Dendritic cells have the option to express IDO-mediated suppression or not

In their recent article,¹  Terness et al contrast their findings with ours (D.H.M. and A.L.M.) regarding indoleamine 2,3-dioxygenase (IDO)–mediated suppression by certain human monocyte–derived dendritic cells (DCs) in vitro.^2,3  We used this model to make the point that there can exist certain types of human antigen-presenting cells in both the macrophage and the DC lineages that have the option of being immunosuppressive due to IDO expression, rather than always being immunostimulatory. Whether suppression or stimulation predominated was found to be regulated by the differentiation conditions, maturation signals, and state of DC activation. The key point from these studies thus was that DCs could be either suppressive or stimulatory, depending on whether the prevailing conditions led to induction of functionally suppressive IDO.

In our studies, we began by replicating the original findings of Hwu and colleagues,⁴  who showed potent IDO-mediated suppression by human monocyte–derived DCs. We then further optimized these conditions to generate consistent, high levels of suppression in our hands.^2,3  Factors enhancing IDO-mediated suppression included use of counterflow elutriation (rather than adherence) to isolate monocytes, use of X-vivo 20 serum-free medium (Cambrex, Baltimore, MD), and appropriate assay conditions to detect suppression. Factors decreasing IDO included bovine serum–containing medium, CD40 ligation, and pretreatment of DCs with interferon-γ in the absence of T cells or B7 ligation. Without Terness and colleagues having first replicated the original system of Hwu et al,⁴  it is difficult to identify specifically which of their modifications resulted in down-regulation of IDO-mediated suppression. However, their system used none of the factors that we found necessary to enhance IDO expression, while using all of the factors mentioned above that reduced IDO expression (plus, in many experiments, the addition of lipopolysaccharide). Thus, it is reasonable that the authors saw little IDO-mediated suppression in their system.

Cultured monocyte-derived DCs have been a useful model in many respects. We have published the details of our methods so that those investigators who wish to study IDO-mediated suppression in human DCs may follow our methods if they desire. Conversely, those investigators for whom IDO would be an undesirable attribute (as with DCs intended for clinical immunotherapy) may simply wish to test and reassure themselves that their system does not generate IDO-mediated suppression. It is interesting to know that the particular system employed by Terness and colleagues does not generate suppressive DCs. However, other investigators, including co-authors of this letter not affiliated with the earlier publications (J.W.Y. and M.R.), find that IDO expression can markedly alter the biology of T-cell responses to human monocyte–derived DCs (manuscript in preparation). The findings of Terness et al are thus in no way inconsistent with the stated conclusions of the earlier studies, and all of these points have been raised and discussed in our recent review.⁵