Decitabine, a new star in epigenetic therapy: the clinical application and biological mechanism in solid tumors

Highlights

• The function and mechanism of DAC in solid tumor treatment.

• The clinical experience and biological activity of DAC therapy in solid tumors.

• The evaluation targets and potential biomarkers for the epigenetic therapy.

• DAC combined with other treatments would be promising in solid tumors.

Abstract

Epigenetic alterations are strongly associated with cancer development and drug resistance. The use of the DNA methylation inhibitor decitabine (Dacogen®) has been approved in the treatment of hematological malignancies, and its clinical effects on solid tumors have gained attention. Here, we present a review of the molecular regulation mechanisms, clinical experiences and biological evaluation for novel decitabine-based therapies in solid tumors. We also discuss the following questions: What is the best administration schedule of decitabine in solid tumors? Is there tumor type specificity for decitabine-based epigenetic therapy? What are the biological function and mechanism of decitabine in suppressing tumor development? Is there a correlation between DNA demethylation and clinical response? Importantly, low-dose decitabine and combined therapy show significant improvement in solid tumor treatment. However, the correlation studies are preliminary, and key biomarkers for prognosis need further investigation.

Introduction

Surgery, radiotherapy and chemotherapy are the mainstays of tumor therapy. Recently, cancer immunotherapy gains worldwide attention and is regarded as one of the most promising cancer treatments in 21st century. Besides targeting the checkpoint inhibitors, such as CTLA-1 and PD-1, modification of other immune system components would also be prospective.

DNA methylation is an epigenetic event that regulates chromatin compaction and repression of gene expression. In cancer cells, a variety of genes are abnormally silenced by DNA methylation, including tumor suppressors, and genes controlling the immune response and drug sensitivity [1], [2]. The inhibition of DNA methylation with cytidine analogues such as decitabine (5-aza-2′-deoxycytidine, Dacogen®, DAC for short) reactivates the expression of genes silenced by hypermethylation. DAC incorporates into replicating DNA and forms irreversible covalent bonds with the active sites of DNA methyltransferase [3].

Various animal model and cell line studies indicated that DAC induced the expression of genes that control cell apoptosis, cell cycle arrest, and cancer testis antigens (CTA), MHC molecules and co-stimulation molecules by DNA hypomethylation. As a result, DAC enhanced the anti-tumor immune response and inhibited tumor development [4], [5], [6]. Clinically, previous observations indicated that higher doses of DAC induce cytotoxicity, while low-dose, prolonged infusion with DAC correlates with clinical response in hematological disorders [7]. DAC has been used as an approved therapy for the treatment of hematological malignancies such as, myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML). The data indicate that approximately 50% of MDS patients demonstrate a clinical response after DAC treatment and that with myelosuppression is the major adverse effect [8].

Over the past two decades, besides treatment of hematological diseases, many clinical trials have focused on utilizing DAC in solid tumors as a lone agent or as part of a combination chemotherapy approach. The clinical effectiveness of the DAC-based therapy in patients with solid tumors has recently been demonstrated. The remarkable response and mild toxicity caused by DAC treatment in certain patients highlighted the excellent prospective use in cancer treatment. This review focuses on the biological mechanism and therapeutic effect of DAC treatment in solid tumors. We summarize the potential marker genes, methylation status, administration regimen and clinical response for this epigenetic therapy. Additionally, we discuss the correlation between gene demethylation and clinical response and provide the potential early targets for DAC-based epigenetic therapy.

Surgery, radiotherapy, chemotherapy, and targeted therapy are widely used as first-line cancer regimens. However, overall survival is still limited by drug resistance. Numerous DNA methylation alterations in the cancer methylome are related to drug resistance. For example, there are genes involved in DNA repair, apoptosis, cell cycle control and drug metabolism [9], [10], [11]. Pretreatment with low-dose DAC could function as a potential sensitizer of chemotherapeutic drugs via the epigenetic modulation of aberrant DNA methylation in cancers.

Hypermethylation and inactivation of the DNA mismatch repair gene MLH1 were observed in 25–35% of ovarian cancer patients following platinum-based chemotherapy. This result was related to a poor survival rate in ovarian cancer patients [12]. In vitro restoration of MLH1 could confer sensitivity of DNA damaging drugs and DAC treatment resensitized tumor cells to chemotherapeutic agents in vivo [13].

The methylation-mediated silence of pro-apoptotic genes such as caspase-8 occurs in multiple solid tumors and could be a key mechanism of resistance to chemotherapy. A study in neuroblastoma patients showed that caspase-8 promoter demethylation and gene expression were observed in 2/7 bone marrow samples following DAC treatment [14].

