the CD4:CD8 ratio is Ͻ1 in CD8.4 mice. intensity of MHC-II-specific CD4/TCR signals is greater than that of MHC-I-specific CD8/TCR signals because of the greater asso- CD4/CD8 lineage choice ciation of intracellular Lck with CD4 than CD8 coreceptors in DP This study refutes two classical models of CD4/CD8 lineage choice, thymocytes. Importantly, this reformulation is also contradicted by the instructional model and the strength of signal model. The instruc- this study because Lck association with CD4 coreceptors is not tional model postulates that coreceptor signals transduced by the CD4 greater than with CD8.4 coreceptors in CD8.4 DP thymocytes be- cytosolic tail specifically direct thymocytes to differentiate into CD4ϩ cause both coreceptors expressed the identical CD4 cytosolic tail. T cells, whereas coreceptor signals transduced by the CD8 cytosolic Nevertheless, MHC-I-specific CD8.4/TCR signals still only gen- ϩ tail specifically direct thymocytes to differentiate into CD8ϩ T cells erated CD8 T cells and MHC-II-specific CD4/TCR signals still ϩ (34). The present demonstration that MHC-I-specific thymocytes dif- only generated CD4 T cells in the CD8.4 thymus. Thus, by dem- ferentiated into CD8 lineage T cells regardless of the identity of the onstrating that CD4/CD8 lineage choice is not determined by the cytosolic tail on CD8 coreceptors directly refutes the instructional extent of Lck/coreceptor associations, this study directly contra- model, a conclusion also drawn by others (25). Similarly, the strength- dicts the duration-of-signal model (23). of-signal model postulates that strong coreceptor signals (such as We think that CD4/CD8 lineage choice is best described by the those transduced by the CD4 cytosolic tail) direct thymocytes to dif- kinetic signaling model which postulates that positive selection ferentiate into CD4ϩ T cells, whereas weak coreceptor signals (such and lineage choice are developmentally distinct events (Fig. 8) (4, as those transduced by the CD8␣ cytosolic tail) direct thymocytes to 46, 47). In the kinetic signaling model, TCR plus coreceptor me- differentiate into CD8ϩ T cells (19–23). In direct contradiction to the diated positive selection signals, regardless of their MHC speci- strength-of-signal model, this study documents that strong CD8.4 co- ficity, terminate CD8 coreceptor gene expression causing signaled 6624 POSITIVE SELECTION AND LINEAGE CHOICE

ϩ Ϫ DP thymocytes to transcriptionally convert into CD4 8 interme- 8. He, X., X. He, V. P. Dave, Y. Zhang, X. Hua, E. Nicolas, W. Xu, B. A. Roe, and diate thymocytes; lineage direction is then determined by whether D. J. Kappes. 2005. The zinc finger transcription factor Th-POK regulates CD4 ϩ Ϫ versus CD8 T-cell lineage commitment. Nature 433: 826–833. positive selection signaling persists or ceases in CD4 8 inter- 9. Sato, T., S. Ohno, T. Hayashi, C. Sato, K. Kohu, M. Satake, and S. Habu. 2005. mediate thymocytes. If TCR-mediated positive selection signaling Dual functions of Runx proteins for reactivating CD8 and silencing CD4 at the persists despite absent CD8 coreceptor expression, intermediate commitment process into CD8 thymocytes. Immunity 22: 317–328. ϩ 10. Sun, G., X. Liu, P. Mercado, S. R. Jenkinson, M. Kypriotou, L. Feigenbaum, thymocytes differentiate into CD4 T cells; if TCR-mediated pos- P. Galera, and R. Bosselut. 2005. The zinc finger protein cKrox directs CD4 itive selection signaling ceases, intermediate thymocytes re-initiate lineage differentiation during intrathymic T cell positive selection. Nat. Immunol. CD8 gene expression and differentiate into CD8ϩ T cells. In this 6: 373–381. 11. Guidos, C. J., I. L. Weissman, and B. Adkins. 1989. Intrathymic maturation of way, positively selected thymocytes determine lineage choice by murine T from CD8ϩ precursors. Proc. Natl. Acad. Sci. USA 86: assessing the CD8 dependence of their TCR. For example, CD8- 7542–7546. independent positive selection signals mediated by MHC-II-spe- 12. Veillette, A., M. A. Bookman, E. M. Horak, and J. B. Bolen. 