Tissue trafficking patterns of effector memory CD4+ T cells in rheumatoid arthritis




Clonal populations of CD4+,CD28− T cells accumulating in rheumatoid arthritis functionally resemble end-differentiated, nondividing, short-lived effector memory cells that reside in peripheral tissues. We undertook this study to examine the tissue niche for CD4+,CD28− T cells and the signals regulating their survival and tissue homing patterns.


Chemokine receptor expression on CD4+,CD28− T cell clones and peripheral blood lymphocytes was assessed by multicolor cytometry. In vitro chemotaxis and transendothelial migration were examined in a Transwell system. In vivo tissue-homing patterns were established by adoptively transferring fluorescence-labeled T cell clones into SCID mice engrafted either with rheumatoid synovium or with human lymph nodes.


CD4+,CD28− T cell clones adoptively transferred into human tissue–SCID mouse chimeras infiltrated rheumatoid synovium but preferentially homed to lymph nodes. Such T cells coexpressed the chemokine receptors CCR7, CCR5, and CXCR4 and migrated in response to both inflammatory chemokines (CCL5) and homing chemokines (CXCL12). T cell receptor crosslinking abrogated chemotactic responsiveness. In contrast, interleukin-12 stimulation induced the up-regulation of CCR5 and a shift in the in vivo homing pattern away from the lymph nodes toward the inflamed synovium.


CD4+,CD28− T cells resemble both short-lived effector memory cells and long-lived central memory cells, and they find a niche both in inflamed synovium and in lymph nodes. Nonspecific cytokine stimulation, not antigen recognition, triggers the transition from the lymph node to the synovium. By maintaining CCR7 expression, these end-differentiated T cells can home to lymphoid organs, enhance their survival, support clonal expansion, and perpetuate autoreactivity.

The current disease model for rheumatoid arthritis (RA) proposes that synovial inflammation results from sustained memory T cell responses. This model is supported by the findings that inflammatory lesions form ectopic lymphoid microstructures such as germinal centers (1–3), and that almost all tissue-infiltrating CD4+ T cells express a memory phenotype (4, 5).

Considerable progress has been made in understanding how memory T cell responses are induced and maintained. Memory T cells are a heterogeneous population consisting of phenotypically distinct subsets. Lanzavecchia and Sallusto have identified two subsets of memory T cells called central memory cells and effector memory cells, respectively (6, 7). Central memory cells reside in secondary lymphoid tissue and have low levels of effector function but are exquisitely capable of proliferating and expanding after antigenic restimulation. Effector memory cells have fully developed effector functions and live and operate in peripheral inflamed tissues (8–11).

The molecular underpinning of the differential tissue specificity lies in the absence or presence of homing receptors. Similar to naive T cells, central memory T cells possess L-selectin; CCR7, the receptor for CCL19 and CCL21; and CXCR4, the receptor for CXCL12. These ligands are present in the T cell zones of lymph nodes and on high endothelial venules, thus pulling CCR7+ T cells to secondary lymphoid tissues. Conversely, effector memory T cells lose CCR7 and respond to the inflammatory chemokines CCL3, CCL4, and CCL5. Upon antigenic restimulation, central memory T cells expand, and their progeny acquire effector function, lose CCR7, up-regulate CCR5, and become tissue-invasive. Human central memory cells proliferate and differentiate into tissue-homing effector cells in response to interleukin-7 (IL-7) and IL-15 (12).

In RA, tissue-infiltrating CD4+ T cells include a population of cells that have lost CD28 expression and display characteristics of senescent lymphocytes (13, 14). Functionally, CD4+,CD28− T cells resemble effector memory cells. They produce high levels of interferon-γ (IFNγ), have cytotoxic capability, and effectively kill endothelial cells (15–17). CD4+,CD28− T cells respond to autoantigens (18), but their responsiveness is regulated by a number of immunoreceptors including killer immunoglobulin-like receptors and NKG2D (19–22).

In contrast to classic effector memory T cells, which are terminally differentiated and exist in peripheral tissues as short-lived, nondividing cells, CD4+, CD28− T cells also display features of central memory T cells. They undergo pronounced clonal expansion and accumulate to large clonal sizes (18). T cell receptor (TCR) sequences expressed by CD4+,CD28− T cells persist over many years and are also easily detected in the peripheral blood (23). This sustained clonal dominance may be caused by apoptotic defects in these cells (24, 25). However, selective recruitment and purging from the peripheral blood would be expected if CD4+,CD28− T cells were typical effector memory cells.

