sign in sign up
<<
Home , Memory B cell

Memory CD4 T Cells That Express CXCR5 Provide Accelerated Help to B Cells Megan K. L. MacLeod, Alexandria David, Amy S. McKee, Frances Crawford, John W. Kappler and Philippa Marrack This information is current as of September 26, 2021. J Immunol 2011; 186:2889-2896; Prepublished online 26 January 2011; doi: 10.4049/jimmunol.1002955 http://www.jimmunol.org/content/186/5/2889 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2011/01/26/jimmunol.100295 Material 5.DC1

References This article cites 46 articles, 18 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/186/5/2889.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 26, 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 © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Memory CD4 T Cells That Express CXCR5 Provide Accelerated Help to B Cells

Megan K. L. MacLeod,* Alexandria David,* Amy S. McKee,* Frances Crawford,* John W. Kappler,*,†,‡ and Philippa Marrack*,‡,x

CD4 help for B cells is critical for effective Ab responses. Although many of the molecules involved in helper functions of naive CD4 T cells have been characterized, much less is known about the helper capabilities of memory CD4 T cells, an important con- sideration for the design of that aim to prime protective memory CD4 T cells. In this study, we demonstrate that mem- ory CD4 T cells enable B cells to expand more rapidly and class switch earlier than do primary responding CD4 T cells. This accelerated response does not require large numbers of memory cells, and similar numbers of primary responding cells provide less effective help than do memory cells. However, only memory CD4 T cells that express the follicle homing molecule, CXCR5, are able to accelerate the response, suggesting that the rapidity of the Ab response depends on the ability of CD4 memory T cells to Downloaded from migrate quickly toward B cells. The Journal of Immunology, 2011, 186: 2889–2896.

he basis for is that Ag-experi- microorganisms by, for example, neutralizing the invader or im- enced respond better than their naive counter- proving uptake by phagocytic cells (10). The faster the Ab re- T parts. Although this phenomenon is widely accepted, there sponse, the more quickly the invading organisms can be con- is still a paucity of understanding of the mechanisms involved in trolled. Therefore, the rapid generation of class-switched Ab is an http://www.jimmunol.org/ enhanced memory responses. Memory cells are generated fol- important consideration for design (6, 7). lowing the initial primary response in which Ag-specific cells first There is some evidence that CD4 memory T cells can provide proliferate, differentiate, and then most, but not all, of these cells accelerated help for Ab responses (11–14); however, this may be undergo apoptosis (1). The surviving memory cells differ from limited to B cell responses directed toward rather than naive cells in two main ways. First, there are more Ag-specific more relevant Ags (15). More importantly, there is currently no cells in the memory as compared with the naive pool (2–4). mechanistic information to explain how CD4 memory T cells Second, memory cells can make an effector response more rapidly could provide an enhanced helper response. after stimulation (1, 5). Which of these two factors is important for T follicular helper (Tfh) cells, the CD4 T cells that provide the improved responses observed upon reactivation is not clearly help to B cells, are defined by their expression of the transcription by guest on September 26, 2021 understood but is an important issue when considering the design factor, Bcl-6, and the cell surface makers ICOS, CXCR5, and pro- of T cell-mediated vaccines. grammed cell death 1 (PD-1) (16–20). Tfh memory cells have Memory CD4 T cells could provide a protective response to been defined in humans by these markers, in particular by the pathogens by helping B cells make a more rapid Ab response (6, expression of CXCR5, the receptor that allows cells to 7). That CD4 help is required for primary responding B cells to migrate toward the B cell area of lymphoid organs (16). Moreover, form germinal centers and produce high-affinity class-switched these cells are able to provide effective help to B cells in vitro Ab is well established (8). These signals are supplied via cell (20). Likewise, CD4 memory cells with similar surface properties surface molecules such as CD40L and ICOS, and by means of have been found in the lymph nodes of mice that harbor persistent soluble molecules such as the IL-4 and IL-21 (9). The Ag (21); however, an enhanced helper activity by these cells has Abs generated by such a response protect the host against invading not been demonstrated. In this study, we compared directly the functions and effec- *Howard Hughes Medical Institute, Integrated Department of Immunology, National tiveness of resting endogenous memory and naive CD4 T cells to Jewish Health, Denver, CO 80206; †Program in Biomolecular Structure, University help primary responding B cells to produce class-switched Ab to ‡ of Colorado Denver, School of Medicine, Aurora, CO 80045; Department of Med- a protein Ag. We found that Ag-specific memory CD4 T cells do icine, University of Colorado Denver, School of Medicine, Aurora, CO 80045; and xDepartment of Biochemistry and Molecular Genetics, University of Colorado Den- indeed stimulate primary B cell responses better than do naive ver, School of Medicine, Aurora, CO 80045 T cells. This is not just a consequence of the increase in Ag-specific Received for publication September 2, 2010. Accepted for publication December 20, cell precursor frequency in the memory as compared with the naive 2010. T cell pool. Instead, this effect is contained within a subpopulation This work was supported by National Institutes of Health Grants AI-18785 and AI- of memory cells that express CXCR5 at high levels, suggesting that 22295 and by the U.S. Department of Defense/U.S. Army Medical Research and Materiel Command (W81XWH-07-1-0550). their better function is due to their ability to migrate more rapidly Address correspondence and reprint requests to Dr. Philippa Marrack and Dr. Megan to B cell follicles. MacLeod, National Jewish Health, 1400 Jackson Street, Denver, CO 80206. E-mail addresses: [email protected] and [email protected] Materials and Methods The online version of this article contains supplemental material. Mice, , and infections

