Granulocyte colony-stimulating factor promotes the generation of regulatory DC through induction of IL-10 and IFN-α



We have recently demonstrated that G-CSF promotes the generation of human T regulatory (TREG) type 1 cells. In this study, we investigated whether the immunomodulatory effects of G-CSF might be mediated by DC. CD14+ monocytes were cultured with serum collected after clinical administration of G-CSF (post-G), which contained high amounts of IL-10 and IFN-α. Similar to incompletely matured DC, monocytes nurtured with post-G serum acquired a DC-like morphology, expressed high levels of costimulatory molecules and HLA-DR, and exhibited diminished IL-12p70 release and poor allostimulatory capacity. Importantly, post-G DC-like cells were insensitive to maturation stimuli. As shown by neutralization studies, IFN-α and, even more pronounced, IL-10 contained in post-G serum inhibited IL-12p70 release by post-G DC-like cells. Furthermore, phenotypic and functional features of post-G DC-like cells were replicated by culturing post-G monocytes with exogenous IL-10 and IFN-α. Post-G DC-like cells promoted Ag-specific hyporesponsiveness in naive allogeneic CD4+ T cells and orchestrated a TREG response that was dependent on secreted TGFβ1 and IL-10. Finally, neutralization of IL-10 and IFN-α contained in post-G serum translated into abrogation of the regulatory features of post-G DC-like cells. This novel mechanism of immune regulation effected by G-CSF might be therapeutically exploited for tolerance induction in autoimmune disorders.


Bone marrow stromal cell




CD40 ligand


5-Carboxyfluorescein diacetate succinimidyl ester


Granulocyte colony-stimulating factor


GM-CSF + IL-4-differentiatedDC


Hematopoietic stem cell


After clinical administration of G-CSF


Before clinical administration of G-CSF


Regulatory DC


Regulatory T


TREG type 1

1 Introduction

G-CSF is routinely administered to cancer patients and healthy individuals to mobilize hematopoietic stem cells (HSC) into peripheral blood for autologous and allogeneic transplantation, respectively, in malignant and non-malignant diseases. Recently, unexpected effects of G-CSF on T cells have been demonstrated 14, and studies in mice and humans haveshown that G-CSF mobilizes Th2-inducing DC (DC2) with up-regulated levels of IL-3 receptor α chain (CD123) and CCR7 chemokine receptor 5, 6. Interestingly, monocytes contained at high frequency within PBMC mobilized with G-CSF express reduced levels of the costimulatory molecule CD86 and might contribute to the suppression of T cell proliferation through the release of IL-10 7, 8.

We have previously shown the induction of human T regulatory (TREG) type 1 (TR1) cells by G-CSF 9. T cells primed in vivo by G-CSF released high amounts of IL-10 and TGF-β1 when activated with allo-Ag in vitro and suppressed the proliferation of nonregulatory T cells primarily by a cell contact-independent but cytokine-dependent mechanism 9. Interestingly, the T cell regulatory phenotype was specifically induced by in vivo administration of G-CSF and not by in vitro exposure to the cytokine. Thus, we hypothesized that the immunomodulatory effects of G-CSF on TR1 cell development may be mediated by as yet uncharacterized cellular and/ or humoral factors.

DC are highly specialized professional APC that initiate primary immune responses 10. DC can be defined by distinctive morphology, by the expression of MHC class I and II Ag and costimulatory molecules, and by the ability to prime allogeneic T cells and to process and present Ag to autologous T cells 10. Our current knowledge suggests that the induction of tolerance as opposed to immunity might be determined by the ratio of immature DC (iDC) to mature DC (mDC) 11. As DC are pleiotropic modulators of T cell activity and are endowed with exquisite plasticity, pharmacological agents that manipulate DC function and favor the development of TREG cells might be exploited for the treatment of autoimmune disorders 12.

In this study, we report that the TR1-like cytokines IL-10 and IFN-α 13, which are produced and released by BMSC upon G-CSF administration (post-G), promote the generation of regulatory DC-like cells that are profoundly impaired in IL-12p70 release and in their ability to stimulate the alloreactivity of naive CD4+ T cells. Furthermore, post-G DC-like cells as well as monocytes matured with exogenous IL-10 and IFN-α induce the differentiation of naive CD4+ T cells into TREG cells that suppress bystander T cell responses via the secretion of TGF-β1 and IL-10.

2 Results

2.1 Cellular origin of IL-10 and IFN-α induced by G-CSF administration

Following our previous observation that TR1 cells can be generated in vitro after clinical provision of G-CSF to healthy volunteers 9, we first measured serum levels of cytokines implicated in the differentiation of TREG cells. Interestingly, both IFN-α and IL-10 were increased after G-CSF administration, whereas TGF-β1 levels were not (Fig. 1A). Highly purified post-G T and B cells, monocytes and NK cells secreted comparable amounts of IFN-α and IL-10 on a cell-per-cell basis with respect to equal numbers of their counterparts before administration of G-CSF (pre-G) (data not shown). Notably, BMSC cultured in the presence of exogenous G-CSF released significantly higher quantities of IL-10 and IFN-α at any time point compared with BMSC maintained in the absence of exogenous cytokines. As shown in Fig. 1B, IL-10 secretion was maximal at 24 h, whereas IFN-α release reached a peak level following 96 h of activation and declined thereafter (data not shown). Thus, BMSC may represent one major source of IL-10 and IFN-α after in vivo G-CSF administration.