We studied the homing patterns of CD4+,CD28− T cells from patients with RA to understand how they fit into the paradigm of central and effector memory T cells and how this relates to their biology in vivo. We found that the majority of CD4+,CD28− T cells coexpress CCR5, CXCR4, and CCR7 and that they home to lymph nodes and synovial lesions. The combination of these functionally distinct phenotypes may be one mechanism by which CD4+,CD28− T cells attain clonal dominance. The balance between trafficking to secondary and tertiary lymphoid structures was shifted by IL-12 and not by TCR stimulation, suggesting that innate immune activation rather than antigenic stimulation regulates tissue migration.


Patients and tissues.

Synovial tissue specimens were obtained from patients with seropositive RA and active synovitis who underwent synovectomy or total joint replacement surgery. These patients had longstanding disease. Peripheral blood mononuclear cells (PBMCs) from 20 additional patients with RA were used for flow cytometric studies and T cell cloning. All patients fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) 1987 revised criteria for the classification of RA (26). Periaortic or cervical human peripheral lymph nodes of normal size and macroscopic appearance were collected from patients requiring vascular surgery. The protocol was approved by the Mayo Clinic Institutional Review Board, and appropriate informed consent was obtained.

T cell clones.

CD4+,CD28− T cells were sorted from PBMCs of patients with RA, activated with anti-CD3 monoclonal antibodies (mAb) in the presence of irradiated PBMCs, and then cloned by limiting dilution in the presence of 50 units/ml IL-2 (Proleukin, Emeryville, CA). Established T cell clones were maintained by biweekly restimulation with anti-CD3 mAb and irradiated Epstein-Barr virus–infected cells and PBMCs in the presence of IL-2 (16, 19). All experiments were performed with T cell clones that had rested for at least 10 days after the last stimulation. All experiments were performed with T cell clones that had been stimulated with immobilized anti-CD3 mAb, 10 ng/ml IL-12, or a combination of both stimuli for 48 hours.

Flow cytometry.

The following antibodies were used for flow cytometric analysis: peridin chlorophyll protein–labeled anti-CD4 (Becton Dickinson, San Jose, CA), allophycocyanin-labeled anti-CD45RO (BD PharMingen, Carlsbad, CA), fluorescein isothiocyanate (FITC)–labeled anti-CCR5 (R&D Systems, Minneapolis, MN), FITC-labeled anti-CXCR4 (R&D Systems), phycoerythrin (PE)–labeled anti-CCR7 (R&D Systems), PE-labeled anti–IL-12 receptor β1 (anti–IL-12Rβ1; BD PharMingen), PE-labeled anti–IL-12Rβ2 (BD PharMingen), and isotype control mAb (Becton Dickinson). Data were collected on a FACScan flow cytometer (Becton Dickinson) and analyzed using WinMDI software (version 2.8, Joseph Trotter; Scripps Research Institute, La Jolla, CA).

Reverse transcriptase–polymerase chain reaction (RT-PCR).

Total RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA), and complementary DNA was amplified by RT-PCR (19). The following primers were used: β-actin, 5′-CATGGTGGTGCCGCCAGACAG-3′ (sense) and 5′-ATGGCCACGGCTGCTTCCAGC-3′ (antisense); IL-12Rβ1, 5′-TGTCACACTCTGGGTGGAATCCT-3′ (sense) and 5′- AGTATCATCATCCTGAGGTCCG-3′ (antisense); IL-12Rβ2, 5′-GAGGGACTGGTACTGCTTAATCGACTC-3′ (sense) and 5′-CCTCACACAGGTTCATTATGTTAATACGAGTG-3′ (antisense); CCR5, 5′-CGTCTCTCCCAGGAATCATCTTTAC-3′ (sense) and 5′-TTGGTCCAACCTGTTAGAGCTACTG-3′ (antisense); CXCR4, 5′-GGTAGCAAAGTGACGCCGAGGG- 3′ (sense) and 5′-GGAGTAGATGGTGGGCAGGAAG-3′ (antisense); and CCR7, 5′-GCTGTCCTGTGTGGGCATCTGG-3′ (sense) and 5′-CTGAGGCAGCCCAGGTCCTTGA-3′ (antisense).

Migration assays.

Transendothelial migration assays were performed in 12-well plates with collagen-coated PTFE Transwell inserts (diameter, 12 mm; pore size, 3 μm) (Costar, Cambridge, MA). Human umbilical vein endothelial cells (American Type Culture Collection, Manassas, VA) were plated at 5 × 105 to 1 × 106/insert and grown to confluent endothelial cell monolayers. T cells (500 μl, 1 × 106/ml) were placed in the upper chamber of the Transwell insert in RPMI 1640 medium with 1% fetal calf serum (Hyclone, Logan, UT), and CCL5 and CXCL12 (both from R&D Systems) were added to the lower chambers. The Transwell chambers were then incubated at 37°C in 5% CO2. After 4 hours, the T cells that had migrated across the endothelial cell monolayer into the lower chamber were recovered and counted. The cell migration percentage was calculated by dividing the number of migrated cells by the total number of cells placed in the upper chamber.

