To define the role of indoleamine 2,3-dioxygenase (IDO) in driving pathogenic B cell responses that lead to arthritis and to determine if inhibitors of the IDO pathway can be used in conjunction with therapeutic B cell depletion to prevent the reemergence of autoantibodies and arthritis following reconstitution of the B cell repertoire.
Immunoglobulin-transgenic mice were treated with the IDO inhibitor 1-methyltryptophan (1-MT) and monitored for the extent of autoreactive B cell activation. Arthritic K/BxN mice were treated with B cell depletion alone or in combination with 1-MT. Mice were monitored for the presence of autoantibody-secreting cells, inflammatory cytokines, and joint inflammation.
Treatment with 1-MT did not affect the initial activation or survival of autoreactive B cells, but it did inhibit their ability to differentiate into autoantibody-secreting cells. Treatment with anti-CD20 depleted the B cell repertoire and attenuated arthritis symptoms; however, the arthritis symptoms rapidly returned as B cells repopulated the repertoire. Administration of 1-MT prior to B cell repopulation prevented the production of autoantibodies and inflammatory cytokines and flare of arthritis symptoms.
IDO activity is essential for the differentiation of autoreactive B cells into antibody-secreting cells, but it is not necessary for their initial stages of activation. Addition of 1-MT to therapeutic B cell depletion prevents the differentiation of autoantibody-secreting cells and the recurrence of autoimmune arthritis following reconstitution of the B cell repertoire. These data suggest that IDO inhibitors could be used in conjunction with B cell depletion as an effective cotherapeutic strategy in the treatment of rheumatoid arthritis.
The inflammatory autoimmune disease rheumatoid arthritis (RA) has classically been thought to be mediated by T cells, either directly, by infiltration of tissues, or indirectly, through the release of inflammatory cytokines (1, 2). It is becoming increasingly apparent that B cells also play a critical role in driving inflammatory autoimmunity in RA (3). In addition to producing pathogenic autoantibodies, B cells can trigger autoimmune responses through the presentation of self-reactive antigens to T cells and the production of inflammatory cytokines. The most convincing evidence supporting the role of B cells in RA is the recent success of B cell–mediated therapies (4). However, the factors important in initiating and maintaining autoreactive B cell responses remain unknown.
A promising strategy for treating RA relies on B cell depletion using a chimeric monoclonal antibody directed against the B cell–specific cell surface marker CD20 (rituximab) (4). The addition of rituximab to the treatment regimen was shown to reduce autoantibody levels and improve clinical signs and symptoms in the majority of RA patients, with some showing complete resolution of inflammation (5). Similarly, B cell depletion has been shown to be effective in several mouse models of arthritis (6, 7). Unfortunately, the primary limitation of B cell depletion therapy in both humans and mice is that eventually the B cells return, and the repopulation of the B cell repertoire correlates with the return of arthritis symptoms in many individuals (8, 9). A cotherapeutic strategy aimed at inhibiting the activation of autoreactive B cells upon repopulation would help to lengthen the effectiveness of the therapeutic window and could improve clinical outcomes in RA patients.
Researchers in our laboratory recently identified indoleamine-2,3-dioxygenase (IDO) as an important factor in driving the initial stages of B cell–mediated autoimmune responses (10). IDO is an interferon-γ (IFNγ)–inducible enzyme that catalyzes the initial and rate-limiting step in tryptophan degradation (11). Elevated tryptophan degradation has been shown to correlate with disease activity in RA patients (12). Likewise, we have shown that IDO activity is highest during the acute phase of disease in the K/BxN mouse model of inflammatory joint disease (10). Inhibition of IDO activity in K/BxN mice with the pharmacologic inhibitor 1-methyltryptophan (1-MT) led to reduced levels of inflammatory cytokines, diminished titers of autoantibodies, and an attenuated course of disease. This alleviation of arthritis was not due to a reduction in regulatory T cells or an altered T helper cell phenotype, but rather, it resulted from a diminished autoreactive B cell response (10). This work demonstrated a previously unappreciated role of IDO in stimulating B cell responses; however, the role of IDO in B cell activation remained unknown.
