Development of curcumin as an epigenetic agent


  • Siqing Fu MD, PhD,

    Corresponding author
    1. Department of Investigational Cancer Therapeutics, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    • Department of Investigational Cancer Therapeutics, Unit 0455, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030
    Search for more papers by this author
    • Fax: (713) 745-3855

  • Razelle Kurzrock MD

    1. Department of Investigational Cancer Therapeutics, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author


The clinical benefits of curcumin as a single agent were demonstrated in patients with advanced pancreatic cancer in a phase 2 study despite pharmacokinetic analysis showing a much lower plasma concentration of curcumin in humans than in vitro. The diverse and broad biological activities of curcumin are mediated through direct interaction of curcumin with target proteins as well as epigenetic modulation of target genes, supported by evidence that curcumin modulates gene expression in a time- and concentration-dependent manner in human cancer cells. This review delineates the novel mechanisms of curcumin as an epigenetic agent through its interaction with histone deacetylases, histone acetyltransferases, DNA methyltransferase I, and microRNAs. Accumulating data support curcumin's functionality in modulating multiple biological processes at low concentrations through its activity as an epigenetic agent. The development of curcumin as an epigenetic agent warrants further preclinical and clinical studies to explore its diversity and efficacy in cancer treatment and in combination with other anticancer agents. Cancer 2010. © 2010 American Cancer Society.

Curcumin as a single agent displayed clinical benefit in patients with advanced pancreatic cancer1 despite pharmacokinetic analysis showing its much lower plasma concentration in humans than in vitro. A possible explanation for its activity in humans is that accumulation of curcumin and its metabolites in cells during daily exposure occurs because it is its highly hydrophobic nature.2 Another explanation may be that curcumin exerts its biological activities through epigenetic modulation, which requires lower concentrations. This review summarizes the current data surrounding curcumin-mediated epigenetic modulation and its role in anticancer treatment.

Curcumin (diferuloylmethane), a yellow hued polyphenol, is an active component of the perennial herb Curcuma longa, commonly known as turmeric.2 The major components in the curcuminoid complex are curcumin (approximately 80%), demethoxycurcumin (approximately 17%), and bisdemethoxycurcumin (approximately 3%).3 Curcumin is minimally soluble in water but quite soluble in organic solvents. Its enolic and beta diketone forms are stable in acidic pH, but unstable in neutral and basic pH solutions, where they are degraded to ferulic acid and feruloylmethane. Curcumin is, however, more stable in 10% fetal calf serum and human blood, with <20% of curcumin decompensation in the first hour after administration.4

Curcumin has diverse molecular targets associated with numerous biochemical and molecular cascades with activity exerted via direct interaction with and/or epigenetic modulation of gene expression. To date, curcumin has been found to physically bind to as many as 33 different proteins, including thioredoxin reductase, COX2, protein kinase C, 5-lipoxygenase, and tubulin.5 The various molecular targets modulated by curcumin include transcriptional factors, growth factors and their receptors, cytokines, enzymes, and genes regulating cell proliferation and apoptosis.

Cancer epigenetics is characterized by heritable patterns of specific DNA methylation and chromatin modification.6 Epigenetic factors can also affect microRNA expression.7 Neoplasia is characterized by distinct patterns of disrupted pathways involving genetic alterations and epigenetic changes mediated by DNA methyltransferases, methyl-CpG-binding domain proteins, histone acetyltransferase (HAT), histone deacetylase (HDAC), histone methyltransferases, and histone demethylases, resulting in DNA global hypomethylation, and promoter-localized hypermethylation.8 Here we delineate the factors responsible for curcumin's promise as an epigenetic agent for cancer therapy drug development.

Curcumin and Histone Deacetylases

At least 18 HDACs have been identified in humans, primarily occupying 4 classes based on homology with yeast deacetylases.9 HDAC enzymes do not bind to DNA directly but rather interact with DNA through multiprotein complexes that include corepressors and coactivators.

