(BALB/c × A)F1
Carboxyfluoroscein succinimidyl ester
Glucocorticoid-induced tumor necrosis factor receptor family-related gene
Chicken ovalbumin peptide 323–339
Pigeon cytochrome c peptide 88–104
T regulatory cell(s)
CD4+CD25+ regulatory T cells (Tregs) are critical for peripheral tolerance and prevention of autoimmunity. In vitro coculture studies have revealed that increased costimulation breaks Treg-mediated suppression in response to anti-CD3 or antigen. However, it was unclear whether loss of suppression arose from inactivation of Tregs or whether increased stimulationcaused Th cells to escape suppression. We have investigated conditions that allow or override Treg-mediated suppression using DO11.10 TCR-transgenic T cells and chicken ovalbumin peptide 323–339-pulsed antigen-presenting cells. Treg suppression of Th proliferation is broken with potent stimulation, using activated spleen cells and high antigen dose, but is intact at low antigen dose. Costimulation with CD80 and CD86 expressed on activated dendritic cells was essential for Th cell escape from suppression at a high antigen dose. Potently stimulated Tregs were functional since they reducedlevels of IL-2, IFN-γ, IL-4 and Th CD25 expression in cocultures. Furthermore, Tregs responding to high antigen dose and activated splenocytes retained the ability to suppress proliferation, but only of Th cells responding to a sub-optimal dose of independent antigen. Together, our results demonstrate that under conditions of strong antigen-specific stimulation, Tregs remain functional, but Th cells escape Treg-mediated suppression.
Adoptive transfer experiments have clearly demonstrated that CD4+CD25+ T regulatory cells (Tregs) are critical for the maintenance of peripheral tolerance and the prevention of autoimmune disease 1, 2. In rodents, Tregs constitute 5–10% of the spleen and LN CD4+ T cell pool and display a polyclonal TCR repertoire, suggesting that they respond to diverse antigens 3–5. When stimulated through their TCR, Tregs are anergic, but are capable of inhibiting CD4+ Th cell proliferationthrough direct cell contact 5–7. While the mechanism of suppression remains unknown, Tregs inhibit Th IL-2 mRNA expression and IL-2 secretion, thus restricting levels of the cytokine critical for naive Th cell proliferation. In addition, exogenous IL-2 breaks the Treg anergic and suppressive state, demonstrating the pivotal role IL-2 plays in regulating proliferation in coculture experiments 5, 7.
A major question of Treg biology concerns how the balance between Treg-mediated suppression and Th-mediated activation is controlled during an immune response. While TCR stimulation of Tregs appears to be necessary for induction of their suppressor function, experiments with T cells from TCR-transgenic mice have demonstrated that Tregs responding to one antigen can suppress Th cells responding to a different antigen, so-called bystander suppression 5, 8. For example, in the presence of ovalbumin peptide 323–329 (OVA), Tregs from DO11.10 TCR-transgenic mice effectively suppressed the proliferative response of BOG TCR-transgenic Th cells to an independent ovalbumin peptide (OVA 271–285). Furthermore, Treg suppressor function was elicited withmuch lower peptide doses than required for Th stimulation 5. Combined with the observations that CD4+CD25+ Tregs require high TCR affinity for thymic selection and display a memory phenotype, these data suggest that Tregs may have TCR with high affinity for self antigen and lower requirements for costimulation 9, 10. Thus itis possible that populations of Tregs may be continually stimulated by self antigen, and that Treg-mediated suppression must be overcome for Th to mount an immune response to foreign pathogens.
Several stimuli have been shown to influence the balance between Treg-mediated suppression and Th proliferation in vitro. In addition to exogenously added IL-2, suppression is abrogated with increased costimulation, such as addition of anti-CD28 mAb or in the presence of CD86-transfected P815 cells as APC 3, 5, 7. These findings are consistent with the observation that suppression is more efficient in the presence of B cells compared to DC, the latter expressing higher levels of CD80 and CD86 11. However, it was unclear in these studies whether loss of suppression was due to inactivation of Treg suppressor function or due to escape of Th cells from the effects of otherwise functional Tregs. Anti-CD28 mAb or IL-2 treatment are known to have direct effects on Tregs, since Tregs proliferate rather than remain anergic to TCR stimulation in their presence. However, with removal of these reagents, Tregs revert to an anergic and suppressive state 5, 8, 12.
