|Home | About | Journals | Submit | Contact Us | Français|
T cell recognition of peptide/MHC complexes is flexible and can lead to differential activation, but how interactions with agonist (full activation) or partial agonist (suboptimal activation) peptides can shape immune responses in vivo is not well characterized. We investigated the effect of stimulation by agonist or partial agonist ligands during initial CD4+ T cell priming, and subsequent T-B cell cognate interactions, on antibody production by anti-chromatin B cells. We found that autoantibody production required TCR recognition of an agonist peptide at the effector stage of B cell activation. However, interaction with a weak agonist ligand at this effector stage failed to promote efficient autoantibody production, even if the CD4+ T cells were fully primed by an agonist peptide. These studies suggest that the reactivity of the TCR for a target self-peptide during CD4+ T-B cell interaction can be a critical determinant in restraining anti-chromatin autoantibody production.
T cells recognize antigen in the context of a peptide presented by an MHC class I/II molecule. Many studies have demonstrated the flexibility of the TCR for recognizing peptide/MHC class II complexes that can lead to complete (agonist) or incomplete (partial agonist) activation . For example, CD4+ T cell clones stimulated in vitro by partial agonist peptides could produce cytokines in the absence of robust proliferation [2, 3]. Similarly, in vivo immunization with partial agonist peptides led to altered cytokine responses [4, 5]. Furthermore, a recent study of CD8+ T cells in Listeria monocytogenes-infected mice showed that partial agonist ligands induced the development of effector and memory cells, but failed to promote full expansion despite early normal proliferation . These data have shown that T cells can respond differentially to agonist and partial agonist ligands, but how these differing responses affect distinct phases in the development of an immune response remains to be understood.
We were interested in determining how modulating the quality of TCR activation affects autoantibody production. CD4+ T cell-dependent antibody responses require the initial activation or priming of the T cell, typically by dendritic cells, followed by interactions with the B cell [7, 8]. At both steps, the CD4+ T cell receives unique signals provided by the cell that presents the antigenic peptide in the context of MHC class II molecules . How TCR recognition of agonist and partial agonist peptides, both at the time of initial CD4+ T cell priming and at the point of T-B cell collaboration (effector phase), impact the development of an autoantibody response is the focus of this study.
We have established a transgenic CD4+ T and B cell third-party transfer model where the lymphocytes and antibody production can be tracked in an immmunocompetent mouse (Fig. 1A) . We used CD4+ T cells from two lineages of TCR transgenic mice that recognize analog peptides from variant influenza viruses. TS1 CD4+ T cells recognize the PR8 HA (hemagglutinin) as an agonist, but are weakly reactive with the HA from another influenza virus (SW) which differs from PR8 by 2 amino acid residues in the S1 determinant . Conversely, TS1(SW) CD4+ T cells recognize the SW HA as an agonist, and are weakly reactive with the PR8 HA [11, 12]. We confirmed that the PR8 HA protein expressed as a transgene is recognized as an agonist for TS1 CD4+ T cells and as a partial agonist for TS1(SW) CD4+ T cells, when expressed as a neo-self antigen in vivo (Supporting Information Fig. 1).
As a source of anti-chromatin B cells, we used site-directed-(sd)-VH3H9 kappa deficient (κ−/−) mice . To establish a model system in which we can provide cognate CD4+ T cell help for anti-chromatin B cells, the sd-VH3H9.Igκ−/− mice were bred to an additional transgenic lineage that expresses the PR8 HA as a neo-self antigen under the control of the MHC class II promoter (HACII) . In this case the PR8 HA is produced selectively by class II bearing cells, including B cells .
Using this model, CD4+ T cells from TS1.RAG2−/− or TS1(SW).RAG2−/− mice were first transferred into CB17 mice (Igb allotype) along with either PR8 or SW virus to initiate their activation either by an agonist or a partial agonist. The following day, B cells from sd-VH3H9.HACII.Igκ−/− mice were transferred into these primed mice (Fig. 1A, B). These anti-chromatin Iga allotype B cells have been primed through their BCR by self-antigen and are anergized, but can produce antibody upon receipt of help from HA-specific CD4+ T cells [10, 13, 14]. The studies here aimed to determine how activation of CD4+ T cells by either agonist or partial agonist peptides at the priming or the effector phase (when the CD4+ T cells interact with anti-chromatin B cells) affects their ability to provide help for autoantibody production (Fig. 1A, B).