Among 33 carboxylic acid derivatives, curcumin was the most effective HDAC inhibitor at 50 to 500 μM in HeLa nuclear extracts at an IC50 of 115 μM. Molecular docking studies performed for human HDAC8 showed that curcumin has a free binding energy (−8.55 ΔG kcal/mol) and inhibition constant (Ki = 539 nM) comparable to Trichostatin A and vorinostat (−8.59 and −7.65 ΔG kcal/mol, and Ki = 504 nM and 2.47 μM, respectively), and that it is more potent than valproic acid and sodium butyrate (−4.41 and −4.65 ΔG kcal/mol, and Ki = 564 and 365 μM, respectively). Analysis of curcumin's molecular docking in a complex with human HDAC8 revealed that the ligand adopted a stable binding conformation extended toward the entrance domain.10

One study showed that Raji cells, a cell line of Epstein-Barr virus-transformed lymphocytes with surface Fc receptors, treated with different concentrations of curcumin for 0, 24, 36, 48, 60, and 72 hours inhibited cell proliferation in a dose- and time-dependent manner with an IC50 of 36 hours at 24 μM.11 Another study revealed that the expression of HDAC1, 3, and 8 was significantly higher in untreated Raji cells compared with curcumin-treated cells, leading to significantly greater levels of acetylated histone H4 in cells treated with 25 μM of curcumin for 24 hours. Western blots showed that HDAC1, 3, and 8 protein levels decreased in a dose-dependent manner with curcumin exposure,12 but did not show whether curcumin could directly inhibit HDAC enzyme activity. Increased acetylated histone H4 was associated with decreased levels of HDAC1, 3, and 8 proteins. Another study found that curcumin at nanomolar concentrations restored corticosteroid function in monocytes exposed to oxidants by maintaining HDAC2 activity via preventing oxidant-induced degradation of HDAC2 through down-regulating gene expression associated with protein degradation.13 Although controversy exists about the ability of curcumin to inhibit HDAC enzymes, it likely modulates HDAC activity through regulating gene expression. Further research is required to validate this hypothesis.

Curcumin and Histone Acetyltransferases

HATs including at least 25 members are organized into 4 families based on primary structure homology.14 The disparate family of HATs, acetylates, histones, and nonhistone targets is involved in diverse processes such as transcription activation, gene silencing, DNA repair, and cell cycle progression.15

Recent enzymatic studies dealing with the p300 HAT domain support the existence of a Theorell-Chance catalytic mechanism combined with broad substrate specificity independent of a specific Asp/Glu general base mechanism. Despite evidence that p300 acts as a tumor suppressor, some studies showed that when p300 was down-regulated and inhibited, tumor regression resulted, along with activation of a senescence checkpoint, which is logical given the finding that p300/CBP is important for the G1/S cell cycle transition.16 This suggests that the role of p300/CBP in cancer biology is versatile and context-dependent, supporting curcumin's role as a specific inhibitor of p300/CBP capable of inducing apoptosis in cancer cells without producing cytotoxic effects in healthy cells. Thus, p300/CBP HAT is an excellent target for novel cancer treatment drug development.17

Curcumin has been identified as a specific inhibitor of p300/CBP HAT activity in vitro and in vivo, but not of p300/CBP-associated factor, a HAT that associates with p300/CBP. Filter binding and gel HAT assays showed that acetylation of histones H3 and H4 by p300/CBP was strongly inhibited covalently by curcumin, with an IC50 of approximately 25 μM, whereas p300/CBP-associated factor HAT activity did not change even in the presence of 100 μM curcumin.18 These findings demonstrate curcumin's high specificity for p300/CBP HAT activity, but not for other enzymes that have histone substrates. The binding site on p300/CBP was specific for curcumin, and binding led to a conformational change, resulting in a decrease in the binding efficiency of histones H3 and H4 and acetyl CoA.19 These findings are significant in cancer because of the complex role of HATs in cancer, as is specifically explained later in this review.

Kinetic studies of enzymatic activity showed curcumin binding to p300, but not as an allosteric inhibitor with binding to active sites of histone or acetyl CoA.18 In addition, curcumin was a potent and specific inhibitor of p300/CBP HAT activity-dependent chromatin transcription. Curcumin induced cervical cancer cell apoptosis at concentrations of 50 to 100 μM associated with inhibiting acetylation of histones and p53 in vivo through specific inhibition of p300/CBP.20 In brain glioma cell lines, curcumin induces histone hypoacetylation to activate poly adenosine diphosphate ribose polymerase- and caspase-3–mediated apoptosis.21 In addition, inhibition of p300 mediated by curcumin decreases acetylation of RelA, which attenuates interaction with IκBα, leading to decreased IκBα-dependent nuclear export of the complex through a chromosomal region maintenance-1–dependent pathway.22 Therefore, curcumin's ability to suppress p300/CBP HAT activity, may account, at least in part, for its potent nuclear factor kappa B (NF-κB) inhibitory activity.

