Absence of CD28 Humoral Immune Responses in the Major Costimulatory Role in The Inducible Costimulator Plays the

Absence of CD28 Humoral Immune Responses in the Major Costimulatory Role in The Inducible Costimulator Plays the
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  of July 6, 2015.This information is current as Responses in the Absence of CD28Costimulatory Role in Humoral Immune The Inducible Costimulator Plays the Major Ohashi and Tak W. Mak Okada, Andrew Wakeham, Bernhard Odermatt, Pamela S.Arda Shahinian, Suzanne Plyte, Gordon S. Duncan, Hitoshi Woong-Kyung Suh, Anna Tafuri, Nancy N. Berg-Brown, 10.4049/jimmunol.172.10.59172004; 172:5917-5923; ;  J Immunol References, 21 of which you can access for f ree at: cites 48 articles This article Subscriptions is online at: The Journal of Immunology Information about subscribing to Permissions copyright permission requests at: Email Alerts free email-alerts when new articles cite this article. Sign up at: Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2004 by The American Association of 9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc., is published twice each month by The Journal of Immunology   b  y g u e  s  t   on J   ul   y 6  ,2  0 1  5 h  t   t   p :  /   /   w w w . j  i  mm un ol   . or  g /  D o wnl   o a  d  e  d f  r  om  b  y g u e  s  t   on J   ul   y 6  ,2  0 1  5 h  t   t   p :  /   /   w w w . j  i  mm un ol   . or  g /  D o wnl   o a  d  e  d f  r  om   The Inducible Costimulator Plays the Major CostimulatoryRole in Humoral Immune Responses in the Absence of CD28 1 Woong-Kyung Suh, 2 * † Anna Tafuri, 3 * † Nancy N. Berg-Brown, 4† Arda Shahinian,* † Suzanne Plyte,* † Gordon S. Duncan,* † Hitoshi Okada,* † Andrew Wakeham,* † Bernhard Odermatt, ‡ Pamela S. Ohashi, † and Tak W. Mak 2 * † CD28 plays crucial costimulatory roles in T cell proliferation, cytokine production, and germinal center response. Mice that aredeficient in the inducible costimulator (ICOS) also have defects in cytokine production and germinal center response. Because thefull induction of ICOS in activated T cells depends on CD28 signal, the T cell costimulatory capacity of ICOS in the absence of CD28 has remained unclear. We have clarified this issue by comparing humoral immune responses in wild-type, CD28 knockout(CD28 KO), and CD28-ICOS double-knockout (DKO) mice. DKO mice had profound defects in Ab responses against environ-mental Ags, T-dependent protein Ags, and vesicular stomatitis virus that extended far beyond those observed in CD28 KO mice.However, DKO mice mounted normal Ab responses against a T-independent Ag, indicating that B cell function itself was normal.Restimulated CD4  DKO T cells that had been primed in vivo showed decreased proliferation and reduced IL-4 and IL-10production compared with restimulated CD4  T cells from CD28 KO mice. Thus, in the absence of CD28, ICOS assumes themajor T cell costimulatory role for humoral immune responses. Importantly, CD28-mediated ICOS up-regulation is not essentialfor ICOS function in vivo.  The Journal of Immunology,  2004, 172: 5917–5923. A n optimal T cell response requires two types of signals(1). One is delivered by the TCR upon recognition of specific peptide-MHC complexes displayed on the sur-face of APCs. The second signal is provided by stimulatory orinhibitory coreceptors that bind to their cognate ligands on APCs.CD28 is a key stimulatory coreceptor expressed on naive T cells.Ligation of CD28 by its ligand, CD80 (B7-1) and CD86 (B7-2),augments and sustains TCR signaling and leads to increased pro-duction of IL-2, cell cycle progression, and enhanced cell survival(1). The crucial role played