Abbreviations used in this article: MFI, mean fluorescence intensity; PD-1, pro- + grammed cell death 1; PNA, peanut agglutinin; Tfh, T follicular helper; Tg, trans- Female B6 and B6.PL-Thy1a/CyJ (Thy1.1 ) mice were obtained from The genic. Jackson Laboratory. 508 TCR transgenic (Tg) mice expressing a TCR specific for I-Ab/3K (22) were bred at National Jewish Health. All mice Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 were maintained in a specific pathogen-free environment in accordance www.jimmunol.org/cgi/doi/10.4049/jimmunol.1002955 2890 HELPER RESPONSES OF CD4 MEMORY T CELLS with institutional guidelines in the Animal Care Facility at National Jewish stored at 280˚C. Sections were thawed, rehydrated, and blocked with 3% Health. Mice were age-matched within experiments. Mice were immu- goat serum and treated using an avidin/biotin blocking kit (Vetor Labo- nized with 10 mg 3K (FEAQKAKANKAVD) supplied by The Molecular ratories) according to the manufacturer’s instructions. Sections were Resources Center at National Jewish Health or by JPT Peptide Technol- stained with anti-B220 FITC, anti-CD4 allophycocyanin, and anti-Thy1.2 ogies (Berlin, Germany) and 7 mg LPS (Escherichia coli; Difco) i.v. or, biotin, washed, and then incubated with streptavidin-Cy3 before they after the transfer of larger numbers of TCR Tg cells, recipients were im- were mounted using VectaShield (Vector Laboratories). Slides were pho- munized with 20 mg 3K plus 10 mg LPS. For 3K or GP61–66 OVA con- tographed using the Marianas system (3i) and analyzed using Slidebook jugations, 3K or GP61 peptide (GLNGPDIYKGVYQFKSVEFD) were version 5 (3i) . TCR Tg cells were identified as Thy1.2+CD4+. The distance conjugated to malemide-activated OVA as described by the manufacture between the TCR Tg cell and the nearest B cell follicle was measured (Pierce). Peptide/protein conjugates were tumbled with alum (Alhydrogel; using the Slidebook software. Brenntag Biosector) for 2 h at room temperature, and each mouse received 1 mg peptide/protein with 0.2 mg alum i.p. Mice were infected with Statistical analysis a single dose of 2 3 105 PFU LCMV (Armstrong) i.p. Data are presented as indicated in the figure legends, and statistical sig- Cell transfer nificance was determined using either a Student two-tailed t test or Mann– Whitnety U test with GraphPad Prism software version 4. and lymph nodes (brachial, axillary, inguinal, popliteal, lumbar) were taken from naive 508 TCR Tg mice that had been crossed to a Rag- negative background. A single-cell suspension was prepared and RBCs Results were lysed. Cells were washed and injected i.v. to give a starting population Memory CD4 T cells accelerate primary Ab responses equal to that found in the 3K-memory mice in the particular experiment. Alternatively, to track the location of the memory cells, congenic hosts were We first asked whether memory CD4 T cells could enhance a 3 6 given 2.5 10 cells. To sort memory cells from LCMV-infected mice, primary B cell response to a protein Ag. To determine this we Downloaded from spleens and lymph nodes were taken 8 wk postinfection. A single-cell + needed to prime CD4 T cells to an Ag that did not prime a B cell suspension was prepared and CD4 cells were presorted using the Milte- nyi Biotec CD4 T cell isolation kit as per the manufacturer’s instructions. response. To achieve this, we primed C57BL/6 (B6) mice with the These cells were then stained with Abs to CD4, CD44, CXCR5, B220, I-Ab–binding peptide, 3K, delivered with the TLR agonist, LPS, as MHC class II, and CD8, washed, and then stained with fluorescently la- an adjuvant. The 3K peptide was used because T cells specific for beled streptavidin. After washing, these cells were sorted using a MoFlo it can be tracked with I-Ab MHC class II tetramers containing 3K. XDP.

Moreover, we wanted to use a soluble Ag and adjuvant combi- http://www.jimmunol.org/ Flow cytometry nation so that there would be little, if any, persistent Ag. This

b b combination clearly primed CD4 T cells that could be stained with PE-labeled I-A /3K was produced as described (23). PE-labeled I-A /GP66–77 b tetramer (referred to as I-Ab/GP66 in the text) was supplied by the Na- I-A /3K tetramer at early and late time points (Fig. 1A). tional Institutes of Health Tetramer Core Facility. Single-cell suspensions Ten to 14 wk later, the 3K-primed mice (hereafter 3K-memory from the spleens of immunized or infected mice were stained with tetra- mice) and naive age-matched mice were immunized with 3K mer at 37˚C for 2 h. Abs to surface markers were added and the cells peptide that had been chemically conjugated to OVA protein (3K- incubated for a further 20 min at 4˚C. Tetramer+ cells were defined by + hi OVA). In the 3K-memory mice, we would initiate a gating on live CD4 CD44 cells that were CD8, B220, F4/80, MHC class II negative. For Bcl6 staining, cells were fixed with 4% paraformaldehyde response to the 3K peptide but a primary T and B cell response to after surface staining. OVA-Alexa 488 (Invitrogen) was incubated with OVA protein, whereas there would be a primary T and B cell re- by guest on September 26, 2021 cells for 2 h at 37˚C. These cells were washed and then incubated sponse to both Ags in the naive mice. The peptide/protein conjugate with 24G2 (made at National Jewish Health) for 20 min before staining was delivered with alum, as this adjuvant induces a strong Ab re- with cell surface Abs. Abs were either from BD Biosciences, eBioscience, or BioLegend, and anti-Bcl6 was from Santa Cruz Biotechnology and sponse. The anti-OVA Ab response was analyzed over time in the peanut agglutinin (PNA) was from Vector Laboratories. Two to 5 million sera of the immunized animals, with day 0 indicating the time point events were collected on a CyAn ADP (Beckman Coulter) and data an- at which the mice were immunized with 3K-OVA plus alum. Im- alyzed using FlowJo version 8.8 (Tree Star). munized 3K-memory mice had levels of anti-OVA IgG above those Intracellular staining in control mice at days 6 and 8 (Fig. 1B). However, by day 15 the levels of anti-OVA Ab were the same in the two groups. Therefore, For analysis of cytokine production by intracellular staining, splenocytes Ag-specific memory CD4 T cells accelerated the production of m m were activated ex vivo with 10 g/ml GP66 and 1 g/ml of GolgiPlug (BD class-switched anti-OVA Ab but did not increase the peak response. Biosciences) for 6 h. Cells were stained with surface Abs and then fixed and permeabilized using the BD Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer’s instructions. Ag-specific cytokine was de- Naive CD4 T cells provide less effective help than do similar fined as staining above the level of staining from splenocytes cultured in numbers of memory CD4 T cells + the absence of peptide. Cells were gated on live CD4 cells that were CD8, There are a number of explanations for this accelerated response. B220, and MHC class II negative. The most obvious is that the 3K-memory mice contain larger OVA-specific ELISAs numbers of CD4 T cells specific for 3K than do naive animals. ELISAs were performed as described (24). Briefly, 96-well Immulon plates Naive B6 mice contain approximately seventy 3K-specific cells (4) (Thermo Fisher Scientific) were coated with OVA protein at 100 mg/ml in and an unknown number of cells that are specific for the CD4 PBS overnight at 4˚C. Following washing, the plates were blocked with found within OVA. The 3K memory mice contain ∼1–5 3 10% FCS/PBS before serum samples were added and titrated down the 103 3K-specific memory cells (data not shown) and presumably plate. To determine relative units, we used a positive control serum sample the same number of OVA-specific T cells as do naive mice. from B6 mice that contained OVA-specific Th2 and Th1 isotypes on each plate. The samples were incubated overnight at 4˚C. Plates were washed To test the idea that precursor frequency alone was responsible and alkaline phosphatase-conjugated anti-IgG1 or anti-IgG2c (BD Bio- for the efficiency of B cell help in 3K-memory mice, we examined sciences) detection Abs were added. Plates were washed again and p- the B cell response in animals that contained an equivalent number nitrophenyl phosphate substrate diluted in glycine buffer was added to of naive and memory Ag-specific cells. We first counted the number each well. Plates were allowed to develop and 405 nm absorbance values were collected on an Elx808 microplate reader. of 3K-specific memory cells in a cohort of 3K-memory mice and then transferred 10-fold this number of naive T cells bearing an I- Immunofluorescence Ab/3K–specific Tg TCR into naive B6 mice. The 10-fold increase Spleens were frozen in OCT-embedding medium and frozen on dry ice. was used based on the well-established finding that only ∼10% of Tissue sections (7 mm) were cut onto glass slides, fixed in acetone, and T cells survive the transfer (4, 25, 26). The Journal of Immunology 2891