Other conditions have been shown to modulate suppressor function through direct delivery of signals to Tregs. For example, agonistic anti-glucocorticoid-induced tumor necrosis factor receptor family-related gene (GITR) mAb abrogate suppression in vitro by signaling through GITR, which is expressed at high levels on Tregs 13, 14. In contrast, blockade of CTLA-4 on Tregs inhibits Treg-mediated suppression in vitro and elicits autoimmune disease in vivo12, 15. However, some effects of anti-CTLA-4 mAb may result from actions on the Th cells 16. Thus there may be several ways in which the balance between Treg suppression and Th activation can be shifted, including signals delivered to Tregs that inhibit or augment their suppressor function, signals to Th that promote escape from suppression, or alterations in the quality of signals delivered by APC to either cell type.
In the current study, we have examined the influence of antigen dose and APC type in modulating the balance between Treg-mediated suppression and Th activation in vitro. Using DO11.10 TCR-transgenic T cells, we show that Tregs fail to suppress proliferation of Th cells responding to activated APC and high dose antigen, but that suppression is favored when B cells compared to DC are used as APC. Costimulatory blockade on activated DC revealed that signals from CD80 and CD86 are critical for Th cells to escape from suppression at high antigen dose. Despite their failure to suppress Th proliferation under conditions of potent antigenic stimulation, the presence of Tregs leads to decreased Th CD25 cell surface expression, and reduced levels of IL-2, IFN-γ, and IL-4 in coculture supernatants. Also, potently stimulated DO11.10 Tregs retain the ability to suppress AND TCR-transgenic Th cells proliferating in response to an independent antigen, pigeon cytochrome c (PCC). Thus we demonstrate that in this in vitro system, potently stimulated Tregs are functional, but suppression of proliferation is lost because Th cells escape Treg-mediated effects. These findings have important implications for the in vivo manipulation of Tregs in efforts to control autoimmune disease or transplant rejection.
2.1 Flt3-L treatment increases the number of functional CD4+CD25+ Tregs in the spleen and MLN of TCR-transgenic mice
We sought to establish an antigen-specific experimental system to analyze conditions that alter the balance between Treg-mediated suppression and Th activation. However, a major limitation of studying Tregs is their rarity: CD4+CD25+ cells comprise only 3–5% of CD4+ T cell cells in DO11.10 mice. Human CD4+ T cells with regulatory properties can be expanded by coculture with immature DC in vitro17, 18. Multiple DC subpopulations are greatly expanded in the spleens and LN of Flt3-L-treated mice 19, 20. To determine whether in vivo expansion of DC also supports Treg expansion, we compared the number and cell surface phenotype of CD4+CD25+ cells recovered from the spleen and MLN of untreated or Flt3-L-treated DO11.10 TCR-transgenic mice. The fraction of CD4+ T cells that co-express CD25 (± SD) is significantly increased in the spleen (Flt3-L 7.0±0.9% vs. control 4.2±1.1%, p<0.001) and MLN (Flt3-L 8.3±5.4% vs. control 3.5±1.5%, p<0.01) following Flt3-L treatment (Fig. 1A). Tregs from Flt3-L-treated and control mice had identical cell surface phenotypes as assessed by staining for CD45RB, CD69 (Fig. 1B), as well as CD62L and OX-40 (data not shown). In addition, the absolute number of CD4+CD25+ increased more than threefold with Flt3-L treatment (Fig. 1C). We next tested the functional activity of CD4+CD25+ cells from Flt3-L-treated mice in vitro (Fig. 1D). CD4+CD25+ T cells from control and Flt3-L-treated DO11.10 mice were anergic to stimulation and potently suppressed OVA-specific proliferation of CD4+CD25– T cells across a wide antigen-dose range (5 nM–5 μM). Titration of Tregs in the coculture assay confirmed that suppressor activity of Tregs from Flt3-L-treated mice was comparable to controls (Fig. 1E). Thus, Flt3-L treatment expands the number of functional Tregs that can be isolated from TCR-transgenic DO11.10 mice.