Curcumin and DNA Hypomethylation

Molecular docking of the interaction between curcumin and DNA methyltransferase I suggested that curcumin covalently blocks the catalytic thiolate of C1226 of DNA methyltransferase I to exert its inhibitory effect.23 Curcumin exists predominantly in solution as the enol form, which serves as a Michael acceptor to covalently block the catalytic thiol group in DNA methyltransferase I through the C3 keto-enol moiety of the curcumin compounds, demethoxycurcumin, and bisdemethoxycurcumin.

To validate the results of a DNA methyltransferase I homology modeling study, inhibition of the enzymatic activity of M.Sss1 (methyltransferase Sss1, an analog of DNA methyltransferase I) by curcumin was assessed in vitro. The methylation level of the double-strand oligonucleotide directly correlated with the enzymatic activity of M.Sss1 with exposure to various concentrations of curcuminoids from 1 nM to 100 μM. The apparent IC50 of curcumin with respect to M.Sss1 inhibition was identified as 30 nM.23

Genomic DNA was extracted from a leukemia cell line exposed to curcuminoids at concentrations of 0, 1, 3, and 30 μM for 72 hours. Analyses of global DNA methylation levels showed stability at 1 μM curcumin, but decreased by approximately 15% to 20% at 3 μM and 30 μM curcumin compared with untreated basal methylation levels, which was equivalent to decitabine-induced decreases in global DNA methylation levels.23 These data show that curcumin is a potent DNA hypomethylating agent, which is consistent with its broad activity in inflammation, cancer, and many other diseases, while remaining relatively safe in normal healthy cells. Further exploration and validation of curcuminoids as DNA hypomethylating agents is warranted. Coincidentally, a phase 1 trial with curcumin administered several days before docetaxel in patients with metastatic breast cancer resulted in 5 partial remissions and stable diseases in 3 of 8 patients.24 This unexpectedly high response rate might result from the clever sequential delivery of the 2 agents, which capitalized on and maximally used curcumin's epigenetic activity for cancer treatment.

Curcumin and Epigenetic Changes Impacting microRNAs

MicroRNAs are small, noncoding regulatory RNAs ranging in size from 17 to 25 nucleotides, which are matured by Dicer/Drosha (an RNase) from hairpin-structured precursors.25 MicroRNAs post-transcriptionally repress gene expression by recognizing complementary target sites in the 3′-untranslated regions of target mRNAs. Currently, >500 microRNAs have been identified in humans.26 Many microRNAs exhibit characteristic expression patterns.27 Expression of a specific microRNA can be measured by Northern blot, RNase protection assay, primer extension assay, and global microRNA expression profiles.28 MicroRNAs are aberrantly expressed in diverse cancers, differentially characteristic for specific cancer phenotypes, stages, and other clinical variables, and play important roles in cell cycling, programmed cell death, cell differentiation, tumor development, invasion, metastasis, and angiogenesis.29

Aberrant expression of microRNAs can arise through numerous mechanisms, including genomic abnormalities, transcriptional regulation, and processing of microRNAs.7 Epigenetic factors also affect the expression of microRNAs.25 In many cancers, hypermethylation of CpG islands in promoter regions and histone modifications result in heritable transcriptional silencing of tumor suppressor genes. Some microRNAs are up-regulated by hypomethylating agents30 and/or inhibition of histone deacetylation.31 In addition, microRNAs may counteract CpG methylation and the epigenome.32 For example, miR-29 directly targets DNMT-3A and -3B. Consistent with this observation, ectopic expression of miR-29 resulted in a global reduction of DNA methylation, abrogating the expression of tumor suppressor genes previously silenced by promoter methylation in cancer cells.33

Interestingly, curcumin modulated microRNA expression in human pancreatic cells, up-regulating microRNA-22 and down-regulating microRNA-199a* through an unknown mechanism.34 Combined with the recent finding that curcumin is a DNA hypomethylating agent, epigenetic modulation of microRNA expression may be an important mechanism underlying curcumin's biological effects. Microarray analyses of microRNA expression in pancreatic cancer cells exposed to curcumin at concentrations of up to 10 μM for 72 hours revealed 11 microRNAs that were significantly up-regulated and 18 microRNAs that were significantly down-regulated. MicroRNA-22 demonstrated tumor suppressor activity through inhibition of at least 50 target genes, showing that a key effect of curcumin on cancer cells was mediated by epigenetic modulation of microRNAs, which, in turn, regulated many target genes involved in tumorigenesis, development, and metastasis. Modulating microRNA expression is yet another novel strategy that ultimately can be embraced as a potential therapy for the prevention and treatment of human cancer.29