Membrane transporters play an important role in the uptake of the chemotherapeutic agents. Therefore, the reduced expression of these genes may be another critical factor causing drug resistance. For example, the ATP-binding cassette (ABC) drug transporters control tumor multidrug resistance [15], [16]. Strikingly, administration of DAC dramatically enhances the expression of transporter CTR1, ABCB1 and ABCG2 in solid tumors [17].

Zeller et al. detected 9 genes that acquired methylation in ovarian tumors at relapse after chemotherapy or in chemoresistant cell lines. These genes included the following: ARHGD1B, ARMCX2, COL1A, FLNA, FLNC, MEST, MLH1, NTS and PSMB9 [6].

The anti-epidermal growth factor receptor (EGFR) therapy is a common regimen for patients with wild-type (wt) KRAS metastatic colorectal cancer (mCRC). However, a substantial portion of these patients are also resistant to treatment and most responders will become resistant to anti-EGFR therapy probably by EGFR hypermethylation and silencing.

Scartozzi et al. observed a higher response rate to anti-EGFR therapy in the absence of EGFR promoter methylation [18]. DAC induces EGFR expression and restores the sensitivity to EGFR inhibitors. A recent phase I/II clinical trial of DAC in combination with panitumumab in patients with wt KRAS mCRC was performed. Of note, 2 of the 20 patients tested had a partial response, and 10 patients had stable disease (3 patients longer than 16 weeks). Furthermore, the time to progression (TTP) was longer in 7 patients when compared to their previous TTP on ceruximab [19]. It is possible that the combination regimen resensitized patients to the anti-EGFR treatment. However, the roles and mechanism of DAC in overcoming resistance of mAbs against EGFR in CRC cells need further elucidation.

The cancer stem cell theory indicates that although chemotherapy kills most cells in a tumor, the tumor stem cells survive and contribute to drug resistance. This hypothesis is based on the evidence suggesting the tumor initiating cell may be innately resistant to various therapies. As a result, the cancer stem cells were able to survive cytotoxic or targeted therapies and caused tumor development and relapse.

Many studies have observed a limited clinical response using only a demethylating drug. However, improved effects were found when followed by other subsequent treatments such as chemotherapy, immunotherapy, and targeted therapy. This result suggests there is a “priming” ability of the epigenetic agent [20]. Transient low doses of DNA demethylating drugs exert durable antitumor effects on hematological and epithelial tumor cells. The demethylating drugs might inhibit the tumor initiating or cancer stem-like and self-renewing cells, and provide a memory for the reprogramming-like status [21]. A recent study showed that low-dose 5-azatidine elicited a DNMT3B-dependent activation of p53 target genes and also caused both DNA damage and DNA demethylation. The result was the downregulation of the pluripotency genes including NANOG, SOX2, GDF3 and Myc target genes in NT2/D1 cells [22]. Thus, epigenetic therapy combined with chemotherapy would be a potential new strategy for tumor treatment and could overcome drug resistance by killing both cancer cells and cancer stem cells.

Demethylating drugs regulate multiple pathways for tumorigenesis such as cell cycle control, DNA replication, mRNA splicing and translation. It was recently demonstrated that there was a significant enrichment for immunomodulatory pathways in tumor cell lines after treatment with 5-azacitidine. The pathways included interferon signaling, antigen processing and presentation, cytokines/chemokines, and the upregulation of cancer testis antigens [23]. An analysis of gene expression profiles of PBMCs or biopsies during DAC treatment in solid tumor patients confirmed the regulation of immune activity. George et al. detected that genes enriched on day 7 post-decitabine play roles in synaptogenesis, B cell activation, antigen processing and peptide presentation, and inhibition of cell growth. The day 28 samples also included genes with biogenic amine catabolism, regulation of interferon gamma, calcium mediated signaling and regulation of cdc42 GTPase activity [14].

Several large studies have demonstrated that 5-azacytidine activates antitumor immunity in different ways. In addition to regulating cancer testis antigens such as MAGE family [24], the demethylating drug up-regulated other tumor immune stimulating molecules, including MHC-I antigens, beta-2-microglobulin (B2M), CD58, TAP1, and the immune-proteasome subunits PSMB9 and PSMB8 [25], [26], [27], and the transcription factors IRF7 and IRF5 [24], [28]. Furthermore, deletion of DNMT1 and subsequent DNA methylation influenced the proliferative potential of antigen-specific CD8+ T cells [29], and 5-azacytidine treatment (5 µM or 20 µM) promoted an inhibitory T cell phenotype and impaired immune mediated anti-leukemic activity [30]. Hence, methylation modification and silencing of specific genes may be critical during T cell maturation and commitment. Treatment with 5-azacitidine may sensitize tumors to immune checkpoint therapy by targeting PD-L1/PD-1 interaction. Thus, the combination of epigenetic therapy and immunotherapy augments the clinical response in tumor patients [27], [31].