FIGURE 1. Memory cells provide accelerated help for primary responding B cells. A, B6 mice were immunized with 3K peptide and LPS i.v. and the Downloaded from percentages of CD4+ cells that bound to I-Ab/3K MHC tetramer were examined 8 or 70 d later in the spleens of these animals and naive age-matched mice. Cells are gated on live CD4+ lymphocytes that were negative for B220, CD8, MHC class II, and F4/80. The number indicates the percentage of cells in the indicated I-Ab/3K tetramer+CD44hi gate. B, After 10–14 wk these animals (3K memory mice) and naive age-matched mice were immunized with 3K-OVAplus alum i.p. on day 0. The mice were bled and the relative amount of OVA-specific IgG1 was measured. Each point shows the mean of the group, and error bars are SEMs. The data are representative of three experiments with at least five mice per time point. *p , 0.05, **p . 0.001. C, 3K memory mice, age-matched naive mice that had received 5 3 104 3K-specific TCR Tg cells on day 21, and naive B6 mice were immunized with 3K-OVAplus alum on day 0. The number of 3K- specific cells in the spleen was examined. Each point shows the mean of the group, and error bars are SEMs. The gray line indicates the level of background http://www.jimmunol.org/ staining of splenocytes from naive mice. D, The mice in C were bled and the relative amount of OVA-specific IgG1 was measured. Each point shows the mean of the group, and error bars are SEMs. The data are representative of three experiments with at least five mice per time point. **p . 0.001.

One day following the transfer, the TCR Tg recipients, naive more quickly than naive T cells could. Thus, the accelerated help nontransferred mice, and the remaining 3K-memory mice were afforded by CD4 memory T cells is not just a matter of increased immunized with 3K-OVA plus alum. We have previously reported numbers of T cells; that is, CD4 memory T cells have some in- that primary responding T cells undergo more rounds of pro- trinsic property that renders them more effective at early stages in liferation than do reactivated memory cells (4). Therefore, as the B cell response. expected, there were more I-Ab/3K–specific cells present in the by guest on September 26, 2021 Activated naive and memory cells upregulate markers mice that had received naive TCR Tg cells than in the memory associated with Tfh cells mice 8 d after (Fig. 1C). Given this greater number of T cells, it may have been expected that the Ab response would Reactivated memory cells are known to differentiate into effector be highest in these mice. However, at early time points, the pri- cells more rapidly than do primary responding cells (5). To find out mary anti-OVA response was greatest in the 3K-memory mice whether reactivated memory cells expressed proteins associated (Fig. 1D). The presence of the TCR Tg cells did increase the anti- with Tfh cells more rapidly than did naive T cells, we tracked the OVA response compared with the nontransferred primary re- expression of CXCR5, ICOS, and PD-1 on the activated cells. sponding mice at day 8 (Fig. 1D). Importantly, memory TCR Tg Many of the specific cells increased expression of these markers, cells could also provide accelerated help to primary responding and the kinetics of expression were similar in either primary B cells (Supplemental Fig. 1), demonstrating that the TCR Tg (endogenous or TCR Tg) or secondary responding cells (Fig. 3A– memory cells could act in an equivalent fashion to endogenous C). It was possible that there were differences in Tfh cell marker memory cells. Interestingly, the peak Ab response was equivalent expression at a time point earlier than day 6. There were not b + in all groups, suggesting that the limiting factor for the total Ab enough primary responding endogenous I-A /3K tetramer cells response was the number of B cells that can differentiate into to examine their expression of Tfh cell markers before day 6. plasma cells rather than any limitation on T cell help. However, the expression of CXCR5 and ICOS was equivalent in primary responding TCR Tg cells and endogenous memory cells OVA-specific, class-switched B cells are at day 4 (Fig. 3D). The secondary responding memory cells did increased in CD4 memory mice express slightly more PD-1 than did the primary responding TCR To examine the B cell response more closely, we counted the Tg cells at day 4 (Fig. 3D). However, there seems to be little number of OVA-specific B cells that had class-switched in the requirement for PD-1–PDL1 interactions at this early time point different groups of mice after immunization. Spleen B cells were (27), suggesting that this small difference is unlikely to contribute stained with fluorescently labeled OVA, anti-IgG1, and PNA to to the accelerated OVA-specific Ab response in mice containing identify OVA-specific, class-switched germinal center cells (Fig. the memory CD4 T cells. Taken together, these data suggested that 2A). The previously naive mice that had not been given naive TCR the memory and primary responding cells differentiate into Tfh Tg cells contained a small number of class-switched OVA-specific cells at a similar rate following activation. B cells 8 d after immunization. The mice that had received TCR To confirm that the activated primary and memory cells Tg cells had more of such B cells. However, the greatest per- expressed higher levels of CXCR5, ICOS and PD-1 and were not centage and number (Fig. 2B) of class-switched B cells was found just binding more of the Abs nonspecifically because they were in the 3K-memory mice. These data suggest that the memory larger cells, we compared the level of binding of these Abs to the T cells could induce B cells to enter the germinal center response I-Ab/3K tetramer+ cells to that on CD44lo CD4 T cells (naive) and 2892 HELPER RESPONSES OF CD4 MEMORY T CELLS