by CD28 in the proliferation and dif-ferentiation of T cells has been highlighted by studies of CD28knockout (KO) 5 mice. CD28-deficient T cells hypoproliferateupon antigenic stimulation, and CD28 KO mice have reduced Abresponses with impaired germinal center (GC) formation (2–4).However, some immune responses remain intact in the absence of CD28, perhaps because a prolonged TCR signal overcomes theneed for costimulation (5, 6), and/or other costimulatory moleculescan substitute for CD28 (7). CD28-mediated costimulation is reg-ulated by the CD28-related molecule CD152 (CTLA-4), which isexpressed on the T cell surface after T cell activation. CD152 bindsB7-1 and B7-2 with greater affinity than CD28 and down-regulatesT cell activation by inhibiting TCR/CD28-mediated signalingpathways (1).Another CD28-related molecule, inducible costimulator (ICOS),is induced in activated T cells (8). ICOS binds to a member of theB7 family called ICOS ligand (ICOSL; also known as B7RP1 (9),B7-H2 (10), B7h (11), GL50 (12), and LICOS (13)). Unlike CD80and CD86, whose expression is largely restricted to APCs, ICOSLis expressed in multiple nonlymphoid organs (11, 12). ICOS liga-tion enhances T cell proliferation and the production of variouscytokines, including IFN-   , IL-4, and IL-10 (8, 10, 14, 15). How-ever, in contrast to CD28, ICOS ligation augments only a minimalamount of IL-2 that is readily consumed by proliferating T cells(14, 15). Although ICOS-ICOSL interaction appears to providecostimulatory signals to both CD4  and CD8  T cells (16), anal-yses of mice and humans deficient in ICOS or ICOSL highlight thecritical role of ICOS costimulation for Ab isotype switching, af-finity maturation, and GC response (17–23).Interestingly, CD28, ICOS, and CD152 are not only all struc-turally related, but also cross-regulate each other’s expressionand/or functions. CD152 ligation inhibits the induction of ICOSand ICOS-mediated signaling, although this effect can be readilyover-ridden by IL-2 (15). Furthermore, CD28 ligation up-regulatesICOS expression   8- to   20-fold above the level achieved byTCR stimulation alone (15, 24, 25). The CD28-mediated ICOSup-regulation is more prominent in CD4  T cells than in CD8  Tcells (25). Thus, the impact of CD28 deficiency in CD28 KO miceshould be mediated by two components: the lack of CD28 itself,and the compromised ICOS induction. One prediction would bethat inactivation of the  Icos  gene in CD28 KO mice might havelittle further impact on CD28 KO phenotypes. *Advanced Medical Discovery Institute, Ontario Cancer Institute, Toronto, Ontario,Canada;  † Departments of Medical Biophysics and Immunology, University of To-ronto, Toronto, Ontario, Canada; and  ‡ University Hospital Zurich, Department of Pathology, Zurich, SwitzerlandReceived for publication December 12, 2003. Accepted for publication March1, 2004.The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked  advertisement   in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by the Canadian Institute of Health Research, CanadianNetwork for Vaccines and Immunotherapeutics of Cancer and Chronic Viral Dis-eases, and the Cancer Research Institute. 2 Address correspondence and reprint requests to Dr. Tak W. Mak, Advanced MedicalDiscovery Institute, 620 University Avenue, Suite 706, Toronto, Ontario, CanadaM5G 2C1. E-mail address: 3 Current address: Hoˆpital Necker, Institut National de la Sante´ et de la RechercheMe´dicale, Unite´ 580, 161 rue de Se`vres, 75015 Paris, France. 4 Current address: L121 Division of Science, Medicine Hat College, Medicine Hat,Alberta, Canada T1A 3Y6. 