Ambivalent effects of dendritic cells displaying prostaglandin E2-induced indoleamine 2,3-dioxygenase



Prostaglandin E2 (PGE2), an abundantly produced lipid messenger in mammalian organisms, has been attributed to possess potent albeit ambivalent immunological functions. Recently, PGE2 has been reported to stimulate the commonly believed immunosuppressive indoleamine 2,3-dioxygenase (IDO) pathway in human dendritic cells (DCs), but without promoting DC immunosuppressive activity. Here, we report that PGE2 used as a DC maturation agent apparently has more diverse functions. PGE2-matured DCs acquired powerful IDO activity, which was sustained even after removing PGE2. These IDO-competent DCs were able to stimulate allogeneic T-cell proliferation, but achieved inhibitory activity as their content in DC/T-cell co-cultures increased. The DC inhibitory activity was reversed upon blockade of IDO activity, confirming that the suppressive effect was in fact mediated by IDO and occurred in a dose-dependent fashion. IDO-mediated T-cell suppression was restored upon re-stimulation of T cells in the absence of IDO activity, confirming its reversibility. T cells stimulated by PGE2-matured IDO-competent DCs were sensitized to produce multiple cytokines, comprising Th1, Th2, and Th17 phenotypes. Collectively, these data suggest that T cells stimulated by PGE2-matured DCs are not terminally differentiated and their ultimate type of response may be formed by microenvironmental conditions.


The intracellular enzyme indoleamine 2,3-dioxygenase (IDO) has a key function in tryptophan metabolism along the kynurenine pathway [[1]]. IDO's metabolic activity has been perceived to play a significant role in immune regulation. On the cellular level, the complementary effects of IDO-mediated metabolism, tryptophan starvation, and accumulation of immunomodulatory kynurenines at the antigen-presenting cell (APC)/T-cell synapse were proposed to effect regulation of T-cell activation and proliferation [[1, 2]]. In fact, in the last decade, multiple studies corroborated the role of IDO in immune regulation and induction of tolerance (reviewed in [[3, 4]]). IDO competence, i.e. IDO expression and activity [[5, 6]], has been suggested to play a critical role in numerous clinical conditions entailing immune regulation, including tolerance induction in pregnancy [[7]], transplantation [[8, 9]], and tumor immune evasion [[10]].

Effective induction of IDO competence is, by and large, restricted to cells of the monocyte/macrophage lineage and, in humans is predominantly found in dendritic cells (DCs) [[4, 11]]. Numerous agents, most prominently interferon (IFN)-γ, have been described to induce IDO activity in DCs [[12]]. IDO induction is generally believed to represent a feedback mechanism of APC activation [[8]]. This understanding is supported by the observation that neopterin, a guanosine triphosphate (GTP) cleavage product [[13]], has been found to be closely linked to IDO activity in a wide spectrum of conditions involving immune activation [[14-16]].

In some recent reports, also prostaglandin E2 (PGE2) was described to up-regulate IDO competence in human DCs [[17, 18]]. Induction of IDO competence through PGE2 occurred, like in human DC maturation [[19, 20]], in concert with TNF-α. This observation was quite intriguing since PGE2 routinely was part of the standard cytokine cocktail utilized to generate DCs for cancer vaccines. DC cancer vaccines strongly rely on effective T-cell activation. Thus, IDO-mediated DC immunoregulatory activity [[18]] in PGE2-matured DCs would obviously constitute an undesired counteractive effect. However, Krause et al. [[21]] reported that PGE2-matured DCs possess high T-cell stimulatory capacity despite IDO induction.

PGE2 is the most abundantly produced prostaglandin in human organisms and is synthesized from arachidonic acid through cyclooxygenase-2 (COX-2). It is a short-lived lipid messenger molecule signaling via binding of its cognate receptors, EP1–EP4 [[22]]. In immunology, research on PGE2 has so far revealed a puzzling picture comprising seemingly opposing functions, immune stimulation, and immune suppression [[22-24]]. PGE2 utilized as a maturation agent of human DCs was reported to inhibit proinflammatory cytokine production and to favor a Th2 polarization of T cells [[25, 26]].

In this study, we demonstrate that PGE2-matured DCs displaying strong IDO competence can indeed act as potent inhibitors of T cells in an allogeneic setting in vitro. However, allogeneic T-cell inhibition by PGE2-matured DCs is critically dependent on cell culture conditions and is immediately reversible once T cells are re-stimulated by APCs in the absence of IDO competence. These re-stimulated T cells are sensitized to produce multiple cytokines consistent with a mixed effector phenotype.