Figure 1.

 G-CSF administration enhances IFN-α and IL-10 release. (A) Serum levels of IL-10, TGF-β1 and IFN-α were measured by ELISA. Mean and SD from 12 individual samples analyzed in duplicate are shown. *p<0.01 compared with pre-G values. (B) Release of IL-10 and IFN-α from BMSC exposed to exogenous G-CSF at 20 ng/ml. Prior to cytokine ELISA, G-CSF-treated BMSC were activated for 24 h at 37°C with 10 ng/ml PMA (Sigma Chemical Co.). *p<0.01 compared with BMSC cultured in the absence of exogenously added G-CSF.

2.2 Post-G monocytes cultured with post-G serum acquire the phenotypical and morphological features of DC

The finding of increased IL-10 and IFN-α serum levels prompted us to explore how post-G serum would affect monocyte differentiation. Highly purified post-G CD14+ monocytes cultured in the presence of post-G serum, and further activated with TNF-α, acquired a DC-like phenotype (Fig. 2A; Table 1) and morphology (Fig. 2B). Specifically, the monocyte/macrophage Ag CD14, costimulatory molecules, HLA-DR, CD4 and CD40 were highly expressed. Compared with their pre-G counterparts, post-G DC-like cells expressed higher levels of CD1a, CD4, CD86, CD83, BDCA-4 and HLA-DR. CCR6 chemokine receptor, that has been associated with regulatory and/or iDC subsets 14, 15, was detected on 21%±3% and 40%±6% of post-G and pre-G DC-like cells, respectively. Exogenous G-CSF exerted no measurable effects on post-G monocyte differentiation, indicating that G-CSF per se was not implicated in the acquisition of the above reported DC-like phenotype (data not shown).

Similar to post-G DC-like cells, post-G monocytes incubated under serum-free conditions and in the presence of exogenous IL-10 and IFN-α maintained elevated CD14 and CD80/CD86 Ag levels and expressed high levels of the CD83 maturation Ag (Fig. 2). Also, the DC-associated Ag BDCA-4 was abundantly expressed on DC-like cells differentiated in the presence of IL-10 and IFN-α.

Table 1. Effect of post-G serum on expression of DC markers
Priming conditionBDCA-4CD40HLA-DRCD14CD83CD11cCD123CD4CD80CD1aCD86
  1. a) Measured by flow cytometry as % cells stained. MFI = Mean Fluorescence Intensity.

  2. b) *p<0.05 compared with pre-G serum-supplemented DC cultures. §p<0.05 compared with post-G serum-supplemented DC cultures.

BIT %a) 4 ± 2  95 ± 4  97 ± 2  95 ± 4  5 ± 2  95 ± 4  73 ± 8  90 ± 4  54 ± 4  12 ± 4  10 ± 4 
MFI  20 ± 3  50 ± 8 600 ± 582440 ± 12026 ± 10110 ± 14 75 ± 10240 ± 10 60 ± 7  67 ± 7  50 ± 8 
Pre-G serum %a) 54 ± 6  85 ± 6  97 ± 2  98 ± 1  8 ± 3  93 ± 4  72 ± 8  79 ± 5  70 ± 7  2 ± 2  70 ± 5 
MFI  38 ± 7  40 ± 6 400 ± 502460 ± 13030 ± 5 75 ± 10 70 ± 10 80 ± 9  55 ± 6 N.D. 80 ± 15
Post-G serum %a) 93 ± 3*b)  90 ± 5  98 ± 2  95 ± 2  8 ± 4  95 ± 4  62 ± 9  95 ± 4*  84 ± 9  15 ± 4*  92 ± 5* 
MFI 120 ± 10*  80 ± 7* 700 ± 75*2800 ± 245*42 ± 8100 ± 14 47 ± 10140 ± 15* 65 ± 7  55 ± 8 125 ± 15*
Post-G serum %a) 93 ± 3*b) 90 ± 5  98 ± 2  95 ± 2  8 ± 4  95 ± 4  62 ± 9 95 ± 4*  84 ± 9  15 ± 4*  92 ± 5* 
MFI 120 ± 10*  80 ± 7* 700 ± 75*2800 ± 245*42 ± 8100 ± 14 47 ± 10140 ± 15* 65 ± 7  55 ± 8 125 ± 15*
GM-CSF+IL4 %a) 95 ± 3  92 ± 5  97 ± 2  38 ± 10 15 ± 6  80 ± 6  72 ± 6  40 ± 6  85 ± 6  39 ± 6  29 ± 6 
MFI 440 ± 22 100 ± 12450 ± 46 110 ± 24 35 ± 8 40 ± 10 75 ± 8  72 ± 7  84 ± 13 42 ± 10 60 ± 12
GM-CSF+IFNα %a) 97 ± 2  92 ± 6  97 ± 1  86 ± 10 18 ± 4  79 ± 8  90 ± 5  81 ± 6  85 ± 6  27 ± 6  19 ± 7 
MFI 650 ± 25  70 ± 7 680 ± 782280 ± 80 46 ± 9 50 ± 8 140 ± 12 90 ± 8 130 ± 12 40 ± 12 60 ± 16
IL-10+IFNα %a) 98 ± 1  87 ± 6  96 ± 3  94 ± 3 70 ± 7§ 65 ± 8§ 88 ±5  93 ± 2  94 ± 3  23 ± 5  95 ± 2 
MFI 645 ± 30§b) 45 ± 7 520 ± 522400 ± 50 80 ± 5§ 48 ± 12§160 ± 15§130 ± 18120 ± 20§ 36 ± 14210 ± 25
Figure 2.