5 Abbreviations used in this paper: KO, knockout; GC, germinal center; DKO, doubleknockout; ES, embryonic stem; ICOS, the inducible costimulator; ICOSL, ICOS li-gand; KLH, keyhole limpet hemocyanin; PNA, peanut agglutinin; TNP, 2,4,6-trini-trophenol; VSV, vesicular stomatitis virus; WT, wild type. The Journal of Immunology Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00   b  y g u e  s  t   on J   ul   y 6  ,2  0 1  5 h  t   t   p :  /   /   w w w . j  i  mm un ol   . or  g /  D o wnl   o a  d  e  d f  r  om   CD28 and ICOS appear to play overlapping costimulatory rolesin humoral immune responses. However, the degree of synergybetween CD28 and ICOS in humoral immunity remains unclear.CD28 in fl uences the ICOS expression level in vitro, but whetherthis has critical in vivo consequences is unknown. We addressedthese issues by comparing the Ab responses in wild-type (WT),CD28 KO, and double-knockout (DKO) mice. We found thatDKO mice exhibit enormous defects in Ab responses against en-vironmental Ags, T-dependent protein Ags, and viral infection.These defects are far beyond the scope observed in CD28 KOmice. We show that these defects are due to the impaired prolif-eration and cytokine production of T cells activated in the absenceof both CD28 and ICOS. Thus, CD28 and ICOS act independently,but cooperatively, in T cell costimulation during humoral immuneresponses, and ICOS can assume crucial costimulatory functions inthe absence of CD28. Materials and Methods  Mice The  Icos   /   embryonic stem (ES) cell clone was generated as previouslydescribed (17). We disrupted the second exon of the  Cd28   locus in  Icos   /   ES cells using the same strategy that had been used for generation of thesrcinal CD28 KO mice (2), except that a hygromycin resistance cassette( hygro ) was used in place of the neomycin resistance cassette ( neo ; Fig.1  A ). We bred two strains of mice derived from two independent ES clonesthat had  Icos  ( neo ) and  Cd28   ( hygro ) on the same chromosome (as judgedby the cosegregation pattern of the two alleles) to obtain WT ( Cd28    /   ;  Icos   /   ), DHet ( Cd28    /   ;  Icos   /   ), and DKO ( Cd28    /   ;  Icos   /   ) mice.DKO mice derived from the two strains of mice showed the same pheno-types. We bred one strain of mice derived from one ES clone that had  Icos ( neo ) and  Cd28   ( hygro ) on the opposite chromosomes to generate WT( Cd28    /   ), CD28 Het ( Cd28    /   ), and CD28 KO ( Cd28    /   ) animals.CD28 KO mice from this strain showed the same phenotypes as the src-inal CD28-de fi cient mice (2). For some experiments we used ICOS KOmice (17) as controls. All the animals described above were of a mixed129/Ola  C57BL/6 genetic background and within 2 wk of age differencein all experiments. All live animal experiments were conducted with theapproval of the University Heath Network animal care committee (To-ronto, Canada).  Antibodies All Abs were purchased from BD PharMingen (San Diego, CA), except foranti-ICOS-PE (eBioscience, San Diego, CA).  Ab analysis Basal Ig levels were determined using an ELISA kit (Southern Biotech-nology Associates, Birmingham, AL) according to the manufacturer ’ s in-structions. To measure anti-keyhole limpet hemocyanin (anti-KLH) re-sponses, we injected mice s.c. with KLH emulsi fi ed in IFA (50  g/mouse).Anti-KLH IgM was measured on day 7, and anti-KLH IgG1 and IgG2awere measured on day 14 by ELISA. For anti-2,4,6-trinitrophenol (TNP)-Ficoll responses, we injected mice i.p. with TNP 24 -Ficoll (25   g/mouse;Biosearch Technologies, Novato, CA). Anti-TNP IgM was measured onday 7, and anti-TNP IgG3 was measured on day 22 by ELISA usingTNP 26 -BSA (Biosearch Technologies) as the Ag.  