2.2 Tregs fail to suppress Th cell proliferation elicited by DC and high antigen dose
Tregs from TCR transgenic mice activated by peptide and splenic APC efficiently suppress antigen-specific Th proliferative responses 3, 5, 7, 8, 13, 21. In the spleen, several potential cell types can serve as APC, each with different capacities to stimulate T cell function. Cederbom et al. 11 have previously shown that B cells allow more potent Treg-mediated suppression to anti-CD3 than DC. To extend these findings in an antigen-specific in vitro system, we compared the ability of freshly isolated whole splenocytes, B cells or DC to elicit peptide-induced regulation of Th proliferation. Tregs were anergic to a wide range of OVA doses presented by unstimulated whole spleen (containing B cells, DC, and macrophages), B cell or CD11c+ DC APC (Fig. 2A). Tregs interacting with whole spleen or B cell APC potently suppressed Th cell proliferation across a wide range of OVA doses. Thus, B lymphocytes support potent regulation by CD4+CD25+ T cells in vitro, even though they are less efficient at stimulating Th proliferative responses. DC also strongly supported suppression at lower peptide doses (5 nM and 50 nM). This finding is similar to that in a previous report showing Treg suppressor function at OVA doses up to 100 nM using unactivated spleen cell APC 5. However, suppression of proliferation was lost with increasing antigen dose presented by DC but not B cells. Taken together, our results suggest that Treg-mediated suppression is lost with more potent antigen-specific stimulation.
We next analyzed the suppression elicited by activated whole-spleen APC, B cells (cultured with LPS), and DC (cultured with LPS and GM-CSF). LPS activation increased CD80 and CD86 expression on all three APC types (Fig. 2B) and as expected, the activated APC stimulated more vigorous peptide-specific Th cell proliferation (Fig. 2A). Tregs were anergic to a wide range of OVA doses presented by all three types of activated APC. Tregs interacting with activated B cells (as with unactivated B cells) potently suppressed Th cell proliferation at all OVA doses tested. In contrast, Treg suppression of proliferation was evident at low, but not high antigen doses, when activated whole spleen was used as the APC. Furthermore, suppression of proliferation was lost at a log lower peptide dose when activated instead of unactivated DC were used as APC. This trend was evident in cultures harvested on day 2 and 3, and also in microscale cultures using as few as 3,000 T cells (data not shown), indicating that the relative change in proliferation and suppression was not due to culture overgrowth. Another method to increase potency of T cell stimulation in vitro is to increase DC number. At 50 nM OVA, titration of DC numbers revealed that Treg-mediated suppression was favored with lower numbers of activated or unactivated DC (Fig. 2C). At high antigen dose and activated DC, Tregs were unable to suppress proliferation even at a 1:10 DC:T ratio. These results are consistent with a previous report showing loss of Treg-mediated suppression with anti-CD3 and increasing DC numbers 11. Thus, potent stimulation conditions favor Th activation over Treg suppression.
The murine spleen contains several functionally and phenotypically distinct DC subsets 22, 23. As with B cell APC, Treg suppression was favored when CD11cdimB220+ plasmacytoid DC that express very low levels of costimulatory molecules and poorly induce Th cell proliferation 24, 25, were used as APC (data not shown). Thus, APC such as B cells and plasmacytoid DC that poorly costimulate Th cell function preferentially allow Treg-mediated suppression.
The high levels of CD80 and CD86 expressed by activated DC correlated with loss of suppression at a lower antigen dose (500 nM) compared to unactivated DC. To address the role that costimulation played on escape from suppression in vitro, we blocked CD80 and CD86 on activated DC in cocultures stimulated with OVA (Fig. 2D). At 50 nM of OVA, Th proliferation was suppressed either by blocking CD80 and CD86 or by coculture with Tregs. At 500 nM peptide, on the other hand, Th proliferation was unaffected by costimulatory blockade. This result is consistent with the ability of high antigen dose to induce Th proliferation in the absence of B71/ B72 costimulation 26. However, in Th cocultures with Tregs at 500 nM OVA, CD80/86 blockade resulted in suppression of Th proliferation. Thus, increased costimulation allows Th cells to escape Treg-mediated suppression during an antigen-specific response.