 Phenotypic and morphological features of post-G DC-like cells. (A) Post-G DC-like cells were differentiated as detailed in section 4. One representative experiment out of ten with similar results is shown. The percentage of cells reacting with each mAb is indicated. (B) Cytospins of cultured monocytes were performed by centrifugation of 2×103 cells onto slides (500 rpm), followed by staining with Wright-Giemsa. Post-G monocytes acquired a heterogeneous morphology with irregular nuclear outline and numerous cytoplasmic projections.

2.3 IL-12p70 release by post-G monocytes cultured with autologous post-G serum is inhibited by IL-10 and IFN-α 

Following the observation that G-CSF down-regulates monocytic IL-12 release 16, we tested whether, and to what extent, IL-12 secretion from post-G DC-like cells might be affected. IL-12p70 release was significantly lower from post-G compared with either pre-G DC-like cells or GM-CSF + IL-4-differentiated DC (GM-CSF/IL-4DC) (Fig. 3A). Similarly, DC differentiated from post-G monocytes in the presence of IFN-α and GM-CSF (GM-CSF/IFN-αDC), but especially those obtained from monocytes cultured with IFN-α and IL-10 (IFN-α/IL-10DC), released low levels of IL-12p70. To substantiate our hypothesis that suboptimal induction of IL-12p70 might be favored by IL-10 and IFN-α contained in post-G serum, neutralizing Ab to these cytokines were added individually or in combination at the start of the DC culture. As depicted in Fig. 3A, IL-12p70 release from post-G DC-like cells was restored and equaled that from pre-G DC-like cells. Inhibition of IL-12p70 release exerted by IFN-α and IL-10 was synergistic, since the combination of anti-IFN-α and anti-IL-10 Ab was more effective at reinducing IL-12p70 release than either Ab used alone. However, slightly more elevated IL-12p70 production occurred in cultures containing anti-IL-10 Ab compared with anti-IFN-α Ab (Fig. 3A). To determine whether maturation stimuli could counteract the inhibitory activity of post-G serum on IL-12p70 secretion, either LPS or TNF-α or CD40 ligand (CD40L) were added to post-G DC-like cells. The stimuli that we applied neither enhanced IL-12p70 release compared with the levels measured above (data not shown) nor induced maturation-associated changes in cellular phenotype (Fig. 3B).

Figure 3.

 Functional features of post-G DC-like cells. (A) DC-like cells were activated for 48 h with LPS at 1 μg/ml. IL-12p70 release was measured by ELISA. Neutralizing Ab against IFN-α (§) and IL-10 (°) or control irrelevant isotype-matched Ab were added at the start of the cultures, each at 10 μg/ml. Mean and SD; No. of experiments =4. *p<0.05 compared with post-G DC-like cells differentiated in the absence of neutralizing Ab. **p<0.05 compared with DC differentiated in the presence of GM-CSF and IL-4. (B) Post-G DC-like cells were activated for 48 h with 1 μg/ml LPS and than analyzed by flow cytometry. Ag expression was expressed in terms of Mean Fluorescence Intensity Ratio (MFI of the positive cell population divided by the MFI of the negative cell population). Mean and SD; No. of experiments =4.

2.4 Post-G DC-like cells induce Ag-specific hyporesponsiveness in naive allogeneic CD4+ T cells

DC can be functionally defined by the ability to stimulate primary T cell responses 10. Compared with the other culture conditions established here, and similar to DC differentiated with exogenous IFN-α and IL-10, post-G DC-like cells induced low levels of BrdUrd incorporation by responder CD4+ T cells (Fig. 4A).