Potent and sustained IDO competence is inducible in human DCs by combined activity of TNF-α and PGE2

In order to address the question of whether IDO induction through PGE2 in DCs promotes a suppressive effect on T cells, human monocyte-derived DCs were first exposed to a cytokine cocktail, containing IL-1β, IL-6, and TNF-α, with or without PGE2 for 48 h and examined for IDO protein expression and tryptophan-metabolizing activity. To exclude the possibility that PGE2 affects the development of a mature phenotype of DCs, we ascertained that the exposure of immature DCs (iDCs) to the cyto-kine cocktail with and without PGE2 allowed generation of DCs exhibiting an equally mature phenotype. In fact, DCs displayed similar up-regulation of levels of expression of CD80, CD86, MHC class I and MHC class II, and CD40 upon being exposed to a cytokine cocktail with or without PGE2 for 48 h, compatible with a fully mature phenotype [[27]] (Fig. 1A). Consistent with previous findings, DCs exposed to cytokines with PGE2 up-regulated CCR7, a key factor for supporting the DC migratory capacity [[28]] and the costimulatory molecule OX40L [[29]] (Fig. 1A).

Figure 1.

Activation status and IDO induction in human monocyte-derived DCs upon maturation with cytokine cocktail with and without PGE2. (A) Immature DCs were matured with a cytokine cocktail with or without PGE2 for 48 h and DCs were examined for cell-surface marker expression typically associated with maturation (shaded histograms, unstained control; open histograms, DCs after maturation); one representative donor (experiment) of six is shown. (B) IDO protein expression was analyzed in immature and in differently matured DCs (top). One donor (experiment) representative of three is shown. Tryptophan (TRP, gray bars; left scale) and kynurenine (KYN, black bars; right scale) concentrations were determined in the cell culture supernatant after termination of maturation period by HPLC (bottom). Data are shown as mean + SEM of results pooled from three donors (experiments). (C) IDO protein expression was determined in immature and in differently matured DCs after removal of stimulating agents and re-culture in fresh complete medium for 48 h (top). One donor (experiment) representative of three is shown. Tryptophan and kynurenine concentrations in DC culture supernatants after removal of stimulating agents and re-culture in fresh complete medium for 48 h are shown. Data are shown as mean + SEM of results pooled from three donors (experiments).

Next, with regard to the induction of IDO competence, we observed that the addition of PGE2 to the cytokine cocktail was indispensible for the induction of IDO protein expression. Indeed, iDCs or DCs matured for 48 h by the cytokine cocktail only (in the absence of PGE2) failed to express IDO protein (Fig. 1B, upper panel) and their cell culture supernatants contained high concentrations of tryptophan (35.3 ± 4.1 μM, mean ± SEM) and low kynurenine levels (1.2 ± 0.9 μM). In clear contrast, when DCs were exposed to the cytokine cocktail and PGE2, IDO induction in DC developed gradually by time (B. Jürgens, unpublished observation) and was fully present by 48 h of exposure. By function, these 48-h matured DCs (∼0.5–1 × 106/mL) were competent to effectively metabolize tryptophan to kynurenines in cell culture medium, resulting in a nearly complete tryptophan deplete (1.2 ± 1.0 μM) and kynurenine-enriched milieu (15.9 ± 3.1 μM) (Fig. 1B, lower panel), indicating the potent IDO inducing effect of PGE2. A cytokine cocktail containing IL-1β, IL-6, but not TNF-α, resulted in an incomplete consumption of tryptophan from the cell culture medium of freshly cultured DCs despite the presence of PGE2 (tryptophan levels 12.7 μM) (Supporting Information Fig. 1, upper panel), supporting the previous observation of a combined effect of TNF-α and PGE2 being required to effectively induce IDO competence in human DCs [[17]].

Importantly, once IDO competence was induced through the combined activity of PGE2 and TNF-α, the capability of tryptophan degradation was maintained even upon re-culture of DCs in fresh medium only (Fig. 1C). Indeed, along with prolonged IDO protein expression after re-culture (Fig. 1C, upper panel), tryptophan was likewise nearly completely depleted from the fresh cell culture medium (Fig. 1C, lower panel) within 48 h of re-culture. In contrast, DCs matured with cytokine cocktail only and re-cultured in fresh medium were devoid of tryptophan-metabolizing capacity. The combinatory effect of PGE2 and TNF-α was confirmed in re-culture experiments, appeared to have a combinatory role, since stimulation and re-culture of DCs matured in the presence of PGE2 but absence of TNF-α had an intermediate effect on tryptophan consumption (Supporting Information Fig. 1, lower panel).

Together, these data demonstrate the powerful efficacy of PGE2 to induce strong and, as shown here, durable IDO activity in human DCs. Thus, PGE2-matured DCs are similarly effective in induction of DC IDO activity as was recently shown for DCs exposed to lipopolysaccharide (LPS) and IFN-γ [[6]].

PGE2 stimulates DC IDO competence but not proinflammatory cytokine or neopterin release

IDO induction is commonly believed to represent a feedback mechanism of immune activation [[8]]. In our recent studies, we observed that strong IDO induction as a counterregulatory mechanism mediated APC activation [6, 15], particularly when DCs were polarized by IFN-γ toward a DC1 phenotype [[30, 31]]. Here, consistent with previous literature [[25, 26]], we found that the cytokine and PGE2-matured DCs failed to produce the key cytokines characterizing a DC1 phenotype IL12p70 and IFN-γ [[30, 32]] (Fig. 2A). Instead, DCs matured with the cytokine cocktail only were unable to induce the DC1-associated cytokine IL-12, but released enhanced amounts of transforming growth factor (TGF)-β. Importantly, when PGE2 was added on top of the cytokine cocktail, DCs were induced to release IL-23 along with TGF-β.