FIGURE 2. CD4 memory cells increase the number of Ag-specific primary responding B cells. B6 mice were immunized with 3K plus LPS i.v. and rested for 12 wk, and then these animals, naive mice given 5 3 104 TCR Tg cells on day 21, and naive mice were immunized with 3K-OVA plus alum i.p. on day 0. On day 8, the percentages of CD19+ cells that were PNA+IgG1+ but negative for CD4, CD8, IgM, CD11b, and CD11c in the spleen were examined (A). The number indicates the percentage of cells (PNA+OVA+) in the top right quadrant out of IgG1+CD19+ cells, or the numbers of CD19+ IgG1+OVA+PNA+ cells in the spleens of these mice were examined at various time points after immunization (B). The data are combined from three independent experiments with three to four mice per group per time point. *p , 0.05.

CD44hi CD4 T cells (memory phenotype). Although memory congenic hosts to increase the frequency of Ag-specific cells, b phenotype cells were of a similar size to the activated I-A /3K allowing us to find the cells by immunofluorescence. Once these Downloaded from tetramer+ cells, they bound less of the Tfh cell marker Abs, and cells had developed into memory cells, we examined the location of naive CD4 T cells bound less of the Abs and were also of the memory cells in relationship to B cell follicles and compared a smaller size (Supplemental Fig. 2A). Moreover, to exclude the the location of these cells to that of transferred naive TCR Tg cells possibility that staining with the tetramer affected the expression (Fig. 3E). The memory cells were, in general, found to be in closer of the markers, we examined the level of staining on in vivo- proximity to B cell follicles than were naive T cells (Fig. 3F). This b activated TCR Tg cells identified using either I-A /3K tetramer location and the higher expression of the Tfh cell markers suggest http://www.jimmunol.org/ or a congenic marker (Thy1.2). There were no differences in the that memory CD4 T cells are poised ready to help B cells prior to expression level of any of the makers on TCR Tg cells identified reactivation. by the two different markers (Supplemental Fig. 2B). Therefore, we conclude that the I-Ab/3K tetramer+ cells do indeed express LCMV-specific memory cells express higher levels of Tfh cell increased levels of CXCR5, ICOS, and PD-1. markers than do naive T cells The higher expression of CXCR5 by the memory cells should The best way to test whether the increased expression of the Tfh position these cells close to B cell follicles. To examine the loca- cell markers were sufficient for the better helper activity of CD4 tion of the memory cells, we transferred naive TCR Tg cells into memory T cells would be to purify various populations of T cells, by guest on September 26, 2021

FIGURE 3. Primary and reactivated memory cells upregulate Tfh cell markers. B6 mice were immunized with 3K plus LPS i.v. and rested for 12 wk, and then these animals, naive B6 mice given 5 3 104 TCR Tg cells on day 21, and naive B6 mice were immunized with 3K-OVA plus alum i.p. on day 0. At various time points thereafter, the MFI of CXCR5 (A), ICOS (B), and PD-1 (C) on the I-Ab/3K tetramer+ cells was examined by flow cytometry. Each point represents the mean of the group, and the error bars show the SEMs. The data are representative of three experiments with three to four mice per group per time point. Additionally, the MFI of CXCR5, ICOS, and PD-1 was examined on day 4 of the primary responding TCR Tg cells and on the secondary responding endogenous memory cells (D). These data are combined from two experiments with four mice per group. TCR Tg cells (2.5 3 106) were transferred into congenic hosts that were immunized the following day with 3K peptide and LPS. Thirty days later, spleens were taken from these mice and naive age-matched animals that had received naive TCR Tg cells the day before. Spleen sections were stained with B220, CD4, and Thy1.2, and the distance between the naive or memory TCR Tg cells (Thy1.2+ red and CD4+ blue) and the nearest B cell follicle (green) was examined. A representative photograph taken with a 340 objective is shown from mice containing naive, memory, or no TCR Tg cells (E), or the distance between the T cells and the boarder of the nearest B cell follicle was measured and graphed (F). Data are combined from three to four mice per group, with each point representing a single T cell and the line showing the mean of all the cells. The Journal of Immunology 2893 transfer them into naive animals, and evaluate their ability to provide enhanced help. This experiment would also allow evalu- ation of the functions of the transferred cells in the absence of any other factors that may have been induced in previously immunized animals. This experiment was not feasible using 3K-activated memory cells, as too few memory cells are generated to allow isolation of sufficient cells for transfer. To increase the number of available endogenous Ag-specific memory cells we used mice that had previously been infected with LCMV. In such animals ∼1–2% of CD4 T cells are specific for the immunodominant LCMV peptide (3), GP61–71 (referred to as GP66). Eight weeks postinfection, we stained splenocytes with I-Ab/GP66 MHC tetramer and Abs to PD- 1, ICOS, and CXCR5 and also CCR7. The memory cells expressed higher levels of all Tfh cell markers compared with naive CD4 T cells (Fig. 4A). We could not, however, detect a population of cells that expressed two or three of these markers at a high level that would be equivalent to the Tfh cells found during the primary response (19, 28–30). This is demonstrated in Fig. 4B in which the Downloaded from expression of ICOS or PD-1 on cells that express CXCR5 at a high or low level was examined: these two populations expressed identical levels of ICOS or PD-1. Similarly, regardless of CXCR5 expression, the cells expressed equivalent levels of CCR7. These data suggested that the memory cells that expressed