1988. The CD4 and ϩ Ϫ CD8 T cell surface antigens are associated with the internal membrane tyrosine- cific TCR would persist in CD4 8 intermediate thymocytes de- protein kinase p56lck. Cell 55: 301–308. spite absent CD8 gene expression, resulting in differentiation into 13. Barber, E. K., J. D. Dasgupta, S. F. Schlossman, J. M. Trevillyan, and C. E. Rudd. ϩ 1989. The CD4 and CD8 antigens are coupled to a protein-tyrosine kinase CD4 T cells, whereas CD8-dependent positive selection signals lck ϩ Ϫ (p56 ) that phosphorylates the CD3 complex. Proc. Natl. Acad. Sci. USA 86: mediated by MHC-I-specific TCR would cease in CD4 8 inter- 3277–3281. mediate thymocytes because of absent CD8 gene expression, re- 14. Wiest, D. L., L. Yuan, J. Jefferson, P. Benveniste, M. Tsokos, R. D. Klausner, ϩ L. H. Glimcher, L. E. Samelson, and A. Singer. 1993. Regulation of T cell sulting in differentiation into CD8 T cells. Notably, the kinetic ϩ ϩ receptor expression in immature CD4 CD8 thymocytes by p56lck tyrosine ki- signaling model can explain why MHC-I-specific thymocytes were nase: basis for differential signaling by CD4 and CD8 in immature thymocytes ϩ originally observed to differentiate into CD4 T cells in CD8 co- expressing both coreceptor molecules. J. Exp. Med. 178: 1701–1712. Downloaded from 15. Wiest, D. L., J. M. Ashe, R. Abe, J. B. Bolen, and A. Singer. 1996. TCR acti- receptor transgenic mice encoding chimeric CD8/CD4 coreceptor vation of ZAP70 is impaired in CD4ϩCD8ϩ thymocytes as a consequence of proteins (20): because CD8 coreceptor transgene expression was intrathymic interactions that diminish available p56lck. Immunity 4: 495–504. regulated by heterologous transcriptional control elements (rather 16. Veillette, A., J. C. Zuniga-Pflucker, J. B. Bolen, and A. M. Kruisbeek. 1989. ␣ Engagement of CD4 and CD8 expressed on immature thymocytes induces acti- than endogenous CD8 transcriptional control elements), tran- vation of intracellular tyrosine phosphorylation pathways. J. Exp. Med. 170: scription of transgenic CD8 coreceptors was not terminated by 1671–1680.

TCR-mediated positive selection signals, causing continuous CD8 17. Kim, P. W., Z. Y. Sun, S. C. Blacklow, G. Wagner, and M. J. Eck. 2003. A zinc http://www.jimmunol.org/ expression and persistent signaling by MHC-I-specific TCR which clasp structure tethers Lck to T cell coreceptors CD4 and CD8. Science 301: ϩ 1725–1728. induced intermediate thymocytes to differentiate into CD4 T 18. He, X., and D. J. Kappes. 2006. CD4/CD8 lineage commitment: light at the end cells. Indeed, increasing evidence supports kinetic signaling as the of the tunnel? Curr. Opin. Immunol. 18: 135–142. mechanism of CD4/CD8 lineage determination in the thymus (24, 19. Matechak, E. O., N. Killeen, S. M. Hedrick, and B. J. Fowlkes. 1996. MHC class II-specific T cells can develop in the CD8 lineage when CD4 is absent. Immunity 46, 48–50). 4: 337–347. 20. Itano, A., P. Salmon, D. Kioussis, M. Tolaini, P. Corbella, and E. Robey. 1996. Conclusions The cytoplasmic domain of CD4 promotes the development of CD4 lineage T cells. J. Exp. Med. 183: 731–741. This study reveals that coreceptor signal strength is a key regulator 21. Hernandez-Hoyos, G., S. J. Sohn, E. V. Rothenberg, and J. Alberola-Ila. 2000. of positive selection, determining positive selection efficiency and Lck activity controls CD4/CD8 T cell lineage commitment. Immunity 12: by guest on September 27, 2021 thereby the CD4:CD8 ratio, but coreceptor signal strength does not 313–322. 22. Basson, M. A., U. Bommhardt, M. S. Cole, J. Y. Tso, and R. Zamoyska. 