Figure 2.

Cytokine and neopterin release by differently matured human monocyte-derived DCs. (A) Immature DCs were matured with LPS and IFN-γ, cocktail or cocktail containing PGE2 for 48 h. Cytokine release into cell culture supernatant was analyzed with a cytokine bead array. The asterisk (*) indicates below detection limit; n.a., not available because IFN-γ was present during DC maturation. (B) In parallel, cell culture supernatants were analyzed for concentrations of kynurenine (KYN, gray bars; left scale) and neopterin (NEO, black bars; right scale) released by DCs after 48 h of maturation (top) and after re-culture in fresh complete medium for 48 h (bottom). Data are shown as mean + SEM of results pooled from three donors (experiments).

To further pursue the question whether IDO activity induced by PGE2 correlated with DC activation, DCs were matured with cytokine cocktail in the presence or absence of PGE2 for 48 h and neopterin production and tryptophan metabolism were examined in parallel. These analyses showed that DCs activated by LPS and IFN-γ [[6]] concomitantly to acquiring IDO competence released high levels of neopterin into cell culture supernatants (Fig. 2B). This was in contrast to DCs matured with cytokine cocktail and PGE2. Despite displaying effective IDO competence as indicated by high-level kynurenine release, these DCs produced only negligible levels of neopterin. DCs matured with cytokine cocktail only, in addition to their inability of metabolizing tryptophan, also produced low levels of neopterin (Fig. 2B, upper panel). Similar observations were made, when DCs were re-cultured in fresh medium as in Fig. 1C (Fig. 2B, lower panel).

Collectively, these findings show that the strong and sustained IDO competence in human DCs achieved by PGE2 occurs in the absence of overt DC activation or proinflammatory cytokine release, but is accompanied by production of cytokines IL-23 and TGF-β, which were previously described to favor Th17 development [[33-35]].

PGE2-matured IDO-competent DCs act as potent inhibitors of allogeneic T-cell responses

To precisely assess whether PGE2 induced IDO activity in human DCs affects their stimulatory capacity, we pursued an experimental approach in which DCs were matured with cytokine cocktail in the absence or presence of PGE2 and were subsequently added in graded amounts to allogeneic CD3+ T cells (mixed leukocyte reaction, MLR). In these experiments, we made the following important observations: (i) DCs matured by cytokine cocktail in the absence of PGE2 were potent stimulators of total allogeneic T cells, resulting in a mean proportion of 90% and 88% of proliferating T cells—carboxyfluorescein succinimidyl ester (CFSE) negative T cells as shown in Fig. 3A—in cell cultures containing 10% and 20% DCs, respectively. Increasing the DC content to 50% slightly reduced the proportion of proliferating T cells (55 ± 5%, CFSE negative cells, mean ± SEM) (Fig. 3B, upper panel). (ii) Likewise, PGE2-matured, IDO-competent DCs were potent stimulators of allogeneic T-cell proliferation [[21]] (Fig. 3B, upper panel). However, when the DC content was gradually increased, the percent CFSE negative T cells was reduced. In fact, the per cent CFSE negative T cells stimulated by PGE2-matured IDO-competent DCs was slightly lower (75 ± 3%, p < 0.01) but strongly reduced to 22 ± 3%, (p < 0.0001) when the DC content was increased to 20% and 50%, respectively (mean ± SEM of five experiments), as compared with the per cent CFSE negative cells stimulated by cytokine-matured IDO-incompetent DCs. (iii) Inhibition of T-cell responses by PGE2-matured IDO-competent DCs equally affected both major T-cell subsets, CD4+, and CD8+ cells (Fig. 3B, lower panels). (iv) To unambiguously explore whether the inhibition of T-cell proliferation was IDO dependent, the previously described highly potent IDO inhibitor methyl-thiohydantoin-tryptophan (MTHT) was added to cell cultures in which the strongest inhibition of T-cell proliferation by IDO-competent DCs was observed (DC/T cell ratio 1:1) (Fig. 3C). MTHT has been described to powerfully and similarly to the more frequently used IDO inhibitor 1-methyl-DL-tryptophan (MDLT) inhibit tryptophan consumption (Supporting Information Fig. 2, [[6, 36]]). Addition of MTHT (50 μM) slightly increased the proliferative capacity of CD4+ cells but not of CD8+ cells in MLRs stimulated by cocktail-matured, IDO-incompetent DCs (Fig. 3C). However, MTHT potently reversed the inhibitory effect of IDO-competent DCs and enabled allogeneic CD4+ as well as CD8+ T cells to proliferate similarly as if they were stimulated by IDO-incompetent DCs (p = n.s.). This observation strongly supported that the inhibition of T-cell proliferation by PGE2-matured DCs was IDO dependent.

Figure 3.

