|Home | About | Journals | Submit | Contact Us | Français|
Bone marrow transplantation (BMT) under costimulation blockade induces mixed chimerism and tolerance in rodent models. Recent data, predominantly from in vitro studies, suggest that in addition to blocking the CD28 costimulation pathway CTLA4Ig also acts through upregulating the tryptophan-catabolizing enzyme indoleamine-2,3-dioxygenase (IDO). Here we demonstrate that even though CTLA4Ig is critically required for the induction of chimerism and tolerance in a murine model of non-myeloablative BMT, IDO activity is not. No significant differences were detectable in the kynurenine to tryptophan ratios (indicative of IDO activity) in sera of BMT recipients treated with CTLA4Ig (tolerant group) versus BMT recipients treated without CTLA4Ig (non-tolerant group) versus naïve controls. In vivo inhibition of IDO immediately after BMT with CTLA4Ig or several months thereafter did not block achievement of chimerism and tolerance. Thus, IDO does not play a critical role in the induction or maintenance of chimerism and tolerance in a CTLA4Ig-based BMT model.
The dimeric fusion protein CTLA4Ig (abatacept) has been rationally designed with the intent to prevent T cell activation by blocking the CD28 T cell costimulation pathway (through saturating binding to CD28’s only known ligands CD80 and CD86) (1). Indeed, CTLA4Ig has been shown to effectively modulate T cell responses in numerous in vivo transplantation and autoimmune disease models (1-5) and has recently been approved by the FDA for the treatment of rheumatoid arthritis (6).
It has been suggested that CTLA4Ig and CTLA4-expressing T regulatory cells (T regs) have an additional mode of action by inducing the immuno-regulatory enzyme indoleamine-2,3-dioxygenase (IDO) in antigen-presenting cells (APC) by triggering a signal through CD80/CD86 leading to IFNγ-dependent induction of IDO (7;8). IDO is a tryptophan-catabolizing enzyme impairing T cell proliferation by tryptophan starvation and by the emergence of toxic downstream products (e.g. kynurenine) (9). The IDO inhibitor 1-methyltryptophan (1-MT) prevented maternal tolerance towards fetal alloantigen (10), and abrogated the graft-prolonging effect of CTLA4Ig in a pancreatic islet transplantation model (11) which is to date the only in vivo allotransplant model where the benefit of CTLA4Ig was shown to critically depend on IDO. Overall, the biologic relevance of IDO-mediated immunoregulation, however, remains a matter of debate (12;13).
A promising application of costimulation blockade is its use as a component of various BMT protocols for the induction of mixed chimerism and donor-specific skin graft tolerance (14-19). The immunological mechanisms underlying induction and maintenance of tolerance in such protocols have not been fully delineated yet. In particular, the role of IDO in mixed chimerism models has not been investigated so far. Here we demonstrate that induction of chimerism and tolerance in such a murine BMT model occurs independent of IDO, even though the regimen critically depends on CTLA4Ig.
First, in order to determine whether the success of this tolerance model depends on CTLA4Ig, groups of C57BL/6 (H2b) mice received non-myeloablative total body irradiation (TBI, 3 Gy, d-1), approximately 20×106 unseparated, fully mismatched Balb/c (H2d) bone marrow cells (d0) and costimulation blockade consisting of an anti-CD154 mAb (MR1; 1mg d0) with or without CTLA4Ig (0.5mg d+2) as previously described (20). The majority of mice treated with CTLA4Ig developed high levels of chimerism which remained stable in all tested leukocyte lineages for the length of follow-up (up to 46 weeks), whereas chimerism was not successfully induced in most recipients treated without CTLA4Ig (8/10 with vs. 3/10 without CTLA4Ig at week 12; p<0.05; Figure 1A-B, data from one representative experiment). Also, permanent donor skin graft survival was observed in the majority of CTLA4Ig-treated BMT recipients (7/10, Figure 1C), while almost all donor grafts were rejected in the group treated without CTLA4Ig (1/10 acceptors, p<0.01). Third party grafts were promptly rejected in all groups (Figure 1C) (pooled data of several similar experiments are shown in Table 1). Hence, CTLA4Ig plays a critical role in inducing chimerism and tolerance in this fully mismatched BMT model.
Costimulation-based mixed chimerism models critically require anti-CD154 mAb, and fail if this component is eliminated from an otherwise successful regimen (15;16;19). In contrast, costimulation blockade-based BMT models not requiring CTLA4Ig have been described (using anti-CD154 alone), but usually involve additional interventions such as donor-specific transfusion or CD8 depletion (21;22). Since the CTLA4Ig-based protocol used in the present studies crosses full major and minor antigen barriers it is a very stringent model with clinical relevance. The rate of successful chimeras achieved with the described protocol in the present study is consistent with previous experience (14;20). Likewise, we have previously observed that not all chimeras accept donor skin grafts permanently, the reason for which is currently under investigation, and might be related to tissue-specific antigens that are disparate between donor and recipient in this fully mismatched model.
Two approaches were employed to determine whether the effect of CTLA4Ig in this chimerism-tolerance model depends on IDO: 1) assessment of IDO activity by determining the ratio of kynurenine to tryptophan in serum (denoted as kyn/trp in this study) (23); and 2) in vivo inhibition of IDO by implantation of pellets releasing 1-methyltryptophan (1-MT) (10;11).