CD4+ T cells from TS1.RAG2−/− and TS1(SW).RAG2−/− mice underwent comparable division when analyzed by CFSE dilution three days after priming with their agonist viruses (PR8 or SW respectively) (Fig. 2A), and the relative numbers of CD4+ T cells that accumulated after 3 or 8 days did not differ significantly (Fig. 2B). These CD4+ T cells also upregulated ICOS, a critical co-stimulatory molecule that supports germinal center formation and extrafollicular antibody-secreting cell development (Fig. 2C) [15–17]. By contrast, CD4+ T cells from TS1(SW).RAG2−/− mice underwent less proliferation and possessed lower levels of ICOS when activated by PR8 virus (which they recognize as a partial agonist) (Fig. 2A–C). Moreover, CD4+ T cell proliferation, ICOS induction, and autoantibody production were greatly diminished in the absence of virus (Supporting Information Fig. 2A, B). Thus, efficient priming of CD4+ T cells was promoted by an agonist ligand, while a partial agonist induced less efficient expansion, accumulation, and co-stimulatory molecule up-regulation that together are likely to limit the ability to provide help for B cell responses.
We evaluated the ability of the primed CD4+ T cells to provide help for autoantibody responses by examining the numbers of Iga B cells recovered from recipient mice and the anti-chromatin IgMa serum antibody levels, eight days post-transfer. As expected, the TS1(SW).RAG2−/− CD4+ T cells that had been primed by the PR8 partial agonist had promoted little or no increase in the recovery of Iga B cells (Fig. 2D), and IgMa anti-chromatin antibodies were undetectable in the serum (Fig. 2E). The TS1(SW).RAG2−/− CD4+ T cells that had been primed with the SW virus induced efficient recovery of Iga B cells (Fig. 2D), but notably, the levels of serum IgMa autoantibody that were induced were significantly lower in these mice than in mice receiving TS1.RAG2−/− CD4+ T cells that had been primed by their respective agonist ligand (Fig. 2E). Thus, despite efficient activation of TS1(SW).RAG2−/− CD4+ T cells by the agonist SW virus at the priming stage, these cells were significantly less able to provide help for autoantibody production when interacting with the partial agonist PR8 HA at the effector stage (i.e. when PR8 HA is presented by the anti-chromatin B cells) than TS1.RAG2−/− CD4+ T cells that recognized the PR8 HA as an agonist during both the priming and the effector phase.
To examine how CD4+ T cells that have undergone differentiation to acquire a polarized phenotype interact with B cells expressing an agonist or partial agonist peptide, we cultured CD4+ T cells from TS1 and TS1(SW) mice in vitro with their respective agonist peptides under conditions that promote Th1 cell differentiation prior to transfer into recipient mice. We induced Th1 cell differentiation because Th1 cells and IgG2a antibody responses are characteristic of lupus-prone mice that develop spontaneous anti-chromatin autoantibody responses [18, 19]. As had been observed with undifferentiated CD4+ T cells, the transferred Th1 cells generated from TS1(SW) mice survived poorly in response to the partial agonist PR8 virus (Fig. 3A), promoted less efficient recovery of anti-chromatin B cells (Fig. 3A), and failed to induce detectable anti-chromatin IgMa or IgG2aa serum antibody responses (Fig. 3B). Also resembling the findings with undifferentiated T cells, the Th1 cells from TS1(SW) and TS1 mice primed with agonist peptides in vivo survived well and were similar in their ability to promote B cell recovery (Fig. 3A). In this case, Th1 cells from TS1(SW) mice activated with their agonist in vivo were able to support higher levels of anti-chromatin IgMa production than were found in unimmunized mice. However, the anti-chromatin antibody production was significantly lower compared to mice that received Th1 cells from TS1 mice that are able to recognize their agonist peptide during priming and T-B cell contact (Fig. 3B). Additionally, the Th1 TS1(SW) cells did not produce any detectable serum anti-chromatin IgG2aa antibodies (Fig. 3B). The chromatin ELISA and immunohistochemical analysis revealed that only the Th1 TS1 cells that recognized agonist peptide during priming and the effector phase could induce class switching to the IgG2a isotype (Fig. 3B–D). While IgG2aa switched autoantibodies were induced, and the cells found in the extrafollicular areas, there was no evidence that they participated in a germinal center reaction (Fig. 3E).
We have shown that the ability of a CD4+ T cell to recognize a target peptide as an agonist or a partial agonist can affect different stages of an autoantibody response. CD4+ T cells that had been primed by a partial agonist exhibited impaired proliferation, reduced ICOS levels, and provided inefficient help for autoantibody responses, contrasting a recent report in which low affinity antigens were found to induce CD8+ T cell effector functions . More strikingly, CD4+ T cells that had been activated by an agonist could enhance the survival of autoreactive B cells by interacting with a partial agonist, but were inefficient at supporting the production of IgM autoantibodies, and as differentiated Th1 cells failed to support class switching to the IgG2a subtype. Agonist peptides are characterized by their ability to support sustained interactions with the TCR , and such interactions may be necessary to recruit the full range of effector functions that a CD4+ T cell requires in order to promote autoantibody production. It is currently uncertain whether this is a general feature of T-B cell collaboration, or if autoreactive B cells have a unique requirement for highly activated CD4+ T cells in order to receive help. In either case, the studies here suggest that CD4+ T cells that evade central tolerance induction because they have low affinities for self-antigens may have a limited capacity to induce autoantibody production, even if they have been activated by pathogens with which they are strongly reactive.