Inhibitory effect of prostaglandin E2 (PGE2)-matured DCs on allogeneic T-cell proliferation depending on the amount of DCs in MLR. (A) DCs were matured with cocktail in the absence or presence of PGE2 and added to allogeneic T cells at the indicated ratios. The proliferative response of triplicate cultures was measured by CFSE dilution at day 7 of the MLR and is depicted as percent of CFSE negative CD3+ T cells. One donor, representative of five donors (experiments), is shown. (B) Cumulative results of (A). Gray squares: MLR with DCs matured with cocktail. Black squares: MLR with DCs matured with cocktail and PGE2. Data are shown as mean + SEM of results pooled from five donors (experiments). ***p < 0.001, Student's t-test. (C) DCs were matured as in (A) and were co-cultured with allogeneic T cells at a 1:1 ratio in triplicates in the absence or presence of the IDO inhibitor methyl-thiohydantoin-tryptophan (MTHT). T-cell proliferation was analyzed on day 7 of MLR and is indicated as mean percent CFSE negative T cells. Gray bars: MLR with DCs matured with cocktail. Black bars: MLR with DCs matured with cocktail and PGE2. Data are shown as mean + SEM of results pooled from three donors (experiments). ***p < 0.001; *p < 0.05; ns, not significant; Student's t-test.

Altogether, the present data clearly evidence that PGE2-matured, IDO-competent DCs are enabled to function as potent inhibitors of allogeneic T-cell responses in an IDO-dependent manner. However, similar to previous observations of our [[6]] and other laboratories [[37, 38]], the capability of IDO expressing DCs to suppress T-cell responses apparently highly depends on cell culture conditions.

The inhibition of allogeneic T-cell proliferation by PGE2-matured IDO-competent DCs is reversible

Based on the so far generated results, we next sought to explore whether the suppression of allogeneic T cells by stimulation with IDO-competent DCs is sustained and antigen-specific. To address this question, DCs were matured as above and used as stimulators of allogeneic T cells. After termination of the MLR by day 7, T cells were recovered from MLR, extensively washed, stained with CFSE, and re-stimulated in fresh medium with freshly prepared IDO-incompetent monocytes from either the same or third party donors.

In these re-stimulation experiments, the remarkable observation was that T cells stimulated with PGE2-matured, IDO-competent DCs and inhibited to proliferate in the first MLR, were able to fully proliferate to the same and third party donors upon re-stimulation by IDO-incompetent APCs (Fig. 4). The impact of this observation is that the inhibition of T-cell responsiveness by PGE2-matured IDO-competent DCs is neither antigen-specific nor durable. In other words, the IDO-mediated suppression of allogeneic T-cell responses functions in a reversible manner.

Figure 4.

Rapid restoration of T-cell proliferative capacity upon re-stimulation. DCs were matured as in Fig. 3A and were co-cultured with allogeneic T cells at a 1:1 ratio. T cells were then recovered on day 7 of the MLR, stained with CFSE and re-stimulated with freshly prepared monocytes from the same donor as in the primary MLR or monocytes from a third party donor for 4 days at a 1:1 ratio in duplicates. T-cell proliferation is indicated as per cent CFSE negative T cells. Gray bars: MLR with DCs matured with cocktail. Black bars: MLR with DCs matured with cocktail and PGE2. Data are shown as mean + SEM of results pooled from three donors (experiments).

PGE2-matured IDO-competent DCs sensitize allogeneic T cells for cytokine production

Finally, to get a better insight into the functional characteristics of T cells stimulated by PGE2-matured IDO-competent DCs, we examined the polarization of T cells by cytokine profiling.

T cells were co-cultured with differently matured DCs as above. At day 7 of MLR, cells were harvested and sorted for CD3+ T cells by flow cytometry to obtain a population without residues of dead cells or DCs. Subsequently, T cells were stimulated with anti-CD3 and anti-CD28 monoclonal antibodies (mAbs) to initiate cytokine production. At 24 h of culture, cell culture supernatants were collected and analyzed for a broad panel of cytokines.

These analyses revealed that T cells stimulated by DCs that had been matured by cytokine cocktail containing PGE2 were endowed with the capability of producing high levels of multiple cytokines upon re-stimulation with anti-CD3/anti-CD28 mAbs. In fact, the pattern of cytokine production of these T cells comprised cytokines associated with a Th1, Th2, and, intriguingly, Th17 polarization phenotype (Fig. 5A) and their cytokine-producing capability was significantly (by >70%) enhanced in either of the examined cytokines as compared with T cells stimulated by cocktail-matured DCs in the absence of PGE2.

Figure 5.

T cells stimulated by DCs matured with cytokine cocktail containing PGE2 display a mixed polarization phenotype. (A) DCs were matured and co-cultured with allogeneic T cells as in Fig. 4. After 7 days, T cells were recovered from MLR and sorted as CD3+ cells by flow cytometry and stimulated in triplicates with plate-bound anti-CD3/anti-CD28 mAb for 24  h. The amount of cytokines released was quantified in the cell culture supernatant by flow cytometry using a cytokine bead array. Gray bars: MLR with DCs matured with cocktail. Black bars: MLR with DCs matured with cocktail and PGE2. Data are shown as mean + SEM of results pooled from seven donors (experiments). *p < 0.05; **p < 0.005; ***p < 0.001; Student's t-test. Note the different scales on the x-axis. (B) T cells were cultured as in (A). After 24 h of re-stimulation, T cells were analyzed for intracellular cyto-kine release by flow cytometry. T cells that had not been prestimulated in an MLR were used as control. One donor (experiment), representative of seven donors (independent experiments), is shown.