Anti-vesicular stomatitis virus (anti-VSV) responses We infected mice with VSV (Indiana strain; 2  10 5 PFU i.v.) and bled theanimals on days 4, 8, and 12. Neutralizing Ab titers were determined by aplaque formation assay as previously described, using 2-fold serial dilutionof the sera (4). Serum concentrations that provide 50% protection of the fi broblast monolayer were considered to be speci fi c titers. On day 12postinfection, spleens were frozen in liquid nitrogen. The spleen sectionswere stained with peanut agglutinin (PNA), anti-CD4 Ab, and anti-B220Ab to evaluate GC reaction as previously described (4). T cell analysis For CD154 staining, we cultured total lymph node cells in RPMI 1640medium containing 10% FCS plus antibiotics (complete medium) in thepresence of 1   g/ml anti-CD3 Ab (2C11; BD PharMingen) and anti-CD154-PE Ab or isotype control-PE Ab for 16 h. After stimulation, cellswere washed and then stained with anti-CD4-allophycocyanin Ab. CD154expression was assessed on CD4  T cells by  fl ow cytometry. For intra-cellular cytokine analysis, we injected mice s.c. with KLH-IFA (50   g/ mouse) and isolated the draining lymph nodes 10 days later. We preparedsingle-cell suspensions in complete medium and then depleted CD8  andB220  cells using Ab-conjugated magnetic beads (Dynal Laboratories,Chantilly, VA). Cells were then cultured at 5    10 6 cells/ml/well in 12-well plates in the presence of 100   g/ml KLH. After 90-h incubation, thecells were washed and restimulated with PMA (10 ng/ml) plus ionomycin FIGURE 1.  Disruption of the  Cd28   locus in an  Icos   /   ES cell clone.  A ,Targeting strategy. The targeting vector was designed to replace the secondexon of the  Cd28   gene ( f ) with the hygromycin resistance gene ( hygro ).S,  Sal I; X,  Xba I; RI,  Eco RI.  B , Southern blot analysis. Genomic DNAsfrom WT ( Cd28    /   ;  Icos   /   ), double-heterozygote (DHet;  Cd28    /   ;  Icos   /   ), and double-homozygote (DKO;  Cd28    /   ;  Icos   /   ) mice weredigested with  Xba I and subjected to Southern blotting using the probeshown in  A .  C  , Flow cytometric analysis. Total splenocytes from WT,DHet, and DKO mice were stimulated with Con A (1   g/ml) for 48 h andstained with anti-TCR  -FITC plus anti-CD28-PE, or anti-TCR  -FITCplus anti-ICOS-PE. Activated T cells (TCR   ) were gated based on highforward scattering. 5918 CRITICAL ROLE OF ICOS IN CD28-DEFICENT MICE   b  y g u e  s  t   on J   ul   y 6  ,2  0 1  5 h  t   t   p :  /   /   w w w . j  i  mm un ol   . or  g /  D o wnl   o a  d  e  d f  r  om   (100 ng/ml) for 4 h in the presence of GolgiStop (BD PharMingen). Cellswere then stained with anti-CD4-FITC and anti-CD25-PE, followed byallophycocyanin-conjugated Abs to IFN-   , IL-4, and IL-10 using an intra-cellular cytokine staining kit (BD PharMingen). For proliferation assays,cells were prepared as described above at 2    10 5  /ml/well in 96-well,U-bottom plates with or without KLH (100   g/ml) and pulsed with[ 3 H]thymidine (1   Ci/well) for the last 8 h of a 2-, 3-, or 4-day cultureperiod. Statistical analysis Student ’ s  t   test was used to determine the statistical signi fi cance of differ-ences between genotypes. Results Generation of   Cd28   /   ;  Icos   /   mice The genes encoding CD28, CD152, and ICOS are clustered on thesame chromosome: chromosome 1 in mice (26) and chromosome2 in humans (27). The mouse  Cd28   and  Icos  genes are separatedby only  1.5 cM (26), precluding the generation of DKO animalsby conventional breeding steps. We therefore used a hygromycinresistance cassette ( hygro ) to target the  Cd28   gene in an  Icos   /   ES clone in which one  Icos  allele had already been disrupted byinsertion