2.3 With potent stimulation, the presence of Tregs leads to decreased Th cell CD25 expression and cytokine levels in coculture
As shown above, Tregs fail to control Th proliferation to high doses of antigen and activated APC in vitro. It is possible that under these conditions, Treg suppressor function is inactivated, or that Treg suppressor function is intact but Th cells escape regulation. To investigate this matter, we first analyzed cellular proliferation and CD25 expression on CFSE-labeled Th cells with high and low dose of OVA in the presence or absence of Tregs (Fig. 3). Consistent with our results in Fig. 2, Tregs greatly inhibited Th cell division to low dose (50 nM) but not high dose (5 μM) peptide (Fig. 3A). While CFSE labeling indicated that some initial Th proliferation occurred in response to 50 nM of OVA, Treg-mediated suppression was intact by day 4. This observation is consistent with previous time-course studies using anti-CD3 stimulation which showed some initial IL-2 production that was then shut down in Treg cocultures at later time points 5, 7. With 5 μM of OVA, Th cell division was unaffected by the presence of Tregs. Despite unhindered proliferation at this OVA dose, Th cells showed decreased levels of CD25 in the presence of Tregs (Fig. 3B). As expected, the loss of CD25 expression was more pronounced on Th stimulated with low dose OVA. These results suggest that although Tregs cannot control proliferation of potently stimulated Th cells, they can reduce Th CD25 surface expression, and thus have not been rendered completely non-functional by stimulation with high-dose antigen.
We next analyzed cytokine levels in cultures of Tregs and/or Th cells responding to OVA and freshly isolated or activated splenic APC. Tregs failed to secrete substantial amounts of IL-2, IFN-γ, or IL-4 when stimulated using unactivated or activated APC (Fig. 4A). There was an OVA dose-dependent increase in IL-10 present in supernatants when Tregs were cultured with activated APC, suggesting that Tregs did retain some function under conditions of strong stimulation. These results are consistent with the findings that anti-CD3-stimulated Treg cultures increase IL-10 but not IL-2 mRNA expression 5, 7. Notably, levels of IL-2, IL-4, and IFN-γ in both high and low dose of OVA-stimulated Th cultures were significantly reduced by the presence of Tregs. Despite the lower cytokine levels, we confirmed in the same experiment that Tregs did not suppress Th proliferation in response to activated APC and high dose of OVA (Fig. 4B). These data suggested that potently stimulated Tregs may be functionally capable of suppressing Th cytokine secretion, but fail to adequately suppress T cell proliferation. However, it is also possible that Tregs may consume cytokines and contribute to the decrease in their levels in cocultures. Although IL-2 levels in Th cultures were markedly decreased by the presence of Tregs, significant amounts of IL-2 were present in cocultures stimulated with LPS-activated APC 0.5 and 5 μM of OVA (5.1 and 21.1 ng/ml, respectively). In fact, addition of as little as 3 ng/ml IL-2 will override Treg-mediated suppression of Th proliferation in vitro7. Thus, these results suggest that potently stimulated Th cells may produce sufficient levels of IL-2 to override Treg-mediated suppression and to drive their own proliferation.
