Volume 36, Issue 9, Pages 1057-1072 (September 2008)

2 of 18

ABSTRACT

FULL TEXT

FULL-TEXT PDF (782 KB)

A cell stress signaling model of fetal hemoglobin induction: what doesn't kill red blood cells may make them stronger

Received 18 May 2008; received in revised form 25 June 2008; accepted 27 June 2008.

A major goal of hemoglobinopathy research is to develop treatments that correct the underlying molecular defects responsible for sickle cell disease and β-thalassemia. One approach to achieving this goal is the pharmacologic induction of fetal hemoglobin (HbF). This strategy is capable of inhibiting the polymerization of sickle hemoglobin and correcting the globin chain imbalance of β-thalassemia. Despite this promise, none of the currently available HbF-inducing agents exhibit the combination of efficacy, safety, and convenience of use that would make them applicable to most patients. The recent success of targeted drug therapies for malignant diseases suggests that this approach could be effective for developing optimal HbF-inducing agents. A first step in applying this approach is the identification of specific molecular targets. However, while >70 HbF-inducing agents have been described, neither molecular mechanisms nor target molecules have been definitively verified for any of these compounds. To help focus investigation in this area, we have reviewed known HbF-inducing agents and their proposed mechanisms of action. We find that in many cases, current models inadequately explain key experimental results. By integrating features of the erythropoietic stress model of HbF induction with data from recent intracellular signaling experiments, we have developed a new model that has the potential to explain several findings that are inconsistent with previous models and to unify most HbF-inducing agents under a common mechanism: cell stress signaling. If correct, this or related models could lead to new opportunities for development of targeted therapies for the β-hemoglobinopathies.

Article Outline

• Abstract

• Therapies aimed at the underlying molecular causes of SCD and β-thalassemia

• Potential of targeted drug therapy

• Defining “pharmacologic induction of fetal hemoglobin”

• Experimental systems for studying HbF induction

• DNA methyltransferase inhibitors

• Other cytotoxic agents

• Short-chain fatty acids

• Histone deacetylase inhibitors

• Other inducing agents

• Role of cell signaling in HbF induction

• A unified model of HbF induction

• Implications of a stress signaling model of HbF induction

• Summary

• Acknowledgment

• References

• Copyright

Sickle cell disease (SCD) and β-thalassemia are the most prevalent of the human β-hemoglobinopathies and they, along with the α-thalassemias, comprise the single largest class of human genetic diseases [1]. The high prevalence of the SCD and β-thalassemia mutations is attributed to the fact that, in the heterozygous state, both appear to decrease the severity of malarial infection 2, 3. However, when present in the homozygous state, these conditions shorten the lives of affected individuals.

SCD results from a glutamate to valine substitution at position 6 of the β-globin protein [4]. When incorporated into the hemoglobin tetramer, the resulting sickle hemoglobin (HbS, α2βS2) can polymerize under specific physiologic conditions. These HbS polymers in turn alter red cell structure and function leading to a complicated cascade of events affecting not only erythrocytes, but also neutrophils, platelets, endovascular cells, and the coagulation system, resulting in vascular occlusion and the characteristic signs and symptoms of SCD (reviewed in [5]). Even with the most advanced medical care, including optimal red cell transfusions and hydroxyurea (HU) therapy, SCD patients are subject to many debilitating disease complications, including anemia, life-threatening infections, central nervous system thrombosis, acute and chronic pain syndromes, retinopathy, pulmonary hypertension, heart failure, renal dysfunction, priapism, and skin ulcerations. Data reported in the mid-1990s showed that the median survival for SCD patients in the United States was only into their 40s [6], while in Sub-Saharan Africa, it has been estimated that only half of affected children live beyond 5 years of age [7].

