Generation of myeloid-derived suppressor cells using prostaglandin E2

Myeloid-derived suppressor cells (MDSCs) are natural immunosuppressive cells and endogenous inhibitors of the immune system. We describe a simple and clinically compatible method of generating large numbers of MDSCs using the cultures of peripheral blood-isolated monocytes supplemented with prostaglandin E2 (PGE2). We observed that PGE2 induces endogenous cyclooxygenase (COX)2 expression in cultured monocytes, blocking their differentiation into CD1a+ dendritic cells (DCs) and inducing the expression of indoleamine 2,3-dioxygenase 1, IL-4Rα, nitric oxide synthase 2 and IL-10 - typical MDSC-associated suppressive factors. The establishment of a positive feedback loop between PGE2 and COX2, the key regulator of PGE2 synthesis, is both necessary and sufficient to promote the development of CD1a+ DCs to CD14+CD33+CD34+ monocytic MDSCs in granulocyte macrophage colony stimulating factor/IL-4-supplemented monocyte cultures, their stability, production of multiple immunosuppressive mediators and cytotoxic T lymphocyte-suppressive function. In addition to PGE2, selective E-prostanoid receptor (EP)2- and EP4-agonists, but not EP3/1 agonists, also induce the MDSCs development, suggesting that other activators of the EP2/4- and EP2/4-driven signaling pathway (adenylate cyclase/cAMP/PKA/CREB) may be used to promote the development of suppressive cells. Our observations provide a simple method for generating large numbers of MDSCs for the immunotherapy of autoimmune diseases, chronic inflammatory disorders and transplant rejection.


Biology of myeloid-derived suppressor cells
Dendritic cells (DCs) are key initiators and regulators of immune responses [1][2][3]. Therapeutic programming of DCs to suppress their function has been shown beneficial in autoimmunity and transplantation [4][5][6]. In contrast to DCs, suppressive macrophages [7] and myeloid-derived suppressor cells (MDSCs), originally shown to accumulate at the site of tumors, suppress the ability of CD8 + T cells to mediate effective responses against cancer cells, but can be beneficial in controlling autoimmune phenomena or transplant rejection [8][9][10].
The presence of prostaglandin E 2 (PGE 2 ) at early stages of DC development was shown to suppresses the differentiation of human monocytes into functional T helper (Th)1-inducing CD1a + DCs [29]. Additionally, PGE 2 is needed for the development of tumor-associated suppressive macrophages [30][31][32]. Our two recent reports [33,34] demonstrate that PGE 2 is both required and sufficient to redirect the differentiation of human dendritic cells into monocytic MDSCs. It also mediates the induction of MDSC-associated suppressive factors in human MDSCs [21] in a mechanism involving the establishment of a positive feedback loop between PGE 2 and cyclooxygenase (COX)-2 [33], the key regulator of PGE 2 production [35]. Additionally, PGE 2 has been shown to enhance the numbers of MDSCs in mouse models and induce their expansion ex vivo [36][37][38].

In vitro generation of myeloid-derived suppressor cells
Recent work in mice demonstrated that functional MDSCs can be generated in vitro from mouse embryonic stem cells and bone marrow hematopoietic stem cells, resulting in two subpopulations -CD115 + Ly-6C + (equivalent to the monocytic Gr-1 + CD115 + F4/80 + MDSCs found in tumorbearing mice) and CD115 + Ly-6Ccells (resembling the granulocyte/macrophage progenitors) [37,[39][40][41]. Adoptive transfer of these MDSCs prevented graft-versus-host disease mediated by alloreactive T cells. While granulocytic MDSCs may induce non-specific immune suppression and suppress the effector phase of the allogeneic immune response at an early stage, the monocytic MDSCs emerge as the key subset needed to promote T reg development and to establish long-term antigen-specific tolerance [37,[39][40][41]. Another source of MDSCs is the bone marrow, which harbors a large reservoir of MDSCs. Recent studies have demonstrated an efficient growth factor/cytokine (granulocyte macrophage colony stimulating factor (GM-CSF) + G-CSF or GM-CSF + IL-6 or IL-13)-induced expansion of MDSCs populations in vitro, utilizing bone marrow cells from either mice or human sources [42,43] to generate IL4Rα + MDSCs. In mice these cells were able to impair the priming of CD8 + T cells, and enabled long-term acceptance of pancreatic islet allografts [43]. Furthermore, bone marrow progenitor cells can be induced by lipopolysaccharide to develop into CD11b + Gr1 int F4/80 + cells that, when adoptively transferred, suppressed allergen-induced airway inflammation in recipient mice [44]. Due to the massive accumulation of MDSCs in the spleens of tumor-bearing mice, the spleen is considered to be a reservoir of MDSCs and their precursors [45]. The drawback of these reported initiatives to develop MDSC-based therapeutic strategies is the lack of a reliable source of MDSCs.
For human treatment regimens the control of MDSCs in vitro by manipulating recipient myelomonocytic precursor cells appears most applicable. While there are low frequency and total numbers of MDSCs in peripheral blood (approximately 5% of cells in healthy subjects), peripheral blood constitutes a very convenient source of myelomonocytic precursor cells for MDSC generation. Apart from the recently described cytokine regimens that showed the feasibility of in vitro expansion of bloodisolated MDSCs populations [46] the induction of human MDSCs has been proven a feasible in vitro approach for the generation of CD14 + HLADR neg/low MDSCs by differentiation of isolated CD14 + cells in the presence of IL-4 + GM-CSF and tumor-derived microvesicles [46]. Alternatively, functional MDSCs can be induced in peripheral blood mononuclear cell (PBMC) cultures supplemented with several cytokine induction combinations, produced by tumor cell lines [47].
Our current data provides evidence for the feasibility of generating large numbers of monocytic MDSCs for the immunotherapy of autoimmune and inflammatory diseases, or transplant rejection by using a single common determining factor -PGE 2 , a common inflammation-associated master regulator of immune responses -that can redirect the development of CD1a + DCs to CD14 + CD33 + CD34 + monocytic MDSCs [48].