1998. determine CD4/CD8 lineage choice. That coreceptor signal CD3 ligation on immature thymocytes generates antagonist-like signals appro- strength influences positive selection without affecting lineage priate for CD8 lineage commitment, independently of T cell receptor specificity. choice is precisely in accord with the kinetic signaling concept that J. Exp. Med. 187: 1249–1260. 23. Germain, R. N. 2002. T-cell development and the CD4-CD8 lineage decision. positive selection and lineage choice are discrete and separable Nat. Rev. Immunol. 2: 309–322. events during thymocyte differentiation. 24. Bosselut, R., L. Feigenbaum, S. O. Sharrow, and A. Singer. 2001. Strength of signaling by CD4 and CD8 coreceptor tails determines the number but not the lineage direction of positively selected thymocytes. Immunity 14: 483–494. Acknowledgments 25. Kao, H., and P. M. Allen. 2005. An antagonist peptide mediates positive selection We thank Larry Granger and Tony Adams for expert flow cytometry, and and CD4 lineage commitment of MHC class II-restricted T cells in the absence Drs. Yousuke Takahama, Remy Bosselut, and Hyun Park for critically of CD4. J. Exp. Med. 201: 149–158. reading the manuscript. 26. Shores, E. W., T. Tran, A. Grinberg, C. L. Sommers, H. Shen, and P. E. Love. 1997. Role of the multiple T cell receptor (TCR)-␨ chain signaling motifs in selection of the T cell repertoire. J. Exp. Med. 185: 893–900. Disclosures 27. Mamalaki, C., J. Elliott, T. Norton, N. Yannoutsos, A. R. Townsend, P. Chandler, The authors have no financial conflict of interest. E. Simpson, and D. Kioussis. 1993. Positive and negative selection in transgenic mice expressing a T-cell receptor specific for influenza nucleoprotein and endog- References enous superantigen. Dev. Immunol. 3: 159–174. 28. Fung-Leung, W. P., M. W. Schilham, A. Rahemtulla, T. M. Kundig, 1. Kaye, J., M. L. Hsu, M. E. Sauron, S. C. Jameson, N. R. Gascoigne, and ϩ M. Vollenweider, J. Potter, W. van Ewijk, and T. W. Mak. 1991. CD8 is needed for S. M. Hedrick. 1989. Selective development of CD4 T cells in transgenic mice development of cytotoxic T cells but not helper T cells. Cell 65: 443–449. expressing a class II MHC-restricted antigen receptor. Nature 341: 746–749. 29. Crooks, M. E., and D. R. Littman. 1994. Disruption of T positive and 2. Sha, W. C., C. A. Nelson, R. D. Newberry, D. M. Kranz, J. H. Russell, and negative selection in mice lacking the CD8 ␤ chain. Immunity 1: 277–285. D. Y. Loh. 1988. Selective expression of an antigen receptor on CD8-bearing T 30. Grusby, M. J., R. S. Johnson, V. E. Papaioannou, and L. H. Glimcher. 1991. lymphocytes in transgenic mice. Nature 335: 271–274. ϩ 3. Teh, H. S., P. Kisielow, B. Scott, H. Kishi, Y. Uematsu, H. Bluthmann, and Depletion of CD4 T cells in major histocompatibility complex class II-deficient H. von Boehmer. 1988. Thymic major histocompatibility complex antigens and mice. Science 253: 1417–1420. the ␣␤ T-cell receptor determine the CD4/CD8 phenotype of T cells. Nature 335: 31. Lakso, M., J. G. Pichel, J. R. Gorman, B. Sauer, Y. Okamoto, E. Lee, F. W. Alt, 229–233. and H. Westphal. 1996. Efficient in vivo manipulation of mouse genomic se- 4. Singer, A., and R. Bosselut. 2004. CD4/CD8 coreceptors in thymocyte develop- quences at the zygote stage. Proc. Natl. Acad. Sci. USA 93: 5860–5865. ment, selection, and lineage commitment: analysis of the CD4/CD8 lineage de- 32. Soriano, P., C. Montgomery, R. Geske, and A. Bradley. 1991. Targeted disrup- cision. Adv. Immunol. 83: 91–131. tion of the c-src proto-oncogene leads to osteopetrosis in mice. Cell 64: 693–702. 5. Ellmeier, W., S. Sawada, and D. R. Littman. 1999. The regulation of CD4 and 33. Zamoyska, R., P. Derham, S. D. Gorman, P. von Hoegen, J. B. Bolen, CD8 coreceptor gene expression during T cell development. Annu. Rev. Immunol. A. Veillette, and J. R. Parnes. 1989. Inability of CD8 ␣Ј polypeptides to associate 17: 523–554. with p56lck correlates with impaired function in vitro and lack of expression in 6. Kioussis, D., and W. Ellmeier. 2002. Chromatin and CD4, CD8A and CD8B gene vivo. Nature 342: 278–281. expression during thymic differentiation. Nat. Rev. Immunol. 2: 909–919. 34. Seong, R. H., J. W. Chamberlain, and J. R. Parnes. 1992. Signal for T-cell dif- 7. Bosselut, R. 2004. CD4/CD8-lineage differentiation in the thymus: from nuclear ferentiation to a CD4 cell lineage is delivered by CD4 transmembrane region effectors to membrane signals. Nat. Rev. Immunol. 4: 529–540. and/or cytoplasmic tail. Nature 356: 718–720. The Journal of Immunology 6625