β-thalassemia is characterized by decreased or absent production of β-globin protein. More than 200 β-thalassemia mutations affecting transcriptional and posttranscriptional processes have been described [8]. The lack of β-chain production leads to accumulation of free intracellular α-chains, which are thought to cause oxidative damage to the red cell membrane and apoptosis of erythroid precursors [9]. Inefficient iron utilization has also been linked to ineffective erythropoiesis [10]. These effects lead to severe anemia due to the combination of ineffective erythropoiesis, hemolysis, and hypersplenism, which in turn cause several pathologic effects, including splenomegaly, skeletal abnormalities, and growth retardation. Most patients with β-thalassemia die as a result of cardiac failure from iron overload due to long-term transfusion therapy. Despite advances in iron chelation, including introduction of effective oral agents, median survival has recently been estimated to be 49 years for patients who adhere to optimal chelation regimens while a median survival of only 28 years was estimated for “typical” adherence to chelation schedules [11].

Therapies aimed at the underlying molecular causes of SCD and β-thalassemia

While most significant advances in the treatment of SCD and β-thalassemia have resulted from preventing or treating the complications of these diseases, a long-held goal of modern hemoglobinopathy research is to target the underlying molecular events that are responsible for the clinical features of these diseases—sickle hemoglobin polymerization and globin chain imbalance. Three distinct strategies have shown promise for achieving this goal and offer opportunities to prevent or lessen most disease complications.

Allogeneic hematopoietic stem cell transplantation (HSCT) has been successfully applied to several hundred SCD and thalassemia patients (reviewed in 12, 13) producing overall long-term disease-free survival rates of >70% with rates from 82% to 87% for selected thalassemia patients less than 17 years of age [14] and event-free survival of 82% to 86% for SCD patients 15, 16, 17. Despite these promising results and ongoing research, HSCT is likely to be limited to a small proportion of hemoglobinopathy patients for the foreseeable future. So far, the best results have been obtained with younger patients and those who have not already developed significant disease complications making many current patients less than optimal or ineligible candidates for HSCT [18]. Most successful transplants have used stem cells from matched sibling donors. Estimates indicate that only 14% to 18% of SCD HSCT candidates have such donors available [12]. While small trials using matched unrelated donors and alternative preparative regimens have produced encouraging results, these trials have shown decreased rates of disease-free survival and increased rates of graft-vs-host disease compared to sibling donors [12]. Finally, HSCT for the hemoglobinopathies is currently available at only a limited number of highly specialized centers worldwide.

Transfer of γ- or β-globin genes into the hematopoietic stem cells of patients is another option for eliminating the underlying molecular defects of the β-hemoglobinopathies. Due largely to difficulties in achieving efficient transduction of human hematopoietic stem cells and high-level gene expression [19], it was only within the last year that an approved clinical gene transfer protocol for hemoglobinopathy patients has been opened [20]. Even if this initial human trial is successful, widespread application of this approach may be limited by concerns about insertional mutagenesis and the effect of inserted vectors on the expression of nearby genes 21, 22. In the near term, gene transfer therapy, like HSCT, will require access to highly sophisticated medical care and extensive financial resources that are unavailable to most SCD and thalassemia patients.

A third option for treating the underlying molecular causes of the β-hemoglobinopathies is the pharmacologic induction of fetal hemoglobin (PIFH). In a remarkable turn of evolution, humans have been provided with two perfectly good substitutes for the mutated or inactive β-globin genes of SCD and β-thalassemia. As shown in Figure 1A, the two fetal β-like globin genes are located upstream of the adult δ- and β-globin genes. Fetal Hb (HbF, α2γ2) is normally expressed at high levels only during the fetal and early postnatal periods (Fig. 1B) [23]. Studies from the 1950s and 1960s showed that people who co-inherited β-hemoglobinopathies and additional mutations that allowed γ-globin gene expression in adults (hereditary persistence of fetal Hb) had less severe clinical courses 24, 25. This was subsequently shown to be due to the ability of HbF to inhibit HbS polymerization in erythroid cells of patients homozygous for the βS gene [26] and to lessen the degree of the globin chain imbalance through the interaction of γ-globin chains with the excess α-globin chains of β-thalassemia [27]. These findings led to the proposal that pharmacologic reactivation of the fetal globin genes in patients with β-hemoglobinopathies could provide an effective treatment strategy [28]. Ground-breaking work by Desimone, Ley, Dover, and their colleagues, first in baboons and then in hemoglobinopathy patients, demonstrated the ability of the nucleoside analogue 5-azacytidine (5-Aza) to induce transient fetal globin gene expression and HbF production 29, 30, 31, 32. As will be discussed here in more detail, in the more than 25 years since these reports, other studies have shown that HU, butyrate derivatives, and another DNA methyltransferase (DNMT) inhibitor, 5-Aza-2-deoxy-cytidine (decitabine) are also able to induce HbF production in hemoglobinopathy patients.

