Synthesis of several chemokines but few cytokines by primed uncommitted precursor CD4 T cells suggests that these cells recruit other immune cells without exerting direct effector functions


  • Li Yang,

    1. David H. Smith Center for Vaccine Biology and Immunology, and the Department of Microbiology and Immunology, University of Rochester Medical Center, Rochester, USA
    Search for more papers by this author
  • Tim Mosmann

    Corresponding author
    1. David H. Smith Center for Vaccine Biology and Immunology, and the Department of Microbiology and Immunology, University of Rochester Medical Center, Rochester, USA
    • David H. Smith Center for Vaccine Biology and Immunology, University of Rochester Medical Center, 601 Elmwood Avenue, Box 609, Rochester, NY 14642, USA Fax: +1-5852732452
    Search for more papers by this author


Antigen-stimulated naive CD4 T cells may differentiate into effector T cells such as Th1 and Th2 cells, or may remain as proliferating but uncommitted, primed, precursor cells (Thpp cells) that can subsequently differentiate into Th1 or Th2 cells in appropriate cytokine environments. To examine potential Thpp effector functions, we compared the genes expressed by mouse Thpp, naive, Th1 and Th2 cells, using Affymetrix GeneChip and RNase Protection assays. Similar to naive CD4 T cells, Thpp cells expressed IL-2 but not the cytokines characteristic of differentiated Th1 or Th2 cells, such as IFN-γ, IL-4, or IL-5. However, Thpp, Th1 and Th2 cells, but not naive cells, expressed several CC chemokines including CCL1/TCA3, CCL5/RANTES, CCL3/MIP-1α, CCL4/MIP-1β, and CCL9/MIP-1γ. Secretion of the corresponding proteins was confirmed by ELISA and Elispot. Consistent with this chemokine expression, supernatants of activated Thpp, Th1 and Th2 cells but not naive CD4T cells induced pertussis toxin-sensitive chemotaxis of B and T cells. Supernatants of Thpp cells did not bias differentiation of naive CD4 T cells towards either Th1 or Th2 cells. The secretion of several chemokines, but few cytokines, by primed uncommitted Thpp cells suggests that their activation during an immune response may recruit effector cells without directly polarizing effector functions.


Enzyme-linked spot assay




Pertussis toxin

Thpp cells:

Primed precursor CD4 T cells

1 Introduction

Naive CD4 T cells have limited cytokine expression and secrete only IL-2 before they differentiate into Th1 or Th2 effector cells upon antigen stimulation. Th1 cells secrete IL-2, IFN-γ and lymphotoxin (LT) and participate in responses to intracellular immune pathogens, whereas Th2 cells mainly secrete IL-4, IL-5, IL-10 and IL-13 and mediate anti-helminth and allergic responses 1, 2.

However, development into Th1 and Th2 effector phenotypes is not the only potential fate of naive CD4 T cells. Upon antigen stimulation in vitro in the presence of TGF-β and anti-IFN-γ Ab, naive CD4 T cells proliferate and differentiate into cells expressing high levels of CD44 and secreting IL-2, but not IL-4 or IFN-γ 3. These primed precursor CD4T cells (Thpp cells) 4 can further differentiate into either Th1 or Th2 effector cells in appropriate cytokine environments 3. Similar to naive cells, Thpp cells required 3–5 days of differentiation before expressing Th1 or Th2 phenotypes 4. After adoptive transfer, Thpp-like cells (non-polarized T cells) migrated preferentially to lymph nodes, whereas Th1 and Th2 cells homed preferentially to spleen and nonlymphoid tissues 5. Thus, Thpp cells are primed (CD44high) CD4 T cells that produce IL-2 but not IFN-γ or IL-4, home selectively to lymph nodes and retain the ability to differentiate into either Th1 or Th2 cells.

Cells similar to the in vitro-defined Thpp cells may also exist in vivo in mice. Jenkins et al. 6 showed that CD4 T memory cells in lymph nodes preferentially produced IL-2, and not IFN-γ. We showed 7 that CD4 T cells primed by several different antigens became CD44high cells that produced IL-2 but not IFN-γ or IL-4, were found preferentially in lymph nodes rather than spleen and could differentiate into either IFN-γ- or IL-4-producing cells in vitro, as shown by split cloning experiments. The IL-2-only phenotype can be retained even after extensive proliferation in vivo (X. Wang and T. R. Mosmann, unpublished data). Thus, Thpp-like, IL-2-secreting, uncommitted cells can be found in mouse lymphoid tissue in vivo.

Similar IL-2-producing cells may exist in humans. Sallusto et al. 8 identified two types of CD4 T memory cells in human peripheral blood. CCR7high CD45RA CD4+ T cells (central memory, Tcm) produced IL-2, but not IFN-γ or other effector cytokines. This population contained precursors of both IL-4- and IFN-γ-producing cells. The CCR7low CD45RA CD4+ cell population produced both IL-4 and IFN-γ in addition to IL-2. Using a modified two-color enzyme-linked spot (Elispot) assay, we have identified substantial numbers of human T cells producing IL-2, but not IFN-γ, in response to recall antigens (A. Divekar and T. R. Mosmann, unpublished data). Although these human IL-2-producing T cells share some properties with the mouse Thpp cells described above, other groups 911 have found that CCR7 does not clearly discriminate IFN-γ producers from non-IFN-γ producers among either CD4 or CD8 T cells. Thus, some primed human CD4 T cells may be equivalent to the mouse Thpp cells, but the expression of CCR7 on these cells is controversial.

The Thpp cells may provide an expanded pool of cells that add flexibility to host defense by subsequently differentiating into Th1 or Th2 cells, as required. However, because these cells are antigen-specific, it may be advantageous for the immune system if they expressed additional functions to enhance immune responses. Previous studies showed that Thpp cells do not express typical Th1 or Th2 cytokines such as IL-4, IL-5, and IFN-γ 3, and so we wished to analyze the expression of additional effector proteins of these cells.

To examine a much wider range of potential effector proteins, we have analyzed Thpp cell gene expression by Affymetrix gene chips, particularly focusing on genes selectively expressed after TCR-mediated activation. Activated Thpp cells expressed few cytokine genes, but did express several chemokine genes at high levels. We confirmed that Thpp cells secreted these chemokine proteins and that activated Thpp cell supernatants potently induced pertussis toxin (PT)-sensitive lymphocyte migration. However, Thpp cell supernatants did not strongly influence naive CD4 T cell differentiation. Thus, Thpp cells, in contrast to naive CD4 T cells, may attract other immune cells into a site of antigen reactivity, without strongly biasing the type of the ensuing effector response.

2 Results

2.1 In vitro production of allospecific primed precursor CD4 T cells

To obtain large numbers of well-defined Th1, Th2 and Thpp cells suitable for gene expression analysis, short-term allospecific CD4 T cells were generated in vitro. Allostimulation was chosen because this stimulates a high percentage of naive CD4 T cells, allowing rapid generation of large numbers of normal T cells with diverse T cell receptors. TCR-transgenic T cells were not used, as the expression of a single TCR affinity by all cells might result in skewed differentiation that would not represent the range of phenotypes in normal populations.

Naive mouse CD4 T cells (CD4+, CD44low and CD62Lhigh) isolated from normal C57BL/6 (H-2b) spleen and lymph node cells were stimulated with irradiated TA3 (H-2d/k) cells in the presence of TGF-β and anti-IFN-γ Ab to induce differentiation into Thpp cells 3. Control Th1 and Th2 populations were produced by culturing with IL-12 or IL-4, respectively. After 6 days, stimulation of the resulting cell populations (>99% CD4+, unpublished data) with plate-bound anti-CD3 and anti-CD28 Ab for 24 h induced secretion of IL-2 by the Thpp cultures; IFN-γ and IL-2 by the Th1 cultures; and IL-4 by the Th2 cultures (Fig. 1A). These cytokine secretion patterns were confirmed by Elispot assay (unpublished data).

To confirm that these in vitro-derived Thpp cells had the ability to further differentiate, cells grown in anti-IFN-γ Ab + TGF-β for 6 days were restimulated under either Th1 (IL-12) or Th2 (IL-4) conditions. Fig. 1B indicates that IL-12 and IL-4 induced further differentiation of Thpp cells into cells secreting IFN-γ or IL-4, respectively. These results demonstrate that the allospecific Thpp cells express the properties described previously and have a high purity appropriate for analysis of gene expression by microarray analysis.

Figure 1.

 Characterization of T cell populations used for gene expression analysis. (A) Primary differentiation: naive CD4 T cells were stimulated with irradiated TA3 cells under Th1, Th2 and Thpp conditions. After 6 days, aliquots of the cells were restimulated with plate-bound anti-CD3 + anti-CD28 Ab for 24 h. Cytokine levels in supernatants were measured by ELISA. (B) Secondary differentiation: differentiated Thpp cells (A) were restimulated with irradiated TA3 under Th1 or Th2 conditions. After 7 days, these cells were washed and restimulated with plate-bound anti-CD3 + anti-CD28 Ab for 24 h. Cytokine levels were measured by ELISA.

2.2 Activated Thpp cells express few cytokine genes

To identify induction-specific genes in Thpp cells, we compared gene expression in resting and activated Thpp, Th1 and Th2 cells as well as sorted naive (CD44low, CD62Lhigh) CD4 T cells. For each cell type, gene chip hybridization revealed that the vast majority of genes were expressed at similar levels before and after activation. A small number of genes, mostly cytokine or chemokine genes, showed large increases after cell stimulation. Consistent with the results in Fig. 1 and our previous study 7, Thpp and naive T cells expressed IL-2, but not IFN-γ (Fig. 2A). Interestingly, both Thpp and naive T cells expressed moderate levels of LT. As expected, activated Th1 cells, but not Th2 cells, expressed high levels of all three genes.

Thpp (and naive) cells also expressed negligible levels of mRNA for cytokines expressed strongly by Th2 cells (IL-4, IL-5, IL-6, IL-10, and IL-13, Fig. 2B) and for cytokines expressed by both Th1 and Th2 cells (IL-3, GM-CSF, and IL-9, Fig. 2C). TNF-α was expressed at low but significant levels by both naive and Thpp cells. A striking difference between naive and Thpp cells was that Thpp cells, but not naive T cells, expressed high levels of the neuropeptide precursor pre-proenkephalin, which is also strongly expressed in activated Th1 and Th2 cells 12.

Figure 2.

Activated Thpp cells express few cytokine genes. Specific cytokine gene expression was evaluated by gene chip analysis of the cell populations described in Fig. 1. (A) Th1-specific genes, IL-2, LT and IFN-γ. (B) Th2-specific cytokine genes, IL-4, IL-5, IL-6, IL-10 and IL-13. (C) Additional Th1/Th2 genes, IL-3, IL-9, GM-CSF, TNF-α, LTβ and pre-proenkephalin. The expression levels are represented by fluorescence units.

2.3 Activated Thpp cells express high levels of certain CC and C chemokine genes

In contrast to their lack of expression of most cytokine genes, Thpp cells (as well as Th1 and Th2 cells, but not naive cells) expressed substantial levels of mRNA for several CC and C chemokine genes, including CCL3/MIP-1α, CCL4/MIP-1β, CCL1/TCA3 and XCL1/lymphotactin (Fig. 3). Two more CC chemokines, CCL5/RANTES and CCL9/MIP-1γ, were also expressed at moderate levels in Thpp and Th2 cells, and at higher levels in Th1 cells. Although significant levels of CCL4/MIP-1β and XCL1/lymphotactin mRNA were detected in naive cells, these levels were much lower than in all other T cell types and could conceivably represent low levels of contaminating effector cells. For all expressed chemokines, activation by anti-CD3 + anti-CD28 Ab increased expression, in many cases more than 100-fold. The expression of mRNA for these CC and the C chemokine was confirmed by RNase protection assay or RT-PCR (data not shown).

Many other CC chemokines, all CXC chemokines represented on the gene chip and the CX3C chemokine C3XCL1/Fractalkine were not significantly expressed by any of the four T cell types analyzed (all p>0.01), except for low expression of CXCL2/MIP-2 by activated Th2 cells. Table 1 summarizes these expression patterns and also shows the close chromosomal linkage between the five CC chemokine genes expressed by Thpp, Th1 and Th2 cells (between 47.3 and 47.6 cM on chromosome 11).

Figure 3.

 Activated Thpp cells express some chemokine genes. Gene chip hybridization was used to evaluate chemokine gene expression by activated and non-activated CD4 T cells. The expression levels are represented by fluorescence units.

Table 1. Chemokine gene expression profile of Thpp, naive, Th1 and Th2 CD4 T cells
  1. a) Average fluorescence values of hybridization to gene chip chemokine probe sets. Values in bold represent significant expression (p<0.01).

original image

2.4 Activated Thpp cells secrete high levels of CC chemokines

The results above show that activated Thpp cells expressed mRNA for CC and C chemokines. As some genes are regulated at the translational level, we used ELISA and Elispot assays to test whether Thpp cells also synthesized and secreted these chemokines. CCL3/MIP-1α, CCL4/MIP-1β and CCL1/TCA3 were detected at high levels in the supernatants of activated Thpp, Th1 and Th2 cells, but not naive CD4 T cells (Fig. 4A). The effectiveness of polarization of each T cell population was shown by the levels of cytokines secreted: Th1 and Th2 cells produced more than 100-fold greater amounts of IFN-γ and IL-4, respectively (Fig. 4A). Th1 cells, and to a lesser extent Thpp and Th2 cells, also produced CCL5/RANTES and CCL9/MIP-1γ. Elispot analysis (Fig. 4B) showed that a high frequency of Thpp cells secreted CCL3/MIP-1α and CCL4/MIP-1β and a lower frequency secreted CCL5/RANTES. Thus, activated Thpp cells secreted significant amounts of all five of the CC chemokines identified by RNA analysis.

Figure 4.

 Activated Thpp, Th1 and Th2 cells, but not naive cells, secreted CCL1/TCA3, CCL3/MIP-1α, CCL4/MIP-1β, CCL9/MIP-1γ, and CCL5/RANTES. (A) Naive and in vitro-differentiated Thpp, Th1 and Th2 cells were restimulated by plate-bound anti-CD3 + anti-CD28 Ab for 24 h. Chemokines and cytokines in the supernatants were quantified by ELISA. (B) Naive CD4 T cells and in vitro-differentiated Thpp, Th1 and Th2 cells were restimulated with alloantigen (TA3 cells) in chemokine Elispot assays. These results are representative of more than three experiments.

2.5 Supernatants of activated Thpp cells contain chemoattractant activity but do not bias T cell differentiation

To determine whether chemokines secreted by Thpp cells were biologically active, we tested the ability of supernatants of stimulated T cells to induce migration of normal spleen cells. Supernatants of activated Thpp, Th1 and Th2 cells were more chemoattractive for spleen cells than activated naive T cell supernatants or control migration medium only (Fig 5a, solid bars) or non-activated T cells (data not shown). Surface antigen analysis of the migrating cells showed that Thpp, Th1 and Th2 cell supernatants all induced significant migration of CD4, CD8 and B cells (data not shown).

To test whether this chemoattractant activity was due to chemokines, target spleen cells were pretreated with PT. ADP-ribosylation of Gαi subunits by PT abrogates signaling through Gαi-coupled receptors, including the chemokine receptors. PT reduced chemotaxis in response to Thpp, Th1 and Th2 cell supernatants (gray bars, Fig. 5A), indicating that the majority of the chemotactic activity in each supernatant was due to ligands acting through G protein-coupled receptors of target cells. Thus, activated Thpp cells produce Gαi-dependent chemoattractant activity, consistent with their secretion of chemokines detected by ELISA and Elispot.

The lack of secretion of IL-4 or IFN-γ by Thpp cells suggests that activated Thpp cells would not strongly influence naive CD4 T cell differentiation. However, chemokines may also affect differentiation under some circumstances 13, so we tested the ability of T cell supernatants to regulate the differentiation of naive CD4 T cells. As expected, Th1 and Th2 cell supernatants strongly biased the differentiation of naive CD4 T cells towards cells secreting more IFN-γ or more IL-4, respectively. In contrast, Thpp cell supernatants did not enhance the levels of either IL-4 or IFN-γ. Thpp cell supernatants also did not prevent IL-12 or IL-4 from inducing naive T cells to differentiate into Th1 or Th2 cells, respectively (Fig. 5B).

Figure 5.

 Biological activities of supernatants from activated Thpp cells. (A) Migration of untreated (solid bars) or PT-treated (gray bars) spleen cells in response to activated T cell supernatants (2%) was tested in triplicate in transwell migration assays. Background migration values of 0.98% (untreated) and 0.47% (PT-treated) target cells have been subtracted. These results are representative of more than three experiments (untreated cells) or two experiments (PT-treated cells) at T cell supernatant concentrations ranging from 0.08% to 10%. (B) Naive CD4 T cells were induced to differentiate in the presence of the indicated combinations of activated CD4 T cell supernatants (tested at 2%) and cytokines. After 5 days, cells were washed and stimulated with anti-CD3 + anti-CD28 Ab. Then cytokine levels in the supernatants were measured by ELISA. These results are representative of three experiments, at T cell supernatant concentrations ranging from 0.4% to 10%.

3 Discussion

We have previously suggested that primed but uncommitted CD4 T cells expressing IL-2 but few other cytokines 3, 7 may function as an expanded pool of uncommitted, antigen-specific cells that provide flexibility for future immune responses by subsequently differentiating into effector T cell phenotypes. The present study confirmed that the in vitro-derived, primed, uncommitted Thpp cells do not express IL-4, IL-5 or IFN-γ, and also showed that Thpp did not express other effector cytokines, including IL-3, IL-6, IL-9, IL-10, IL-13 or GM-CSF. However, a more active role of Thpp cells in immune responses is now suggested by our finding that Thpp cells synthesize several chemokines at levels similar to those of Th1 and Th2 cells. These chemokines appeared to be biologically active, as supernatants of activated Thpp (but not naive) cells induced PT-sensitive chemotaxis of T and B lymphocytes. Thus, the primed but uncommitted Thpp cells express additional functions compared to naive cells but do not express the strongly polarized functions typical of effector cells such as Th1 and Th2.

The gene chip results for chemokines and cytokines showed good agreement with additional RNA assays (RNase Protection Assay and RT-PCR), as expected for genes that were induced very strongly (10–1,000-fold). Protein assays (ELISA and Elispot) also revealed very similar chemokine and cytokine expression patterns, suggesting that transcription or RNA stability is the main regulatory mechanism in T cells for these cytokine and chemokine genes. Most of these genes were expressed at very low levels in uninduced T cells, except that moderate levels of CCL5/RANTES mRNA were detected in uninduced Th1 cells. CCL5/RANTES is expressed for an extended period after activation 14, so this mRNA may reflect the continued expression of CCL5/RANTES from a previous antigen stimulation.

Although individual functions vary, in aggregate, the CC chemokines expressed by Thpp cells (CCL3/MIP-1α, CCL4/MIP-1β, CCL9/MIP-1γ, CCL5/RANTES and CCL1/TCA3) induce migration of macrophages/monocytes, granulocytes, eosinophils, DC, and CD4 T, CD8 T, and NK cells 15, 16. Migration can be induced directly or, in the case of CCL1/TCA3, indirectly by up-regulating CCL3/MIP-1α production 17. The C chemokine XCL1/lymphotactin is also chemotactic for T and B lymphocytes, NK cells and neutrophils, but not monocytes or DC 18, 19. Thus, the expression of these CC and C chemokines predicts that activation of Thpp (and Th1 or Th2) cells will attract lymphocytes and other immune cells into the site of antigen activation.

Consistent with the mRNA and protein data showing chemokine expression, supernatants from activated Thpp, Th1 and Th2 cells, but not naive cells, were chemoattractive for lymphocytes. Some cytokines, e.g. IFN-γ or TNF-α, also facilitate T cell migration. However, a substantial part of the chemoattractant activity in supernatants of activated Thpp cells was lost when target cells were pretreated with PT, strongly implicating chemokine receptors signaling through G proteins.

All the CC chemokine genes expressed by Thpp, Th1 and Th2 cells are clustered tightly on mouse chromosome 11, between map locations 47.32 and 47.6. The CC chemokine CCL6/C10, although located in the middle of the expressed gene cluster at 47.51, is not expressed in any of the three differentiated T cell types. This separate regulation of CCL6/C10 is consistent with other studies showing that CCL6/C10 is not expressed coordinately with other CC chemokines such as CCL3/MIP-1α 20. Another five adjacent CC chemokine genes on chromosome 11 (between 46.5 and 47.0) are not detectably expressed by these T cells, nor are the CC chemokine genes on other chromosomes.

LT (TNF-β) was expressed by both induced and noninduced Thpp, naive and Th1 cells. As both Thpp and naive CD4 T cells migrate preferentially to lymph nodes 5, consttutive production of LT by these two cell types may assist lymphoid organogenesis, especially the development of follicular dendritic cell (FDC) clusters and germinal centers 21.

Pre-proenkephalin, highly expressed by Thpp, Th1 and Th2 cells, but not naive T cells, is a precursor protein that is cleaved to release endogenous opioid peptide enkephalins. Enkephalins modulate responses to painful stimuli 22 and enhance natural killer cell cytotoxicity in vivo23. It is not yet clear whether the pre-proenkephalins produced by activated T cells provide regulatory signals within the immune system or whether they locally reduce pain in the region of an immune reaction.

Human Thpp-like `central memory' cells, producing IL-2 but not IL-4 or IFN-γ, expressed high levels of CCR7 8. However, effector/memory T cells can also transiently express CCR7 following TCR stimulation 24, and several groups 911 suggest that CCR7 expression may not tightly associate with non-IFN-γ-producing, non-polarized CD4 and CD8 T cells. Nevertheless, naive mouse CD4 T cells and our in vitro-defined mouse Thpp cells both expressed CCR7 mRNA (unpublished data), whereas neither Th1 nor Th2 cells expressed significant levels unless activated. Thus, even though increased expression of CCR7 in human central memory cells may be partially obscured by additional factors, our results support the existence of a primed, IL-2-producing CCR7+ CD4+ T cell population.

We have previously demonstrated the existence of primed, IL-2-secreting, uncommitted CD4 T cells in vivo7. Our recent data (manuscript in preparation) shows that asubpopulation of memory CD4 T cells from spleen and especially lymph nodes of C57BL/6 mice can secrete both IL-2 and CCL3/MIP-1α, but not IL-4 or IFN-γ. These cells may be the in vivo counterparts of the Thpp cells we have defined in vitro.

Taken together with previous information, these results suggest that Thpp are primed but uncommitted CD4 T cells that, when stimulated, can express IL-2 and some chemokines and can subsequently differentiate into Th1 or Th2 cells. As these cells migrate preferentially to lymph nodes, the activation of Thpp cells may attract other lymphocytes, which may include committed T cells that willpolarize the local environment. Thus, the Thpp cells are uncommitted with respect to their eventual effector phenotype and their effector function, as the mediators secreted by Thpp may enhance butnot polarize the immune response.

4 Materials and methods

4.1 Mice

Female 6–8-week-old C57BL/6 mice were obtained from The Jackson Laboratory (Bar Harbor, ME).

4.2 Antibodies and cell lines

Purified anti-mouse CD3ϵ (145–2c11), CD28 (37.51), IL-2 (JES6–1A12), IFN-γ (R4–6A2), CCL4/MIP-1β (A65–2), biotinylated rat anti-mouse IL-2 (JES6–5H4), IL-4 (BVD6–24G2), IFN-γ (XMG1.2) and CCL4/MIP-1β (polyclonal) Ab were obtained from BD PharMingen (San Diego, CA). Purified anti-mouse CCL3/MIP-1α (39624.11), CCL9/MIP-1γ (62105), CCL1/TCA3 (148113), CCL5/RANTES(53405.111), biotinylated anti-mouse CCL3/MIP-1α, CCL9/MIP-1γ, CCL1/TCA3 and CCL5/RANTES Ab were from R&D systems (McKinley Place, NE). Additional rat anti-mouse cytokine Ab were purified on a protein G column (Pharmacia, Bridgewater, NJ) from the supernatants of 11B11 (IL-4 ), XMG1.2 (IFN-γ ), and AN18 (IFN-γ) hybridoma cell lines. The following R-PE-, Cychrome-, FITC-and biotin-conjugated Ab were also purchased from BD PharMingen: anti-CD4 (RM4–5), anti-CD44 (IM7), anti-CD45R/B220 (RA3–6B2), anti-CD8a (53–6.7), anti-CD3ϵ (1215–2C11), anti-CD28 (37.51), anti-CD62L (Mel-14). The cell line TA3, which expresses H-2Id/k25 on the surface, is a hybrid between a B cell lymphoma and a normal B cell.

4.3 T cell purification and differentiation

C57BL/6 CD4 T cells were enriched (to 85–95%) by the removal of B cells and CD8+ T cells from splenocytes on MACS columns (Miltenyi Biotec, Bergisch Gladbach, Germany). Naive CD4 T cells (CD44low, CD4high, CD62Lhigh) were purified from CD4-enriched cells on a FACS Vantage cell sorter (Becton Dickinson, Franklin Lakes, NJ), using subsaturating amounts of anti-CD4 Ab. Naive CD4 T cells (5×104 cells/ml) were stimulated with irradiated (10,000 rad) alloantigenic TA3 cells (5×104 cells/ml) in a 96-well tissue culture plate in 200 μl/well standard medium (RPMI + 8% FBS + 50 μM 2-ME [Sigma, St Louis, MO]) with the following supplements: Th1, IL-2 (1 ng/ml) + IL-12 (1 ng/ml); Th2, IL-2 (1 ng/ml) + IL-4 (2 ng/ml);Thpp, IL-2 (1 ng/ml) + TGF-ß (2 ng/ml) + anti-IFN-γ Ab (XMG1.2, 50 μg/ml). Medium (100 μl/well) was removed 2 days later, and 100 μl fresh medium containing IL-2 (Figs. 13) or all supplements (Figs. 4, 5) was added. For some experiments, the CD4 T cells were restimulated for 6more days (secondary culture) with irradiated TA3 cells (5×104 cells/ml) in the presence of the above supplements.

4.4 Cytokine assays

T cells were stimulated (1×106 cells/ml, 100 μl/well) with plate-bound anti-CD3 (2 μg/ml) + anti-CD28 (1 μg/ml) Ab for 24 h. Cytokines and chemokines produced in the supernatant were assayed by sandwich ELISA. Capture Ab (1 μg/ml in PBS) were coated onto 96-well microtrays (Corning Glass Incorporated, Corning, NY) overnight at 4°C. After washing with PBS + 0.1% Tween-20 (PBST), diluted samples and standards (in 30 μl/well standard medium) were added for 1 h at room temperature (RT). After washing, 0.2 μg/ml of biotinylated Ab (see above) were applied for 45 min at RT, followed by incubation with a 1:5,000 dilution of horseradish peroxidase-conjugated streptavidin (Jackson ImmunoResearch Laboratories, West Grove, PA) for 40 min at RT. Plates were developed with tetramethyl benzidine substrate (Sigma). H2SO4 (0.18 M) was used to stop the enzymatic reaction after 10–40 min, and plates were read at 450 nm with a reference wavelength of 570nm.

For Elispot assays 7, filter plates (Millititer MAIPN4550, Millipore Corp., Bedford, MA) were pre-coated with anti-IL-2, anti-IL-4, anti-IFN-γ, anti-CCL3/MIP-1α, anti-CCL4/MIP-1β or anti-CCL5/RANTES Ab at 2 μg/ml, 50 μl/well at RT for 2 h. Then the plates were washed three times with standard medium. Serial dilutions of CD4 T cells and constant numbers of irradiated TA3 (allogeneic target) cells were added to the wells and incubated (37°C, 8% CO2) for 20 h. The plates were washed with PBST, then incubated with biotinylated secondary Ab (2 μg/ml) in PBSTB (PBST + 2% BSA), followed by streptavidin-conjugated alkaline phosphatase (Jackson) diluted 1:1,000 in PBSTB, and finally the AEC substrate kit (Vector Laboratories, Burlingame, CA). The plajtes were then dried, and the spots were enumerated using a dissecting microscope.

4.5 Total RNA extraction from resting and stimulated T cells

Half of each cell preparation (resting cells) was pelleted by centrifugation, and RNA was directly extracted using the Qiagen mini kit (Qiagen, Valencia, CA). The remaining cells were stimulated with plate-bound anti-CD3 + anti-CD28 Ab for 9 h (when cytokines are accumulating rapidly in the supernatant) before RNA was extracted. RNA concentration and purity were determined from 260/280 nm absorbance readings. The integrity of total RNA was checked by gel electrophoresis and ethidium bromide staining.

4.6 Chip hybridization and data analysis

Samples were analyzed by the University of Rochester Medical Center Microarray Core Lab on the mouse U74A microarray (Affymetrix, Santa Clara, CA) containing 12,654 probe sets representing more than 10,000 genes. Each gene was represented by 16 to 20 pairs of 25mer oligonucleotides. Probe sequences and the chip hybridization protocol were as described at The biotinylated cRNA samples were hybridized to Affymetrix U74A mouse gene chips. The chips were washed, stained with streptavidin-PE (SAPE) and scanned for fluorescence (Scan1, S1). Subsequently, the chips were washed and stained with biotinylated anti-PE, followed by restaining with SAPE and measurement of fluorescence (Scan2, S2). Data was processed using Affymetrix Microarray Suite to obtain average fluorescence values for each probe set and the probability of expression of each gene.

Preliminary analysis showed that the S2 values for several highly expressed genes (including ribosomal proteins and cytokines) were underestimated due to saturation, whereas the S1 data was less sensitive for detecting low-level gene expression. Therefore, for all samples, we used S2 values for signals below 10,000 fluorescence units and S1 values for more highly expressed genes (above 10,000). Using only S1 or only S2 values altered some of the values, but did not change any of the conclusions described below. Chip analysis was performed on one RNA sample for each condition, whereas all subsequent analyses of RNA (RNase Protection assays) and protein (ELISA, Elispots and bioassays) were performed at least three times.

4.7 Transwell assay

Chemotaxis assays were performed in 24-well transwell plates (Corning Life Sciences) essentially as described 26. Total C57BL/6 mouse spleen cells were incubated overnight in migration medium (MM = RPMI 1640 + 0.5% BSA [fraction V; Sigma]) in 8% CO2 at 37°C, washed and resuspended at 107 cells/ml in MM. An aliquot of these spleen cells was washed and treated with 100 ng/ml PT (List Laboratories, Campbell, CA) at 3×107 cells/ml for 2 h at 37°C in 5% CO227. Dilutions of activated CD4 T cell supernatants or controls in MM (0.6 ml) were added to each outer well, and 0.1 ml of MM containing 106 PT-treated or untreated spleen cells was added to each insert. The cells were allowed to migrate through the 5-μm-diameter pores in 8% CO2 at 37°C for 4–5 h, then cells in the upper and lower chambers were collected separately and stained for CD4, CD8 and B220. A predetermined number of polystyrene beads (15 μm; Polysciences, Warrington, PA) was added to the cells, and the mixture was analyzed by FACScan. The ratios of beads to cells were used to calculate the numbers and typesof cells in each compartment.


This work was supported by NIH grant AI48604. We thank Dr. Andrea Bottaro, Steven Dewhurst, Deborah Fowell, Alexandra Livingstone and Jim Miller for comments on the manuscript.


  1. 1


  2. 2


  3. 3


  4. 4


  5. 5