To ascertain whether post-G DC-like cells induced Ag-specific hyporesponsiveness in alloreactive T cells, we used a two-step culture system, in which allogeneic T cells were initially activated with either immunogenic GM-CSF/IL-4DC or with putative regulatory post-G DC-like cells and then restimulated in a second co-culture using immunogenic GM-CSF/IL-4DC from the same donor. As expected, CD4+ T cells co-cultured with pre-G DC in the primary MLR responded robustly to rechallenge with immunogenic, same-donor GM-CSF/IL-4DC (Fig. 4B). Conversely, CD4+ T cells primed with post-G DC-like cells were hyporesponsive to restimulation with immunogenic GM-CSF/IL-4DC from the same donor, but not with immunogenic GM-CSF/IL-4DC from a third-party donor (Fig. 4B). Notably, CD4+ T cells challenged with post-G DC-like cells during the primary MLR maintained lower levels of proliferation compared with CD4+ T cells activated with pre-G DC-like cells, even after costimulation with exogenous IL-2 (Fig. 4C).

As depicted in Fig. 4D, naive CD4+ T cells stimulated with post-G DC-like cells or with IFN-α/IL-10DC expressed CD25 Ag at lower levels compared with CD4+ T cells primed with immunostimulatory GM-CSF/IL-4DC or GM-CSF/IFN-αDC. In sharp contrast, pre-G DC-differentiated CD4+ T cells acquired a robust activation phenotype (more than 65% CD4+CD25+ T cells) that correlated with their preserved ability to proliferate in response to allo-Ag. The lower levels of CD25 on CD4+ T cells activated with post-G DC-like cells supported the view that consumption of IL-2 by responder CD4+ T cells was not primarily responsible for suppression of proliferation. Additional comparative analyses of Ag expression on developing T cells indicated that CD4+ T cells co-cultured with post-G DC-like cells expressed lower levels of CD62L compared with pre-G DC-primed CD4+ T cells (Fig. 4D), thus resembling ex vivo-activated post-G T cells 9. Of potential interest, CD4+ T cells challenged with post-G DC-like cells expressed detectable levels of intracellular CTLA-4/CD152 (Fig. 4D), an Ag that has been associated with the acquisition of regulatory features by T cells 17. The phenotypic profile of naive umbilical cord blood (UCB) CD4+ T cells challenged with DC differentiated under the other culture conditions established here is illustrated in Fig. 4D.

Figure 4.

 T cell allostimulatory capacity and induction of Ag-specific hyporesponsiveness by post-G DC-like cells. (A) DC-like cells differentiated as detailed in section 4 were further matured for 24 h with TNF-α and plated in allogeneic MLR. Mean and SD; No. of experiments =10. *p<0.05 compared with CD4+ T cells primed with pre-G DC-like cells and with GM-CSF/IL-4DC or GM-CSF/IFN-αDC. (B) Pre-G DC-primed CD4+ T cells (open circles) and post-G DC-primed CD4+ T cells (filled circles) were rechallenged with same-donor or with third-party-donor immunogenic GM-CSF/IL-4DC for 7 days. Mean and SD; No. of experiments =6. *p<0.01 compared with CD4+ T cells primed with pre-G DC during the primary MLR and restimulated with same-donor GM-CSF/IL-4DC. §p<0.01 compared with CD4+ T cells primed with post-G DC during the primary MLR and restimulated with third-party-donor GM-CSF/IL-4DC. (C) Effect of exogenous IL-2 at 50 IU/ml on the proliferation of CD4+ T cells recovered from the primary MLR. Mean and SD; No. of experiments =3. (D) Phenotypic features of T cells primed with DC in the primary MLR. Mean and SD; No. of experiments =6. *p<0.05 compared with pre-G DC-primed CD4+ T cells. §p<0.05 compared with GM-CSF/IL-4DC-primed and GM-CSF/IFN-αDC-primed CD4+ T cells.

2.5 IFN-α and IL-10 contained in post-G serum mediate induction of post-G DC-like cells with regulatory properties

We reasoned that a cytokine(s) present in post-G serum rendered post-G DC-like cells incapable of optimal T cell priming. In line with this hypothesis, post-G DC-like cells differentiated with neutralizing Ab against IFN-α and IL-10 recovered proliferation of allogeneic CD4+ T cells. As suggested by the data in Fig. 5, IFN-α and IL-10 contained in post-G serum acted synergistically, since the restoration of T cell responses effected by either Ab alone was lower than that mediated by the Ab in combination. Also, neutralization of IL-10 was more effective at restoring T cell proliferation than neutralization of IFN-α alone. As expected, pre-G DC-like cells efficiently primed responder CD4+ T cells, and the addition of neutralizing Ab to IFN-α and IL-10 to the cultures did not significantly affect T cell proliferation (Fig. 5).

Figure 5.