These findings indicated the possibility that the demethylating drugs might play an immune stimulatory role in cancer therapy by sensitizing the patients to immune responses. The potential molecular mechanisms of DAC function in cancer therapy are summarized in Fig. 1, while the detailed regulation pathways and molecular networks need further investigation (Fig. 1). In addition, identifying the crosstalk between immunotherapy and chemotherapy will be promising.

In the past 10 years, clinical trials treating solid tumors with DAC have improved. We collected the recent clinical studies in the NCBI database in May 2014 that administered DAC to patients with solid tumors. There are 13 phase I or II clinical trials (305 patients, 18 tumor types) using DAC agent at different doses examined in this review. The studies are shown in Table 1.

Several trials were performed with different schedules of DAC-based therapy to determine the tolerable administration dose and time in solid tumor patients. A phase I study of 19 solid tumor patients used DAC 20–40 mg/m²/day for a 72-hour continuous intravenous infusion on days 1–3 of a 28-day cycle. The authors reported that the tolerated dose of DAC in these patients was 30 mg/m²/day during the 72-hour continuous infusion [32]. In another clinical trial, Samlowski et al. applied lower doses of DAC (2 mg/m²/d) via long-term 7-day continuous infusion to patients with refractory solid tumors. They found that the level of toxicity was suitable for further DNA methylation evaluation. However, patients with extensive radiotherapy to marrow-containing bones or extensive nucleoside analog pre-treatment should be excluded, due to the possible adverse effects [33].

DAC showed a clinical effect in hematological diseases, but the combination with other therapies would be more effective in solid tumors. Appleton reported that combination administration of DAC (90 mg/m²/day) as a 6-hour infusion on day 1 followed by the chemotherapeutic agent carboplatin on day 8 in a 28-day cycle showed a favorable response rate [34]. Recently, our group proposed the regimen of extreme low-dose DAC (7 mg/m²/day) as a half-hour infusion per day for 5 consecutive days. This therapy could also be combined with chemotherapy or adoptive immunotherapy (for example, cytokine-induced killer cells (CIKs)) in solid tumor patients. Strikingly, clinical benefits were found in 60% of patients. Moreover, the median progression free survival (PFS) was prolonged compared with the PFS of previous treatments. Furthermore, we observed a significant correlation between the PFS of previous treatment and clinical response [35]. A phase I trial used DAC in combination with the HDAC inhibitor vorinostat in patients with advanced solid tumors. Stathis et al. compared the sequential schedule (DAC 10 mg/m²/day for 5-day infusion, combined with vorinostat 200 mg 3 times a day on days 6–12) and a concurrent schedule (vorinostat 200 mg 2 times a day on days 3–9). The authors found that the sequential schedule were easier to deliver and was more tolerable [36].

Stewart observed no significant correlation between the clinical responses and DAC doses or tumor characteristics when using different doses of DAC in distinct patient cohorts [17]. In pediatric patients with neuroblastoma and other solid tumors, the maximum-tolerated dose (MTD) of DAC was 5 mg/m²/day for 7 days when administered in combination with doxorubicin and cyclophosphamide. The patients showed significant myelosuppression as the primary toxicity. There were only limited clinical responses observed in this study because the doses of DAC needed to produce clinically relevant biologic effects were not well tolerated with this combination in children.

The first-line chemotherapy regimen for ovarian cancer patients generally consists of a platinum compound (cisplatin or carboplatin) and a taxane, while many patients will relapse and acquire resistance to the current treatment and resistance is a key limiting factor in patient survival [37]. Numerous preclinical studies have suggested that the development of epithelial ovarian cancer is correlated with the accumulation of aberrant DNA methylation. Thus, hypomethylating agents can reverse platinum resistance in ovarian cancer cells [38]. Several phase I/II clinical trials used low-dose DAC combined with carboplatin in platinum-resistant ovarian cancer patients. The Nephew group has observed DAC tolerability at a dose of 10 mg/m² daily for 5 days followed by carboplatin administration on day 8 in ovarian cancer patients [39]. Using this regimen, 35% of 17 patients showed clinical response with progression-free survival (PFS) at 10.2 months, and 9 patients (53%) had PFS at 6 months [40].