View Large Image
Download to PowerPoint

Figure 1 β-like globin gene expression during human development. (A) Structure of the human β-globin gene locus including the embryonic (ɛ), fetal (γ), and adult (δ and β) β-like globin genes and the 5 DNase hypersensitive sites that comprise the upstream locus control region. (B) Developmental patterns of β-like globin gene expression. Figure 1B is redrawn from reference [23].

Unfortunately, none of these clinically active HbF-inducing agents exhibit the optimal combination of efficacy, safety, and ease of use that would make them applicable to most hemoglobinopathy patients worldwide. The DNMT inhibitors are potent inducers of fetal hemoglobin. They are active in SCD patients resistant to HU [33] and have provided significant clinical benefits to people with severe or end-stage SCD [34] and β-thalassemia [35]. However, as currently formulated, these agents must be administered by injection or infusion over several days each month, they cause hematopoietic suppression, and there is little experience with the side effects of long-term administration. There is concern that because they alter the expression of genes in a nonspecific fashion, chemically modify DNA and change the epigenetic structure of cycling cells, they might increase the risk of cancer when used over the extended periods of time necessary to treat hemoglobinopathy patients [36].

Butyrate derivatives have also been shown to be effective inducers of HbF in patients. However, the widespread application of these agents has been limited by inconvenience of administration and poor patient compliance (due to the need for daily IV infusions or many pills), loss of effectiveness with some administration schedules and suppression of erythropoiesis (reviewed in [37]). Like the DNMT inhibitors, these agents also alter epigenomic structure, but in this case global histone acetylation is increased. The long-term effects of such epigenetic changes are not known.

HU can be taken orally, is relatively inexpensive and has proven benefits for SCD patients, but it too has shortcomings. HU only increases HbF in approximately half of SCD patients [38] and is less effective in increasing HbF for β-thalassemia patients 39, 40, 41. While HU does decrease the frequency of painful crises and acute chest syndrome, it has not eliminated these complications and does not appear to decrease the incidence of stroke [42]. It produces low blood counts and thus requires close clinical monitoring and dose adjustment [42]. Concerns remain that there may be long-term side effects, including increased risk of malignancy [43]. The conclusion that long-term HU therapy decreases mortality in SCD patients [44] has been questioned [45]. Despite the fact that HU is approved by the US Food and Drug Administration (FDA) for use in SCD patients and has well-documented clinical benefits, recent independent studies have found significant underutilization of HU in Maryland and Florida 46, 47. Research from Nigeria, the African country with the highest incidence of SCD, shows that HU has not been accepted by patients or physicians due to concerns over safety, compliance, drug availability, cost, risk of tuberculosis reactivation, carcinogenesis, and teratogenicity [48].

While trials utilizing DNMT inhibitors, butyrate and HU have proven that PIFH can be an effective therapeutic strategy, it is clear that none of these agents are optimal inducers of HbF. The current challenge for investigators is to produce drugs that are effective, have acceptable short and long-term side effect profiles, and offer the convenience of use and affordability that will make them of benefit to all hemoglobinopathy patients, including those who live in areas that lack modern medical facilities and financial resources. While this may seem a daunting task, the effectiveness of oral pharmacologic regimens for patients with AIDS and tuberculosis in less advantaged areas of the world attest to the feasibility of these goals 49, 50.