In the present study, we used Ig-transgenic mice to define the stage at which B cell activation is influenced by IDO. We demonstrated that IDO activity is involved in the differentiation of autoreactive B cells into antibody-secreting cells (ASCs) but is not required for the initial stages of B cell activation or germinal center formation. This suggests that IDO plays a role in establishing the autoreactive B cell profile at the initiation of the autoimmune response. Thus, inhibitors of IDO activity should be most useful therapeutically at the initiation of autoreactive B cell responses. We proposed that inhibition of IDO activity at this critical stage would prevent the establishment of the autoreactive B cell profile, thereby reducing subsequent joint inflammation and damage. To test this hypothesis, we combined 1-MT treatment with therapeutic B cell depletion using antibodies to CD20. We demonstrated that the addition of 1-MT inhibits the differentiation of autoantibody-secreting cells following B cell depletion therapy and prevents the recurrence of autoimmune arthritis. These data suggest that inhibition of the IDO pathway could be an effective strategy for use in conjunction with B cell depletion therapy in the treatment of RA.
MATERIALS AND METHODS
The KRN T cell receptor–transgenic and VH147 Ig–transgenic mice on a C57BL/6 background that were used in this study have been described previously (13, 14). NOD mice were purchased from The Jackson Laboratory. To obtain arthritic mice, KRN and VH147/KRN-transgenic C57BL/6 mice were crossed with NOD mice to yield KRN and VH147/KRN (C57BL/6 × NOD)F1 mice, which were designated K/BxN and VH147 K/BxN, respectively. All mice were bred and housed under specific pathogen–free conditions in the animal facility at the Lankenau Institute for Medical Research (LIMR). Studies were performed in accordance with the National Institutes of Health and the Association for Assessment and Accreditation of Laboratory Animal Care guidelines, with approval from the LIMR Institutional Animal Care and Use Committee.
Administration of 1-MT and anti-CD20 Ig.
Mice were given 400 mg/kg per dose (100 μl total volume) of D/L-1-MT (Sigma) diluted in Methocel/Tween (0.5% Tween 80, 0.5% methylcellulose [volume/volume] in water) or carrier alone (Methocel/Tween) twice daily by oral gavage, a single 200-μg (200 μl total volume) intraperitoneal injection of anti-CD20 Ig (clone 5D2; Genentech) or isotype control antibody (mouse IgG; Jackson ImmunoResearch) diluted in phosphate buffered saline, or a combination of 1-MT plus antibody.
Determination of arthritis incidence.
Swelling of the ankle joint was used as an indicator of arthritis. The rear ankles of K/BxN mice were measured axially across the ankle joint, using a Fowler Metric Pocket Thickness Gauge. Ankle thickness was rounded to the nearest 0.05 mm. At the termination of the experiment, ankles were harvested, fixed in 10% buffered formalin for 48 hours, decalcified in 14% EDTA for 2 weeks, embedded in paraffin, sectioned, and stained with hematoxylin and eosin.
Enzyme-linked immunospot (ELISpot) assay.
Cells derived from draining lymph nodes were plated at 4 × 105 cells/well and diluted serially 1:4 in MultiScreen-HA plates with mixed cellulose ester membranes (Millipore) that had been coated with glucose-6-phosphate isomerase (GPI)–His (5 μg/ml). The cells were incubated for 4 hours at 37°C on the antigen-coated plates. Ig secreted by the plated cells was detected by alkaline phosphatase–conjugated goat anti-mouse total Ig secondary antibody (SouthernBiotech) and visualized using nitroblue tetrazolium /BCIP substrate (Sigma).
Enzyme-linked immunosorbent assay.
Serum samples were plated at an initial dilution of 1:100 and diluted serially 1:5 in Immulon II plates that had been coated with GPI-His (5 μg/ml). Horseradish peroxidase–conjugated donkey anti-mouse total Ig (Jackson ImmunoResearch) was used as a secondary antibody. Antibody was detected using ABTS substrate (Fisher). The serum titer was defined as the reciprocal of the last dilution that gave an optical density >3 times the background.
Measurement of cytokine secretion.