To measure lymphocyte chemotaxis in the absence of an endothelial cell monolayer, T cells were added to the upper chambers of Transwell plates with polycarbonate membranes (Costar), and chemokines were added to the lower chambers. After 2 hours, T cells in the lower chamber were recovered, and the percentage of migrated T cells was determined.

Generation of human synovium–SCID mouse chimeras and adoptive transfer experiments.

NOD-SCID mice (NOD.CB17-Prkdcscid/J; The Jackson Laboratory, Bar Harbor, ME) age 6–8 weeks were engrafted with human tissues as previously described (1). Pieces (20–30 mm3) of human synovial tissue and human lymph nodes were placed into subcutaneous pockets on the upper dorsal midline. T cells (2 × 107/ml) were labeled with 4 μl PKH26 red fluorescent cell dye (Sigma, St. Louis, MO) in 2 ml PKH26 diluent C (Sigma). Cell viability as determined by trypan blue exclusion was >95%. Flow cytometry confirmed PKH26 labeling efficiency before adoptive transfer. T cell clones (1 × 107 cells in 0.5 ml phosphate buffered saline [PBS]) were adoptively transferred by intraperitoneal injection. In selected experiments, the T cells were preincubated with 200 μg/ml anti-CCR5 mAb before the adoptive transfer. The mice were killed after 48 hours, and the human tissue and murine spleen were harvested and embedded in OCT compound (Tissue-Tek; Sakura Finetek, Torrance, CA) for immunohistochemical analysis.

Histologic analysis.

Serial cryosections (10 μm) were mounted on Vectabond-coated slides (Vector, Burlingame, CA) and dried overnight at room temperature. Sections assigned for analysis of PKH26-positive cells were washed for 10 minutes in PBS (pH 7.6) and mounted using aqueous mounting media (Immunofluor; ICN Pharmaceuticals, Costa Mesa, CA). For each specimen, 5 representative nonserial sections were analyzed using a fluorescence microscope (Carl Zeiss Instruments, Thornwood, NY). For each section, 25 high-power fields (hpf; 20× objective) were analyzed, and the mean cells/hpf was calculated.

Statistical analysis.

The Mann-Whitney rank sum test and Student's t-test were used as appropriate (SigmaStat; SPSS, Chicago, IL). Results are shown as box plots with medians if the Mann-Whitney rank sum test was used and as the mean ± SD if Student's t-test was used.


In vivo trafficking pattern of CD4+,CD28− T cells.

In vivo trafficking patterns of CD4+,CD28− T cells were examined in human tissue–SCID mouse chimeras. Chimeras were generated by implanting human lymph nodes or inflamed synovial tissue subcutaneously into SCID mice. CD4+,CD28− T cell clones were isolated from 3 patients with RA. Labeled T cell clones were adoptively transferred into the chimeras 8 days after tissue implantation. Forty-eight hours later, the grafts were harvested and analyzed for infiltration of labeled T cells. As shown in Figure 1, the human T cell clones displayed a clear preference for the human tissues, and very few labeled T cells were detected in the mouse spleen. CD4+,CD28− T cells migrated into lymph nodes, where they accumulated in the T cell zones and avoided the follicles. In the synovial lesion, labeled T cells were spread as a diffuse infiltrate. In 3 of 3 experiments, lymph nodes outperformed synovial tissues in attracting the adoptively transferred T cell clones (Figure 1). Lymph nodes from different donors were comparable in their abilities to recruit and retain CD4+,CD28− T cells. Similarly, synovial tissues derived from different donor patients had equal potency in recruiting adoptively transferred T cells.

Figure 1.

In vivo trafficking pattern of effector memory CD4+,CD28− T cells. SCID mice were implanted with synovial tissue harvested from human rheumatoid joints or with normal human lymph nodes. CD4+ memory T cell clones isolated from patients with rheumatoid arthritis were labeled with PKH26 and adoptively transferred into the chimeras. Tissue migration was measured after 48 hours. A, Number of labeled T cells. At least 20 high-power fields were counted in the tissue sections. Representative results from 1 of 3 experiments are shown as box plots, where the boxes represent the 25th and 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the 10th and 90th percentiles. Very few human T cells were detected in the mouse spleen. CD4+ memory T cells accumulated both in synovial lesions and in lymph nodes, with a preference for lymphoid organs. B, Fluorescence microscopy. Labeled CD4+ T cells diffusely infiltrated the synovial lesions, homed to the T cell zones of lymph nodes, and were very rare in the murine spleen. (Original magnification × 200.)