Efficient generation of human myeloid-derived suppressor cells using prostaglandin E 2
The development of functional MDSCs requires the inhibition of development of immunostimulatory antigen presenting cells and concomitant induction of suppressive functions [8]. The expansion of iMCs can be induced by factors such as GM-CSF, IL-6, or vascular endothelial growth factor [24,[49][50][51]. The upregulation of MDSCassociated immunosuppressive factors and establishment of their immunosuppressive function can be induced by such factors as IL-1β, IFNγ, PGE 2 , or Toll-like receptor ligands [8]. While the above MDSC-activating factors have apparently diverse character and functions, they all share the ability to induce COX2 expression and PGE 2 production [52][53][54], suggesting the key role of COX2 and PGE 2 in MDSCs development.
The differentiation of monocytes into functional CD1a + DCs could be redirected into CD1a -CD14 + CD80 -CD83 -MDSCs by their exposure to PGE 2 only at early stages of DC development (that is, from day 0, PGE 2 d0 ) [29] but not at later time points (that is, at day 6, PGE 2 -conditioned DCs d6 ).
While the immunosuppressive phenotype of the PGE 2induced MDSCs proved to be PGE 2 concentrationdependent ( Figure 1C) [29], it was independent of the presence of IL-4, indicating a key role for PGE 2 , but not for IL-4, in inducing MDSCs.
Exposure to PGE 2 induced the expression of endogenous COX2 in differentiating monocytes, leading to the establishment of a PGE 2 -COX2-mediated positive feedback loop, and the induction of IDO1, NOS2, IL-10, or IL-4Rα -the typical MDSC-associated factors ( Figure 1C). PGE 2 -induced cells displayed a suppressive phenotype, marked by the expression of inhibitory molecules -inhibitory receptor Ig-like transcript (ILT)2, ILT3, ILT4 and programmed cell death 1 ligand 1 (previously implicated in the suppressive functions of myeloid cells [27,28]), produced the immunosuppressive factors IDO1, IL10 and PGE 2 and exerted suppressive functions, blocking the proliferation and development of CD8 + T cells into granzyme B (GrB) high cytotoxic T lymphocytes [33].

Therapeutic potential of ex vivo induced myeloidderived suppressor cells
Anti-inflammatory activity of MDSCs in a variety of physiological settings and their therapeutic promise in transplantation [57] suggest that these cells may provide a novel cell-based immunotherapy in transplantation [40,58] and autoimmune diseases [59].
While the spontaneously arising endogenous MDSCs present in many forms of autoimmune diseases appear to be defective and ineffective in controlling the disease (reviewed in [60]), it was shown that adoptive transfer of MDSCs can limit autoimmune pathology [61][62][63], providing a rationale for the development of methods to expand or induce MDSCs ex vivo.
Transfer of MDSCs can prevent graft-versus-host disease [42], and prolong the survival of allo-skin [64] and allo-kidney transplants [65], and play an essential role in an allogeneic cardiac transplantation model [57]. Adoptively transferred MDSCs, isolated from synegeic tumorbearing mice, can prevent the onset of type 1 diabetes in non-obese diabetic mice [63] and ameliorate the symptoms of inflammatory bowel disease [59]. In a mouse model of alopecia, adoptively transferred MDSCs have been shown to promote partial restoration of hair growth [62].
From the therapeutic standpoint, it is important to identify central regulatory pathways that maintain the suppressive functions of MDSCs mediated by different suppressive molecules (arginase 1 [42], ILT-2 [66], heme-oxygenase (HO-1) [64], and iNOS [65]). Our data [48,67] -showing that the exposure of differentiating monocytes to PGE 2 (and the establishment of a positive feedback between PGE 2 and COX2) is both required and sufficient for MDSC stability and their ability to produce all MDSC-associated suppressive mediators and suppress CD8 + T cell function [48] provides evidence for a feasible and clinically compatible method of generating suppressive cells for immunotherapeutic purposes.

Conclusions
Due to their ability to suppress T cell responses in multiple diseases [65,68,69], MDSCs represent a promising population of cells for use in tolerogenic therapies. Our recent observations demonstrating the feasibility of using PGE 2 to promote the development of MDSCs from monocytic precursors provide a clinically feasible system of generating large numbers of MDSCs ex vivo, facilitating the development of new therapies for autoimmune diseases and transplant rejection.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions NO and PK conceived the work and wrote the manuscript. Both authors read and approved the final manuscript.