of a neomycin resistance cassette ( neo ; Fig. 1  A ). As de-scribed in  Materials and Methods , we obtained three ES clonesthat were used to derive WT, CD28 KO, and DKO mice.Disruption of the  Cd28   allele in DKO mice was demonstratedby Southern blot analysis (Fig. 1  B ). The  Cd28   and  Icos  mutationswere con fi rmed to be null, as assessed by  fl ow cytometric analysisof the expression of CD28 and ICOS proteins on the surface of activated T cells (Fig. 1 C  ). Examination of T, B, NK, and NKTcell populations in primary and secondary lymphoid organs re-vealed normal differentiation and distribution of these cells in theabsence of CD28 and ICOS (data not shown). However, the per-centage of CD44 high CD62L low cells (activated/memory T cells) inthe total CD4  T cell population in the peripheral blood decreasedwith the loss of both genes (mean  SD for WT, 17.7  3.5%; forCD28 KO, 5.3    0.8%; for DKO, 2.3    1.2%; for ICOS KO,8.7  3.5% (three mice per genotype at 3 mo of age;  p  0.05 forWT-CD28 KO, CD28 KO-DKO, WT-DKO, and WT-ICOS KOcomparisons). These results show that CD28 and ICOS are bothpositive regulators of peripheral T cell activation.  Drastically reduced basal IgG1 in DKO mice Basal serum Ig levels provide an unbiased indication of the ef  fi -ciency of immune responses evoked by environmental Ags. De fi -ciency of either CD28 or ICOS results in a reduction of basalserum IgG1. Previous reports have shown that CD28 KO miceretain   20% of the mean WT IgG1 concentration (2), whereasICOS KO mice show  25 – 30% WT IgG1 (17, 18). In this studywe compared serum concentrations of IgM, IgG1, IgG2a, IgG2b,IgG3, and IgA in 3-mo-old WT, CD28 KO, and DKO mice (Fig.2). Among these isotypes, a statistically signi fi cant difference fromthe WT mean was apparent only for IgG1. CD28 KO mice retained  10% of the mean WT IgG1 level, whereas DKO mice showedonly   1.2% of WT IgG1. Other Ab isotypes tended to be de-creased in CD28 KO and DKO mice, but these differences lackedstatistical signi fi cance. These data indicate that in the absence of CD28, ICOS plays a critical role in promoting Ab isotype switch-ing to IgG1. Consistent with this idea, the bulk of ICOS  CD4  Tcells in nonmanipulated mice express IL-4, the key cytokine forisotype switching to IgG1 (28). Profound defect in Ab isotype switching induced by aT-dependent Ag Although CD28 is the major costimulatory molecule for responsesto many immunological challenges, the magnitude of the impact of CD28 de fi ciency varies depending on the immunization protocol(2, 29, 30). For this study we chose an immunization protocol thatrevealed a partial defect in CD28 KO mice and allowed us tomeasure the impact of CD28-ICOS double de fi ciency. WhenCD28 KO mice were immunized with KLH emulsi fi ed in IFA, wefound that a small, but signi fi cant, reduction in anti-KLH IgG1production occurred relative to the WT mean (Fig. 3  A ). Impor-tantly, this partial anti-KLH IgG1 response was completely abol-ished in DKO mice. Under the same conditions, CD28 KO miceproduced normal levels of IgG2a, whereas this isotype was nearlyundetectable in the sera of DKO mice (Fig. 3  A ). In addition,whereas the IgG2a defect in DKO mice became much less signif-icant after secondary immunization, the IgG1 defect remained pro-found (Fig. 3  B ).The level of IgM in DKO mice was not reduced compared withthat in WT or CD28 KO mice when the animals were nonmanipu-lated or actively immunized (Fig. 2, IgM; Fig. 3  A , IgM). Thesedata thus suggested that B cell function itself was not affected bythe combined loss of CD28 and ICOS. To con fi rm this, we im-munized WT and DKO mice with the T-independent Ag, TNP-Ficoll. This Ag directly stimulates B cell responses without theneed for thymus-derived T cell help (31). As predicted, DKO miceproduced normal levels of anti-TNP IgM and IgG3 (Fig. 3 C  ).Taken together, our results show that Ab isotype switching isheavily dependent on ICOS-mediated T cell help when CD28 isnot present. FIGURE 2.  