To understand why transferred CD4+ T cells homed to lymphoid nodes and inflamed synovium, we determined the chemokine receptor profile by RT-PCR and flow cytometry. Results representative of 26 CD4+,CD28− T cell clones are shown in Figure 2A. All of the CD4+ T cell clones expressed CCR5, 46% coexpressed CCR7, and 54% were positive for CXCR4. CCR7 and CXCR4 were frequently coexpressed; only 30% of the T cell clones showed the typical pattern of effector memory cells exclusively expressing CCR5. Clones that homed to both tissues in the adoptive transfer experiments coexpressed CCR5 and CCR7.

Figure 2.

Increased frequencies of CD4+,CCR5+,CCR7+ T cells in rheumatoid arthritis (RA). Twenty-six CD4+ T cell clones were established from the peripheral blood of 3 RA patients by limiting-dilution cloning. All CD4+ T cell clones had a memory phenotype and expressed CCR5. Twelve CD4+ T cell clones also expressed CCR7, and 14 T cell clones coexpressed CXCR4. A, A representative T cell clone expressing all 3 chemokine receptors. B, Peripheral blood lymphocytes were isolated from 16 RA patients and 17 age-matched controls, and CD4+ memory T cells were analyzed for the expression of the chemokine receptors CCR5 and CCR7 by 4-color flow cytometry. In RA patients, CD4+,CCR5+,CCR7+ T cells were expanded compared with those in healthy controls (P < 0.01). Conversely, CD4+,CCR5+,CCR7− T cells were diminished in RA patients compared with those in healthy controls (P = 0.02). Results are shown as box plots, where the boxes represent the 25th and 75th percentiles, the lines within the boxes represent the median, and the lines outside the boxes represent the 10th and 90th percentiles.

Tissue culture conditions influence T cell differentiation and may imprint a response pattern on T cell clones that is not representative of the cells in vivo. This is of primary importance for cloning of noncommitted naive T cells, but it may also influence already differentiated effector cells. We therefore examined whether CD4+,CCR5+,CCR7+ T cells are present in vivo and analyzed peripheral blood lymphocytes from 16 RA patients and 17 age-matched controls by 4-color flow cytometry. As previously reported, RA patients had elevated frequencies of CD4+,CD28− T cells (17, 18) (data not shown). CD4+ memory T cells expressing both CCR7 and CCR5 were detected at low levels in controls but were clearly expanded in patients (P < 0.01) (Figure 2). The expected higher frequency of CCR5+,CD4+ T cells in the CCR7− effector memory cells was found only in healthy individuals, while the opposite was true in RA patients (Figure 2B). In patients, only one-third of the CCR5+ T cells were CCR7−.

CD4+,CD28− T cells respond to inflammatory and homing chemokines.

CCL5, a CCR5 ligand, is abundantly expressed in rheumatoid synovitis and facilitates CCR5+ T cell chemotaxis (27–29). Homing to the lymph nodes is achieved through CCR7 ligands CCL19 and CCL21 and the CXCR4 ligand CXCL12 (30, 31). To confirm that the majority of terminally differentiated CD4+,CD28− effector T cells respond to both inflammatory and lymph node homing chemokines, we tested chemotaxis and transendothelial migration of these clones in a modified Boyden chamber system. For transendothelial migration experiments, the chemokine source and T cells were separated by a contiguous endothelial monolayer. As shown in Figure 3, CD4+, CD28− T cells coexpressing CCR5, CCR7, and CXCR4 were responsive to both the inflammatory chemokine CCL5 and the homing chemokine CXCL12. Likewise, transmigration across the endothelial barrier was markedly enhanced by both chemokines. These in vitro studies confirmed the coexistence of functional chemokine receptors that are usually distributed on mutually exclusive T cell subsets.

Figure 3.

Chemotactic responses and transendothelial migration of CD4+,CCR5+,CCR7+,CXCR4+ T cells. The chemotactic response of CD4+ T cell clones was determined in a modified Boyden chamber system. Increasing concentrations of the homing chemokine CXCL12 and of the inflammatory chemokine CCL5 were added to the lower chamber, and migration was measured after 2 hours (left panel). Transendothelial migration was assayed by placing an endothelial monolayer between the chambers (right panel). CD4+,CCR5+,CCR7−,CXCR4+ T cell clones did not respond to CXCL12 (data not shown). CD4+,CCR5+,CCR7+,CXCR4+ T cells were responsive to both CXCL12 and CCL5. Results are shown as the mean ± SD of triplicatecultures and are representative of 7 experiments.

