Dendritic Cells in Myelodysplastic Syndromes

From Pathogenesis to Immunotherapy

Nathalie Kerkhoff; Hetty J Bontkes; Theresia M Westers; Tanja D de Gruijl; Shahram Kordasti; Arjan A van de Loosdrecht


Immunotherapy. 2013;5(6):621-637. 

In This Article

DCs as a Target for Immunotherapy in MDS

Bone marrow transplantation is the only potential cure for MDS. However, most patients develop the disease at a later age, with the median onset in the seventh decade of life, and are often not eligible to undergo such an intensive therapy. To improve long-term survival and remission duration, therapeutic alternatives are needed. With the increased insight into immunological processes and disturbances that are involved in pathogenesis and progression of MDS, strategies that directly target these mechanisms are being developed. These new immunotherapeutic approaches interfere with and modulate the immune system in order to generate either a tolerogenic response (in low-risk disease) or a strong immunological reaction (in high-risk disease). Modified DCs have been used for immunotherapy of different types of tumors with variable outcomes, often depending on the disease stage.[96–100] However, as MDS is a myeloid disease and perhaps influence the function and maturation of DCs due to possible clonal involvement, the cell compartment is still an attractive target in this disease. Restoring DC function by immunotherapeutic agents may be beneficial and could normalize immune functions. The proposed mechanisms of different immunotherapeutic and immunomodulatory agents are outlined below, with the main focus on the role of DCs as targets in treatment regimens for low- and high-risk MDS.

Epigenetic-modifying Drugs

Epigenetic modifications, such as DNA methylation, are often involved in cancer.[101,102] DNA methylation has a critical function in regulating transcriptional activity. Methylation of CpG islands within a promoter, catalyzed by DNA methyltransferase, causes silencing of that specific gene. In addition, in different hematological malignancies, including MDS, it is known that a variety of epigenetic changes can contribute to the outgrowth of the malignant clone.[101] In MDS, several tumor suppressor genes are often hypermethylated and important control mechanisms, such as DNA repair, cell-cycle control and apoptosis, are affected.[103] Methylation patterns have a high impact and can even predict prognosis, survival and response to therapy in MDS patients.[4,104] Gene inactivation by methylation can be reversed with targeted therapies. Hypomethylating agents, such as azacitidine and decitabine, inhibit the enzyme methyltransferase by direct incorporation into DNA and induce re-expression of several genes.[105] As a therapy for high-risk MDS patients it has been shown to recover hematopoiesis, reduce marrow blast percentages, normalize karyotypes, improve quality of life and increase overall survival compared with conventional therapies.[106–108] These agents have also been shown to possess immunoregulatory capacities. For example, different genes that play a role in cytokine production, T-cell polarization and Treg formation are regulated by epigenetic mechanisms.[109,110] It has been demonstrated that azacitidine demethylates the Foxp3 promoter and induces overexpression of Foxp3.[111,112] This was confirmed by Bontkes et al. who showed in vitro expansion of functional Tregs after exposure to azacitidine.[113] Despite increased Foxp3 expression, these Tregs lack the regulatory function and cytokine profile of Tregs. On the contrary, a 'Treg-like' subset of cells was identified, which expresses increased levels of Foxp3 and, at the same time, secretes high levels of the proinflammatory cytokine IL-17.[111] The different effects of azacitidine on the immune system can be beneficial in high-risk MDS. As was shown previously, Treg frequencies are increased in high-risk MDS patients.[13] Azacitidine may reduce the number of functional Tregs and induce Th17-like responses, which could help eliminate the tolerogenic environment in high-risk disease.

The impact of azacitidine on DC function has also been studied recently.[114] Phenotypically, DCs treated with azacitidine display an increased upregulation of maturation and costimulatory molecules (CD40 and CD86). Moreover, these azacitidine-treated DCs secrete lower levels of IL-10 and IL-27. Furthermore, it can be hypothesized that treated DCs also affect T-cell polarization differently compared with nontreated DCs. In patients receiving azacitidine treatment, the frequency of IL-4-producing CD4+ T cells is reduced, whereas the IL-17A- and IL-21-secreting cells are increased. This suggests a Th17-like response in the peripheral blood of patients treated with azacitidine and may indicate a long-term T-cell activation as a possible mechanism of action (Figure 2A).[114] Another way in which DCs could be targeted and modulated by hypomethylating agents is via the re-expression of the tumor suppressor gene, p15Ink4b. This gene usually controls cell-cycle progression through the G1 phase and is involved in normal myelopoiesis.[115,116] Silencing of p15Ink4b, by hypermethylation of its promoter, has been found in 50% of MDS patients and correlates with leukemic transformation and a poor outcome.[4,117,118] This gene also participates in DC development as it is strongly induced during differentiation and activation of DCs. In p15Ink4b-knockout mice, conventional DCs show impaired expression of MHC II and costimulatory molecules.[119] Furthermore, these DCs are less efficient in taking up antigens and stimulating T cells. After re-expression of p15Ink4b, all defects are restored, indicating a crucial role for this gene in DC function. Additionally, reactivation of the gene elevates phosphorylation of Erk1/Erk2 protein kinases and subsequently increases the activity of the PU.1 transcription factor. This transcription factor is highly involved in DC development and differentiation. It was demonstrated that PU.1 expression is often reduced in MDS patients and that hypomethylating agents could upregulate transcription with myeloid differentiation as a result.[120] The demethylation of p15Ink4b was also induced by decitabine treatment in responding MDS patients and p15 expression was upregulated to normal levels.[121] Therefore, hypomethylating agents could also indirectly contribute to restored DC development and function in MDS patients and provide an alternative mechanism of modulating the immune response.