Lymph node cells were harvested and cultured (2 × 106 cells/ml) for 24 hours in media alone or in phorbol myristate acetate (50 ng/ml) plus ionomycin (500 ng/ml). The supernatants were then harvested, and the levels of cytokines were determined by cytometric bead array (BD Biosciences). The samples were stained according to manufacturer's instructions and analyzed on a FACSCanto II flow cytometer using FACSDiva software (both from BD Biosciences). Cytokine concentrations were calculated by comparison to standard curves using FACSArray analysis software (BD Biosciences).
Spleen cells or joint draining lymph node cells (1 × 106) were stained with recombinant GPI-His and annexin V or with antibodies to B220, CD25, CD69, CD80, CD86, and class II major histocompatibility complex (MHC). Peripheral blood was collected from control, 1-MT–treated, anti-CD20–treated, and anti-CD20 plus 1-MT–treated K/BxN mice and stained with antibodies to pan-CD45 and the B cell–specific isoform B220. Samples were analyzed on a FACSCanto II flow cytometer using FACSDiva software. Data were analyzed using CellQuest software (BD Biosciences). Gating on live lymphocytes was based on forward and side scatter, with 100,000 events collected for each lymph node and spleen sample and 2,000 events for each blood sample.
Statistical significance was determined using Student's unpaired t-test or the Mann-Whitney nonparametric test with the use of InStat Software (GraphPad Software).
No effect of 1-MT on initial activation or survival of anti-GPI B cells.
Our previous work demonstrated that the anti-arthritic effect of 1-MT in K/BxN mice was due to its inhibition of the pathogenic B cell response, suggesting that IDO plays an important role in driving B cell–mediated autoimmunity (10). To define the mechanism by which IDO shapes the autoreactive B cell response that leads to arthritis, it is important to determine the step(s) at which 1-MT affects B cell activation and/or differentiation. Several steps in the B cell response could be affected by 1-MT treatment, and each of them could result in a diminished autoantibody response. These include the initial activation of the autoreactive B cells, their recruitment into germinal centers, or their terminal differentiation into antibody-secreting cells (ASCs), which occurs later in the response.
In the K/BxN mouse model, the autoantigen targeted by the pathogenic B cells is the glycolytic enzyme GPI (15). To follow anti-GPI B cells specifically, we used VH147 Ig–transgenic mice, which have an increased frequency of anti-GPI B cells in their preimmune repertoire (14). Use of these mice allows anti-GPI B cells to be tracked prior to activation, through the initial stages of activation, as well as once the immune response is under way, an impossibility without the Ig transgene (16). We previously showed that anti-GPI B cells in nonautoimmune mice were not rendered tolerant (14). Instead, the anti-GPI B cells in the recirculating follicular/lymph node B cell pool remained naive, although they showed clear evidence of antigen encounter. Importantly, these anti-GPI B cells could be induced to secrete high levels of autoantibodies in response to cognate T cell help when bred onto the autoimmune K/BxN background (VH147 K/BxN mice) (14).
To determine if 1-MT treatment would affect the initial activation of anti-GPI B cells, VH147 K/BxN mice were treated with 1-MT or with carrier alone. At the peak of arthritis, the transgenic anti-GPI B cells from the spleen and joint draining lymph nodes were analyzed by flow cytometry, gating on GPI-binding B cells. We found no differences in the percentages of GPI-binding B cells in the spleen (mean ± SEM 25.4 ± 2.3% in carrier-treated mice versus 25.9 ± 3.2% in 1-MT–treated mice; P > 0.9) or lymph nodes (8.6 ± 1.1% in carrier-treated mice versus 8.4 ± 2.1% in 1-MT–treated mice; P = 0.9) of carrier-treated mice as compared to 1-MT–treated mice. Compared to nontransgenic B cells (Figure 1) and anti-GPI B cells from nonarthritic mice, which remain naive (10), anti-GPI B cells from arthritic VH147 K/BxN mice showed signs of activation. They expressed elevated levels of the early activation markers CD25 and CD69, the costimulatory molecules CD80 and CD86, and class II MHC molecules (Figure 1A). Anti-GPI B cells from 1-MT–treated VH147 K/BxN mice also expressed elevated levels of these activation markers, and the levels were indistinguishable from those in carrier-treated mice (Figure 1A), indicating that 1-MT did not block this initial stage of B cell activation.