CXCR5 may not be a memory equivalent of the Tfh cells described http://www.jimmunol.org/ in primary responses (19, 28–30). To investigate this further, we examined the levels of the transcription factor Bcl6 in the memory cells. Memory and activated Ag-specific cells did express higher levels of Bcl6 than did naive CD4 T cells, with the recently ac- tivated cells expressing slightly more (Fig. 4C). The highest level of Bcl6 is thought to be in germinal center B cells (31), and IgG1+ PNA+CD19+ cells from recently immunized mice did express more Bcl6 than did any of the T cell populations. When we compared expression of Bcl6 in CXCR5hi and CXCR5lo tetramer+ by guest on September 26, 2021 memory cells there were no differences between the groups as illustrated by the mean fluorescence intensity (MFI) of Bcl6 in these populations (Fig. 4D). Interestingly, in primary responding cells, there appeared to be slightly more Bcl6 in the CXCR5lo tetramer+ cells than in the CXCR5hi cells. Whether this small difference has a biological significance remains to be determined. FIGURE 4. LCMV-activated memory cells express Tfh cell markers. A, Regardless, it is difficult to conclude whether the CXCR5hi mem- B6 mice were infected with LCMV i.p. and 8 wk later spleen cells were ory cells are a Tfh memory cell equivalent. stained with I-Ab/GP66 tetramer and the expression of CXCR5, ICOS, and The accelerated helper activity of the memory CD4 T cells is PD-1 on these cells and naive T cells was examined. Cells were gated on live lymphocytes that were CD4+ and negative for CD8, B220, MHC class contained within the CXCR5hi subset II, and F4/80 in the top panels, with the number indicating the percentage Even though we could not clearly identify a population of Tfh of CD4 cells in the indicated gate (I-Ab/GP66 tetramer+ CD44hi), and then Ag-specific memory cells in the LCMV memory mice, we noted on the I-Ab/GP66 tetramer+ CD44hi (red line) or naive CD4+CD44lo cells a broad expression level of CXCR5 on these cells. The first step (gray filled line) in the bottom panel. B, The expression of CCR7, ICOS, or PD-1 on the CXCR5hi (red line) and CXCR5lo (blue line) I-Ab/GP66 that CD4 T cells must take to provide help to B cells is to migrate + toward the follicle. We therefore decided to find out whether ex- tetramer cells was examined. For anti-CCR7 staining, B cells are shown as a negative control. C, Spleen cells from mice either infected with pression of CXCR5, the protein that guides T cells to follicles, was LCMV 8 wk previously (memory) or immunized with GP66-OVA plus correlated with the accelerated helper activity. alum 8 d earlier (activated) were stained with I-Ab/GP66 tetramer and the This experiment could not be done with prior sorting of the Ag- intracellular levels of Bcl6 were examined. Cells are gated on I-Ab/GP66 specific T cells, since staining with MHC class II tetramers requires tetramer+ memory cells (red line), I-Ab/GP66 tetramer+ activated cells incubation with tetramer at 37˚C for at least 1 h; such conditions (blue line), IgG1+PNA+CD19+ cells (black line, germinal center [GC] may induce TCR signaling and/or affect the phenotype of the cells), or naive CD4 cells (gray filled line). D, The experiment was per- cells. Instead, CXCR5hi and CXCR5loCD44hiCD4+ cells were formed as in C, but the MFI of Bcl6 in the indicated populations was sorted from the spleens and lymph nodes of mice infected 8 wk examined. Each line represents a mouse and the line shows the mean of the previously with LCMV without regard to their Ag specificity. A group. The data are representative of two experiments. small fraction of the sorted populations were stained with I-Ab/ GP66 MHC class II tetramers to count the number of Ag-specific us to estimate the percentage of CXCR5hi and CXCR5lo cells that cells in each population (Fig. 5A). Of note, both populations were I-Ab/GP66-specific, allowing us to transfer the same num- bound to the tetramer at an equivalent level, suggesting that they bers of Ag-specific T cells into naive hosts. Because the memory were of similar affinities (Fig. 5B). The tetramer staining allowed cells expressed a range of CXCR5, we were unable to separate the 2894 HELPER RESPONSES OF CD4 MEMORY T CELLS

CXCR5 high and low cells cleanly, instead enriching for either supporting our findings in the 3K-OVA experiments (Fig. 6). The high or low expression (Fig. 5B). magnitude of the Ab response was similar in mice that received Memory cell recipients, as well as age-matched mice that had CXCR5lo memory cells and mice that had not received any not received any cells, were immunized with OVA that had been transferred T cells. In contrast, mice that received CXCR5hi chemically conjugated to the GP66 peptide plus alum 1 d after memory cells had greater levels of OVA-specific IgG1, at levels transfer. As a positive control, we also immunized a cohort of the that were similar to those in mice previously infected with LCMV LCMV-infected mice (LCMV memory mice) that contained a (Fig. 6A). This improved response was not evident for IgG2c anti- greater frequency of GP66-specific cells: LCMV memory mice OVA Abs (Fig. 6B), even though the CXCR5hi cells could make contained 1–4 3 105 I-Ab/GP66 tetramer+ cells while the adoptive IFN-g (Supplemental Fig. 4). Perhaps this reflects the influence of transfer recipients contained ∼100–400 I-Ab/GP66 tetramer+ cells the numerically larger recipient T cell response that is primed in after the transfer (data not shown). the type 2 environment following alum immunization. Using the congenic marker to identify the transferred cells, we found that the number of Ag-specific transferred cells in the re- Discussion cipients 8 d after immunization was low but similar in mice that Memory responses are considered to be better than primary received CXCR5hi or CXCR5lo cells (Fig. 5C). The percentages of responses, but what is meant by “better” for CD4 memory T cells I-Ab/GP66 tetramer+ cells out of total transferred cells increased is not well defined. In this study we conclusively show that dramatically following immunization, from 1–2% on the day of memory CD4 T cells primed either by Ag and adjuvant or by transfer to 10–30% at day 8, indicating that both populations had a viral infection provide accelerated help for B cell responses. been reactivated and responded to the immunization (Fig. 5A,5D). We demonstrate that this acceleration of the B cell response is Downloaded from When we examined the phenotype of the reactivated memory due to an intrinsic difference between naive and memory cells and cells, a similar percentage expressed PD-1. Many of the CXCR5lo that this differences correlates with high expression of the B cell cells had upregulated CXCR5, and slightly more of them follicle homing molecule, CXCR5, on the memory CD4 T cells. expressed ICOS than the CXCR5hi-enriched cells (Fig. 5E). Adaptive immune responses to infections and vaccines lead to Transfer of the memory cells did not inhibit the host response, as the generation of memory cells that can protect the host from b + similar numbers of I-A /GP66 tetramer host cells were found in subsequent infections. The advantages of memory CD8 T cells and http://www.jimmunol.org/ all the primary responding mice (Supplemental Fig. 3). B cells are obvious: faster responses that generate effector cells or To find out whether the transferred memory subsets could molecules that directly clear or neutralize the pathogen or its toxins provide accelerated help to the host primary responding B cells, we (10, 32). Understandably, these cell types have been found to be examined the anti-OVA Ab response 8 d after immunization. The long-lived, although human memory CD8 T cells may have rel- LCMV-immunized mice made a mixed IgG1/IgG2c response that atively shorter lifespans than do those in mice (33–35). CD4 was greater than the response in primary responding animals, memory T cells, on the other hand, have a shorter lifespan both in by guest on September 26, 2021