IL-12 modifies the tissue migration pattern of CD4+,CD28− T cells.

CCR7+ central memory T cells are considered to be the reservoir for effector memory cells. They reencounter antigens in the secondary lymphoid tissue and then home to peripheral tissue sites as effector memory T cells. Interestingly, CCR7+ memory CD4+ T cells can be induced to proliferate and differentiate into a CCR7– population, using a cocktail of cytokines (12). One major cytokine produced by activated dendritic cells in the T cell zones of secondary lymphoid tissues is IL-12. We therefore screened 26 CD4+,CD28− T cell clones established from 3 patients for the expression of the IL-12 receptor β1 and β2 chains. Eighty percent of the derived T cell clones produced IL-12Rβ1–specific transcripts (Figure 4A), and 60% also expressed IL-12Rβ2. We screened 30 additional T cell clones generated from 5 RA patients, and only 4 CD4+,CD28− T cell clones were identified that lacked expression of IL-12Rβ1 and IL-12Rβ2 (data not shown).

Figure 4.

Responsiveness of CD4+,CCR5+,CCR7+ T cells to interleukin-12 (IL-12). Twenty-six CD4+ T cell clones isolated from patients with rheumatoid arthritis (see Figure 2) were screened for the expression of IL-12 receptor (IL-12R) β-chains by polymerase chain reaction. A representative sample of 5 clones is shown in A. Clones 1, 2, and 4 spontaneously expressed both IL-12R β-chains. Clone 3 essentially lacked both IL-12R β-chains, and clone 5 had only IL-12Rβ1–specific transcripts. B–G, IL-12R+ and IL-12R– CD4+ T cell clones were cultured in IL-12 for 48 hours and then analyzed by flow cytometry. The expression levels of CD161, CCR5, CXCR4, and CCR7 were compared before (dashed line) and after (solid line) exposure to IL-12. Shaded areas indicate control antibody. In IL-12R–expressing clones, up-regulation of the IL-12–inducible gene CD161 demonstrated responsiveness to IL-12 (B). Results of the induction of chemokine receptors are shown for one CD4+,CD28–,IL-12R+ T cell clone (CE) and for one CD4+,CD28–,IL-12R− T cell clone (F and G) and are representative of 7 experiments. IL-12 increased the surface expression of CCR5 in IL-12R+ clones (C), while expression of CXCR4 (D) and CCR7 (E) remained unchanged. IL-12 did not affect the expression of CD161 or CCR5 in IL-12R– T cell clones (F and G).

To determine whether the IL-12 receptor was functional, CD4+,CD28–,IL-12R+ T cell clones were stimulated with recombinant IL-12 for 48 hours. IL-12 responsiveness was demonstrated by the up-regulation of the IL-12–inducible genes CD161 (32) and CCR5 (Figures 4B and C). The expression of CXCR4 was essentially unaffected by IL-12 stimulation (Figure 4D); expression of CCR7 was unchanged or, in some clones, minimally enhanced (Figure 4E). Culture of IL-12Rβ2–negative T cell clones with recombinant IL-12 did not induce the up-regulation of CD161 or CCR5 (Figures 4F and G).

To examine whether the IL-12–induced expression of CCR5 was functionally significant, we tested the chemotactic response of T cell clones to the CCR5-binding chemokine CCL5 and the CXCR4-binding chemokine CXCL12. As shown in Figure 5, preconditioning of CD4+,CD28–,IL-12R+ T cells with IL-12 enhanced migration toward CCL5 but reduced chemotaxis toward CXCL12. Indeed, pretreatment with IL-12 essentially abrogated chemotactic response toward CXCL12 even though CXCR4 continued to be expressed. Treatment of CD4+,CD28–,IL-12R− control T cell clones left the chemotactic response unaffected (Figure 5A).

Figure 5.