Curcumin and Gene Expression

The diverse and broad biological activities of curcumin are mediated through direct interaction of curcumin with target proteins as well as epigenetic modulation of target genes, supported by evidence that curcumin modulates gene expression in a time- and concentration-dependent manner in human cancer cells.35 Gene expression changes in human colon cancer cells in response to curcumin revealed early response genes, defined as differentially expressed genes after exposure to curcumin for 3 to 6 hours, involved in the cell cycle (p16INK4, BUB1B, STK6, STK12, PLK, Rb, p53, cyclin G1, and cyclin E1), signal transduction (STAT3, STAT5b, MAPK, AKT, VEGF, and FGFR1), DNA repair (hMLH1, MSH3, and ERCC2), gene transcription (HDAC1, ATF4, and EGF1), cell adhesion (annexin and integrins), and xenobiotic metabolism (GSTT2, GSTM4, and CYP1B1) while the cells accumulated in the G2/M phase with increased apoptosis. Expression of metallothionein genes changed in response to curcumin at 12 to 24 hours. Tubulin genes were down-regulated after 3 hours of exposure to curcumin at 25 μM, but up-regulated after 48 hours of exposure to curcumin at 100 μM.36

When RNA expression profiles were compared, diverse cell lines had variable responses to curcumin in a time- and concentration-manner, suggesting that clinical responses to curcumin differ because of substantial physiologic differences in cancer cells that contribute to diverse gene expression profiles.37 In the LNCaP prostate cancer cell line, the number of genes regulated by curcumin with a >4-fold increase in expression were 8, 73, 181, 3, and 0 at 3, 6, 12, 24, and 48 hours, respectively. At 12 hours, curcumin up-regulated 181 genes and down-regulated 13 genes by >4-fold in the androgen-responsive LNCaP prostate cancer cell line, and 27 and 13 genes in the C4-2B androgen-refractory prostate cancer cell line, respectively.37 Therefore, curcumin-induced gene expression was dependent on dose and cell context.

Treatment of cancer cells with curcumin results in a dose- and time-dependent inhibition of cell proliferation associated with time-, dose-, and cell context-dependent changes in gene expression. However, the precise mechanisms by which curcumin modulates gene expression have not been delineated. One theory is that curcumin modulates gene expression through direct interaction with diverse intracellular signal transduction pathways. Here we propose the hypothesis that curcumin produces an integrated modulation of signal transduction pathways and epigenetic modulation (Fig. 1). In addition to direct and indirect interactions with intracellular signal transduction pathways regulating gene expression, curcumin inhibits both DNA methyltransferase I and HATs to epigenetically modulate gene expression, as shown in Table 1.

Figure 1.

Schematic shows curcumin's roles in gene expression through direct interaction with transcriptional factors such as nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB), epigenetic modulation via inhibiting DNA methyltransferase I (DNMT1), and histone acetyltransferase (HAT) and histone deacetylase (HDAC) complexes. Cancer cell death induced by curcumin is mediated by reactivating expression of tumor suppressor genes and/or inactivating expression of oncogenes via microRNA (miRNA) modulation.

Table 1. Epigenetic Modulation by Curcumin
 Biological Activities of Curcumin
  1. HDAC indicates histone deacetylase.

DNA methylation1. Covalently blocks the catalytic thiolate of C1226 of DNA methyltransferase I23
 2. Inhibits methyltransferase M.SssI at an IC50 of 30 nM23
 3. Induces global genomic DNA hypomethylation23
Histone acetyltransferases1. Directly inhibits p30018-21, 48-53
 2. Down-regulates p30011, 54
 3. Inhibits GCN5 associated with hypoacetylation of histone H355
 4. Promotes p300 degradation19
HDACs1. Acts as an HDAC2 activator56
 2. Increases HDAC2 protein expression13
 3. Acts as an HDAC8 inhibitor10
 4. Down-regulates HDAC1, 3, and 811, 12, 57