 Post-G DC-like cell regulatory function is mediated by IFN-α and IL-10. Post-G DC-like cells were differentiated in the presence or in the absence of neutralizing Ab to IFN-α and/or IL-10 at 10 μg/ml, matured with TNF-α for 24 h and plated in allogeneic MLR. (A) Proliferative response of CD4+ T cells to pre-G DC-like cells differentiated in the absence or in the presence of anti-IFN-α and anti-IL-10 Ab. Mean and SD; No. of experiments =6. (B) Proliferative response of CD4+ T cells to post-G DC-like cells differentiated in the absence or in the presence of anti-IFN-α and anti-IL-10 Ab. Mean and SD; No. of experiments =6. °p<0.05 compared with CD4+ T cells primed with post-G DC-like cells differentiated in the absence of anti-IFN-α Ab. §p<0.05 compared with CD4+ T cells primed with post-G DC-like cells differentiated in the absence of anti-IL-10 Ab and *p<0.05 compared with CD4+ T cells primed with post-G DC-like cells differentiated in the presence of anti-IFN-α Ab only. §p<0.05 compared with CD4+ T cells primed with post-G DC-like cells differentiated in the presence of both anti-IFN-α and anti-IL-10 Ab.

Figure 6.

 Cytokine secretion profile of CD4+ T cells primed with post-G rDC. The release of prototypic Th1, Th2, Th3 and TR1 cytokines in primary MLR was investigated with ELISA following 24 h activation with 0.1 μg/ml anti-CD3 mAb and 10 ng/ml PMA (both from Sigma Chemical Co.). *p<0.05 compared with MLR cultures containing pre-G DC-like cells; §p<0.05 compared with MLR cultures containing post-G rDC. Mean and SD; No. of experiments =6. GM-4 = GM-CSF + IL-4; GM-α = GM-CSF + IFN-α; 10-α = IL-10 + IFN-α.

2.6 Cytokine secretion profile and regulatory activity of T cells primed with post-G DC-like cells

Several reports suggest that T cells that acquire an anergic phenotype may exert regulatory and/or suppressive activity 18. We then sought to determine whether hyporesponsive CD4+ T cells induced by post-G regulatory DC (rDC) were endowed with TREG activity against nonregulatory CD4+ T cells. As shown in Fig. 6, CD4+ T cells primed with post-G rDC released higher amounts of TGF-β1, IL-10 and IL-5 compared with CD4+ T cells primed with pre-G DC-like cells or with immunogenic GM-CSF/IL-4DC or GM-CSF/IFN-αDC. Furthermore, CD4+ T cells challenged with post-G rDC secreted low amounts of IL-4 and IL-2, consistent with polarization toward a TR1-like profile. In accordance with the reduced ability of post-G DC to release IL-12p70 after in vitro activation with LPS (Fig. 3), lower levels of IL-12 were released by post-G DC compared with their pre-G counterparts upon interaction with T cells (Fig. 6).

When plated in co-culture, CD4+ T cells activated with post-G rDC suppressed by 60% on average the proliferation of nonregulatory CD4+ T cells from an HLA-mismatched, cytokine-untreated healthy volunteer. Interestingly, inhibition was still measurable in split-well cultures, indicating that the diffusion of soluble factors from tolerated CD4+ T cells was responsible for down-regulating the alloresponse of naive third-party CD4+ T cells (Fig. 7). Neutralization studies with anti-TGF-β1 and anti-IL-10 Ab demonstrated that suppression exerted by tolerated CD4+ T cells was dependent on TGF-β1 and IL-10 (Fig. 7). Our finding that CD4+ T cells challenged with post-G rDC differentiated in the presence of anti-IFN-α and anti-IL-10 Ab developed no regulatory activity against third-party CD4+ T cells (Fig. 7) confirmed that IFN-α and IL-10 contained in post-G serum were implicated in the development of TR1 cells through effects on DC.

Figure 7.

 Regulatory activity of naive CD4+ T cells primed with cytokine-differentiated DC. Naive CD4+ T cells were activated with DC differentiated as reported for Fig. 6 (primary MLR). Post-G DC-like cells were also differentiated in the presence of neutralizing anti-IFN-α and anti-IL-10 Ab, each at 10 μg/ml. Secondary MLR were performed as detailed in section 4. Neutralizing anti-TGF-β1 and anti-IL-10 Ab were both used at 10 μg/ml. *p<0.05 compared with the other culture conditions. Mean and SD; No. of experiments =4.

2.7 Freshly isolated post-G DC precursors induce T cell hyporesponsiveness

In a final set of experiments, DC precursors were purified from G-CSF-treated patients based on BDCA-4 expression and were plated in MLR with allogeneic CD4+ T cells (Fig. 8A). CD4+ T cells challenged with post-G DC precursors exhibited a lower proliferative response compared with CD4+ T cells activated with pre-G DC (Fig. 8B). Furthermore, addition of post-G serum to cultures containing DC precursors from G-CSF-untreated donors and allogeneic CD4+ T cells translated into reduction of T cell proliferation (Fig. 8C), indicating that, similar to monocyte-derived post-G DC-like cells, preformed post-G DC precursors were poor activators of an alloactivation response.

Figure 8.