As a next step, to dissect whether T cells stimulated by PGE2-matured DCs are polyfunctional, that is whether T cells can simultaneously produce multiple cytokines or distinct subpopulations with different cytokine-producing capabilities were induced, we examined T cells by intracellular cytokine staining. In pursuing this question, we first observed that in the applied experimental conditions, IL-2-, IFN-γ-, and IL-17-producing T cells were readily detectable by flow cytometry (Fig. 5B), while IL-10- or IL-4-producing T cells were below the detection limit (data not shown). We also noted substantial interindividual variability among the seven subjects included (Table 1). The important observations we made in these experiments were: (i) cytokine production was localized in a limited fraction of T cells. (ii) Prestimulation by allogeneic DCs prior to stimulation with anti-CD3/anti-CD28 increased, albeit not statistically significant, the number of cytokine-producing cells as compared with fresh T cells. (iii) The statistically significant increase of cytokine release by T cells stimulated with PGE2-matured DCs (Fig. 5A) was not significantly correlated with an increased proportion of cytokine-producing T cells. In detail, while in T cells stimulated by PGE2-matured DCs, the fraction of cytokine-producing cells was increased by 1.8-fold, 1.6-fold, and 1.7-fold, the amounts of cytokines released were increased by 7.9-fold, 6.2-fold, and 3.9-fold for IFN-γ, IL-2, and IL-17, respectively (Supporting Information Table 1). Thus, the significantly enhanced cytokine-producing capability of T cells stimulated by PGE2-matured DCs resulted from an enhanced cytokine-producing capability of individual cells rather than from an increase of the fraction of cytokine-producing cells. (iv) Finally, the fraction of T cells concomitantly producing several cytokines was low and was only slightly increased in T cells after stimulation with PGE2-matured DCs. (Fig. 5B; compare numbers in upper right quadrants, Supporting Information Table 2). This finding indicates that cytokine production to a major extent resides in distinct subpopulations while only a minority of T cells are polyfunctional, that is able to produce multiple cytokines in parallel.

Table 1. T cells stimulated by DCs matured with cytokine cocktail with/without PGE2 show large interindividual variability of cytokine producing cells.a
  1. a

    T cells were cultured and stimulated as described in Figure 5. Numbers in brackets indicate the range of the percent positive cells among the tested individuals. Statistics:

  2. b

    Fresh T cells (n = 3) compared to T cells prestimulated with DCs having been matured with cytokine cocktail (n = 7).

  3. c

    Fresh T cells compared to T cells prestimulated by DCs having been matured with cytokine cocktail containing PGE2 (n = 7).

  4. d

    T cells prestimulated by DCs having been matured with cytokine cocktail compared to T cells prestimulated by with DCs having been matured with cytokine cocktail containing PGE2.

PrestimulationPercent positive cellsp-valuePercent positive cellsp-valuePercent positive cellsp-value
 median(range) median(range) median(range) 
None3.9(0.7–7.0)0.34 b26.4(5.1–31.9)0.68 b0.5(0.5–2.2)0.90 b
Cocktail6.4(1.7–25.6)0.05 c22.5(11.5–46.5)0.12 c1.2(0.2–1.5)0.36 c
Cocktail + PGE220.3(4.3–30.9)0.16 d44.4(18.0–61.6)0.10 d1.5(0.3–3.2)0.11 d

All together, these observations suggest that allogeneic T-cell stimulation with PGE2-matured, IDO-competent DCs increases the capacity of T cells of producing multiple cytokines comprising Th1, Th2, and Th17 phenotypes. PGE2-matured DCs have only minor effects on enhancing the fractions of polyfunctional T cells.


The large number of observations assessing the role of PGE2 utilized for DC maturation in shaping immune responses so far has failed to provide a clear pattern. PGE2 comprises powerful though divergent effects on DCs, mediating immunostimulatory as well as immunoregulatory functions [[22-24]]. Our data, in composite, support the view of a diverse and ambivalent effect of PGE2-matured DCs on stimulated T cells.

PGE2, included into the cytokine cocktail used for maturation of human DCs, has, in addition to its known effect on enhancing DC migration, recently been demonstrated to be followed by the induction of IDO competence [[17, 18]]. The underlying mechanism, that is which of the two prostaglandin receptors active in human DCs, EP2, or EP4, is involved in IDO induction, has so far yielded opposing results [[17, 21]], and deserves further extensive and systematic studies for clarification. Irrespectively, our results confirm previous observations that a prolonged activation period (48 h) and a combinatory effect of PGE2 and TNF-α [[17]] enable DCs to become IDO competent and effectively deplete the cell culture medium of tryptophan. In addition, we here extend these findings showing that DC IDO protein expression along with IDO activity induced by PGE2 is remarkably durable; even after removal of the cytokine cocktail and PGE2, DCs continue to abundantly express IDO protein and effectively degrade tryptophan and generate kynurenine. By these findings, we identify PGE2 as being equally potent in IDO induction in human DCs as the so far reported most potent IDO-inducing molecule, IFN-γ [[6, 12]].