Development of Curcumin as an Epigenetic Agent

Epidemiological evidence has shown that people who incorporate high doses of curcumin in their diets have a lower incidence of several common cancers such as breast, colorectal, and prostate cancer. Numerous preclinical reports demonstrated curcumin's anticancer activity through induction of apoptosis as well as sensitization of cancer cells to other anticancer agents. Phase 1 studies of curcumin as a single agent showed that it can be administered safely at oral doses of up to 8 g per day.38, 39 In a phase 2 study of curcumin in 21 evaluable advanced pancreatic cancer patients, single-agent curcumin achieved 1 partial response with a 73% tumor regression, and 1 patient had stable disease for approximately 2.5 years.1 Although these data show curcumin's promise as a single agent for cancer control, its further development in combination with other anticancer agents is warranted. One strategy would be to target the NF-κB pathway for chemosensitization.40 Another strategy would be to develop curcumin as a hypomethylating agent and HAT inhibitor.

DNA methylation plays an essential role in regulating normal biologic processes as well as carcinogenesis.41 DNA methylation is a heritable, DNA methyltransferase-induced modification of DNA structure that does not alter the specific sequence of base pairs responsible for encoding the genome, but that can directly inhibit gene expression.8 Two patterns of DNA methylation have been observed in cancer cells: global hypomethylation across the genome, and localized hypermethylation at specific CpG islands within the gene promoter regions of specific genes.42 Decreased methylation because of global hypomethylation could facilitate the expression of previously quiescent proto-oncogenes and prometastatic genes and promote tumor progression. Alternatively, an aberrant increase in methylation patterns at previously unmethylated sites, such as the promoter regions of tumor suppressor genes, could result in transcriptional silencing and inability to control tumorigenesis.42 Thus far, the US Food and Drug Administration has approved azacitidine and decitabine as hypomethylating agents for treating myelodysplastic syndrome. Preclinically and clinically, azacitidine and decitabine have demonstrated an ability to sensitize cancer cells to chemotherapeutic agents as well as biological targeted agents. Therefore, exploring curcumin as a hypomethylating agent to induce cancer cell chemosensitization is a particularly attractive anticancer strategy.

The reversible process of histone acetylation occurring at the ϵ-amino group of lysine residues in the N-terminal tails of core histones mediates conformational changes in nucleosomes, and is controlled by HATs and HDACs. The anticancer effects of HDAC inhibitors are well known and have been shown in clinical trials. In contrast, the chemotherapeutic potential of HATs has not yet been established. The role of HATs in cancer is complex. Because maintaining a specific histone acetylation process is crucial to various cell functions such as cell cycle progression, DNA repair, cell proliferation and differentiation, mutations and/or chromosomal translocations involving HAT genes result in development of malignancies. Loss of p300 heterozygosity and p300 mutations is associated with glioblastoma and colorectal cancer. CBP knockout mice had an increased incidence of cancers as well as defects in cell proliferation. Thus, p300/CBP appears to exert both tumor suppressor and tumor promoter properties, depending on the specific cellular context. Experiments with HAT inhibitors showed a clear correlation between the HAT-inhibitor potency of these compounds and their effects on cell proliferation, suggesting that HAT inhibitors have a potential role in cancer therapy.43, 44 In addition, coexposure of an HDAC inhibitor such as Trichostatin A improves the anticancer activity of low concentrations of curcumin in human leukemia cells,45 providing evidence that inhibition of HDACs and HATs together may be a novel strategy for cancer treatment.

When HATs and HDACs are physically linked as a complex they interact with other regulators.46 One might hypothesize that optimal therapeutic regimens can be developed by modulating histone acetylation along with other therapeutic approaches such as inhibition of DNA methylation, intracellular signal transduction pathways, and transcriptional factors, and induction of apoptotic pathways.47 In addition to its direct interaction with intracellular signal transduction pathways, curcumin has enormous potential to be developed as an epigenetic agent. Specifically, its suppression of methylation and HATs warrants further exploration.

Although curcumin is able to directly interact with multiple intracellular signal transduction pathways and other target proteins, accumulating data support the concept that epigenetic modulation by curcumin might play a major role in cancer treatment. Thus, an understanding and application of the findings cited in this review will facilitate clinical development of curcumin in combination with other anticancer agents for cancer treatment.


We thank Joann Aaron in the Department of Investigational Cancer Therapeutics at The University of Texas M. D. Anderson Cancer Center for editing the article.


The authors made no disclosures.