Potential of targeted drug therapy

Hematologists and hematology researchers are well aware of the recent advances in targeted drug development. Rituximab targeting of CD20, Imatinib targeting of Bcr/Abl, c-kit, and platelet-derived growth factor tyrosine kinases and development of agents targeting Flt3 and JAK2 kinase molecules are but a few examples of targeted therapies that are approved or under development for the treatment of patients with a wide variety of hematologic malignancies or myeloproliferative disorders. A key first step in developing targeted drugs is identification of specific molecules whose function can be manipulated by small molecules or antibodies. One approach to identifying targets is to first define molecules within pathways that mediate specific functions and to then screen for, or rationally design molecules that alter the function of the targets to produce a desired effect. Despite years of research by many investigators, most of the key molecules and pathways that mediate normal and induced production of HbF remain to be defined. A second approach to target identification is to screen libraries or test promising compounds for the ability to produce a desired cellular effect and to then determine the pathways and molecules targeted by effective agents. In the case of PIFH, >70 compounds have already been identified that increase fetal globin gene expression and/or HbF production in a variety of experimental systems (Table 1). These considerations suggest that the most direct route to identifying specific molecules for targeted drug development may involve taking advantage of these lead compounds and the research that has been performed to determine their mechanisms of action. In the remainder of this article, we review what has been learned about the mechanisms of action of known inducers of HbF, with a focus on specific pathways and potential molecular targets for drug development.

Table 1.

Agents that induce γ-globin gene expression and/or fetal hemoglobin productiona


Agent		Reference(s)
DNA methyltransferase inhibitors (n = 4)
5-Azacytidine	c, m, p, hc, hv	29, 30
Decitabine	c, m, p, hc, hv	[136]
5,6 dihydro-5-Azacytidine	hv	[137]
S110	p, hc	[138]
Cytotoxic agents (n = 23)
DNA alkylators
Busulfan	p, hv	[139]
Cisplatin	c	[140]
Streptozotosin	c	[141]
Nucleoside analogue
Cytosine arabinoside	c, p, hc, hv	[63]
Inosine monophosphate dehydrogenase inhibitors
Ribavirin	c	[142]
Mycophenolic acid	c	[142]
Tiazofulin	c	[142]
Ribonucleotide reductase inhibitors
Didox	c, m	141, 143
Hydroxyurea	c, m, p, hc, hv	See text
Resveratrol	hc	[144]
Trimidox	c	[141]
DNA intercalating agents
Aclarubicin	c	[145]
Chromomycin	c	[146]
Distamycin	c	[147]
Doxorubicin	c	[145]
Mithramycin	c, hv	[146]
Tallimustin	c	[147]
Psoralens ± UVA irradiation
Angelecin	c, hc	[148]
5-Methoxypsoralen	c, hc	[149]
Trimethyl angelicin	c, hc	[148]
Dihydrofolate reductase inhibitor
Methotrexate	p	[150]
Microtubule inhibitor
Vinblastine	p	151, 152
Protein synthesis inhibitor
Anisomycin	c	[105]
Short chain fatty acids and derivatives (n = 25)
Butyrate	c, m, p, hc, hv	See text
Phenyl butyrate	c, hc, hv	[73]
α-aminobutyric acid	c, p, hc	[70]
di-methylbutyric acid	c, m	[153]
Tributyrin	c	[154]
Acetate	c, m, p, hc	[155]
Phenylacetate	c, hv	[75]
Phenoxyacetic acid	c, m	[156]
Butyryl-hydroxamate	c, m, hc	[157]
α-Methylhydrocinnamic acid	c, m	[156]
Caproate	c	[158]
Heptanoic acid	hc	[159]
Hexanoic acid	hc	[159]
Isobutyramide	m, p, hc, hv	[78]
Nonanoic acic	hc	[159]
Octanoic acid	hc	[159]
Pentanoic acid	p, hc	[159]
Propionic acid	p, hc	[159]
Dimethoxyphenyl propionic acid	c	[153]
Propional hydroxamate	c, m, hc	[87]
RB7	c, hc	[134]
RB4, RB9, RB29	c	[160]
Valproic acid and derivatives	c, hc, hv	79, 80
Histone deacetylase inhibitors (n = 11)
Apicidin	c, hc	88, 89
Compounds “24” and “29”	c, hc	[161]
FK228	c, hc	[162]
Helminthsporium toxin	c, hc	[84]
MS-275	c, hc	[162]
SAHA (Vorinostat)	c, hc	87, 89
SBHA	c, hc	[87]
Scriptaid	c, hc	90, 163
Trapoxin	c, hc	84, 90
Trichostatin A	c, hc	84, 164
Imunomodulatory drugs (n = 3)
Thalidomide	hc	[93]
Revlimid	hc	[92]
Pomalidomide	hc	[92]
Hormonal agents (n = 2)
Nomegestrol	hv	[94]
Progesterone	hv	[95]
Cytokines (n = 3)
Erythropoietin	c, m, p, hc, hv	[96]
Stem cell factor	p, hc	97, 98, 99
TGF-β	hc	[100]
mTOR inhibitors (n = 2)
Rapamycin	c, hc	117, 118
Everolimus	c, hc	[119]
Miscellaneous (n = 4)
Zileuton (5-lipoxygenase inhibitor)	c	[165]
Vanadate (phophatase inhibitor)	hc	[166]
FG-2216 (HIF-prolyl hydroxylase inhib)	p, hc	[167]
CysNO (nitric oxide donor)	c, hc	[113]