A second epigenetic process that is frequently involved in MDS is chromatin modification, such as histone deacetylation. Usually, DNA is tightly bound around histones and is hardly accessible for transcription. Histone modification, for example, by acetylation, will open the chromatin and have a direct effect on transcriptional activation. Histone deacetylases (HDACs) control the process of acetylation and deacetylation and these enzymes can be blocked by HDAC inhibitors, which are used as treatment in MDS patients.[122] It has been shown that the use of HDAC inhibitors in combination with hypomethylating agents has synergistic effects and results in more robust re-expression of hypermethylated genes.[123] HDACs are also involved in normal myeloid cell differentiation as was shown by Youn et al..[124] In this study, histone deacetylases (HDAC-2) play a significant role in pathologic differentiation of myeloid cells in tumor-bearing mice by silencing essential genes that normally regulate the induction of myeloid differentiation. This results in the accumulation of immature myeloid cells with immunosuppressive features, termed myeloid-derived suppressor cells. Blocking HDAC-2 can serve as a selective therapeutic target to induce further myeloid differentiation into functional myeloid cells, such as DCs.

Besides its role in epigenetic regulation, HDAC inhibitors also exhibit immunomodulatory functions, and DCs can be directly targeted with these agents. MoDCs that were treated with the HDAC inhibitor valproic acid display downregulation of several maturation markers and MHC molecules. They are also functionally impaired since the secretion of IL-10 and IL-12p70 is inhibited and the proportion of IFN-γ-producing CD4+ T cells in a MLR is reduced.[125] This potential regulatory role of DCs exposed to HDAC inhibitors is also confirmed by other groups.[126,127] Treatment with LBH589, another HDAC inhibitor, also leads to the lower expression of markers associated with DC maturation, antigen presentation and T-cell costimulation. Furthermore, decreased antigen uptake ability was found and these myeloid DCs are less effective in T-cell activation and cytokine secretion. Impairment of DC migration to the chemoattractant, CCL19, was observed after the DCs were exposed to the HDAC inhibitors MS-275 and valproic acid. Brogdon et al. investigated the effects of LAQ824, a HDAC inhibitor, on DC function by genome-wide gene-expression analysis.[128] They found that different genes are targeted by this inhibitor, which are involved in stimulation and chemotaxis of Th1 lymphocytes. The Th2 axis is not affected. By contrast, the immunological effects of HDAC inhibitors on DCs derived from AML cell lines showed significantly improved differentiation, enhanced expression of HLA-DR and CD83, and high allostimulatory capacities.[129] The exact mechanism underlying this discrepancy is not clear. As mentioned earlier, HDAC inhibitors might induce differentiation pathways in a tumorigenic environment, which results in restored DC function. However, the effect of HDAC inhibitors on DC function in MDS patients and whether it induces tolerogenic DCs or differentiation to mature inflammatory DCs, is still yet to be investigated. Clearly, the effect of HDAC inhibitors on DC function should be considered in their therapeutic application. In circumstances where tolerogenic DCs are preferred, which could be the case in low-risk MDS, HDAC inhibitors might be suitable manipulators to target DC function. On the other hand, in combination with vaccination strategies where MDS DCs are used, pretreatment of these DCs with HDAC inhibitors could yield DCs with stimulatory capacities. This effect would be highly wanted in the setting of high-risk disease.

Immunomodulatory Drugs

For low-risk MDS patients with del(5q) lenalidomide, a derivate of thalidomide, is US FDA approved. Treatment with this immunomodulatory drug (IMiD) is suggested to directly target the del(5q) clone, which results in prolonged transfusion independency and reversion of cytogenetic and cytologic abnormalities.[130,131] In addition, limited studies in high-risk MDS and AML patients with 5q aberrancies showed a potential role for lenalidomide.[132] Recently, Sibon et al. demonstrated that non-del(5q) patients also benefit from lenalidomide therapy in the setting of second-line treatment in patients refractory to erythropoiesis-stimulating agents.[133] Beside its direct effect on the del(5q) clone, an important part of the working mechanism of IMiDs is attributed to their function in immunomodulation. Previously, it was thought that the main function of these drugs relies on anti-TNF-α activity and that the effect was primarily tolerogenic. However, IMiDs also display inflammatory features that can be induced in order to modify immune reactions. Besides their antiangiogenic role, they trigger CD28 phosphorylation, which enhances costimulatory activity.[134,135] Furthermore, Treg function is inhibited due to decreased expression of Foxp3.[136] The effect of lenalidomide on DC function was also investigated.[137] DCs treated with lenalidomide take up antigens more efficiently and express higher levels of the maturation marker CD86. In cross-priming assays, CD8+ T-cell proliferation is increased after priming with treated DCs and, simultaneously, the intracellular IFN-γ production by these T cells is augmented (Figure 2A). These findings suggest a potential role for IMiDs in DC-targeted therapies, including vaccination strategies. Regarding the effect in different risk groups of MDS, IMiDs might also be attractive in high-risk patients as these agents induce enhanced cross-priming and could thereby generate tumor-specific CTL responses. Additionally, combinations with other IMiDs could be valuable in the treatment of various MDS risk groups. Different clinical trials that combine lenalidomide with azacitidine, for example, have already been performed and show encouraging outcomes.[138]