 T cell allostimulatory capacity of DC precursors isolated from G-CSF-treated patients. (A) Circulating DC precursors were purified using a magnetic cell sorter (Miltenyi Biotec) 42. The frequency of BDCA-4+ events (left, unmanipulated sample) and the DC purity (right, sorted sample) from a representative experiment are shown (right). (B) Purified pre-G/post-G DC precursors were cultured for 7 days with CFSE-loaded allogeneic CD4+ T cells at a stimulator : responder ratio of 1:3. The dashed line discriminates between cells that underwent less than (right quadrant) or more than five cell divisions (left quadrant) in a representative experiment out of three with similar results. (C) Effect of pre-G/post-G serum supplementation (10% [v/v]) on the proliferative response of allogeneic CD4+ T cells to BDCA-4+ DC precursors from a cytokine-untreated healthy donor. The dashed line discriminates between cells that underwent less than (right quadrant) or more than five cell divisions (left quadrant) in a representative experiment out of three with similar results.

3 Discussion

The outcome of immunological responses, e.g. induction of immunity as opposed to promotion of tolerance, might not be dictated simply by the differential expression of costimulatory molecules on mDC or iDC, which implies that DC expressing high amounts of MHC class II and costimulatory molecules are not necessarily immunogenic 1922.

Our investigations were prompted by several considerations. First, G-CSF administration has been associated with mobilization of Th2-inducing DC 6. Second, we recently reported that G-CSF promotes the generation of TREG cells that suppress bystander T cells via the release of TGF-β1 and IL-10 9. These findings might have relevance in vivo, as suggested by the delayed recovery of anti-fungal immunity in patients treated with G-CSF after HLA haplotype-mismatched HSC transplantation 16. Third, the simultaneous generation of rDC and TREG cells for clinical use would be a desirable approach for the treatment of T cell-dependent immunopathologies 12. The novel finding of this study is that IFN-α and IL-10, that are markedly increased in the serum after G-CSF administration, promote the spontaneous differentiation of monocytes into rDC. In vitro studies indicated that bone marrow stromal cells (BMSC) might be one major source of IFN-α and IL-10 after the exposure to G-CSF, and provide further insight into the BMSC-mediated suppression of immunological reactivity 23.

Although DC can be spontaneously differentiated from PBMC in the absence of exogenous cytokines 24, to the best of our knowledge, this is the first report to demonstrate that monocytes can spontaneously differentiate into functional DC in the presence of autologous serum and that the DC populations obtained under these culture conditions profoundly affect the outcome of an immunological response. IL-10 and IFN-α provided in combination to post-G monocytes by virtue of post-G serum supplementation demonstrated a previously undescribed ability to polarize monocyte differentiation and favor the generation of rDC. These findings accord with prior research emphasizing that the functional activity of DC can be modulated by cytokines or soluble factors contained at high levels in human serum 25, 26.

Phenotypic changes in post-G rDC were specifically induced by mediators contained in post-G serum and were not evident in DC differentiated after culture with pre-G serum or with conventional cytokine combinations. Of interest, post-G rDC maintained high expression of the monocyte/macrophage Ag CD14, but the functional significance of this finding, if any, remains to be elucidated. Although the cytokine-modified rDC characterized to date reportedly express low to undetectable levels of CD14 27, 28, it must be emphasized that these DC populations were generated in vitro in the presence of IL-4, a cytokine that promotes the down-regulation of CD14 Ag on developing DC. Notably, DC subsets with preserved CD14 expression and low T cell stimulatory potential have been identified in vivo within dermal resident DC 22.

As anticipated from the inhibitory effects of IFN-α and IL-10 on IL-12 production 29, post-G rDC were profoundly impaired in the ability to release IL-12p70. Signals delivered through TNF-α, LPS or CD40L failed to stimulate IL-12p70 secretion and to induce phenotypic changes in post-G rDC, indicating that they were maturation resistant. It is tempting to speculate that self Ag-loaded rDC generated from G-CSF-treated individuals might remain insensitive for maturation signals when transferred in vivo, where they might exert potent regulatory activity and efficiently dampen harmful immunological reactions. Collectively, post-G rDC shared functional properties with semi-mature DC 30, a novel subset or developmental stage of DC that can be distinguished as mature by surface marker analysis but lacks significant secretion of proinflammatory cytokines.

At variance with rDC subsets endowed with an immature phenotype, post-G rDC guided the in vitro generation of anergic T cells, acting as semi-mature DC. The Ag-specific hyporesponsiveness was not reverted by exogenous IL-2, as already reported for the anergic state of IL-10-treated T cells but at variance with anergy induced by costimulation blockade 31. Hyporesponsive CD4+ T cells primed with post-G rDC acquired a TR1-like cytokine secretion profile as well as suppressive features against third-party conventional CD4+ T cells,which consisted in cell contact-independent but cytokine-dependent inhibition of T cell proliferation. CD4+ T cells activated with post-G rDC expressed low amounts of CD25, suggesting thatregulatory activity was not associated with a "professional" TREG cell subset but was rather induced de novo. Furthermore, TGF-β1 and IL-10 released by CD4+ T cells primed with post-G rDC were not dispensable for the induction of suppression, indicating that in vitro-differentiated CD4+ TREG cells were closely related to TR1 cells 9, 31.