It is possible that IDO activity was not necessary for the activation of anti-GPI B cells but instead was required for B cell survival, such that blocking IDO activity with 1-MT resulted in B cell apoptosis. To address this possibility, we measured the percentage of apoptotic GPI-binding B cells in carrier-treated versus 1-MT–treated VH147 K/BxN mice with the use of annexin V and flow cytometry. While a greater percentage of B cells from VH147-transgenic mice were annexin V–positive as compared to those from nontransgenic mice (7.4 ± 4.0% versus 0.9 ± 1.5%), no differences were detected between those from VH147-transgenic mice receiving 1-MT (8.8 ± 0.7%) and those from VH147-transgenic mice receiving carrier (7.4 ± 4.0%). Therefore, neither the initial stages of anti-GPI B cell activation nor their survival is affected by 1-MT treatment.
Inhibition of anti-GPI B cell differentiation into ASCs by 1-MT treatment.
We next addressed whether 1-MT inhibited either the recruitment of anti-GPI B cells into germinal centers or their differentiation into ASCs. Recruitment into germinal centers was measured by flow cytometry using the germinal center marker peanut agglutinin. Germinal center B cells were readily detectable among the GPI-binding B cell population (Figure 1B). Similar percentages of GPI-binding B cells from the spleens (mean ± SEM 21.7 ± 4.6% in carrier-treated mice versus 22.7 ± 5.8 in 1-MT–treated mice; P > 0.9) and draining lymph nodes (18.3 ± 6.2% in carrier-treated mice versus 15.5 ± 6.9% in 1-MT–treated mice; P = 0.3) of carrier-treated and 1-MT–treated VH147 K/BxN mice showed high levels of staining with peanut agglutinin (Figure 1B).
To measure the number of anti-GPI B cells that differentiated into ASCs, joint draining lymph node samples from 1-MT–treated and carrier-treated VH147 K/BxN mice were quantified by ELISpot (Figure 1C). High numbers of GPI-reactive ASCs were present in the draining lymph nodes of carrier-treated VH147 K/BxN mice (mean ± SEM 174 ± 45). In contrast, very few anti-GPI ASCs were detectable in 1-MT–treated VH147 K/BxN mice (23 ± 7). As expected, no anti-GPI ASCs were found in the draining lymph nodes of either carrier-treated (0.3 ± 0.3) or 1-MT–treated (0.2 ± 0.2) control nonarthritic VH147 K/BxN mice.
Taken together, these data demonstrate that IDO activity is not required for the initial activation, survival, or recruitment of GPI-reactive B cells into germinal centers; instead, it exerts its effects later in their differentiation into autoantibody-secreting cells.
Inhibition of arthritis in K/BxN mice by B cell depletion.
The experiments using VH147 Ig–transgenic mice showed that IDO plays an activating role in establishing the autoreactive B cell profile at the onset of the autoimmune response. As such, inhibitors of IDO activity will be most useful prior to the activation of the autoreactive B cell repertoire. Therapeutically, this critical stage in B cell activation can be reinstated using B cell depletion with a chimeric monoclonal antibody directed against the B cell–specific cell surface marker CD20 (17). Because anti-CD20 Ig specifically targets immature, naive, and memory B cells, the B cell repertoire will be depleted starting at the precursor B cell stage, effectively “rebooting” the immune system (17). This is an ideal situation for testing whether 1-MT is able to inhibit the establishment of the autoreactive B cell profile in a therapeutic setting in which the initiation phase of the autoimmune response has been reinstated. We hypothesized that administration of 1-MT would inhibit the reactivation of autoreactive B cells as they are regenerated, thus preventing the recurrence of autoimmune arthritis.