FIGURE 5. Both CXCR5hi and CXCR5lo cells respond following transfer and reactivation in vivo. A, Eight weeks after LCMV infection, spleens and lymph nodes were taken and CD4+CD44hi T cells expressing high or low levels of CXCR5 were sorted as described in Materials and Methods. Some of the sorted cells were stained with I-Ab/GP66 tetramer. Cells are gated as in Fig. 4A, with the numbers indicating the percentages of cells in the I-Ab/GP66 tetramer+ CD44hi gate. B, Intensity of tetramer binding and the expression of CXCR5 on the I-Ab/GP66 tetramer+ CD44hi cells after the sort. C, The sorted cells were transferred into congenically marked mice, with each recipient receiving between 1 and 4 3 104 cells that were I-Ab/GP66 tetramer+ CD44hi. These mice, age-matched mice, and a cohort of LCMV memory mice were immunized with GP66-OVA plus alum 1 d after the transfer (day 0). Eight days later, the numbers of I-Ab/GP66 tetramer+ CD44hi cells that were donor derived were examined. Cells were gated as in Fig. 4A and on the donor cells. Each point represents one mouse, and the line shows the means of the group. The data are combined from three experiments with three to five mice per group. D, Representative FACS plots of the cells in C. Cells are gated on donor+ CD4 cells. The number indicates the percentage in the gate: I-Ab/GP66+ CD44hi. E, ICOS, PD-1, and CXCR5 expression on the donor I-Ab/GP66 tetramer+ CD44hi on day 8. The data are from one to three experiments, with each symbol representing a mouse, and the line shows the mean of the group. The Journal of Immunology 2895

may account for their superior helper function, since other, ap- parently similar CD4 memory T cells, that bear low levels of CXCR5 prior to reactivation were not able to deliver the same degree of help. The CD4 memory T cells only affected early Ab production to the Ag. Later, Ab production was the same regardless of the presence of the memory T cells. This suggests that other factors may limit the differentiation of activated B cells to plasma cells. Such a relatively small shift in timing may nevertheless be sig- nificant, especially in cases in which the is dealing with an infection that rapidly changes its B cell, but not its FIGURE 6. CXCR5+-enriched Ag-specific memory cells provide ac- CD4 epitopes, as is the case for influenza (6). celerated help for IgG1 but not IgG2c responses. The experiment was Recently, the T cells that migrate to B cell follicles to provide performed as described in Fig. 5, but sera was prepared from the of help have been identified as a separate subset of helper T cells, Tfh the GP-OVA plus alum-immunized mice and the amount of OVA-specific cells (9). The CD4 memory T cells produced in the experiments IgG1 (A) or IgG2c (B) was examined 8 d after immunization. Each point described in this study all bore or contained higher levels of represents one mouse and the line shows the means of the group. The data markers characteristic of Tfh cells (ICOS, PD-1, CXCR5, and are combined from three experiments with three to five mice per group. Bcl-6) compared with naive cells (9, 31). However, we could

not separate out a population of Ag-specific memory cells that Downloaded from humans and in mice when compared with CD8 T cells or B cells expressed high levels of all the markers. Only those cells that bore (3, 4, 33, 36–38). This may reflect the critical role that CD4 high levels of CXCR5, regardless of their levels of expression of memory cells play as catalysts for both CD8 T cell and B cell other Tfh cell markers, provided superior help. These results responses in the primary responses, but that, once developed, the suggest that it is an oversimplification to assign T cells to a par- CD8 T cells and B cells probably do not require the presence of ticular subset based simply on their gene expression. In the end,

many specific CD4 memory T cells to be reactivated (39, 40). the important criterion may be function. http://www.jimmunol.org/ However, there are many infectious diseases that cause significant Tfh cells can make either a Th1 or Th2 response (45). The mortality or morbidity against which there are no current vaccines experiments described in this article involved CD4 memory or no effective vaccines; in particular, intracellular organisms that T cells that were primed in vivo in the presence of the infectious are adept at hiding from Ab or CD8 T cells or pathogens that alter virus, LCMV, or LPS, conditions that should generate Th1-like their immune targets (7, 41). For such infections, vaccines are cells, although to a different degree: all GP66-reactive T cells required that create cells that can mobilize a multifaceted attack make IFN-g following LCMV infection, whereas LPS-primed on the pathogen. To make progress in the design of such vaccines, memory cells are less committed to the Th1 lineage, as a smaller a greater appreciation of the in vivo functions of memory CD4 percentage of them make IFN-g (3, 4). In the experiments reported T cells is required. in this article, the only animals that made a type 1 Ab response by guest on September 26, 2021 With this in mind, several studies have investigated the con- (indicated by a switch to IgG2c) were those that had been infected tribution of CD4 memory T cells to B cell responses (11–15, 42). with LCMV before immunization with Ag and the Th2-inducing These studies have proved inconclusive, providing conflicting adjuvant, alum, suggesting that LCMV-primed cells are fully results perhaps because they have examined T cell responses in committed to their Th1 phenotype. The switch to a type 2 (IgG1) artificial environments and have not addressed a mechanism for rather than a type 1 Ab response in the LPS-primed mice probably any enhanced help that was observed. Bone marrow-resident reflects a lesser commitment of these cells to IFN-g production memory CD4 T cells have been shown to improve the affinity and the presence of IL-4 producing primary responding T cells to maturation of an Ab response compared with similar memory epitopes within the OVA protein. In the case of the transferred cells that reside in the spleen (43). Again, however, no mechanism LCMV memory cells, the small number of IFN-g–committed was described for this effect, and any help for primary responding memory cells are reactivated in the presence of a numerically B cells, which reside in secondary lymphoid organs, would be larger primary host response that develops in a Th2 environment. delayed as the memory cells migrate from the bone marrow to the This suggests that the IL-4 from these latter cells overcomes the site. Our studies shed new light on this issue enabled, IFN-g made by the reactivated memory cells, inducing a type 2 perhaps, by the more physiological numbers and sources of cells Ab response from the responding B cells (46). used in our studies. In summary, we show that some memory CD4 T cells have an Our results show that naive B cells do indeed respond and intrinsic ability to provide accelerated help to primary responding switch to the production of Ig isotypes other than IgM more rapidly B cells. Important questions remain as to how best to prime or if they are helped by CD4 memory, rather than primary, T cells. maintain CXCR5hi memory cells, a critical consideration for the This result was not simply a numbers issue, as the same numbers design of vaccines against pathogens that are adept at altering of primary responding T cells were unable to induce early IgG their Ab targets to escape immunodetection. from the responding B cells. The data suggest that the superior effects of the CD4 memory T cells were due to a subset of cells that Acknowledgments bore CXCR5, the receptor for CXCL13, the chemokine that attracts We thank Dr. Dirk Homann for supplying LCMV, the National Institutes of b cells to B cell follicles (44). These CD4 memory T cells also Health Tetramer Core Facility for I-A /GP66–77 tetramer, Josh Loomis and expressed the chemokine receptor that attracts cells to the T cell Shirley Sobus for cell sorting and assistance with microscopy, Catherine zone of lymphoid organs, CCR7, which recognizes CCL19 and Haluszczak for assistance with immunofluorescence, and all members of CCL21 (44). Although we do not directly show that CXCR5 ex- the Kappler Marrack laboratory for intellectual contributions to this project. pression is sufficient for the enhanced helper function, we found that the memory cells were more likely to be located close to Disclosures B cell follicles compared with naive CD4 T cells. This proximity The authors have no financial conflicts of interest. 2896 HELPER RESPONSES OF CD4 MEMORY T CELLS