The generation of rDC with the intent to polarize TREG cell differentiation offers therapeutic perspectives in autoimmune disorders. Intriguingly, G-CSF confers durable protection from experimental allergic encephalomyelitis (EAE) and might be clinically useful in human multiple sclerosis by favoring cytokine-based immune deviation 32. Similarly, myelin basic protein (MBP)-pulsed, in vitro-generated tolerogenic APC can promote the development of CD8+ TREG cells that convey peripheral tolerance and suppress autoimmunity toward MBP when adoptively transferred to recipient mice 33.

A further implication of our findings pertains to human graft-vs.-host disease (GVHD). Naturally occurring TREG cells contained in the allogeneic HSC graft have been shown tomodulate GVH reactivity; accordingly, adoptive transfer of freshly isolated or ex vivo-expanded donor TREG cells significantly delays GVHD and prolongs survival of mice 34, 35. Finally, in vitro-generated rDC reportedly prevent acute GVHD and leukemia relapse in a murine model of allogeneic bone marrow transplantation 36.

In conclusion, a plausible in vivo scenario implies that normally, quiescent monocytes exposed to increased levels of IFN-α and IL-10 as a consequence of G-CSF treatment might behave as rDC and dampen auto(allo)-reactive T cell responses. Preparation of large quantities of rDC from peripheral monocytes should be feasible under GMP conditions in G-CSF-treated subjects 37. Future work will specifically focus on determining whether post-G rDC can realistically be pursued for immunotherapy protocols 38.

4 Materials and methods

4.1 Characteristics of HSC donors and cancer patients

Six healthy volunteers (three men, three women, median age 36 years) and six patients affected by stage IV ovarian cancer received 10 μg/kg per day G-CSF (Granulokine, Amgen, Milan, Italy)subcutaneously for 4 days to mobilize HSC for allogeneic or autologous HSC transplantation 9. Donors and patients gave written informed consent, and the investigations were approved by the Institutional Human Research Committee. Peripheral blood was obtained prior to G-CSF administration (pre-G; day 0) and on the day of HSC leukapheretic collection (post-G; day 4).

4.2 Generation of human BMSC

Bone marrow obtained from healthy individuals by aspiration from the iliac crest under general anesthesia was collected into 10% Normosol R (Abbott Laboratories, North Chicago, IL) supplemented with 10 IU/ml preservative-free heparin. Ficoll-separated mononuclear cells (MNC) were resuspended at 1×106 cells/ml in α-MEM (Gibco BRL, Gaithersburg, MD) supplemented with 10% horse serum, 10% bovine serum and 10–6M hydrocortisone. They were inoculated into tissue culture flasks and incubated at 37°C in a humidified atmosphere supplemented with 5% CO2. Culture medium was replaced on a weekly basis and confluent stromal layers were trypsinized for 4 consecutive weeks.

4.3 Generation of DC from peripheral monocytes

PBMC were separated by the Ficoll-Hypaque density gradient method, as reported 9. CD14+ monocytes were purified by positive selection (Monocyte Isolation Kit, Miltenyi Biotec, Bergisch Gladbach, Germany) and were cultured for 4 days at 37°C under serum-free conditions (10% BIT HCC-9500; StemCell Technologies, Vancouver, BC) or in the presence of autologous pre-G/post-G serum (10% [v/v] final concentration) 39 or one of the following cytokine combinations: 800 IU/ml recombinant human GM-CSF and 500 IU/ml IL-4 40; 800 IU/ml GM-CSF and 1,000 IU/ml IFN-α 41; or 1,000 IU/ml IFN-α and 10 ng/ml IL-10 (all from R&D Systems, Oxon, Cambridge, GB). iDC were incubated for an additional 48 h with 500 IU/ml TNF-α (R&D Systems) or with 10 μg/ml LPS (Sigma Chemical Co., St. Louis, MO) or with 1 μg/ml soluble trimeric CD40L (Immunex, Seattle, WA). In selected experiments, neutralizing anti-IL-10 (R&D Systems) and/or anti-IFN-α Ab (Bender MedSystems, Wien, Austria) were provided at 10 μg/ml to pre-G and post-G serum-supplemented DC cultures.

4.4 T cell isolation and stimulation

CD4+ T cells were purified from UCB samples by positive selection (CD4 MACS MultiSort beads; Miltenyi Biotec) 9. To explore the direct pathway of allorecognition, CD4+ T cells were activated with the mixed leukocyte reaction (MLR). Briefly, graded doses of irradiated (25 Gy) DC were cultured with 5×104 allogeneic CD4+ T cells for 7 days. For secondary MLR, CD4+ T cells recovered from the primary MLR were rechallenged with the original stimulator cells that were cryopreserved at the start of the experiment.