Before we could test the efficacy of combination therapy with anti-CD20 plus 1-MT, we first needed to determine whether anti-CD20 could inhibit arthritis in the K/BxN mouse model, and if so, the time frame during which anti-CD20 depletion remained effective. To do this, K/BxN mice were treated with anti-CD20 Ig or isotype control antibody at 2 different time points, starting either after the onset of arthritis symptoms (4–5 weeks of age) or prior to the initiation of arthritis (3 weeks of age). In the first strategy, K/BxN mice were allowed to develop arthritis and were then treated with anti-CD20 or an isotype control antibody (Figure 2A). With this treatment regimen, anti-CD20 had only a minimal effect on arthritis progression. This is similar to findings using anti-human CD20 to deplete B cells in K/BxN mice expressing a human CD20 transgene (18).
In the second strategy, K/BxN mice were treated with anti-CD20 or isotype control antibody prior to arthritis onset (Figure 2B). Anti-CD20 treatment did not affect the time of arthritis onset; however, within 14 days (5 weeks of age), the anti-CD20 group began to show reduced arthritis symptoms. This reduction in ankle inflammation continued until 6–7 weeks of age, when arthritis dramatically flared. The half-life of anti-CD20 in the serum has been reported to be ∼4–5 days, with the B cell repertoire returning to full strength in ∼4–6 weeks (19). Therefore, it was possible that the arthritis flare seen 4 weeks posttreatment was due to the return of autoreactive B cells to the repertoire.
To determine the effectiveness of B cell depletion and, importantly, the timing of the reappearance of B cells following depletion, B cell percentages in the peripheral blood were monitored by flow cytometry. By 1 week following antibody treatment, B cell numbers were significantly decreased in anti-CD20–treated mice as compared to mice receiving isotype control Ig (Figure 2C). B cells began to approach normal levels ∼3 weeks following anti-CD20 treatment, and by 5 weeks following antibody treatment, they were close to the levels in control mice. This timing of B cell repopulation did indeed coincide with the arthritis flare, as shown in Figure 2B.
The data from these initial experiments demonstrated that while anti-CD20 treatment was unable to prevent the initiation of arthritis, it was able to halt the progression of arthritis, provided that treatment was begun prior to the onset of joint inflammation. Furthermore, these experiments established a time course of B cell depletion in this model, with both B cells and arthritis symptoms reappearing 3 weeks after anti-CD20 treatment.
Inhibitory effect of 1-MT treatment on anti-GPI ASC regeneration following B cell repopulation.
As single-agent therapy administered prior to arthritis onset, 1-MT inhibited B cell differentiation into ASCs. We hypothesized that administration of 1-MT would also inhibit the differentiation of autoreactive B cells as they were regenerated after B cell depletion therapy, thus preventing the recurrence of autoimmune arthritis. To test this hypothesis, K/BxN mice were treated with anti-CD20 prior to arthritis onset. Three weeks later, before the B cell population returned to normal levels, the mice were treated with 1-MT. Administration of 1-MT affected neither the depletion nor the repopulation of B cells in the anti-CD20–treated mice (Figure 3A).
After 3 weeks of 1-MT treatment, GPI-specific ASCs in the joint draining lymph nodes were measured by ELISpot (Figure 3B). As expected, large numbers of anti-GPI ASCs were present in draining lymph nodes from isotype control–treated mice (mean ± SEM 330 ± 70). In contrast to the reduction in ASC formation seen when 1-MT was administered before the development of arthritis (Figure 1 in this study and ref.10), administration of 1-MT alone after arthritis onset did not affect ASC formation (279 ± 116; P = 0.3) (Figure 3B). Treatment with anti-CD20 alone reduced the number of anti-GPI ASCs (99 ± 28; P = 0.03) as compared to the level in control mice. Importantly, administration of 1-MT following anti-CD20 treatment led to an even greater reduction in the number of anti-GPI ASCs present in the draining lymph nodes (42 ± 9; P = 0.05) as compared to mice treated with anti-CD20 alone.
A similar trend in reduced anti-GPI titers was seen in the serum of mice treated with anti-CD20 plus 1-MT as compared to those treated with anti-CD20 alone, although the difference did not reach statistical significance (Figure 3D). These data demonstrate that, similar to its effect prior to arthritis onset, 1-MT inhibits the differentiation of autoreactive B cells as they are regenerated following B cell depletion therapy.