Notably, in contrast to IFN-γ stimulation, PGE2-matured DCs did not display signs of activation in terms of proinflammatory cytokines or neopterin production. This phenomenon is quite unusual [[16]], since IDO is often regarded as a feedback mechanism to inflammation, which may be indicated by proinflammatory cytokines [[6, 8]]. However, this finding corroborates our concept that proinflammatory DC cytokine production and IDO induction are separate processes [[6]].

The role of PGE2 in directing DCs to exert an immune stimulatory or immune regulatory effect on T-cell responses, particularly with respect to PGE2-induced IDO induction in human DCs, has recently been controversially discussed. In prior studies, PGE2 was shown to cooperate with TNF-α to support the development of a mature, highly immunostimulatory DC phenotype [[19, 20]], despite PGE2-induced IDO induction [[21]]. As a molecular correlate, PGE2 was reported to mediate up-regulation of co-stimulatory molecules, for example OX40L and CD70 [[29]]. Similarly, our data show that exposing iDCs to the cytokine cocktail with PGE2 resulted in a mature DC phenotype by up-regulation of co-stimulatory DC molecules, including OX40L. However, in clear contrast to previous findings, our data demonstrate that PGE2-matured DCs are endowed with the ability to inhibit T-cell proliferation in an IDO-dependent manner, yet, this ability was dependent on a high DC content in MLRs. This finding is compatible with a concept of a dose-dependent effect of IDO. Previous studies might have missed an IDO-dependent inhibitory effect on T-cell proliferation by PGE2-matured DCs when the amount of DCs was below a critical threshold [[21, 39]]. Conclusively, consistent with to previous observations, these findings suggest that the actual effects of IDO critically depend on experimental conditions [[6, 11, 37, 39]]. Our herein presented data, however, clearly evidence that DC maturation through PGE2 possesses the capacity to promote immunosuppression by effectively inducing a tryptophan deplete/kynurenine-rich microenvironment.

In addition, our data show two further significant facets of the immunological sequelae of stimulating T cells with PGE2-matured, IDO-competent DCs. First, the T-cell suppressive effect of IDO is dependent on the explicit presence of IDO competence; in other words, once IDO-competent DCs are replaced by IDO-incompetent APCs, T-cell proliferation is recovered. This observation fits into a concept, in which IDO despite essentially participating in immunoregulation does not induce sustained down-regulation of immune reactivity [[8]], that is IDO-mediated T-cell inhibition is rapidly reversible once IDO activity ceases. Second, T cells after stimulation with PGE2-matured DCs are predisposed for production of enhanced levels of multifaceted, cytokines polarizing Th1, Th2, and Th17 cells. The observation of a significantly higher simultaneous production of cytokines with seemingly opposing (pro- and antiinflammatory) effects by T cells after stimulation with cocktail and PGE2-matured DCs is particularly intriguing. PGE2 has previously been shown to directly support the generation of Th17 cells, while concomitantly down-regulating the Th1 associated factor T-bet [[33]]. In line with this observation, murine bone marrow-derived DCs stimulated in the presence of PGE2 were reported to support Th17 and suppress IFN-γ-producing cells [[40]] and this effect was maintained in vivo. Here, we show that PGE2 can indirectly, that is via DCs, polarize T cells toward a Th17 phenotype, possibly through the concurrent cytokine production of IL-23 and TGF-β [[34, 41, 42]]. However, differently from previous findings, cocktail/PGE2-matured DCs propagate IL-17 and IFN-γ production in parallel. Furthermore, a small proportion of T cells stimulated by allogeneic PGE2-matured DCs was able to simultaneously produce IL-17 and IFN-γ, which is compatible with a transitional state of Th17 toward Th1 T cells [[41]]. The implications of these findings are twofold. First, and in line with a recent review [[23]], the original perception that DCs matured with PGE2-containing cytokine cocktail are exclusive promoters of a Th2 polarized response [[26]] cannot be reinforced in our study. Second, PGE2-matured DCs used as stimulators of allogeneic T cells appear to generate T-cell subpopulations with differential cytokine production properties. Particularly in light of the increasingly recognized plasticity of T cells [[41, 43]], this multifaceted cytokine polarization phenotype suggests that the ultimate type of T-cell responses emerging from stimulation by allogeneic cytokine cocktail/PGE2-matured DCs is not terminally defined.