References 23. Crawford, F., H. Kozono, J. White, P. Marrack, and J. Kappler. 1998. Detection of -specific T cells with multivalent soluble class II MHC covalent 1. MacLeod, M. K., E. T. Clambey, J. W. Kappler, and P. Marrack. 2009. CD4 peptide complexes. 8: 675–682. memory T cells: what are they and what can they do? Semin. Immunol. 21: 53– 24. McKee, A. S., M. W. Munks, M. K. MacLeod, C. J. Fleenor, N. Van Rooijen, 61. J. W. Kappler, and P. Marrack. 2009. Alum induces innate immune responses 2. Hataye, J., J. J. Moon, A. Khoruts, C. Reilly, and M. K. Jenkins. 2006. Naive and + through and mast cell sensors, but these sensors are not required for memory CD4 T cell survival controlled by clonal abundance. Science 312: 114– alum to act as an adjuvant for specific immunity. J. Immunol. 183: 4403–4414. 116. 25. Blattman, J. N., R. Antia, D. J. Sourdive, X. Wang, S. M. Kaech, K. Murali- 3. Homann, D., L. Teyton, and M. B. Oldstone. 2001. Differential regulation of + + Krishna, J. D. Altman, and R. Ahmed. 2002. Estimating the precursor frequency antiviral T-cell immunity results in stable CD8 but declining CD4 T-cell of naive antigen-specific CD8 T cells. J. Exp. Med. 195: 657–664. memory. Nat. Med. 7: 913–919. 26. Whitmire, J. K., N. Benning, and J. L. Whitton. 2006. Precursor frequency, 4. MacLeod, M. K., A. McKee, F. Crawford, J. White, J. Kappler, and P. Marrack. nonlinear proliferation, and functional maturation of virus-specific CD4+ T cells. 2008. CD4 memory T cells divide poorly in response to antigen because of their J. Immunol. 176: 3028–3036. cytokine profile. Proc. Natl. Acad. Sci. USA 105: 14521–14526. 27. Good-Jacobson, K. L., C. G. Szumilas, L. Chen, A. H. Sharpe, M. M. Tomayko, 5. Berard, M., and D. F. Tough. 2002. Qualitative differences between naive and and M. J. Shlomchik. 2010. PD-1 regulates germinal center B cell survival and memory T cells. Immunology 106: 127–138. the formation and affinity of long-lived plasma cells. Nat. Immunol. 11: 535– 6. Alexander, J., P. Bilsel, M. F. del Guercio, S. Stewart, A. Marinkovic-Petrovic, 542. S. Southwood, C. Crimi, L. Vang, L. Walker, G. Ishioka, et al. 2010. Universal 28. Cannons, J. L., H. Qi, K. T. Lu, M. Dutta, J. Gomez-Rodriguez, J. Cheng, influenza DNA vaccine encoding conserved CD4+ T cell epitopes protects E. K. Wakeland, R. N. Germain, and P. L. Schwartzberg. 2010. Optimal germinal against lethal viral challenge in HLA-DR transgenic mice. Vaccine 28: 664–672. center responses require a multistage T cell:B cell adhesion process involving 7. McKinstry, K. K., T. M. Strutt, and S. L. Swain. 2010. The potential of CD4 T- integrins, SLAM-associated protein, and CD84. Immunity 32: 253–265. cell memory. Immunology 130: 1–9. 29. King, I. L., and M. Mohrs. 2009. IL-4-producing CD4+ T cells in reactive lymph 8. MacLennan, I. C., A. Gulbranson-Judge, K. M. Toellner, M. Casamayor-Palleja, nodes during helminth infection are T follicular helper cells. J. Exp. Med. 206: E. Chan, D. M. Sze, S. A. Luther, and H. A. Orbea. 1997. The changing pref- 1001–1007. erence of T and B cells for partners as T-dependent responses develop. 30. Zaretsky, A. G., J. J. Taylor, I. L. King, F. A. Marshall, M. Mohrs, and Downloaded from Immunol. Rev. 156: 53–66. E. J. Pearce. 2009. T follicular helper cells differentiate from Th2 cells in re- 9. Fazilleau, N., L. Mark, L. J. McHeyzer-Williams, and M. G. McHeyzer-Wil- sponse to helminth . J. Exp. Med. 206: 991–999. liams. 2009. Follicular helper T cells: lineage and location. Immunity 30: 324– 31. Linterman, M. A., L. Beaton, D. Yu, R. R. Ramiscal, M. Srivastava, J. J. Hogan, 335. N. K. Verma, M. J. Smyth, R. J. Rigby, and C. G. Vinuesa. 2010. IL-21 acts 10. Plotkin, S. A. 2008. Vaccines: correlates of vaccine-induced immunity. Clin. directly on B cells to regulate Bcl-6 expression and germinal center responses. J. Infect. Dis. 47: 401–409. Exp. Med. 207: 353–363. 11. Katz, D. H., W. E. Paul, E. A. Goidl, and B. Benacerraf. 1970. Carrier function in 32. Harty, J. T., A. R. Tvinnereim, and D. W. White. 2000. CD8+ T cell effector anti- immune responses, I: Enhancement of primary and secondary anti- mechanisms in resistance to infection. Annu. Rev. Immunol. 18: 275–308. hapten antibody responses by carrier preimmunization. J. Exp. Med. 132: 261– 33. Hammarlund, E., M. W. Lewis, S. G. Hansen, L. I. Strelow, J. A. Nelson, http://www.jimmunol.org/ 282. G. J. Sexton, J. M. Hanifin, and M. K. Slifka. 2003. Duration of antiviral im- 12. Kontiainen, S. 1971. The effect of carrier-primed cells on an adoptive anti- munity after smallpox vaccination. Nat. Med. 9: 1131–1137. hapten response. Eur. J. Immunol. 