Interleukin-12 (IL-12) and T cell receptor (TCR) triggering modulate the chemotactic response pattern and transendothelial migration of CD4+,CCR5+,CCR7+ T cells. A–F, CD4+,CCR5+,CCR7+ T cells were cultured with or without IL-12 and tested for their chemotactic response as described in Figure 3. CD4+ T cell clones lacking IL-12 receptor (IL-12R) served as controls (A and D). IL-12 enhanced migration in response to CCL5 (B) and reduced CXCL12-induced migration (C). Transendothelial migration was similarly affected by IL-12. Stimulation with IL-12 doubled the frequency of CD4+ T cell clones that crossed the endothelial cell barrier in response to CCL5 (E), and it impaired transendothelial migration toward a CXCL12 gradient (F). Results are shown as the mean ± SD of triplicate cultures and are representative of 11 experiments. G–J, We compared chemotactic responses and transendothelial migration of resting CD4+ cells (G and I) and CD4+ cells activated by anti-CD3 stimulation (H and J). Stimulation with IL-12 enhanced the chemotaxis (G) and transendothelial migration (I) of resting CD4+,CCR5+,CCR7+ T cell clones. Forty-eight hours after crosslinking the TCR with anti-CD3, chemotactic responses (H) and transendothelial migration (J) were minimal, regardless of whether the cells had been exposed to IL-12. Results are shown as the mean and SD and are representative of 5 experiments.

To determine whether IL-12 also affected transendothelial migration, CD4+ T cell clones with or without IL-12R were compared for their ability to cross an endothelial barrier. Only IL-12R+,CD4+,CD28– T cell clones changed their transmigration efficiency. IL-12 stimulation doubled the frequency of CD4+ T cells that transmigrated across the endothelial barrier in response to CCL5 (Figure 5E). Again, IL-12 reduced CXCL12-induced transendothelial migration (Figure 5F). Thus, IL-12 was able to profoundly change the response pattern of CD4+ T cells to chemotactic signals, even in the absence of TCR triggering.

TCR triggering renders CD4+ T cells resistant to IL-12–induced shifts in chemokine responsiveness.

Although CD4+,CD28– T cells obviously do not require antigenic restimulation to express functional IL-12R, they are likely to encounter IL-12 during antigenic restimulation in lymph nodes. We therefore studied the effect of TCR triggering on IL-12–induced T cell chemotaxis and transmigration. As shown in Figures 5G–J, the TCR stimulation had profound effects on T cell chemotactic behavior. Only resting CD4+,CD28–, IL-12R+ T cell clones showed enhancement of CCL5-directed migration after IL-12 stimulation.

In contrast, crosslinking of the TCR with anti-CD3 mAb resulted in a reduction of CCL5-directed chemotaxis, independent of whether the T cells had been cultured in the absence or presence of IL-12. Adhesion to and passage through an endothelial monolayer was similarly affected. Reduced CCL5 responsiveness was seen 48 hours after TCR stimulation (Figure 5), but lasted for at least another 24 hours (data not shown). T cells at these time points were viable and were able to proliferate in response to IL-2. Signaling through the TCR essentially abrogated the ability of such CD4+ T cell clones to follow a gradient of CCL5 and pass through an endothelial cell monolayer. Flow cytometric analysis demonstrated that TCR crosslinking induced rapid down-regulation of cell surface CCR5 (data not shown). After activation, the T cell clones became negative for CCR5 expression. Activated CD4+ T cell clones continued to transcribe CCR5-specific sequences (data not shown), suggesting that regulation of CCR5 expression by TCR triggering occurs posttranscriptionally.

IL-12 changes the tissue distribution of CD4+,CD28– T cells in vivo.

The in vivo effects of IL-12 stimulation were examined in adoptive transfer experiments. Labeled CD4+,CD28– T cell clones were injected into human lymph node–SCID mouse chimeras and human synovium–SCID mouse chimeras. IL-12 pretreatment of the T cells prior to transfer did not change their preferential migration to human tissue over mouse tissue. As shown in Figure 6A, only a small number of labeled human T cells were detected in the murine spleen, irrespective of prior IL-12 stimulation. Triggering with IL-12 more than doubled the number of T cells that infiltrated the synovial grafts, from a mean of 32/hpf to a mean of 70/hpf. IL-12 treatment also affected the trafficking of CD4+,CD28– cells to the lymph node but in the opposite direction. In the absence of IL-12, the lymph node was clearly more attractive to injected T cells, with 4–5-fold as many T cells accumulating in the T cell zones compared with the synovial microenvironment. Following IL-12 stimulation, equal numbers of cells were recruited to the two distinct tissues (Figure 6A). In essence, IL-12 deviated the trafficking pattern away from secondary lymphoid tissue and toward peripheral inflammatory lesions.

Figure 6.