Kyn/trp of plasma or serum samples taken at several time points remained low and were similar in BMT recipients treated with CTLA4Ig (tolerant group) or without CTLA4Ig (non-tolerant group) and in naïve controls, arguing against a systemic CTLA4Ig-induced tryptophan catabolism in vivo (day 3 post-BMT: Kyn/trp ratio 11.8±2.7 with CTLA4Ig vs. 12.1±3.3 without CTLA4Ig vs. 18.3 naïve control, day 4: 10.4±3.4 vs. 13.9±4.6 vs. 17.5, day 11: 16.2±4.3 vs. 17.4±6.3 vs. 15.4; p=n.s. for each time-point; n= 8-10 per group) (Figure 2D). Furthermore, 1-MT pellets (see below) did not cause a change in kyn/trp during pellet release time (measured on day 10 after implantation), neither early nor late after BMT (Figure 2E). Studies investigating CTLA4Ig-mediated IDO mechanisms measured kynurenine levels in supernatants from in vitro or ex vivo cultured splenic DCs but not in the serum from experimental mice (8;11;24). In a Cyclosporine A-based murine cardiac allograft model kyn/trp in plasma was only elevated in untreated, rejecting mice but not in 1-MT treated mice (25). In our CTLA4Ig-based mixed chimerism model kynurenine levels remained low, close to the baseline kynurenine production which is predominantly maintained by the hepatic enzyme tryptophan-pyrrolase (26). These data reveal that CTLA4Ig has no detectable systemic influence on tryptophan catabolism in this model.
While the above described results suggest that IDO does not play an important role in this model, the effect of more localized IDO activity could theoretically be missed by serum measurements of tryptophan and kynurenine. To directly investigate whether IDO activity is necessary for chimerism and tolerance, we inhibited IDO activity with subcutaneously implanted pellets continuously releasing 1-MT, a competitive inhibitor of IDO (10) (7 day release time, 200 mg/pellet; purchased from Innovative Research of America, Sarasota, FL, USA (10))(11;27). An additional group of mice was implanted with placebo pellets. 1-MT pellet implantation (d-1, 7-day release) did not negatively affect chimerism induction, as the rate of successful chimeras (6/8 with 1-MT, Figure 1D) was similar to that of controls without pellets and placebo controls in the same experiment (7/12, not shown; p=n.s.). To rule out an insufficient duration of IDO inhibition, we implanted an additional pellet on day 6 in a separate experiment achieving a 14-day-release period that covers the main phase of tolerance induction during which CTLA4Ig and T regs are active (20). Again, chimerism was comparable to those in untreated mice in the same experiment (Figure 1E). Chimerism among all lineages even tended to be somewhat higher in 1-MT-treated animals (1-MT for 7 days or 14 days) compared to controls without pellets (from the same experiment) throughout follow-up, reaching statistical significance two weeks post-BMT for the myeloid lineage (10.1% vs. 35.8% MAC1+ donor cells; chimeras with CTLA4Ig vs. chimeras plus 1-MT for 14 days, p<0.05). Moreover, although more subtle effects of inhibiting IDO cannot be completely ruled out, there was no significant negative influence on the median survival time (MST) or the rate of donor skin graft acceptance between mice treated with 1-MT and the controls from the same experiments (pooled data shown in Table 1). Thus, in this BMT model the chimerism- and tolerance-promoting effect of CTLA4Ig is not reversed by inhibition of IDO.
The divergent results between islet allograft studies (11) and this BMT model regarding IDO dependency might be due to the different tolerance mechanisms playing a role in these protocols (28). Importantly, tolerance induction in a related costimulation-blockade based mixed chimerism protocol has previously been shown not to depend on IFN-γ (17), the suggested key regulator of the IDO pathway (7). Moreover, we used the recently FDA-approved version of CTLA4Ig (abatacept) which is a human CTLA4-IgG1 construct, which may have different binding and signaling properties than the murine CTLA4-IgG3 which showed IDO-dependency in the pancreatic islet allo-transplantation model. Murine IgG3 is known to self-aggregate via Fc-portions, which could potentially enhance signaling through B7 (29). Notably, we have previously shown that tolerance induction with the protocol used herein requires regulation by CD4 T regs (20), which the present data seem to suggest is not prevented by IDO inhibition.
In order to investigate whether IDO activity has a critical role in the maintenance of tolerance, we implanted 1-MT inhibitor pellets (14-day-treatment; n=5) and placebo pellets (n=3) in long-term chimeras bearing healed-in donor skin grafts. All mice maintained chimerism and only one mouse in the 1-MT group rejected its donor graft (Figure 2 A-C), suggesting that IDO inhibition does not abrogate established tolerance in mixed chimeras. Plasma samples of mice having received 1-MT pellets, placebo pellets or no pellets were taken on day 10, during the 14-day-release time and were analyzed for 1-MT concentrations by HPLC (Figure 2F). All tested mice implanted with 1-MT pellets showed 1-MT levels well above the concentration shown to be effective in inhibiting IDO (11), demonstrating that the 1-MT pellets used for IDO inhibition were active (Figure 2F).
Taken together, these data demonstrate that IDO plays no critical role in the induction and maintenance of chimerism and tolerance in this in vivo model, even though the regimen critically depends on CTLA4Ig. Thus, CTLA4Ig presumably acts by virtue of its costimulation-blocking effect.
We want to thank Franz Winkler for helpful technical assistance, the staff of the Institute of Biomedical Research for expert animal care, Dr. Edgar Selzer for assistance with TBI, and Dr. Andreas Heitger for critically reading the manuscript.
5Grant support: This work was supported by grants from the Jubilee Fund of the Austrian National Bank (9477 to F.L.), the Austrian Science Fund (SFB F2310-B13 to T.W.), and the government of the Austrian province Tyrol (to D.F.).
6Conflict of interest: The authors declare that no conflict of interest relevant to this work exists.