1: 276–280. 34. Schluns, K. S., and L. Lefranc¸ois. 2003. Cytokine control of memory T-cell 13. Leclerc, C., C. Sedlik, R. Lo-Man, B. Charlot, M. Rojas, and E. De´riaud. 1995. development and survival. Nat. Rev. Immunol. 3: 269–279. Stimulation of a memory B cell response does not require primed helper T cells. 35. Tangye, S. G., and D. M. Tarlinton. 2009. Memory B cells: effectors of long- Eur. J. Immunol. 25: 2533–2538. lived immune responses. Eur. J. Immunol. 39: 2065–2075. 36. Cauley, L. S., T. Cookenham, T. B. Miller, P. S. Adams, K. M. Vignali, 14. Tada, T., and K. Okumura. 1971. Regulation of homocytotropic antibody for- + mation in the rat, V: Cell cooperation in the anti-hapten homocytotropic antibody D. A. Vignali, and D. L. Woodland. 2002. Cutting edge: virus-specific CD4 response. J. Immunol. 107: 1137–1145. memory T cells in nonlymphoid tissues express a highly activated phenotype. J. 15. Gupta, S. C., H. Hengartner, and R. M. Zinkernagel. 1986. Primary antibody Immunol. 169: 6655–6658. responses to a well-defined and unique hapten are not enhanced by preimmu- 37. MacLeod, M., M. J. Kwakkenbos, A. Crawford, S. Brown, B. Stockinger, K. Schepers, T. Schumacher, and D. Gray. 2006. CD4 memory T cells survive by guest on September 26, 2021 nization with carrier: analysis in a viral model. Proc. Natl. Acad. Sci. USA 83: and proliferate but fail to differentiate in the absence of CD40. J. Exp. Med. 203: 2604–2608. 897–906. 16. Ansel, K. M., L. J. McHeyzer-Williams, V. N. Ngo, M. G. McHeyzer-Williams, 38. Naniche, D., M. Garenne, C. Rae, M. Manchester, R. Buchta, S. K. Brodine, and and J. G. Cyster. 1999. In vivo-activated CD4 T cells upregulate CXC chemokine M. B. Oldstone. 2004. Decrease in measles virus-specific CD4 T cell memory in receptor 5 and reprogram their response to lymphoid . J. Exp. Med. vaccinated subjects. J. Infect. Dis. 190: 1387–1395. 190: 1123–1134. 39. Hebeis, B. J., K. Klenovsek, P. Rohwer, U. Ritter, A. Schneider, M. Mach, and 17. Nurieva, R. I., Y. Chung, G. J. Martinez, X. O. Yang, S. Tanaka, T. H. Winkler. 2004. Activation of virus-specific memory B cells in the absence T. D. Matskevitch, Y. H. Wang, and C. Dong. 2009. Bcl6 mediates the de- of T cell help. J. Exp. Med. 199: 593–602. velopment of T follicular helper cells. Science 325: 1001–1005. 40. Sun, J. C., M. A. Williams, and M. J. Bevan. 2004. CD4+ T cells are required for 18. Johnston, R. J., A. C. Poholek, D. DiToro, I. Yusuf, D. Eto, B. Barnett, the maintenance, not programming, of memory CD8+ T cells after acute in- A. L. Dent, J. Craft, and S. Crotty. 2009. Bcl6 and Blimp-1 are reciprocal and fection. Nat. Immunol. 5: 927–933. antagonistic regulators of T follicular helper cell differentiation. Science 325: 41. McKee, A. S., M. K. MacLeod, J. W. Kappler, and P. Marrack. 2010. Immune 1006–1010. mechanisms of protection: can adjuvants rise to the challenge? BMC Biol. 8: 37. 19. Yu, D., S. Rao, L. M. Tsai, S. K. Lee, Y. He, E. L. Sutcliffe, M. Srivastava, 42. Duffy, D., C. P. Yang, A. Heath, P. Garside, and E. B. Bell. 2006. Naive T-cell M. Linterman, L. Zheng, N. Simpson, et al. 2009. The transcriptional repressor receptor transgenic T cells help memory B cells produce antibody. Immunology Bcl-6 directs T follicular helper cell lineage commitment. Immunity 31: 457– 119: 376–384. 468. 43. Tokoyoda, K., S. Zehentmeier, A. N. Hegazy, I. Albrecht, J. R. Gru¨n, 20. Schaerli, P., K. Willimann, A. B. Lang, M. Lipp, P. Loetscher, and B. Moser. M. Lo¨hning, and A. Radbruch. 2009. Professional memory CD4+ T lymphocytes 2000. CXC chemokine receptor 5 expression defines follicular homing T cells preferentially reside and rest in the bone marrow. Immunity 30: 721–730. with B cell helper function. J. Exp. Med. 192: 1553–1562. 44. Okada, T., and J. G. Cyster. 2006. B cell migration and interactions in the early 21. Fazilleau, N., M. D. Eisenbraun, L. Malherbe, J. N. Ebright, R. R. Pogue-Caley, phase of antibody responses. Curr. Opin. Immunol. 18: 278–285. L. J. McHeyzer-Williams, and M. G. McHeyzer-Williams. 2007. Lymphoid 45. Reinhardt, R. L., H. E. Liang, and R. M. Locksley. 2009. Cytokine-secreting reservoirs of antigen-specific memory T helper cells. Nat. Immunol. 8: 753–761. follicular T cells shape the antibody repertoire. Nat. Immunol. 10: 385–393. 22. Huseby, E. S., J. White, F. Crawford, T. Vass, D. Becker, C. Pinilla, P. Marrack, 46. Perona-Wright, G., K. Mohrs, and M. Mohrs. 2010. Sustained signaling by ca- and J. W. Kappler. 2005. How the T cell repertoire becomes peptide and MHC nonical helper T cell cytokines throughout the reactive . Nat. specific. Cell 122: 247–260. Immunol. 11: 520–526.