Inhibitory effect of 1-MT treatment on arthritis flare following B cell repopulation.
The effectiveness of 1-MT at inhibiting pathogenic ASC differentiation following B cell recovery after B cell depletion suggested that it might also inhibit the arthritis flare observed in anti-CD20–treated mice following B cell repopulation. As shown in Figure 2B, anti-CD20 treatment did not affect the onset of arthritis but did diminish the severity of joint inflammation until repopulation of the B cell repertoire ∼4 weeks later. At this time, arthritis in anti- CD20–treated mice flared to the level seen in control mice. Strikingly, administration of 1-MT just prior to the repopulation of peripheral B cells (3 weeks following CD20 treatment) prevented the arthritis flare observed in anti-CD20–treated mice (Figure 3C). Administration of 1-MT alone at this time point had no effect on arthritis. Likewise, control mice treated with anti-CD20 followed by carrier or treated with isotype control antibody followed by 1-MT did not show this same inhibition (Figure 3C).
At the termination of the experiment (when the mice were 10 weeks of age), the rear ankles were harvested and examined for histologic evidence of arthritis following staining with hematoxylin and eosin (Figure 4). Isotype control–treated mice showed classic signs of arthritis, with a greatly expanded synovium, pannus formation, and inflammatory cell infiltrates. Mice treated with CD20 alone also showed inflammatory cell infiltrates. In contrast, joints from mice treated with CD20 plus 1-MT showed a reduction in the severity of arthritis, with minimal synovial expansion and few infiltrating inflammatory cells (Figure 4).
The levels of cytokines implicated in arthritic responses were compared in the joint draining lymph nodes of mice treated with anti-CD20 alone and mice treated with anti-CD20 plus 1-MT (Figure 5). Consistent with the reduction in arthritis, mice treated with the combination of anti-CD20 plus 1-MT had lower levels of the classic inflammatory cytokines interleukin-6 (IL-6), IL-17, IFNγ, monocyte chemotactic protein 1, RANTES, and tumor necrosis factor α (TNFα) as compared to mice treated with anti-CD20 alone. However, levels of the other inflammatory cytokines either were not different between the 2 treatment groups (macrophage inflammatory protein 1α [MIP-1α], MIP-1β) or were detectable at minimal levels in both groups (IL-1α, IL-1β). Importantly, levels of several cytokines implicated in B cell differentiation (IL-4, IL-6, IL-9, IL-10, and IL-13) were all significantly lower in mice treated with CD20 plus 1-MT (Figure 5).
Treatment with 1-MT did not affect cytokine levels in control mice not treated with anti-CD20 (Figure 6). Taken together, our data demonstrate that 1-MT is able to inhibit the reactivation of autoreactive B cells and prevent the recurrence of arthritis following B cell depletion therapy.
Autoantibodies are the hallmark of many autoimmune diseases, including RA (20). One of the most promising strategies for controlling pathogenic autoantibodies is B cell depletion using a CD20-specific antibody (4). However, long-term B cell depletion is difficult to maintain, and repopulation of the B cell repertoire is often accompanied by the return of arthritis symptoms (8, 9). New strategies for inhibiting the activation of autoreactive B cells upon repopulation would clearly help to increase the therapeutic effectiveness of B cell depletion with a CD20-specific antibody. Recently, we identified the IDO pathway as a major contributor to autoantibody production in a mouse model of RA (10). Inhibition of IDO with 1-MT attenuated arthritis progression by reducing autoantibody levels, suggesting an important role of IDO in driving autoreactive B cell responses.
In the present study, we used Ig-transgenic mice to show that IDO activity is essential for the differentiation of autoreactive B cells into ASCs but is not necessary for the initial stages of B cell activation. Furthermore, the addition of 1-MT to B cell depletion therapy prevented the reemergence of autoantibody-secreting cells and arthritis symptoms following reconstitution of the B cell repertoire. Our data suggest that IDO inhibitors could be used in conjunction with B cell depletion as an effective cotherapeutic strategy in the treatment of RA.