A clinical trial was conducted in metastatic colorectal cancer (mCRC) patients with wild type (wt)-KRAS. The patients were treated with DAC at 45 mg/m² on days 1 and 15, and panitumumab (mAb against EGFR) 6 mg/kg on days 8 and 22 every 28 days. Excitingly, 10% patients had a partial response and 50% patients had stable disease. This result suggested that the combination of DAC and targeted therapy was well tolerated and showed clinical response in mCRC patients [19].

Epigenetic modifications have an important role in lung cancer tumorigenesis. Several recent clinical trials have combined DAC and chemotherapeutic drugs in patients with non-small cell lung cancer (NSCLC). The patients were treated with escalating doses of DAC (5–15 mg/m²) for 10 days in combination with valproic acid (VPA) (10–20 mg/kg/day) on days 5–21 of a 28-day cycle. However, the clinical response was disappointing due to the unacceptable neurological toxicity caused by VPA. There was stable disease only in one patient. Thus, an alternative combination or schedule with less toxicity should be investigated [41].

The demethylating activity of the DAC regimen was examined in vivo to evaluate its biological effects. DNA from peripheral-blood mononuclear cells (PBMCs), plasma and tumor biopsies was extracted and assessed for global DNA methylation and site-specific DNA methylation, and gene expression profile.

Due to limitations in acquiring tumor biopsies, analyses of plasma DNA or PBMCs were performed. In ovarian cancer patients, the methylation specific polymerase analysis indicated the HOXA11 and BRCA1 DNA promoters were hypomethylated in plasma DNA on day 8 and 15 post-decitabine treatment [39]. Appleton et al. observed that the global and MAGE-1 promoter DNA demethylation levels were maximal during days 8–12 in PBMCs and buccal cells. The demethylation levels then returned to near baseline levels by day 22 [34]. Similarly, Samlowski et al. reported that the levels of global DNA and MAGE-1 promoter methylation were apparent by day 7 (immediately after termination of continuous infusion), but declined 14 days after the start of the first cycle [33].

The immunohistochemistry analysis of available tumor biopsies from non-small cell lung cancer (NSCLC) patients suggested a marked induction of NY-ESO-1, MAGE-3, and p16, after DAC treatment [42]. Nevertheless, significant global demethylation was reduced or not found in tumor biopsies compared to PBMCs. This result may be due to the reduced uptake of DAC and lower proliferation rate in tumor cells compared to PBMCs [40], [42].

In addition to the gene alterations in tumor cells, the levels of fetal hemoglobin (HbF) in lymphocytes increased 8–12 days after DAC treatment and then returned to the initial levels by day 15 in cycle 1. However, HbF levels accumulated in the subsequent cycles of treatment. These results suggest that repeated treatment with DAC is necessary to reverse gene silencing [34].

George et al. reported that DAC had a long-term activity within a 28-day cycle, and they observed that the MAGE-1 promoter was demethylated in the PBMC of 18/21 patient samples on day 28 following DAC treatment. Additionally, 14/16 patients exhibited increased HbF mRNA expression on day 28. Of these patients, 10/14 patients had increased HbF expression on day 7 and the expression increased on day 28 [14]. Thus, different DAC schedules may cause differential DNA methylation and gene expression profiles. The molecular mechanism of DAC action in solid tumor therapies is still unclear.

DAC-based therapy has exhibited favorable effects in some types of solid tumor patients. Therefore, further investigations of prognostic biomarkers for hypomethylating strategies in solid tumors are required. Many studies have focused on the detection of MAGE-1 methylation status in peripheral blood as a potential biomarker for DAC treatment. The percentage of patients who displayed significant demethylation of MAGE-1 in PBMCs during the first course of treatment ranged from 60% to 100% [2], [14].

The increased expression of HbF mRNA was reported to be 88% of blood samples and 100% of bone marrow samples in neuroblastoma patients following DAC treatment [14]. Consistent with these data, HbF protein levels were enhanced by Western blotting analysis post-decitabine [34]. Following treatment with a combination of DAC and VPA, the HbF levels increased in 100% of PBMC samples from NSCLC patients [41].

Analyses of DAC-altered genes in ovarian cancer patients post-decitabine treatment revealed that the epigenetic silencing of tumor suppressor genes, such as DLEC1 [43], OPCML [44], and RASSF1A [45] occurred frequently. A phase I/II clinical trial reported that RASSF1A, AKT1S1, and MLH1 genes were hypomethylated in all patients after DAC therapy. The demethylation of the development-associated transcription factors HOXA10 and HOXA11 was specifically related to ovarian cancer responders with PFS more than 6 months and associated with chemotherapy sensitivity [40].