2.4 Potently stimulated DO11.10 Tregs retain the ability to suppress AND TCR-transgenic Th cells proliferating in response to an independent antigen
We have demonstrated that Tregs fail to suppress proliferation under conditions of potent stimulation, but appear to retain the capacity to decrease Th CD25 expression and cytokine levels in cocultures. To distinguish whether loss of suppression at high peptide doses reflects a deficit in Treg function or if highly stimulated Th cells are refractory to Treg suppression, we developed a two-antigen regulation assay. This assay allows independent titration of specific peptide to Th and Tregs, as similarly described by others 5, 8. In our assay, DO11.10 (I-Ad and OVA-specific) and AND mice (I-Ek and PCC-specific) were used as sources for Treg and Th cells. For APC we used activated [(BALB/c × A)F1 (CAF1), I-Ad/I-Ek expressing] splenocytes capable of presenting antigen to both types of transgenic T cells. As expected, DO11.10 Tregs effectively suppressed DO11.10 Th proliferation to OVA presented by activated CAF1 splenocytes at low, but not high, dose (Fig. 5A, top panel). At low antigen dose (50 nM), the suppressive effect was gradually lost with lower Treg:Th ratios (Fig. 5B, top panel). At 5 μM, however, coculture proliferation was increased with increasing Treg numbers. These data suggest that once Tregs lose the ability to control Th proliferation, their presence actually increases the proliferative response to stimulus, likely because their anergic state is broken. In contrast, DO11.10 Tregs efficiently suppressed PCC-specific Th proliferation across the entire OVA dose range (Fig. 5A, middle panel), in a Treg cell number-dependent fashion (Fig. 5B, bottom panel). Furthermore, Tregs pre-stimulated with activated DC and 5 μM of OVA were more potent suppressors of DO11.10 Th cell proliferation than control Tregs (data not shown). Thus, Tregs stimulated with high dose of OVA and activated APC retain the potential to suppress Th proliferation. We also analyzed the functional activity of Tregs from AND mice, and determined that they efficiently suppress AND Th proliferation to 10 μM PCC and activated APC (data not shown). This PCC dose was chosen because it was a mid-optimal dose for stimulating Th cell proliferation using the CAF1 APC (data not shown). Finally, PCC-activated AND Tregs suppressed DO11.10 Th cell proliferation at low (5–50 nM), but not high (5 μM), OVA doses (Fig. 5A, bottom panel). Thus, with sufficient stimulation, Th cells can escape Treg-mediated suppression of proliferation in this antigen-specific in vitro system.
Our findings suggest that the ability of Tregs to suppress Th proliferation depends on the potency of stimulation by antigen and APC. We show that Th proliferation is refractory to Treg-mediated suppression under conditions of potent antigenic stimulation in vitro, such as when high antigen doses are presented by activated APC. Previous studies in which increased costimulation broke Treg suppression did not clarify whether loss of suppression resulted from Treg inactivation or whether Th cells became refractory to Treg-mediated suppression. In addition, studies analyzing the mechanism for loss of suppression to high dose antigen and specific APC have not yet been reported. We show that increased costimulation allows Th cells to escape Treg-mediated suppression of proliferation during an antigen-specific response. Tregs activated with strong antigen-stimulation conditions function as bona fide suppressor cells because they suppress Th CD25 expression, limit IL-2, IL-4, and IFN-γ Th cytokines in coculture supernatants, and have the potential to suppress proliferation of Th cells. However, with sufficient antigen dose and costimulation, Th cells overcome the suppressive functions of activated Tregs. Thus, the presence of activated Tregs does not preclude Th cell activation, but appears to increase the threshold required to achieve a functional Th cell response.
Understanding the molecular mechanisms that control Treg-mediated suppression has become an intense area of study. Recently, it has been shown that Treg suppressor function can be turned off by signaling through GITR expressed by Tregs 13, 14. Thus, a potential mechanism of escape from suppression could involve the encounter of Tregs to GITR-ligand-expressing cells in the context of an activating immune environment. Our data imply that functional Tregs may also fail to control immune activation when Th cells encounter high antigen concentrations and increased costimulation. Our data echoes findings of a recent study showing that DO11.10 Th cells stimulated with high, but not low, OVA doses can overcome the suppressive effects of alloreactive T cells produced by ex vivo CD40:40L blockade 27. Thus, mechanisms also exist that allow Th cells to escape from regulatory cell suppressor function in order to generate an antigen-specific immune response.
IL-2 mediates the antigen-specific expansion and differentiation of Th cells into effector cells, and thus is critical for the initiation of the adaptive immune response. IL-2 also breaks the anergic and suppressive state of Tregs, and central to the capacity of Tregs to suppress immune responses is their ability to inhibit IL-2 production by naive Th cells 5, 7. Our results show that with strong stimulation, Th cells proliferate and secrete significant amounts of IL-2, and that IL-2 levels remain high (>5 ng/ml) even in the presence of Tregs. Whether the IL-2 came only from Th cells, or if Th cells induced Tregs to secrete IL-2 was not addressed. Notably, exogenous addition of as little as 3 ng/ml of IL-2 is sufficient to block the immune suppressive function of Tregs 7. Together, these results suggest a model in which the amount of IL-2 locally produced in response to potent stimulation is sufficient to allow Th cells to proliferate in the presence of Tregs.