Our data demonstrate that IDO plays a role in driving the differentiation of B cells into ASCs in vivo. However, blocking IDO activity with 1-MT does not inhibit antibody production by purified B cells in vitro (10), suggesting that 1-MT may affect B cell differentiation indirectly by affecting the microenvironment in which the B cells are being activated. In support of this, we found that the levels of IL-4, IL-6, IL-10, and IL-13, all of which are cytokines that are necessary for B cell antibody production, were decreased in 1-MT–treated mice. Treatment with 1-MT inhibited ASC formation at the late/post–germinal center stage but did not appear to affect long-lived plasma cells, since the titers of Ig and the numbers of ASCs were not diminished in mice treated with 1-MT after the onset of arthritis. This stage-specific effect could be advantageous, in that 1-MT treatment would affect only the generation of newly formed ASCs and would not inhibit memory responses to pathogens to which acquired immunity has occurred through vaccination or prior exposure.
A positive role of IDO in driving B cell–mediated autoimmune responses is in contrast to the traditional view of IDO as having a suppressive function in T cell–mediated immunity (21–24). These findings may have implications for the development of the IDO inhibitor 1-MT as a clinical agent. Currently, 1-MT is in early-stage clinical testing as an anticancer therapeutic agent (25). Based on the presumed inhibitory action of IDO against T cells, one concern has been that the use of 1-MT might induce severe autoimmune-based toxicities. The use of 1-MT did exacerbate symptoms in some models of induced autoimmunity (26–28). However, there is no evidence of spontaneous autoimmunity resulting from 1-MT treatment in nonautoimmune mouse models (29), and our findings in the K/BxN mouse model of RA actually showed reduced autoantibody levels and evidence of improvement in inflammatory autoimmune symptoms with 1-MT treatment (10). Taken together with our findings demonstrating the role of IDO in driving autoantibody production, this suggests that the potential application of IDO inhibitors may be more far-reaching than is currently appreciated.
One potential new use of IDO inhibitors that our data point to is in a cotherapeutic strategy to increase the effectiveness of B cell depletion therapy. In K/BxN mice and other mouse models of RA, treatment with anti-CD20 leads to a rapid depletion of B cells from the circulating B cell repertoire (7, 18, 19). However, as the anti-CD20 antibody is cleared from the circulation, the B cell repertoire repopulates, and disease symptoms return (19). This is also seen in RA patients, in whom the reemergence of the B cell repertoire is often accompanied by the return of arthritis symptoms (8, 9, 30). A second treatment cycle with rituximab will sometimes, but not always, reduce the RA flare (31). Even when multiple treatment cycles are possible, long-term B cell depletion therapy may not be desirable. Although the problem is not as significant as that associated with other biologic therapies such as use of TNF-blocking agents, patients receiving B cell depletion therapy exhibit a higher risk of infections (31, 32). Furthermore, as B cells provide critical immune functions outside of their ability to produce antibodies, including cytokine secretion and antigen presentation, depletion of the entire B cell repertoire will adversely affect the immune system as a whole. In support of this, B cell depletion in mouse models has been shown to inhibit T cell function (6, 7).
In summary, our data suggest that adding IDO inhibitors to B cell depletion therapy is an effective way to inhibit the reemergence of autoantibody-secreting cells while allowing the repopulation of the B cell repertoire. Combination therapy with anti-CD20 plus 1-MT therefore has the potential to benefit RA patients by both eliminating pathogenic B cells that are already present and preventing the generation of new ones.
Both authors drafted the article, revised it critically for important intellectual content, approved the final version to be published, and take responsibility for the integrity of the data and the accuracy of the data analysis.
The authors would like to thank Drs. Flavius Martin, Andrew Chan, and Qian Gong (Genentech, South San Francisco, CA) for the murine anti-CD20 Ig, and Drs. Lisa Laury-Kleintop, Alexander Muller, and Lauren Merlo (Lankenau Institute for Medical Research) as well as Dr. Sudhir Nayak (The College of New Jersey, Ewing, NJ) for critical reading of the manuscript and thoughtful input.