35. Dutz, J. P., C. J. Ong, J. Marth, and H. S. Teh. 1995. Distinct differentiative stages 44. Huesmann, M., B. Scott, P. Kisielow, and H. von Boehmer. 1991. Kinetics and of CD4ϩCD8ϩ thymocyte development defined by the lack of coreceptor binding efficacy of positive selection in the thymus of normal and T cell receptor trans- in positive selection. J. Immunol. 154: 2588–2599. genic mice. Cell 66: 533–540. 36. Azzam, H. S., A. Grinberg, K. Lui, H. Shen, E. W. Shores, and P. E. Love. 1998. 45. Myrick, C., R. DiGuisto, J. DeWolfe, E. Bowen, J. Kappler, P. Marrack, and CD5 expression is developmentally regulated by T cell receptor (TCR) signals E. K. Wakeland. 2002. Linkage analysis of variations in CD4:CD8 T cell subsets and TCR avidity. J. Exp. Med. 188: 2301–2311. between C57BL/6 and DBA/2. Genes Immun. 3: 144–150. 37. Suzuki, H., J. A. Punt, L. G. Granger, and A. Singer. 1995. Asymmetric signaling requirements for thymocyte commitment to the CD4ϩ versus CD8ϩ T cell lin- 46. Brugnera, E., A. Bhandoola, R. Cibotti, Q. Yu, T. I. Guinter, Y. Yamashita, eages: a new perspective on thymic commitment and selection. Immunity 2: S. O. Sharrow, and A. Singer. 2000. Coreceptor reversal in the thymus: signaled CD4ϩ8ϩ thymocytes initially terminate CD8 transcription even when differen- 413–425. ϩ 38. Lundberg, K., W. Heath, F. Kontgen, F. R. Carbone, and K. Shortman. 1995. tiating into CD8 T cells. Immunity 13: 59–71. Intermediate steps in positive selection: differentiation of CD4ϩ8int TCRint thy- 47. Singer, A. 2002. New perspectives on a developmental dilemma: the kinetic mocytes into CD4Ϫ8ϩTCRhi thymocytes. J. Exp. Med. 181: 1643–1651. signaling model and the importance of signal duration for the CD4/CD8 lineage 39. Ramsdell, F., M. Jenkins, Q. Dinh, and B. J. Fowlkes. 1991. The majority of decision. Curr. Opin. Immunol. 14: 207–215. ϩ Ϫ CD4 8 thymocytes are functionally immature. J. Immunol. 147: 1779–1785. 48. Bosselut, R., T. I. Guinter, S. O. Sharrow, and A. Singer. 2003. Unraveling a 40. Itano, A., and E. Robey. 2000. Highly efficient selection of CD4 and CD8 lineage revealing paradox: why major histocompatibility complex I-signaled thymocytes thymocytes supports an instructive model of lineage commitment. Immunity 12: “paradoxically” appear as CD4ϩ8lo transitional cells during positive selection of 383–389. CD8ϩ T cells. J. Exp. Med. 197: 1709–1719. 41. Bendelac, A., M. N. Rivera, S. H. Park, and J. H. Roark. 1997. Mouse CD1- specific NK1 T cells: development, specificity, and function. Annu. Rev. Immu- 49. Sarafova, S. D., B. Erman, Q. Yu, F. Van Laethem, T. Guinter, S. O. Sharrow, nol. 15: 535–562. L. Feigenbaum, K. F. Wildt, W. Ellmeier, and A. Singer. 2005. Modulation of 42. Egerton, M., R. Scollay, and K. Shortman. 1990. Kinetics of mature T-cell de- coreceptor transcription during positive selection dictates lineage fate indepen- velopment in the thymus. Proc. Natl. Acad. Sci. USA 87: 2579–2582. dently of TCR/coreceptor specificity. Immunity 23: 75–87. 43. Ernst, B., C. D. Surh, and J. Sprent. 1995. Thymic selection and cell division. 50. Liu, X., and R. Bosselut. 2004. Duration of TCR signaling controls CD4-CD8 J. Exp. Med. 182: 961–971. lineage differentiation in vivo. Nat. Immunol. 5: 280–288. Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021