Male and female mice between 6–16 weeks of age were maintained in specific-pathogen-free conditions at the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC)-accredited Wistar Institute under the supervision of the Institutional Animal Care and Use Committee (IACUC). All mice were on the BALB/c background. TS1 (anti-hemagglutinin (HA) from the PR8 influenza virus), TS1.RAG2−/−, TS1(SW) (anti-hemagglutinin (HA) from the SW influenza virus), TS1(SW).RAG2−/−, HACII, and sd-VH3H9.HACII.Ig κ−/− mice were bred at the Wistar Animal Facility. BALB/c (Iga) and CB17 (Igb) mice were purchased from the National Cancer Institute and Charles RiverLaboratory, respectively.
For naïve T cell experiments, TS1.RAG2−/− and TS1(SW).RAG2−/− mice were used. The spleen and peripheral lymph node cells were sorted on CD4+ cells and CFSE labeled . For the in vitro Th1-deviated experiments, lymph node cells from TS1 or TS1(SW) mice were removed, stained, and sorted for CD4+ CD25− cells .
0.5 × 106 CD4+ CD25− cells from TS1 or TS1(SW) mice were cultured in 24-well plates along with IL-2, 5 × 106 irradiated, red blood cell-depleted BALB/c splenocytes, and 1 μM HA S1(PR8) peptide (residues 110 through 120) or 1 μM S1(SW)(SFEKFEIFPKT), respectively . Additionally, Th1 cultures received anti-IL-4 (clone 11B11, cell supernatant) and murine rIL-12 purchased from Peprotech (Rocky Hill, NJ) . At day eight, cells were harvested and tested for cytokine production as previously described . The average cytokine responses for the two Th1 lines were: TS1, IFN-γ = 54.2% ± 4.0% and IL-4 = 1.7% ± 1.0%; TS1(SW), IFN-γ = 59.2% ± 2.3% and IL-4 = 1.5% ± 1.7% (n=3).
15 × 106 CFSE-labeled lymph node cells  from TS1 or TS1(SW) mice were injected i.v. into HACII or BALB/c mice. Three days later CD4+ 6.5+ (TS1 T cells) or CD4+ Vα8.3+ Vβ10+ T cells (TS1(SW) T cells) were analyzed for proliferation.
0.25 – 1 × 106 non-differentiated Th cells or Th1 cells were injected with 1000 hemagluttinating units of purified PR8 (A/Puerto Rico/8/34) or SW (A/SW/31) virus i.v. [10, 22]. One day after giving T cells and virus, splenocytes from sd-VH3H9.HACII.Igκ−/− mice were depleted of red blood cells, and an aliquot was surface stained to determine the frequency of anti-chromatin B cells (B220+ Igλ1+). CB17 recipient mice were injected with splenocytes containing 4 – 5 × 106 anti-chromatin B cells.
The T and B cell recoveries were determined with the following formula: % T or B cell recovery = (the number of T or B cells injected/the number of transferred T or B cells recovered from the spleen) × 100. This calculation was used to take into account small differences in T or B cell numbers that were injected because the number of cells recovered after the killing of the donor mice and/or purification steps varied. Thus, this calculation allowed us to compare all of our data in all the experiments.
Chromatin ELISAs were performed as previously described .
Splenic cells were surface stained  with the following Abs: anti-CD4-PerCP-Cy5.5 (RM4.5), anti-B220 (RA3-6B2), anti-Vα8.3-biotin (KT50) (BD Pharmingen), anti-Vβ10-PE (B21.5), anti-ICOS-PE (7E.17G9), rat IgG2b-PE isotype control (eB149) and streptavidin-allophycocyanin (eBioscience). The 6.5 biotin (anti-clonotype for the TS1 T cells ) was grown as a supernatantand biotinylated.
Spleens were frozen, sectioned, and stained  with the following antibodies from BD Pharmingen: anti-B220-FITC, anti-IgMa-biotin (DS-1), anti-IgG2aa-biotin (8.3) and Sigma: PNA-biotin. Secondary reagents were anti-FITC-horseradish peroxidase (HRP) and streptavidin-AP (Southern Biotechnologies).
Statistical significance was determined via the Wilcoxon signed-rank test, and significance was ascribed when p < 0.05.
The authors would like to thank Daniel Hussey, David Ambrose, and J. S. Faust from the Wistar Flow Cytometry Facility. Additionally, we thank Dr. Audrey Y. Park for critical review of the manuscript. J. E. is supported by the National Institutes of Health (AI32137 and AR47913) and A. J. C. by the National Institutes of Health (AI24541). Additional support for J. E. and A. J. C. are provided by the National Cancer Institute (CA 10815) and the Commonwealth Universal Research Enhancement Program, Pennsylvania Department of Health.
Conflict of interest: The authors declare no financial or commercial conflict of interest