While Tregs inhibit immune activation and protect the host from autoimmune disease, Treg-mediated suppression does not generally prevent immune responses to harmful pathogens. Like memory T cells, Tregs can be activated with much lower doses of peptide than are required by naive Th cells, and Treg function does not require CD28 signaling 5, 28, 29. Thus, the stimulation threshold required for Treg activation is lower than that for naive Th cells. Under poorly immunogenic conditions, autoreactive Th cells would encounter self-antigens in the absence of strong costimulatory signals. Tregs activated under these conditions would fully suppress immune activation, thereby preventing autoreactivity. However, our results also indicate that potent stimulation allows Th cells to overcome Treg-mediated suppression. During an emerging pathogenic infection, many factors contribute to increase the potency of stimulation. DC responding to pathogen-associated molecular patterns (such as LPS) up-regulate expression of costimulatory molecules and increase their capacity to present antigen. With increasing costimulation, increasing localpathogen-derived antigen concentration, and increasing antigen presentation capacity of APC, Th cells attain the ability to respond in the presence of Tregs. Thus, the incomplete dominance of Treg-mediated suppression would allow for the initiation of an appropriate immune response to pathogens and provide a mechanism to suppress autoreactive effector Th cells.
4 Materials and methods
4.1 Mice and Flt3-ligand (Flt3-L) treatment
All mice were housed at Immunex Corp. (now Amgen Corp., Seattle, WA) under specific pathogen-free conditions following Immunex IACUC guidelines. DO11.10 mice 30 were bred and maintained in house. CAF1 mice and AND mice 31 were from Jackson Laboratory (Bar Harbor, ME). BALB/c mice were from Charles River Laboratories (Wilmington, MA). Flt3-L-treated mice were injected i.p. once daily with 10–20 μg of recombinant human Flt3-L (Immunex Corp.) for 10–16 consecutive days 20. Statistical analysis (Alternate Welch's unpaired t-test) was performed using InStat software (GraphPad Software Inc., San Diego, CA).
4.2 mAb and flow cytometry
The following conjugated mAb were obtained from BD PharMingen (San Diego, CA): purified rat IgG2a, κ and hamster IgG2, κ; APC-anti-CD3; FITC-anti-CD3, CD19, CD45RB, CD69, CD80, CD86; PE-anti-CD11c, and CD25 (PC61); PerCP-Cy5.5-anti-CD4. FITC-conjugated isotype controls were: rat IgG2a (R35–95) for CD45RB and CD86; hamster IgG group 1 (A19–3) for CD69; and hamster IgG group 2 (B81–3) for CD80. Anti-Fc receptor CD16/32 (2.4G2), anti-CD80 (16–10A1) and anti-CD86 (GL1) were affinity purified at Immunex. Single-cell suspensions were incubated in PBS containing 5% FCS (HyClone, Logan, UT) and 2.4G2, followed by conjugated mAb. Following staining, cells were washed and fixed in PBS containing 1% paraformaldehyde. Flow cytometry was performed ona FACSCalibur machine and analyzed using CellQuest software (Becton Dickinson, San Jose, CA).
4.3 Isolation of T cell populations
Single-cell suspensions of MLN and RBC-lysed (Sigma, St. Louis, MO) splenocytes from untreated or Flt3-L-injected AND or DO11.10 mice were washed twice in IMDM (Invitrogen, Carlsbad, CA) containing 5% FCS, 5.5×10–5 M 2-ME (Invitrogen), 1 mM sodium pyruvate, 100 μM nonessential amino acids, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine (cIMDM). Spleen and MLN suspensions were then combined and enriched for CD4+ T cells by negative immunomagnetic selection using a CD4 enrichment cocktail (Stem Cell, Vancouver, BC). Labeled cells were passed over a 0.6-inch pump-feed column (Stem Cell) in a magnetic field (Blue magnet, Stem Cell). CD4-enriched cells were stained with PerCP-Cy5.5-anti CD4 and PE-anti-CD25 and separated into CD4+CD25– (Th) and CD4+CD25+ (Treg) populations using a FACSVantage (Becton Dickinson) or MoFlo (Cytomation, Ft. Collins, CO) cell sorter. In some cases, Th cells were labeled for 10 min at 37°C with 5 μM of carboxyfluoroscein succinimidyl ester (CFSE, Molecular Probes, Eugene, OR) at 3×107/ml in IMDM and washed twice in cIMDM.