Reduced serum IgG1 concentration in CD28 KO and DKOmice. Serum samples were prepared from the tail blood of 3-mo-old WT(  ), CD28 KO (  ), and DKO ( E ) mice, and the concentrations of IgM,IgG1, IgG2a, IgG2b, IgG3, and IgA were determined by ELISA ( n  8 – 16mice/genotype). There were signi fi cant differences in IgG1 concentrationamong the genotypes compared with WT:CD28 KO, CD28 KO:DKO, andWT:DKO (horizontal lines).   ,  p  0.05, by Student ’ s  t   test. 5919The Journal of Immunology   b  y g u e  s  t   on J   ul   y 6  ,2  0 1  5 h  t   t   p :  /   /   w w w . j  i  mm un ol   . or  g /  D o wnl   o a  d  e  d f  r  om   Severe impairment of antiviral humoral immunity in DKO mice The overall effectiveness of humoral immunity against pathogenicchallenges depends not only on the amount of Ab synthesized, butalso on its af  fi nity/avidity. Higher af  fi nity Abs are generatedthrough somatic hypermutation of Ig genes in GC B cells. Al-though individual inactivation of the  Cd28   and  Icos  genes leads toa greatly compromised GC response (3, 4, 17, 18, 23), both CD28KO and ICOS KO mice display only a partial defect in the pro-duction of neutralizing Ab during VSV infection (2, 32). In thisstudy we infected WT, CD28 KO, ICOS KO, and DKO mice withVSV and measured the production of neutralizing IgM and IgGAb. As expected, anti-VSV IgM production ont day 4 was equiv-alent in all four genotypes (Fig. 4  A ). By days 8 and 12, CD28 KOmice produced anti-VSV IgG, but displayed a partial defect in thisresponse. Neutralizing IgG was  8-fold lower than the WT meantiter on both days 8 and 12 postinfection (Fig. 4  A ), con fi rming FIGURE 3.  Impaired Ab isotype switching in CD28KO and DKO micein response to a T-dependent Ag.  A , Primary response. WT (  ), CD28 KO(  ), and DKO ( E ) mice were s.c. injected with KLH-IFA, and serumanti-KLH Ab titers were determined on day 7 (IgM) or day 14 (IgG1 andIgG2a) postinjection ( n  8 mice/genotype).  B , Secondary responses. Themice described in  A  were boosted using the same protocol 4 wk after theprimary injection. Anti-KLH IgG1 and IgG2a titers were determined 7days later.  C  , Intact Ab responses against a T-independent Ag. Four WT(  ) and three DKO ( E ) mice were i.p. injected with TNP-Ficoll, andserum titers of anti-TNP IgM and IgG3 Ab were determined by ELISA.   ,  p  0.05;   ,  p  0.01. FIGURE 4.  Severe defects in anti-VSV humoral immunity in DKOmice.  A , WT ( n  3; f ), CD28 KO ( n  2; u ), DKO ( n  8;  ), and ICOSKO ( n  2; o ) mice were i.v. injected with VSV, and the neutralizing Abtiters were determined on the indicated days by a plaque formation assayas described in  Materials and Methods . #, Below detection limit; arrow,GCs are shown in  B . Results shown are representative of two independentexperiments.  B , Impaired GC response. Spleens were taken on day 12postinfection, and GCs were stained as described in  Materials and Meth-ods . Results shown are PNA-stained spleen sections prepared from themice marked with arrows in  A . Representative sections obtained from twomice per genotype. 5920 CRITICAL ROLE OF ICOS IN CD28-DEFICENT MICE   b  y g u e  s  t   on J   ul   y 6  ,2  0 1  5 h  t   t   p :  /   /   w w w . j  i  mm un ol   . or  g /  D o wnl   o a  d  e  d f  r  om 
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks