T cells stimulated with an analog peptide of type II collagen require the Fc receptor γ-chain to secrete interleukin-4 and suppress autoimmune arthritis in mice




To explore the characteristics of the T cell population that responds to an analog peptide (A9) of type II collagen and regulates autoimmunity, using the collagen-induced arthritis (CIA) model.


Analog peptide A9 is a 26–amino acid peptide analogous to the sequence of a segment of type II collagen (CII245–270) but with substitutions at amino acid positions 260 (alanine for isoleucine), 261 (hydroxyproline for alanine), and 263 (asparagine for phenylalanine). We previously showed that A9 profoundly suppressed CIA and immune responses to type II collagen. In order to determine the mechanism of suppression, we used transgenic mice whose T cells express a type II collagen–specific receptor (T cell receptor) and performed passive cell transfer experiments.


The results demonstrated that suppression of CIA by A9 is dependent on T cells. Using multiparameter flow cytometry, we determined that the cells responsible for suppression were CD4+ and expressed high levels of Fcε receptor Iγ chain (FcRγ). To establish the significance of this finding, we obtained mice genetically deficient in FcRγ in order to perform passive transfer experiments. The resulting FcRγ−/− CD4+ T cells, when primed by culture with A9, could not transfer the suppression of arthritis or secrete cytokines in response to A9.


Taken together, the results of this study suggest that the suppression of arthritis and the Th2 cytokine profile elicited by A9 is dependent on the presence of FcRγ in T cells. These findings are novel and may have therapeutic potential for patients with autoimmune arthritis.

The collagen-induced arthritis (CIA) model of inflammatory arthritis is generated by immunizing susceptible animals with type II collagen (CII), the major structural component of cartilage (1). We used this model to develop a highly specific immunotherapy capable of down-regulating autoimmune arthritis in CIA and the response to CII (2). The immunotherapy was based on devising an analog peptide representing the immunodominant epitope of CII but with several critical modifications. This peptide (A9) is analogous to CII245–270 but with substitutions made for the amino acids at positions 260 (alanine for isoleucine), 261 (hydroxyproline for alanine), and 263 (asparagine for phenylalanine). When used to treat CIA, A9 profoundly suppressed arthritis and the immune response to CII. Other analog peptides were either less effective or completely ineffective (2).

In order to have sufficient numbers of CII-specific T cells with which to study the mechanism of suppression, we used CII-specific T cell receptor (TCR)–transgenic (qCII24) mice. These mice are transgenic for a TCR that recognizes the immunodominant CII epitope contained within CII245–270. When immunized with intact CII, they develop a severe arthritis beginning 18 days after immunization (3). Arthritis in the transgenic mice is efficiently suppressed by A9. Here, we demonstrate that T cells activated by the A9 peptide can passively transfer suppression of arthritis.

Functionally distinct subsets of CD4+ T cells are essential to orchestrate efficient immune responses and regulate immune-mediated inflammatory diseases. Although these subsets were initially defined on the basis of the secretion of specific cytokines (i.e., Th1, Th2, Th17), recent experiments have identified nuclear regulators of T cell differentiation and an array of molecular markers that allow more precise characterization of T cells that perform regulatory functions in autoimmune diseases (4).

Using flow cytometry and specific antibodies, we identified CII-specific CD4+ cells that were capable of suppressing arthritis in transgenic mice and established that these cells had up-regulated expression of Fcε receptor Iγ chain (FcRγ), a molecule known to be associated with the TCR complex, but did not express FoxP3 that is characteristic of Treg cells. Using mice genetically deficient in FcRγ, we demonstrated that FcRγ is required both for A9-induced cytokine secretion and for transferring the suppression of arthritis. We believe that the A9 analog peptide functions by stimulating CD4+ T cells to increase both the expression of FcRγ and the secretion of Th2-type cytokines.


Preparation of tissue-derived CII.

Native CII was solubilized from fetal calf articular cartilage by limited pepsin digestion and purified as described previously (5). The purified collagen was dissolved in cold 0.01M acetic acid at 4 mg/ml and stored frozen at −70°C until used.


DBA/1 mice were obtained from The Jackson Laboratory and raised in our animal facility. A transgenic mouse that expresses a CII-reactive TCR specific for the immunodominant determinant on CII has been developed and bred at our facility (3). Briefly, the Va11.1–Ja17 and Vb8.3–Db1–Jb1.4 gene segments derived from an I-Aq–restricted, CII-specific T cell hybridoma were cloned into T cell expression vectors, coinjected into fertilized C57BL/6 mouse eggs, and the transgenic mice backcrossed with DBA/1 for 12 generations to establish the recombinant TCR gene on the DBA/1 background. These mice were then interbred to establish a strain, designated qCII24, which is indistinguishable from wild- type DBA/1 mice and reproduces without difficulty. Another strain of mice that is genetically deficient in the FcR γ-chain has been bred onto the DBA/1 background for 12 generations (6–8). In some experiments, these mice were intercrossed with qCII24 mice to produce a new strain expressing the TCR transgene but deficient in production of the FcR γ-chain.

All mice were fed standard rodent chow (Ralston Purina) and water ad libitum. The environment was specific pathogen–free, and sentinel mice were tested routinely for mouse hepatitis and Sendai viruses. All mice were kept until the age of 7–10 weeks before being used for experiments, which were conducted in accordance with approved Institutional Animal Care and Use Committee protocols.


Type II collagen was dissolved in 0.01M acetic acid at a concentration of 4 mg/ml and emulsified with an equal volume of Freund's complete adjuvant (CFA) containing 4 mg/ml of Mycobacterium tuberculosis H37Ra (Becton Dickinson) (9). Each mouse received 100 μg of CII emulsified in CFA, injected subcutaneously at the base of the tail, in order to induce arthritis.

Measurement of the incidence and severity of arthritis.

The incidence and severity of arthritis were determined by visually examining each fore paw and hind paw and scoring them on a scale of 0–4, as described previously (5). Scoring was conducted by 2 examiners, one of whom was blinded to the identity of the treatment groups. Each mouse was scored 3 times weekly, beginning 3 weeks postimmunization and continuing for 8 weeks. The incidence of arthritis (number of mice with 1 or more arthritic limbs), the percentage of arthritic limbs (total number of arthritic limbs per group, including both fore limbs and hind limbs, expressed as a percent of the total number of limbs), and the mean severity score (sum of the severity scores for the group on each day/total number of mice in the group) were recorded at each time point.

Passive transfer experiments.

In transfer experiments, splenocytes or inguinal lymph node cells from qCII24-transgenic mice previously immunized with the A9 analog peptide in CFA were collected 8 days after immunization, and various cell subsets were fractionated (I-Aq–positive cells, α/β TCR–positive T cells, γ/δ TCR–positive T cells, or CD8+ T cells) using ferromagnetic beads (Miltenyi Biotech) and subset-specific antibodies for positive selection, according to the manufacturer's protocol. CD4+ cells were collected by negative selection, according to the manufacturer's protocol (Miltenyi Biotech). (In preliminary experiments CD4+ T cells were selected by either positive or negative selection, and the experimental results were similar.) The purity of each cell population was confirmed by flow cytometry to have >95% purity.

In each experiment, 10 or 20 mice were used for collection of cells. An equal number of recipient mice were given cells intravenously using 2 protocols, as follows: 1) 2.5 × 107 I-Aq–positive cells, α/β TCR–positive T cells, or γ/δ TCR–positive T cells, or 2) 5 × 105 CD4+ or CD8+ T cells. The cell number used in the first set of experiments (2.5 × 107) was chosen to approximate the number of cells readily obtained from a single mouse spleen. The cell numbers were reduced in the second set of experiments when dose-response experiments revealed that 5 × 105 purified CD4+ cells were sufficient to suppress arthritis. Controls were given equal numbers of nondepleted spleen cells from CFA-immunized control mice. All recipient mice were immunized with CII either on the day of the cell transfer (prevention protocol) or on day 23 after immunization (therapy protocol) and observed for the development of arthritis.

Measurement of serum antibody titers.

Mice were bled at 4 or 6 weeks after immunization, and sera were analyzed for antibodies reactive with native CII, using a modification of a previously described enzyme-linked immunoassay (5). The results are reported as units of activity, derived by comparison of test sera with the curve derived from the standard serum, which was arbitrarily defined as having 50 units of activity. Reactivity to CII was not detected in sera obtained from the control mice.

Characterization of the cells.

Splenocytes from qCII24-transgenic mice were isolated and cultured with various peptides for 24 hours. The phenotype induced by A9 was determined by multiparameter flow cytometry using an LSR II flow cytometer (BD Biosciences) and specific gating on the CD4+ population. Cells were cultured with fluorochrome-labeled antibodies specific for CD4, TCR Vβ8, CD25, CD71, CD44, and CD62L (BD Biosciences). In some experiments, intracellular labeling was performed using antibodies specific for FcRγ (catalog no. 06-727; Upstate Biotechnology) and FoxP3 (clone FJK-16s; eBioscience).

Measurement of cytokines.

To measure cytokines, several cell populations were used, as follows: 1) CD4+ T cells were isolated from the inguinal nodes of DBA/1 FcRγ−/− mice immunized 14 days previously with CII/CFA; 2) CD4+ T cells were isolated from the spleens of either qCII24 mice or qCII24 FcRγ−/− mice. The CD4+ T cells were cultured (5 × 105 CD4+ T cells/ml) with wild-type antigen-presenting cells (APCs; I-Aq–positive splenocytes) (1:2 ratio) that had been prepulsed with 100 μg/ml of the peptides (A2 or A9). The purity of each cell population was confirmed by flow cytometry to have >95% purity. In some experiments, 100 μg of A2 was added to APCs together with increasing molar amounts of A9 in order to establish a dose response. Supernatants were collected 72 hours later and analyzed for the presence of multiple cytokines (interleukin-4 [IL-4], IL-10, IL-2, interferon-γ, and IL-17), using a Bio-Plex mouse cytokine assay (Bio-Rad) according to the manufacturer's protocol. Values are expressed as picograms per milliliter and represent the mean values for each group.

Statistical analysis.

The incidence of arthritis in various groups of mice was compared using Fisher's exact test. Mean severity scores were compared using the Mann-Whitney test, and antibody and cytokine levels were compared using Student's t-test.


Identification of the cells responsible for the A9 effect in vivo.

We previously reported that A9-stimulated splenocytes are capable of transferring suppression of arthritis in the CIA model (5). In order to identify the cell subset responsible for this effect, we used qCII24 mice, which express a transgenic TCR that is specific for the immunodominant CII epitope (3), to perform passive cell transfer experiments. It was necessary to use these mice to provide adequate numbers of CII-reactive T cells for the studies.

The mice were immunized with A9/CFA, and splenocytes were collected for the isolation of various cell populations, using specific antibodies and ferromagnetic beads. As shown in Figure 1 (left), the severity of arthritis in mice receiving either I-Aq–positive cells (APCs) or γ/δ TCR–expressing cells from the spleens of A9-immunized donors was not different from that in controls (given splenocytes from mice injected with CFA alone). In contrast, arthritis was significantly attenuated in mice receiving the α/β TCR–positive A9-immune splenocytes compared with controls (incidence 14% versus 80%; P < 0.01 by Fisher's exact test). IgG antibodies to CII were also significantly suppressed in mice receiving α/β TCR–positive A9-immune splenocytes (mean ± SD 17.1 ± 8 units, versus 69.5 ± 21 units in controls; P < 0.005). These data showed that suppression of CIA by the A9 analog peptide is dependent on a population of α/β TCR–positive cells rather than APCs or γ/δ TCR–positive T cells.

Figure 1.

Identification of suppressive T cells. Left, Splenocytes from type II collagen–specific T cell receptor (TCR)–transgenic (qCII24) mice immunized with A9/Freund's complete adjuvant (CFA) were collected and fractionated into subsets. Each mouse (n = 10 per group) was given 2.5 × 107 of either γ/δ TCR–positive T cells, α/β TCR–positive T cells, I-Aq–positive cells, or depleted splenocytes from CFA-injected controls. Following immunization with type II collagen, only α/β TCR–positive T cells suppressed collagen-induced arthritis (final severity score 1.7 ± 2.7, versus 6.8 ± 3.3 in controls [P = 0.01]). Right, Splenocytes and inguinal lymph node cells from qCII24 mice were collected, and CD4+ and CD8+ T cells were isolated. Each mouse (n = 10 per group) was given 5 × 105 of either A9-primed CD4+ T cells from spleens, A9-primed CD4+ T cells from lymph nodes, A9-primed CD8+ T cells from spleens, or nondepleted CFA-primed splenocytes. Values are the mean ± SD.

The α/β TCR–positive cells consist of 2 major subpopulations characterized by expression of either CD4 or CD8 cell surface molecules. To distinguish which subpopulation is critical for arthritis suppression, splenocytes and draining lymph node cells from qCII24-transgenic mice were collected after immunization with either A9/CFA or CFA alone, and the CD4+ and CD8+ T cells were isolated and infused intravenously into DBA/1 mice prior to immunization with CII. CD4+ A9-immune T cells from either source (splenocytes or lymph nodes) suppressed the severity of arthritis, while CD8+ T cells had no effect (Figure 1, right). Taken together, these data confirm that CD4+ α/β TCR–positive cells are responsible for the suppression of arthritis induced by the A9 peptide.

Characterization of inhibitory T cells.

Activation of naive CD4+ T cells after interactions with antigen-bearing APCs typically leads to the generation of cells with well-defined effector and memory phenotypes. Naive T cells activated by antigen–APC interaction are expected to expand and differentiate into effector T cells, forming a subpopulation of cells that have a long-lived memory phenotype.

Our approach was to use multiparameter flow cytometry to characterize the phenotype of the CD4+ T cell subset generated by interaction with either A2 or A9. The qCII24-transgenic mouse splenocytes were cultured with either A2 or A9 in the presence of APCs. These cells were then tested for the presence of the activation markers CD25, CD71, CD44, and CD62L. We observed that A2 induced the emergence of a population of large blast-like cells with an activated phenotype (CD25high,CD71high) simultaneous with rapid expansion of the memory phenotype (CD44high and CD62Llow cells). In contrast, splenocytes cultured with A9 peptide and APCs did not diverge in phenotype from unstimulated control cells (Figure 2).

Figure 2.

Characterization of inhibitory T cells. Splenocytes from type II collagen–specific T cell receptor–transgenic (qCII24) mice were cultured for 48 hours in the presence of A2 or A9 and tested for the presence of CD25, CD71, CD44, and CD62L. Peptide A2 induced the emergence of a population of cells with an activated phenotype simultaneous with rapid expansion of the memory phenotype. In splenocytes cultured with A9, no change in the expression of either activation or memory markers was observed. The flow cytometric analyses were performed 3 separate times, with specific gating on the CD4+ cell population.

Treg cells are a well-defined subpopulation of CD4+ cells characterized by expression of the transcription factor FoxP3 and are known to have immunomodulatory functions in autoimmunity (10). Therefore, quantification of the Treg cell subpopulation was performed on CD4+ qCII24-transgenic mouse splenocytes cultured with A2, A9, or medium alone. As shown in Figure 3 (top), a small number of CD4+ cells expressing FoxP3 were identified under each of the 3 culture conditions, using flow cytometry. However, there were no clear variations between the numbers of Treg cells induced by either A2, A9, or control cells cultured with medium alone. The experiments were repeated using CD4+ T cells obtained from spleens 48 hours after intravenous administration of the peptides. No differences between the treatments were observed (10.1%, 9.81%, and 9.78% FoxP3+ cells for A2, A9, and phosphate buffered saline, respectively, with gating specific for the CD4+ cells).

Figure 3.

Top, Splenocytes were isolated from type II collagen–specific T cell receptor–transgenic (qCII24) mice and cultured with A2, A9, or medium. After 48 hours, cells were collected, stained using an antibody specific for FoxP3, and analyzed by flow cytometry. Values above the bars are the percentage of FoxP3-positive cells with gating on CD4+ T cells. The data are representative of results from 3 separate experiments. Bottom, Left, Splenocytes from qCII24 mice were cultured for 48 hours in the presence of A2, A9, or medium (no antigen [Ag]). Cells cultured with A9 exhibited up-regulated expression of Fcε receptor Iγ chain (FcRγ), while expression in cells cultured with A2 was no different from that in unstimulated controls. Right, Culture with A9 caused CD4+ T cells to expand and significantly up-regulate the intracellular expression of FcRγ, creating an FcRγhigh activated phenotype. The results are representative of data obtained from 5 separate experiments. Max = maximum.

Because none of the expected markers of activation and regulation were detectable in A9-stimulated CD4+ cells, we examined the contribution of the structure of TCR itself in mediating these suppressive effects. T cells in disease states can undergo extensive down-regulation of the TCR ζ-chain, with a reciprocal up-regulation of FcRγ and association of the FcR γ-chain with TCR (11). Using an antibody specific for FcRγ, we stained the qCII24 mouse splenocytes previously cultured with A2 or A9 peptide and observed that CD4+ T cells stimulated with A9 and APCs showed a significant up-regulation in the expression of FcRγ (Figure 3, bottom) compared with stimulation with A2 or medium alone. Taken together, these data suggest that activation with A9 induces a T cell that has a phenotype characterized by increased expression of FcRγ. The FcRγ induced by A9 most likely is associated with TCR, although other surface receptors might be involved.

Requirement of FcRγ for the suppression of arthritis.

To investigate the possibility that up-regulation of FcRγ might be a secondary epiphenomenon with no pathophysiologic significance and to determine its importance, we obtained DBA/1 mice that were genetically deficient in FcRγ (7). To confirm that FcRγ/ cells could successfully engraft into DBA/1 mice, we isolated CD4+ T cells from FcRγ/ mice that had been previously immunized with CII/CFA and transferred them into DBA/1 mice that were subsequently immunized with collagen.

Sera were obtained from the recipient mice 4 weeks later to test for the presence of autoantibodies to collagen. The mice that had received CII-specific FcRγ/ cells developed significantly greater antibody titers compared with control mice that did not receive cells (mean ± SD 180 ± 35 versus 65 ± 15; P < 0.05). Similar results were obtained in 2 separate experiments, confirming that the CII-specific cells engrafted with FcRγ−/− cells were viable and capable of secreting cytokines that enhance the production of autoantibodies.

In further experiments using a similar approach, FcRγ/ mice were immunized with A9/CFA. CD4+ T cells were isolated from the draining lymph nodes, and passive transfer experiments were performed as described above. Naive wild-type DBA/1 mice were given A9-primed CD4+ T cells prior to immunization with CII (prevention protocol) and observed for the development of arthritis. As demonstrated in Figure 4 (top), cells from mice genetically deficient in FcRγ did not transfer suppression, while wild-type A9-immune CD4+ T cells significantly down-regulated the severity of arthritis. Similar results were obtained when the cells were administered to mice after arthritis was firmly established (therapeutic protocol) (Figure 4, bottom). The mean ± SD antibody titers to CII obtained 8 weeks after immunization were 17 ± 7 (P ≤ 0.05) in mice given A9-immune CD4+ wild-type T cells, 50 ± 12 in mice given A9-immune CD4+ T cells from FcRγ/ mice, and 53 ± 13 in mice given CII-immune CD4+ T cells from FcRγ/ mice. When the A9-immune CD4+ T cells were infused at an even later time point, i.e., day 28 after immunization, the wild-type A9-immune CD4+ T cells significantly suppressed arthritis (final mean ± SD severity score 3.5 ± 1.2, versus 6.2 ± 1.1 in controls), while the FcRγ/ CD4+ T cells were ineffective (final severity score 5.9 ± 2.1, versus 6.2 ± 1.1 in controls).

Figure 4.

Fcε receptor Iγ chain is required for suppression of arthritis by A9. Top, CD4+ inguinal lymph node cells were collected from wild-type (WT) and FcRγ−/− type II collagen (CII)–specific T cell receptor–transgenic (qCII24) mice immunized with either A9/Freund's complete adjuvant (CFA) or CII/CFA. DBA/1 mice (n = 20) were infused with 5 × 105 of either A9-primed CD4+ T cells from wild-type mice, A9-primed CD4+ T cells from FcRγ−/− mice, or CII-primed CD4+ T cells from FcRγ−/− mice. Recipients were immunized with CII and observed for arthritis. Bottom, CD4+ lymph node cells were collected as described above. On day 23 after immunization, either A9-primed CD4+ T cells from wild-type mice, A9-primed CD4+ T cells from FcRγ−/− mice, or CII-primed CD4+ T cells from FcRγ−/− mice (5 × 105) were infused intravenously. Values are the mean ± SD.

Mechanism of suppression.

One of the functions of T cells is to secrete cytokines. Depending on the specific cytokines secreted, different effects are noted in the host. We previously reported that the A9 analog peptide induces T cells to secrete cytokines that are characteristic of Th2-type T cells (12, 13). Therefore, it was of interest to quantify cytokine secretion by T cells genetically deficient in FcRγ, in response to A2 and A9. FcRγ/ mice were immunized with CII, and the draining lymph nodes were collected for in vitro studies to determine the response to A2 and A9 in the presence of normal APCs. As shown in Table 1, the supernatants from cultures of CD4+ cells deficient in FcRγ that were cultured with A2 contained Th1, Th17, and Th2 cytokines at levels significantly greater than background. In contrast, supernatants from cells cultured with A9 elicited only background levels of Th1, Th2, and Th17 cytokines. These data also showed that FcRγ/ T cells had higher background levels of cytokines compared with wild-type DBA/1 cells, although the reasons for this are unclear.

Table 1. Cytokine responses in FcRγ−/− mice and mice treated with A9-immune CD4+ T cells*
  • *

    DBA/1 mice (n = 3/group) were immunized with type II collagen/Freund's complete adjuvant. Draining inguinal lymph node cells were harvested 14 days after immunization, and CD4+ cells (5 × 105/ml) were isolated and cultured with 1 × 106 I-Aq–positive antigen-presenting cells (purified from the spleens of wild-type [WT] DBA/1 mice) that had been prepulsed in vitro with 100 μg/ml of either the immunodominant wild-type peptide A2, the analog A9, or medium alone. Supernatants were collected 72 hours later and analyzed for the presence of the indicated cytokines. On day 23 after immunization, when arthritis was established, DBA/1 mice were infused intravenously with 5 × 105 A9-primed CD4+ T cells from FcRγ−/− mice, or A9-primed CD4+ T cells from wild-type mice, or phosphate buffered saline (PBS) as a control. Inguinal lymph node cells were harvested 14 days after the cell infusion and cultured (5 × 106 cells/ml) with 100 μg/ml of the imunodominant wild-type peptide. Supernatants were collected 72 hours later and analyzed for the presence of the indicated cytokines. IL-2 = interleukin-2; IFNγ = interferon-γ. Values are the mean ± SD pg/ml.

  • P ≤ 0.05 versus no antigen.

  • P < 0.001 versus no antigen.

 FcRγ−/− mice     
  No antigen116 ± 15378 ± 551,422 ± 18515 ± 5443 ± 52
  A2215 ± 251,041 ± 9210,934 ± 50038 ± 121,053 ± 88
  A9117 ± 13427 ± 421,646 ± 42017 ± 10481 ± 41
 WT mice     
  No antigen66 ± 5346 ± 290269 ± 2512 ± 2108 ± 11
  A2347 ± 256,390 ± 40014,891 ± 22076 ± 83,331 ± 31
  A917 ± 2647 ± 200380 ± 2552 ± 42,640 ± 22
 FcRγ−/− CD4+ cells1,302 ± 122545 ± 226,482 ± 2526 ± 4 
 WT CD4+ cells147 ± 15132 ± 13278 ± 2347 ± 8 
 PBS1,522 ± 132655 ± 266,891 ± 2808 ± 5 

In a second set of experiments, the mice genetically deficient in FcRγ were intercrossed with the qCII24 mice to obtain T cells expressing the collagen-specific TCR together with a genetic deficiency of FcRγ. The resulting FcRγ−/− CD4+ T cells from spleens were used to further establish the T cell responses to A9. A set of competitive inhibition studies was performed in which the A9 analog peptide was cultured at various concentrations relative to A2 using T cells from qCII24 mice in which FcRγ was either deficient or sufficient. When the supernatants were collected and tested (Figure 5), A9 cultured with qCII24 mouse cells successfully competitively inhibited the Th1 and Th17 cytokine responses to A2, while the Th2 cytokine response was enhanced in a dose-dependent manner. The T cells genetically deficient in FcRγ neither competitively inhibited nor enhanced the cytokine responses induced by A2.

Figure 5.

Cytokine responses to A2 and A9. CD4+ T cells were isolated from the spleens of wild-type (WT; solid bars) and FcRγ−/− type II collagen–specific T cell receptor–transgenic (qCII24) (hatched bars) mice and cultured (5 × 105 CD4+ T cells/ml) with wild-type antigen-presenting cells (I-Aq–positive splenocytes) (1:2 ratio) prepulsed in vitro with either 100 μg/ml of A2 or 100 μg of A2 together with increasing molar ratios of A9, for 6 days. Supernatants were tested for cytokines using a multiplexed enzyme-linked immunosorbent assay. Top, When supernatants were tested for inflammatory cytokines (interferon-γ [IFNγ], interleukin-2 [IL-2], and IL-17), A9/I-Aq inhibited inflammatory T cell responses in a dose-dependent manner (solid bars). FcRγ−/− qCII24 mouse CD4+ cells responded very poorly to A9/I-Aq and did not decrease the Th1 and Th17 cell responses (hatched bars). Bottom, When supernatants were tested for Th2 cytokines, A9/I-Aq enhanced qCII24 mouse CD4+ T cell responses to IL-4 and IL-10 in a dose-dependent manner (solid bars). FcRγ−/− CD4+ qCII24 mouse cells responded very poorly to A9/I-Aq and did not enhance the Th2 responses (hatched bars). Values are the mean ± SD results of 3 separate experiments. Ag = antigen.


We have identified a unique mechanism by which an analog peptide prevents the development of inflammatory arthritis and inhibits the immune response to CII. This mechanism differs significantly from previously described immunomodulatory mechanisms in terms of its dependence on the presence of FcRγ and the propensity to stimulate the secretion of Th2 cytokines relative to Th1 or Th17 cytokines. Further experiments will be necessary to determine whether the T cells responsible for this effect express specific markers other than enhanced expression of FcRγ.

The concept of “suppressor” T cells, which are able to suppress antigen-specific responses and transfer tolerance in animal models, was first proposed 40 years ago (14). Since that report, an increasing number of immune cell populations that play a role in the regulation of autoimmunity have been described (15), each of which is characterized by specific markers, different mechanisms of action, and interference with different stages of the immune response. By using cell-specific antibodies and flow cytometry, we have excluded from consideration immune cells previously reported to have regulatory functions, including monocytes, CD8+ T cells, γ/δ T cells, natural killer (NK) cells, double-negative CD3+CD4−CD8− T cells, NK T cells, or Treg cells expressing the transcription factor FoxP3. We report that the A9-induced inhibitory cells have an α/β TCR and express CD4.

There are 3 known functionally distinct subsets of CD4+ T cells, defined on the basis of the secretion of specific sets of cytokines (i.e., Th1, Th2, Th17), and a fourth subset (follicular helper T cells) present in germinal centers, which are essential to orchestrate efficient immune responses and regulate immune-mediated inflammatory diseases (16). The A9-induced cytokine profile of predominantly IL-4 and IL-10 is similar to the well-described Th2 CD4+ subset; however, it differs from that subset in that it does not develop an activated or memory phenotype following exposure to A9 (17). Moreover, the proposed A9-induced inhibitory T cell does not express FoxP3, the characteristic phenotype of Treg cells, nor is it an NK T cell, which is an innate-type T cell that responds prior to antigen challenge (18). Recently, Thomson and coworkers described a population of double-negative murine T cells characterized by intracellular phosphorylation of both FcRγ and Syk, which suppressed immune responses in vitro (19). The A9-induced cell bears functional similarities to these cells, although the phenotypes are not identical.

We are cognizant of the possibility that the TCR expressed by qCII24 mice may not be representative of the T cell repertoire found in wild-type mice. In these studies, we were careful to confirm that T cells from TCR-transgenic mice react like cells from nontransgenic mice in their ability to transfer suppression of arthritis and to secrete predominantly Th2 cytokines (especially IL-4) after stimulation with A9. An animal model becomes most useful when it allows the study of a mechanism that would take years to delineate in humans. In that regard, the T cell transfer model is unique in its ability to allow precise definition of the cell subpopulation with suppressive ability. The greater significance of this work, however, lies in the similarity of the cells studied/T cells reported to play a critical role in human autoimmune disease.

Some patients with lupus have effector CD4+ T cells with reduced expression of CD3ζ so that they phosphorylate Syk rather than ZAP-70 (11, 20, 21). Syk is also expressed at high levels in some human CD4+ effector T cells induced by antibodies directed against CD4 (22). Recently, Chauhan et al (23) reported that human T cells that express FcγRIII (i.e., γ/δ, NKT, and CD4+ T cells) can be induced by circulating immune complexes to become activated, differentiate, and proliferate via a Syk-mediated T cell pathway, in which FcRγ binds to FcγRIII. These data bear striking similarities to our data, although FcRγ interacted with the TCR in our A9-induced signaling pathway (Park JE, et al: unpublished observations). More work must be done to clarify which surface receptors and antigenic triggers are necessary to induce this alternate pathway (23).

In this study and in our previous work, we demonstrated that A9-immune CD4+ T cells secrete both IL-4 and IL-10 (Table 1). These data suggest that A9 induces and expands a population of inhibitory T cells that secrete predominantly Th2 cytokines. We believe the induction and expansion of these inhibitory cells explain the profound suppression of arthritis observed following treatment with analog A9. The shift to a Th2 cytokine profile is quite significant, because Th2 cytokines are known to have inhibitory effects on autoimmune arthritis (24–26). Both IL-10 and IL-4 suppress CIA when they are administered to mice (27). Findings of our previous study (28) are consistent with the concept that IL-4 has a unique role in the suppression of arthritis that is only partially duplicated by other Th2-type cytokines in the absence of IL-4.

Despite this observation, the molecular mechanism by which FcRγ mediates the suppressor capacity is unknown. We hypothesize that the structure of TCR is altered in response to A9 such that the TCR ζ-chain is replaced by FcRγ. The fact that FcRγ is known to bind to Syk and not ZAP-70 implies that binding between Syk and FcRγ rewires the cells to produce a different outcome. We previously provided evidence supporting this hypothesis, as follows: 1) expression of Syk but not ZAP-70 is enhanced in response to A9, and 2) inhibition of Syk suppresses downstream events such as TCR-induced cytokine secretion and induction of the Th2-associated nuclear factor GATA-3 (29). The differential expression of TCR ζ-chain and FcRγ in T cell subsets might explain differential functional outcomes of TCR signaling. Therefore, we propose that FcRγ can be up-regulated in T cells following exposure to certain carefully designed peptides and could serve as an important molecular target for controlling autoimmune arthritis.

In summary, we have described an inhibitory CD4+ cell that is induced by exposure to analog peptide/major histocompatibility complex. Following activation, the T cells up-regulate FcRγ, and intracellular signaling leads to the secretion of predominately Th2 cytokines rather than Th1 or Th17 cytokines. This inhibitory T cell develops a phenotype characterized by the intracellular up-regulation of FcRγ without an increase in typical activation or memory markers or FoxP3. Our data suggest that the induction and expansion of these inhibitory cells lead to a profound suppression of arthritis, dependent on the intracellular presence of FcRγ to suppress arthritis and secrete cytokines.

One of the fundamental—and most challenging —goals of immunologic research is to devise a treatment that suppresses immunity to a particular antigen, thus halting an injurious autoimmune process without disrupting the beneficial functions of the immune system such as surveillance for opportunistic infections and tumors. To this end, analog (or altered) peptide ligands are a particularly desirable type of antigen-specific immunotherapy and are well suited for treating autoimmune diseases (30, 31). Although peptide therapy for human disease is still in its infancy (30, 32, 33), recent reports together with our new data suggest that peptide ligands can modulate autoimmune arthritis by inducing regulatory or inhibitory T cells (34–36).


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Myers had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Myers, Brand, Stuart, Kang.

Acquisition of data. Cullins, Brand, Kleinau.

Analysis and interpretation of data. Myers, Cullins, Brand.