DC Vaccination Strategies

Many tumors display immune evasion features that manipulate DC function and prevent adequate induction of CTL responses.[84] For instance, tumor cells and their environment interfere with DC maturation via secretion of anti-inflammatory cytokines (e.g., IL-10).[139] In addition, the generation of tolerogenic DCs through the interaction with stromal cells has been previously described.[140] Furthermore, DCs can directly promote survival and angiogenesis in different tumors.[141] This altered DC behavior in tumors make them ideal candidates for immune-targeted therapies in order to restore their function and favor antitumor immunity (Figure 2B). One of the strategies to create a specific antitumor response via DCs is provided by DC vaccination.[142] In several tumor types, including hematological malignancies, DC vaccination has been shown to induce tumor-specific T-cell responses.[143] To effectively generate such vaccines, a variety of conditions influence the potential antitumor capacities. These include type of DC used, maturation state, antigen-loading strategies, source of antigens, administration routes and adjuvants that are able to make these vaccines much more effective.[143] For instance, MoDCs can be loaded ex vivo with TAAs to induce tumor-specific immune responses. Many sources of antigens are available to load these DCs, such as peptides derived from known TAAs, tumor lysates or apoptotic tumor cells.[144] Important disadvantages of using peptides as a source of TAAs are the restricted range of antigens, the variable expression of a single TAA between patients or even within one tumor clone over time and selection of the best TAA present. In AML, for example, several antigens have been identified that show variable expression rates between patients.[145,146] By contrast, tumor lysates or apoptotic tumor cells provide a broad range of known and unknown TAAs and selection is not required. However, only a small fraction of the entire antigen content is tumor-specific, while the majority consists of normal tissue antigens, which may induce autoimmunity. In the hematological field, DC vaccination research and clinical trials have been performed, for instance in a minimal residual disease (MRD) setting in AML patients.[96,147,148] The presence of MRD after intensive therapy can predict the chance of relapse.[149] With DC vaccination strategies, attempts are made to eliminate these MRD cells in order to prevent relapse. In this perspective, an alternative source of DCs has been investigated. It is suggested that DCs could be generated from clonal AML blasts with the advantage that these cells already harbor the full repertoire of known and unknown leukemia-associated antigens.[150,151] These DCs exhibit migratory capacities and induce CD8+ T-cell responses.[152,153] One of the current ongoing trials is a Phase I study that investigates the feasibility of DCOne™ (DCPrime BV, Leiden, The Netherlands), an allogeneic AML-derived DC vaccine, in patients with AML after having achieved complete remission or in patients having stable disease.[201] Another study that is being performed examines the tolerance of a combination strategy, including decitabine and DC vaccination, for patients with relapsed AML.[202] It has been shown previously that hypomethylating agents, such as decitabine, enhance expression of TAAs, which may have beneficial effects on treatment outcome.[154]

An alternative option to create antitumor immunity via vaccination strategies is offered by the use of primary AML cells or, in the context of MDS, primary MDS cells. These cells, themselves, are able to induce stimulating responses against a broad collection of leukemic antigens in ideal circumstances as they express high levels of MHC I and II molecules and adhesion molecules, and they often express the costimulatory molecule CD86. From this perspective, these AML cells could act as 'professional' APCs to induce cytotoxic activity. However, they often lack the crucial costimulatory molecule CD80 (B7.1), express high levels of CLIP and produce immunosuppressive cytokines, which make them low in immunogenicity.[155,156] The group of Mufti and Farzaneh has therefore generated AML cells that are potent immune stimulators by transduction with a self-inactivating lentivirus encoding CD80 and IL-2.[157] They demonstrated that stimulation of effector cells with transduced AML cells increased their cytotoxic activity and production of IFN-γ. This response also showed high specificity for leukemic cells.[158,159] A Phase I study with these cells is currently ongoing in AML patients.[203]

Until now, little clinical experience regarding vaccination treatments in MDS patients has been achieved. However, it was demonstrated that MDS-derived DCs could successfully be generated under serum-free conditions.[160,161] These DCs are able to prime cytotoxic T cells directed against the dysplastic clone, suggesting a new immunotherapeutic option for MDS patients.[162]