Our in vitro data suggest that in vivo the outcome of T-cell responses stimulated by cytokine cocktail/PGE2-matured DCs—displaying IDO activity and generating multifaceted cytokine-producing T cells—may depend on the balance of the contingent microenvironment conditions, where DCs interact with T cells. For example, in the presence of sufficient numbers of IDO-competent DCs, abundant IDO activity may impair immune effector responses. On the other hand, PGE2-mediated polarization of T cells toward a Th1 and Th17 phenotype may effectively combat pathogen invasion or spreading of tumor cells. Currently, the use of PGE2 for maturation of DCs used for cancer vaccines is ambivalently perceived. In principle, PGE2-matured DC loaded with tumor antigens has been shown to effectively increase the number of circulating tumor-specific T cells [[44]]. PGE2 originally was added to DC maturation strategies in cancer vaccine trials to equip DCs with a migratory potential upon adoptive transfer, in particular by up-regulation of CCR7 [[28]] and IL-12p40 secretion [[45]]. On the other hand, however, PGE2-matured DCs have been shown to be unable to prime naïve T cells [[46]] and to favor the attraction of regulatory T cells [[47]]. Previous studies have suggested that in vivo particularly plasmacytoid DCs are potent mediators of IDO activity [[48-50]], which can promote regulatory T cells [[51]]. For human monocyte-derived DCs stimulated by LPS, IFN-γ or, as shown here, by PGE2 in vitro, it is currently unknown, whether a specific subset of DCs or the total DC population is induced to express IDO competence. Likewise, it is not known whether the in vitro induced IDO competence persists upon adoptive cell transfer in vivo, nor whether, as previously reported in mice [[40]], the capacity of T cells stimulated by PGE2-matured DCs to differentiate into diverse T-cell polarization phenotypes will be retained. Yet, our data, particularly by the finding of sustained IDO activity, point at the potential of PGE2-matured DCs to impair T-cell responses and/or induce regulatory T cells through IDO [[6, 52]]. These potentially adverse effects of PGE2 maturation of DCs in cancer vaccine trials may represent critical factors for the as yet limited success of this type of cell therapy [[53-55]].

In conclusion, our herein presented findings are compatible with a concept suggesting that the effect of PGE2-matured DCs stimulating allogeneic T cells is diverse and may be highly context-dependent. PGE2-matured DCs, provided that local IDO activity is low, may effectively induce stimulatory immune responses. However, once the enzymatic activity of IDO predominates, either by a locally high content of IDO competent DCs or by a relatively increased DC-to-T cell ratio, a tryptophan deplete/kynurenine-rich microenvironment may effectively suppress T-cell responsiveness or promote regulatory T-cell responses [[6, 52]]. Moreover, the type of response, which T-cell populations promote upon stimulation by PGE2-matured DC, may vary dependent on which cytokine polarization prevails. These diverse effects may underlie the as yet unpredictable efficacy of DC vaccine trials using PGE2-matured DCs.

Material and methods


All experiments were performed with human blood obtained from healthy volunteers or from blood donors at the blood bank of the General Hospital of Vienna once informed consent was given, in accordance with the Declaration of Helsinki. Approval was obtained from the Children's Cancer Research Institute Institutional Review Board for these studies. Venipunctures were performed by experienced medical staff under standard sterile conditions.

Cell culture medium

DC differentiation and maturation was performed in AIM V cell culture medium (Gibco; Life Technologies Ltd., Paisley, UK) containing 2% human plasma (Octaplas; Octapharma, Vienna, Austria) and 1% vol/vol L-glutamine (PAA Laboratories, Pasching, Austria), hereafter termed complete medium. T-cell stimulation assays were performed in complete medium supplemented with 25 mM Hepes (PAA Laboratories). All cultures were maintained in humidified air containing 5% CO2 and 37°C.

Cell culture

Monocytes were enriched by adherence of peripheral blood mononuclear cells (PBMCs) or by counterflow centrifugal elutriation (Elutra Cell Separation System; Gambro BCT, Lakewood, CO, USA), which resulted in a population of 80% CD14+ cells. These monocyte-enriched cell populations were plated in culture flasks (Iwaki; Sterilin; Thermo Fisher Scientific, NY, USA) at a density of 0.3–0.5 × 106 cells/cm2. iDCs were generated by culture in complete medium supplemented with 1000 U/mL GM-CSF (PeproTech, London, UK) and 400 U/mL IL-4 (PeproTech) for 5 days. For DC maturation, iDCs were exposed to a cytokine mix containing of IL-1β (1 × 104 U/mL), IL-6 (103 U/mL), TNF-α (103 U/mL) (all PeproTech) in the presence or absence of PGE2 (1 μg/mL) (Sigma-Aldrich, St. Louis, MO, USA) for 48 h. In some experiments, iDCs were matured with 50 ng/mL LPS (Calbiochem, La Jolla, CA, USA) and 1000 U/mL IFN-γ (Boehringer Ingelheim, Ingelheim, Germany). The DC quality was monitored by visual and flow cytometric evaluation of a typical DC morphology and expression of cell-surface markers, respectively.

Highly enriched T cells were prepared as described [[56]]. In brief, total CD3+ T cells were isolated from total PBMCs by magnetic cell sorting (Miltenyi Biotec Inc., Bergisch Gladbach, Germany) according to manufacturer's instructions, routinely resulting in a greater than 95% enrichment of the targeted cell population. Purity and viability were monitored by flow cytometry.