c = immortalized erythroid cell line; hc = human, primary cell culture; hv = human, in vivo; m = murine model; mTOR = mammalian target of rapamycin; p = nonhuman primate, in vivo; TGF-β = tumor promoting graft factor-beta; UVA = ultraviolet A.

Selected references are included.

Defining “pharmacologic induction of fetal hemoglobin”

It has been appreciated for many years that HbF production is not limited to the fetal and early postnatal periods of human development. A median of 2.7% of normal adult human erythrocytes contain HbF, with an average of approximately 20% HbF per cell [51]. Thus, HbF comprises <1% of total Hb in normal adult blood. Studies of Hb production during in vitro differentiation of human erythrocytes demonstrated that γ-globin mRNA and HbF are preferentially produced in early erythroid precursors, while HbA is preferentially produced in more mature erythroid cells 52, 53. This effect is depicted in Figure 2A. The peak amount of γ-globin mRNA is roughly one-tenth that seen for β-globin mRNA. Figure 2B shows the effect of 5-Aza when applied daily during in vitro human primary cell differentiation [54]. This treatment not only increases the maximal steady-state level of γ-globin mRNA but also extends the period of high-level γ-globin mRNA to the later phases of erythropoiesis. This results in an overall increase in the γ-globin mRNA available for translation. We have seen similar effects with butyrate and decitabine (R.M., C.L., unpublished results). These results suggest that HbF-inducing agents do not act by reversing the developmental globin gene switch, as this process has sometimes been referred to, but by extending normal expression of the γ-globin genes to the later phases of adult erythropoiesis.

View Large Image
Download to PowerPoint

Figure 2 β-like globin gene expression during in vitro differentiation of human adult erythroid cells. (A) Expression patterns of the human γ- and β-globin genes during in vitro erythroid differentiation of primary human CD34+ peripheral blood cells. Note that the curves present normalized data to highlight the time course of gene expression. Absolute levels of γ-globin mRNA are roughly 10-fold less than for β-globin. (B) The effect of 5-azacytidine treatment on γ-globin mRNA levels during in vitro erythroid differentiation. Drug was added to 500 nM daily from days 10 to 20 of differentiation. Figures based on data originally published in Blood. Mabaera et al. Neither DNA hypomethylation nor changes in the Kinetics of erythroid differentiation explain 5-azacytidine's ability to induce human fetal hemoglobin. Blood. 2008;111:411–420. © the American Society of Hematology.

While the great majority of experiments on Hb induction have focused on the increase in γ-globin gene expression, there is also evidence that posttranscriptional processes are involved. For example, butyrate has been shown to increase translational efficiency of γ-globin mRNA [55], 5-Aza induces increases in HbF production that are disproportionately larger than increases in γ-globin mRNA steady-state levels [54] and evidence indicating that there are posttranscriptional mechanisms that inherently regulate the relative proportions of HbF and HbA during adult erythropoiesis has been presented 56, 57. These studies suggest that posttranscriptional mechanisms should also be considered in development of models to explain HbF induction.