The effect of interleukin-12 (IL-12) on the in vivo migration pattern of CD4+,CCR5+,CCR7+ T cells in rheumatoid arthritis. CD4+,CCR5+,CCR7+ T cell clones were labeled with PKH26 and adoptively transferred into human synovium– and human lymph node–SCID mouse chimeras. A, Migration into the tissue was determined by quantifying fluorescent T cells in tissue sections. IL-12 stimulation prior to the transfer increased migration into the synovial graft (P < 0.001 versus control) and reduced accumulation of T cells in the lymph nodes (P = 0.005 versus control). Human T cells were rarely detected in the murine spleen either with or without prior IL-12 stimulation. B, Incubation of IL-12–treated CD4+ T cells with 200 μg anti-CCR5 monoclonal antibodies prior to the adoptive transfer essentially prevented tissue invasion (P < 0.001 versus IL-12+ control Ig). Numbers of tissue-infiltrating cells per high-power field are shown as box plots, where the boxes represent the 25th and 75th percentiles, the line inside the boxes represents the median, and the lines outside the boxes represent the 10th and 90th percentiles. Results are representative of 3 experiments.

To test whether the bias of IL-12–treated CD4+,CD28– T cells toward the synovial tissue was related to the up-regulation of CCR5, we coadministered anti-CCR5 mAb with the labeled CD4+,CD28– T cells. Data shown in Figure 6B summarize the results from 3 adoptive transfer experiments. Pretreatment with IL-12 markedly enhanced the tissue accumulation of labeled T cells. The addition of 200 μg of anti-CCR5 mAb not only abrogated the IL-12–induced enhancement of tissue trafficking, but almost completely blocked tissue infiltration of CD4+,CD28– T cells. The inhibitory effect was specific; control Ig did not alter the density of labeled cells in the synovial grafts.


Patients with RA accumulate an unusual population of CD4+ effector T cells that have permanently lost CD28 and costimulate through newly acquired immunoreceptors such as killer immunoglobulin-like receptors and NKG2D. Such T cells have a peculiar tissue distribution pattern; they accumulate in the rheumatoid lesions but are also clonally dominant in the peripheral blood (18). Previous studies in synovial fluid and tissue have shown that such clones represent a considerable portion of the oligoclonal T cell repertoire in the synovium, and that therefore clones with identical TCR sequences can be isolated from either PBMCs or synovium (33). Such clones typically have a long survival time and can be tracked over years. Data from the current study suggest that the persistence of these T cells and their activity in chronic inflammation may be related to an unusual chemotactic behavior.

The data support 3 major conclusions. First, CD4+,CD28– T cells from patients with RA break the paradigm of distinct central and effector memory T cells and display characteristics of both. They express the lymphoid homing chemokine receptors CCR7 and CXCR4, but they also express CCR5, which makes them responsive to inflammatory chemokines. As a consequence, they home to lymph nodes as well as to inflamed synovia. Thus, T cells with effector function invade the niche usually occupied by central memory cells. Second, exposure to IL-12 in the absence of antigenic stimulation redirects the trafficking pattern of these CD4+ memory cells away from the lymph node and toward the synovial lesions. Third, triggering of the TCR paralyzes CD4+,CCR5+,CCR7+ T cells and abrogates their spontaneous and IL-12–induced chemotactic responsiveness to CCL5. In essence, innate immune activation, not antigen restimulation, determines the influx of these proinflammatory CD4+ T cells into the autoimmune lesions.

Lanzavecchia and Sallusto and their colleagues' concept of distinguishing two functional subpopulations of memory CD4+ T cells, central and effector memory T cells, has provided a framework for understanding how the immune system accomplishes rapid eradication of tissue-residing antigens without risking the loss of the memory populations after antigen rechallenge (8, 9). Central memory T cells circulate through secondary lymphoid tissue and survey for their respective antigens. They lack effector functions, are presumably long-lived, and are the reservoir for peripheral effector cells. Effector memory T cells migrate into the peripheral inflamed tissue, cannot recirculate, and are assumed to be short-lived. This division in function guarantees that the number of potentially tissue-injurious cells is low while a large reservoir of memory T cells with proliferative potential is maintained.

The expression of CCR5 on disease-relevant CD4+ effector T cells is within the existing paradigm. CCR5 is an important marker of T cells selectively recruited to the rheumatoid lesions (29, 34). CCR5, together with CCR4 and CXCR3, has been identified as a critical chemokine receptor in RA (27, 35). CCL3, CCL4, and CCL5 are produced in the synovial tissue, probably by distinct cell types, and they all can attract CCR5-expressing memory cells (28, 36). All these data support the hypothesis that CCR5-mediated recruitment is critical in establishing and maintaining rheumatoid lesions in the joints.

Campbell et al demonstrated that this compartmentalization is not as strict as originally stated (37). Phenotypic studies on circulating and tissue lymphocytes found that CCR7 is expressed not only on the vast majority of peripheral blood T cells but also on tissue T cells in the lung, liver, and skin. Campbell et al also had the opportunity to analyze tissue from 1 patient with RA and found CCR7+ T cells in the synovium. This is not unexpected, because the synovium contains ectopic lymphoid follicles (38, 39) and CCL19 (40) and CCL21 (3) chemokines that attract naive and central memory cells are locally produced.