Flow cytometric analysis and cell sorting

Cells were stained with the following mAbs: anti-CD3-perCP, anti-CD4-allophycocyanin, anti-CD25-PE-Cy7, anti-CD86-allophycocyanin, anti-CD14-perCP-Cy5.5, anti-OX40L-PE, anti-CD40-FITC (all from BD Biosciences, San Jose, CA, USA); anti-CD8-PE, anti-HLA class II-FITC, anti-HLA class I-PE (DakoCytomation, Glostrup, Denmark); anti-CD80-PE (Immunotech; Beckman Coulter, Orange County, CA, USA); anti-CCR7-allophycocyanin (eBioscience, San Diego, CA, USA).

Intracellular cytokine staining was performed with the following mAbs: anti-IL-4-FITC, anti-IL-10- allophycocyanin (BD Biosciences); anti-IL-2-Brilliant Violet, anti-IL-17A-PE-Cy7 (Biolegend); anti-IFN-γ-PE (eBioscience).

Flow cytometry analyses used an LSR II flow cytometer (BD Biosciences). T cells were sorted on a FACSAria flow cytometer (BD Biosciences), routinely resulting in a greater than 99% enrichment of the targeted cell population. Data were analyzed with FlowJo software (Treestar, Ashland, OR, USA).

T-cell stimulation

For MLR, graded amounts of differently matured DCs were co-cultured with a constant number of CFSE-labeled T cells (2.5 × 105 cells in 48-well plates (Iwaki)). T cell proliferation was calculated as the per cent CFSE negative cells [[57]] after 7 days of co-culture. Where indicated, cell cultures were supplemented with 50 μM methyl-thiohydantoin-tryptophan (MTHT; Sigma-Aldrich) [[36]] or 300 μM 1-methyl-DL-tryptophan (MDLT; Sigma-Aldrich) to block IDO activity.

For re-stimulation experiments, unlabeled T cells were used for the first MLR. At the termination of MLR, nonadherent cells were recovered, thoroughly washed, and stained with CFSE. A total of 2.5 × 105 T cells were re-stimulated with in 48-well plates with monocytes from the same donor as in the primary MLR or monocytes from a third-party donor for 4 days at a 1:1 ratio.

IDO expression and activity

IDO protein expression in DCs was investigated by immunoblot analysis by using a mouse monoclonal antihuman IDO antibody, kindly provided by O. Takikawa (National Institute for Longevity Sciences, National Center for Geriatrics and Gerontology, Aichi, Japan) [[58]].

IDO enzymatic activity was determined by measuring the levels of tryptophan and kynurenine in the cell culture supernatants by high-pressure liquid chromatography as previously described [[15]]. Briefly, tryptophan was detected by its natural fluorescence at 286-nm excitation and 366-nm emission wavelengths. 3-nitro-L-tyrosine, used as an internal standard, and kynurenine were determined by ultraviolet absorption at 360 nm. An albumin-based external standard mix was prepared and included 50 μM tryptophan (Serva, Heidelberg, Germany), 10 μM kynurenine (Sigma-Aldrich), and a frozen serum pool. Upon the addition of 25 μL of 2 M trichloroacetic acid (Merck, Darmstadt, Germany), the reaction vials were immediately vortexed and centrifuged at 12,000 g (13,000 rpm) for 6 min at room temperature to precipitate protein. The concentration of the components was calculated according to peak heights and was compared with 3-nitro-L-tyrosine as reference standard. Neopterin release by DCs was quantified by ELISA (BRAHMS Diagnostica, Berlin, Germany).

Intracellular cytokine staining

To assess polarization of T cells co-cultured with DCs matured with cocktail in the absence or presence of PGE2, intracellular cytokine staining was performed. T cells were cultured with differently matured DCs in a primary MLR. Cells were recovered and sorted for CD3+ T cells by FACS and re-stimulated with anti-CD3 (125 ng/mL, Sigma) and anti-CD28 (250 ng/mL, Sigma-Aldrich) mAb for 24 h. Subsequently, T cells were stimulated with PMA and ionomycin for another 4 h. T cells were harvested and stained for CD4 and intracellular IL-2, IL-4, IL-10, IL-17A, and IFN-γ according to manufacturer's instructions.

Cytokine production

T cells were cultured with differently matured DCs in a primary MLR. Cells were recovered and sorted for CD3+ T cells by FACS and re-stimulated with anti-CD3 and anti-CD28 mAb for 24 h. Cytokine release into cell culture supernatants was analyzed by a cytokine bead array (FlowCytomix; eBioscience) according to the manufacturer's instructions.

Statistical analysis

All statistical analyses were performed by the use of Student's t-test. A p-value below 0.05 was considered to indicate statistical significance.


We are grateful to Marion Zavadil for carefully reading the manuscript. This work was supported in part by the Austrian Science Fund (FWF), grant P20865-B13, to A.H.

Conflict of interest

The authors declare no financial or commercial conflict of interest.


immature dendritic cell


indoleamine 2,3-dioxygenase




prostaglandin E2