4.5 Immunological markers

Freshly isolated or cultured cells were incubated for 20 min at 4°C with pre-titrated amounts of the following FITC-, PE-, PerCP- or PE-Cyanine 5 (TRICOLOR, TC)-conjugated mAb: CD1a, CD4, CD8,CD11c, CD14, CD25, CD40, CD40L (CD154), CD45RB, CD62L, CD80, CD86, CD83 (Caltag Laboratories, Burlingame, CA), CD123, CCR6, CD28, HLA-DR (Becton Dickinson, Mountain View, CA), CTLA-4 (Serotec Ltd, Oxford, GB), BDCA-4, BDCA-3 (both from Miltenyi Biotec), or with fluorochrome-conjugated isotype-matched irrelevant mAb to establish background fluorescence. Cells stained with the primary anti-CTLA-4 mAb were further incubated with FITC-conjugated goat anti-mouse mAb (Caltag Laboratories) prior to flow cytometric analysis.

4.6 Split-well cultures and assays for TREG function

Co-culture experiments for the assessment of regulatory properties of DC-primed CD4+ T cells were performed with transwell systems (MilliCell inserts, 0.4 μM, Millipore Ltd., Watford, GB). Both chambers of each transwell received irradiated allogeneic monocytes as stimulator cells 9. Briefly, the proliferation of nonregulatory CD4+ T cells plated in the lower chamber of the transwell was monitored in the absence of direct contact with 5×105 DC-primed CD4+ T cells which were placed in the upper compartment (primary co-cultures). In selected co-culture experiments, neutralizing anti-TGF-β1 (20 ng/ml) and/or anti-IL-10 Ab (10 μg/ml; both from R&D Systems) were added as indicated in the figure legends. After 7 days, the basket was removed, and proliferation of nonregulatory CD4+ T cells was measured as detailed below.

4.7 Analysis of cytokine production

IL-2, IL-4, IL-5, IL-10, IL-12p70, TNF-α, IFN-α and TGF-β1 levels were quantified by ELISA, using commercially available kits (R&D Systems, Bender MedSystems) 9.The limits of detection were as follows: 1 pg/ml IL-10 (Ultra Quantikine HS); 7 pg/ml TGF-β1; 3.2 pg/ml IL-12p70; 4.8 pg/ml IFN-α; and 10 pg/ml IL-2, IL-4 and IL-5.

4.8 Proliferation assays

During the last 24 h of each MLR, cells were pulsed with BrdUrd (25 μM; Sigma Chemical Co.). The incorporation of BrdUrd into actively proliferating cells was monitored by flow cytometry using an anti-BrdUrd mAb (clone BR-3; Caltag Laboratories), as already detailed 9.

4.9 Cell tracking experiments

Freshly isolated, unstimulated CD4+ T cells were labeled with 5-carboxyfluorescein diacetate succinimidyl ester (CFSE) (2.5 μM; Molecular Probes, Eugene, OR) 42. After washings in PBS supplemented with 3% FCS, cells were used for MLR experiments as above detailed. CFSE fluorescence profiles were analyzed by flow cytometry using the Proliferation Wizard of the ModFit LT software (Verity House Inc., Topsham, ME).

4.10 Flow cytometry, immunofluorescence and data analysis

Samples were run through a FACScan flow cytometer (Becton Dickinson) with standard equipment 9. Ag expression was quantified using the CellQuest® software (Becton Dickinson) in terms of both percentage of positive cells and mean fluorescence intensity (MFI). The probability of significant differences between the distribution of test and control histograms was calculated with the Kolmogorov-Smirnov test. Details on instrument settings and data analysis were published elsewhere 9.

4.11 Statistical analysis

The approximation of data distribution to normality was preliminarily tested with statistics for kurtosis and symmetry. Results were presented as mean and SD. All comparisons were performed with the Student's t-test for paired or unpaired determinations or with the analysis of variance (ANOVA), as appropriate. The criterion for statistical significance was defined as p<0.05.


The authors are grateful to Dr. Francesco Fagnoni (Medical Oncology Division, Scientific Institute of Pavia, Fondazione Salvatore Maugeri, Pavia, Italy) forcritical reading of the manuscript and for helpful discussions, and to Dr. Annabella Procoli (Department of Gynecology, Catholic University, Rome, Italy) for technical assistance. The authors also acknowledge the kind contribution of Mr. Enrico Guadagni (Department of Human Anatomy, Catholic University Medical School, Rome, Italy) to image acquisition and editing. This study was supported by "Stem Cell Project", Fondazione Cassa di Risparmio di Roma, Rome, Italy; CNR-MIUR, Rome, Italy; "Progetto Finalizzato Oncologia", University of Bologna, Italy (Fondi Ex 60%).


  1. 1


  2. 2


  3. 3


  4. 4


  5. 5


  6. 6


  7. 7


  8. 8