In healthy individuals, effector T cells do not express CCR7 and do not home to lymph nodes. A breakdown in the compartmentalization of such effector T cells is predicted to have detrimental consequences for the immune system. Our data demonstrate that this is exactly the case in RA patients. A large fraction of autoreactive CD4+,CD28– T cells have sustained their ability to home to lymphoid tissue. These cells lack CD27 expression, are capable of producing large amounts of IFNγ, and have cytotoxic activity even without prior restimulation; however, they express CCR7, CCR5, and often also CXCR4. This duplicity of chemokine receptor expression is of functional importance, as demonstrated in the in vivo studies. Many of these cells migrate to the inflamed synovium; most home to the lymph node. Consequently, the majority of the expanded CD4+ clonotypes should be found in patients' secondary lymphoid organs.

One obvious consequence is that CD4+ effector T cells carry potentially tissue-injurious mechanisms into lymphoid tissues. They recognize ubiquitous autoantigens (18), and they may thereby affect the integrity of the lymph node structure. Production of IFNγ in the lymph node could amplify immune responses, and cytotoxic activity would compromise immunocompetence, possibly contributing to the accelerated immunosenescence described for RA.

Conversely, homing of CD4+,CCR5+,CCR7+ T cells to lymph nodes obviously extends the space in which these cells can live and increases their ability to compete with other T cell populations for survival. Homing to secondary lymphoid tissues is a key component in T cell homeostasis, critical for the persistence and long-term survival of central memory cells (24, 41). This mechanism may contribute to the clonal dominance and longevity of CD4+,CD28– T cells in RA. Continuous expression of CCR7 may allow tissue-infiltrating synovial CD4+ T cells to recirculate to lymphoid organs. Such a mechanism would allow for recycling of memory T cells between peripheral inflammatory infiltrates and organized lymphoid organs. The reentry of autoimmune memory CD4+ T cells into the lymph nodes would perpetuate pathogenic immune responses and may be one mechanism that maintains chronic disease in RA.

Expansion and relocalization of memory T cells from central lymphoid organs to peripheral tissues are elicited by antigenic restimulation. However, elegant work by Geginat and colleagues (12, 41) has shown that cocktails of cytokines are sufficient to drive CCR7+ central memory T cells into CCR5+ effector memory-like cells in vitro. Cytokine-dependent, instead of antigen-dependent, regulation of memory T cell function and trafficking seems to be particularly important for patients with RA. Shifting of CCR5+ memory T cells toward the peripheral lesion was accomplished only when the TCR was not triggered. IL-12 emerged as a potent factor in regulating the whereabouts of CD4+ memory T cells.

The main producers of IL-12 are phagocytes and dendritic cells, often in response to microbial stimulation (42, 43). The IL-12 receptor, composed of the IL-12Rβ1 and IL-12Rβ2 chains (44), is mainly expressed by activated T cells and natural killer cells (44). The current paradigm holds that T cell expression of IL-12Rβ2 is confined to Th1 cells (45, 46). CD4+,CD28– T cells have all the characteristics of Th1 cells; however, they do not require TCR stimulation to be IL-12–responsive. Patient-derived CD4+,CD28– T cells responded to IL-12 with a shift in their trafficking pattern. IL-12 could not abrogate the tendency of these CD4+ T cells to behave like central memory cells, but it redirected them toward CCL5-producing tissues.

An important aspect of the current study was that IL-12 stimulation was functional only in redirecting CD4+ T cells away from secondary lymphoid structures to peripheral tissue if the TCR was not stimulated. Therefore, IL-12 would have the most functional impact in the absence of antigen. Any immune response that results in IL-12 production in lymph nodes would have the potential to redirect the migration pattern of CD4+,CD28– T cells and enhance the inflammatory response in the joints. In contrast, the recognition of their respective autoantigens in the lymph node might provide a survival signal, but it would not immediately lead to increased inflammatory activity in the autoimmune lesion. This model fits the clinical observation that patients with RA can present with disease flares when fighting an unrelated infection (47). Inhibiting IL-12 may provide a new opportunity to treat RA with a mechanism of action that goes beyond the inhibition of IL-12–dependent effects in the inflammatory lesions. Anti–IL-12 therapy could be effective in preventing the influx of memory CD4+ T cells into the joint lesions, thus disrupting an antigen-nonspecific amplification mechanism of chronic inflammation.


We thank Dr. S. Pryshchep for help with preparing the figures and T. Yeargin for editorial support.