4.4 Isolation and preparation of APC populations
RBC-lysed BALB/c splenocytes were washed twice and resuspended in cIMDM for use directly in proliferation assays (in the case of whole-spleen APC) or enriched for B cells by negative immunomagnetic selection using a B cell enrichment cocktail (Stem Cell). More than 97% of cells eluted in the flow-through fraction were CD19+. Activated whole-spleen or B cells were obtained following 2 days of culture in 10 μg/ml of LPS (Sigma). Whole spleen or B cell APC were irradiated (3,000 rad from a 137Cs source) prior to use in proliferation assays. For DC-enriched APC, single-spleen cell suspensions from Flt3-L-treated mice were subjected to Nycodenz gradient centrifugation (Invitrogen), followed by immunomagnetic positive selection of Nycodenz-buoyant cells using CD11c microbeads (Miltenyi Biotec) and an AutoMACS cell separator (Miltenyi Biotec). More than 97% of eluted cells were CD11c+. DC were either used fresh or cultured overnight in 10 μg/ml of LPS. Murine GM-CSF (20 ng/ml, Immunex Corp.) was included in LPS-stimulated DC cultures to promote DC survival and maturation 32.
4.5 Proliferation assays
Th (1.5–2.5×104), Treg (1:1, 1:3 or 1:10 ratio to Th), APC (whole-spleen or B cell at a 3:1 ratio to Th, or CD11c+ DC at a 1:2 ratio to Th, except where indicated otherwise), and peptide (5 nM–5 μM OVA, 10 μM PCC) in cIMDM were added to 96-well U-bottom plates (Corning, Corning, NY) for a total of 200 μl/well. Some cultures were incubated with 10 μg/ml of anti-CD80 and anti-CD86, or 10 μg/ml of hamster IgG2 and rat IgG2a isotype controls. Cultures were maintained at 37°C, 5%CO2 for the indicated time periods. OVAand PCC were synthesized and purified by HPLC at Immunex. Th CSFE content and CD4 and CD25 expression were analyzed after 2–4 days of culture. For analysis of [3H]dTdh incorporation, plates were incubated with 1 μCi [3H]dTdh (Amersham Biosciences, Piscataway, NJ) per well during the last 24 h of a 96-h culture period. Contents of each well were transferred to Filtermat Aglass fiber filters (Wallac, Turku, Finland) using the Brandel harvester (Brandel, Gaithersburg, MD) and read on a TriLux 1450 MicroBeta counter (Wallac). Data are presented as the mean of triplicate cultures, with the S.D. indicated for each condition.
4.6 Analysis of cytokine production
Levels of IL-2, IL-4, IL-10, and IFN-γ in 72-h culture supernatants from T cell proliferation assays were determined using the Beadlyte Mouse Multi-Cytokine Detection System (Upstate Biotechnology, Lake Placid, NY) and the Luminex100 plate reader (Luminex Corporation, Austin, TX) according to manufacturers' instructions. Quantification of cytokines was performed by regression analysis from a standard curve generated from cytokine standards included in the kit. Lower limits of detection were: 2 pg/ml for IL-2, 0.2 pg/ml for IL-4, 35 pg/ml for IL-10, and 3 pg/ml forIFN-γ. Data are presented as the mean of triplicate cultures, with the S.D. indicated for each condition.
We thank Steve Brady, Julie Hill and Daniel Hirschstein for expert flow cytometry assistance, Heidi Jessup and Marina Martinez for help with cell isolations, the Immunex DC Biology group for input on DC studies, and Drs. Mark Gavin, Thibaut De Smedt, David Fitzpatrick, and Charlie Maliszewski for critical review of the manuscript.