IDO production, adaptive immunity, and CTL killing

Comment on Braun et al, page 2375, and comment on Potula et al, page 2382

Indoleamine 2,3-dioxygenase, an enzyme induced in antigen-presenting cells by specific proinflammatory mediators such as prostaglandin E₂ (PGE₂) or viral infection, can inactivate specific T-cell functions.

Dendritic cells (DCs) are key regulators of immune responses capable of promoting and suppressing T-cell activation. This functional plasticity is the net result of DCs integrating a plethora of signals in their local microenvironment. Most DC studies examine signals required for inducing protective immunity rather than tolerance. An interesting candidate for the latter is indoleamine 2,3-dioxygenase (IDO), a rate-limiting enzyme in tryptophan catabolism, which is expressed by antigen-presenting cells (APCs) such as macrophages and DC populations.

IDO production can suppress adaptive immunity and has a complex role in immunoregulation, such as T-cell tolerance to cancer and allografts, protection of the allogeneic fetus, resistance to autoimmune diabetes,¹  and regulation of fungal pathogenicity.²  Tolerance induction by IDO appears, at least in part, to be attributed to depletion of the amino acid tryptophan, and resultant increase in metabolites (eg, kynurenine). Interferon-γ (IFN-γ) or cytotoxic T-lymphocyte antigen 4 (CTLA-4)/B7 engagement induce IDO in APCs.¹  In this issue of Blood, 2 groups report novel insights that further elucidate the regulatory role and potential pharmacologic intervention of IDO activity in immunity.

Braun and colleagues describe a 2-step induction of IDO in human monocyte-derived DCs (MoDCs), a widely used in vitro model for studying DC function. Affymetrix gene array analysis indicated that DCs matured with tumor necrosis factor-α (TNF-α) and prostaglandin E₂ (PGE₂), expressed 100-fold higher IDO mRNA. Interestingly, although PGE₂ was responsible for IDO transcriptional up-regulation, other stimuli (such as TNF-α or Toll-like receptor [TLR] ligands) were required for its functional activation. Finally, PGE₂ mediated IDO induction via the E prostanoid-2 (EP2) receptor and cyclic adenosine monophosphate (cAMP) signaling.

These findings shed new light on IDO regulation by inflammatory mediators and explain why mRNA expression did not strictly correlate with IDO activity in previous reports. They also highlight a novel role of PGE₂ (and perhaps other cAMP agonists such as adenosine triphosphate [ATP]) in modulating DC functional capacity via IDO induction. PGE₂ (and ATP) can induce DC migration towards lymph node-directing chemokines,^3-5  attenuate interleukin-12p70 (IL-12p70) and IL-27 production and induce IL-23 production for memory T-cell activation.⁶  PGE₂ is frequently used in maturing MoDCs used for cancer immunotherapy, and so PGE₂ induction of IDO has implications for this strategy. However, do these findings indicate that PGE₂-matured DCs are immunosuppressive or rather that they attenuate the magnitude of responses to avoid excessive tissue damage? Although the functional consequences of DC-derived IDO on T cells was not addressed by this group, previous reports by Terness and colleagues⁷  demonstrate that inhibition of T-cell activation by IDO⁺ DCs may not be a pre-ponderant effect but may depend on the activation context.

What are the consequences of APCs producing IDO? A second paper appearing in this edition may shed light on this question. Potula and colleagues report that HIV-1 infection induces IDO⁺ macrophages that are subsequently protected from cytotoxic T lymphocyte (CTL) killing. Their persistence in the brain likely causes HIV-encephalitis (HIVE). Furthermore, pharmacologic inhibition of IDO enhanced CTL function and clearance of virally infected brain macrophages, highlighting that virally induced IDO in APCs may represent a unique viral escape mechanism. It is unclear whether IDO inhibits CTL development per se, since increased CTL numbers were found with IDO⁺ macrophages in HIVE brains. Munn et al recently showed that IDO⁺ plasmacytoid DCs directly suppress T-cell function via activation of the general control kinase 2 (GCN2) kinase pathway.⁸  Thus, IDO⁺ APCs may inactivate CTL function and escape CTL lysis.

Although APC resistance to CTL lysis appears deleterious in HIVE, it may be beneficial in certain contexts such as protecting antigen (Ag)-loaded, migratory DCs (activated via PGE₂) from CTL attack en route to lymph nodes. It is unclear from current studies whether IDO production is continuous or transient upon stimulus withdrawl. Closer examination of IDO⁺ APC and CTL interactions are required to more fully understand the implications. However, these 2 studies indicate that therapeutic inhibition of IDO may prove promising for enhancing (antiviral and tumor) or attenuating (autoimmune and transplantation) immune responses in a context-dependent manner. A better understanding of the regulation of IDO expression within the immune system may open new avenues for manipulating immunity in health and disease. ▪