Experimental systems for studying HbF induction

A variety of experimental systems have been used to study HbF induction and to identify new inducing agents. These have included immortalized cell lines, normal and genetically manipulated mice, primary cell cultures and in vivo nonhuman primate and human studies. While only certain nonhuman primates exhibit a fetal to adult developmental β-globin locus gene switch equivalent to that seen in humans, each of these approaches has produced insights into normal and induced globin gene regulation. One issue relevant to development of new HbF-inducing agents is which preclinical models are representative of HbF induction in humans. In the case of K562 cells, the most commonly used immortalized cell line, it is often not clear whether observed increases in γ-globin mRNA or HbF are the result of erythroid differentiation with the activation of many erythroid genes or are equivalent to the induction of γ-globin genes seen in differentiating human erythroblasts, which already express a wide array of erythroid genes. While only about 1 in 5 HbF-inducing agents has been tested in human in vivo experiments, each of these agents is also active in immortalized human cell lines (Table 1). This is also true for compounds tested in primary human in vitro erythroid differentiation assays, suggesting that while human primary cells and nonhuman primates represent the most appropriate models for definitive mechanistic and immediate preclinical studies, cell lines are still valuable for candidate drug evaluation and preliminary mechanistic experiments.

DNA methyltransferase inhibitors

Based on the ability of 5-Aza to induce DNA hypomethylation in tissue culture cells and the fact that the promoter regions of the fetal globin genes are hypermethylated in adult erythroid cells, Desimone and colleagues initiated studies demonstrating the ability of this agent to stimulate γ-globin gene expression and HbF production, first in baboons [29] and then in a person with β-thalassemia [30]. Despite encouraging follow-up clinical studies, concerns about potential carcinogenicity [58] caused enthusiasm for these agents to wane from the mid 1980s through the 1990s. Then, largely through the efforts of the Desimone group, a series of trials demonstrating the ability of decitabine to induce HbF in baboons and humans were performed [36]. These results and other factors, including studies suggesting that decitabine was not carcinogenic (reviewed in [36]), increased interest in drugs that affect the epigenome and the approval of 5-Aza and decitabine for the treatment of patients with myelodysplastic syndromes, all contributed to renewed interest in these agents as inducers of HbF.

Discussions concerning the mechanism of action of 5-Aza played a key role in the early development of the field and in development of HU as part of the current standard of care for SCD (reviewed in [59]). Because 5-Aza was a known inhibitor of the DNA methylation, Desimone and colleagues proposed that it caused demethylation of γ-globin promoter CpGs, leading to activation of these genes in differentiating adult erythroid cells [29]. Evidence for this hypothesis was based on the demonstration of 5-Aza–induced hypomethylation of a specific CpG located approximately 50-bp upstream of the transcriptional start site of the γ-globin genes [31] and subsequent experiments showing that demethylation of this site facilitated the binding of an activating protein complex 60, 61. Stamatoyannopoulos and colleagues 59, 62 provided an alternative hypothesis based on the cytotoxic properties of 5-Aza. They produced considerable evidence, based largely on in vivo treatment of human subjects and primates, that altered erythroid growth and differentiation kinetics were associated with increased fetal globin gene expression and HbF production 59, 62. This led them to propose that suppression of a cohort of differentiating erythroid cells by a cytotoxic agent resulted in a compensatory increase in the rate of differentiation following the removal of the drug. This more rapid differentiation was proposed to allow fetal globin gene expression to persist into the later phases of erythroid differentiation resulting in increased HbF production [62]. This hypothesis was supported by experiments showing that other cytotoxic agents, which were not DNMT inhibitors, including HU, also increased HbF production [63]. We have recently questioned both of these models, based on findings that HbF induction in primary human erythroid differentiation experiments could be achieved at doses of 5-Aza that do not alter erythroid differentiation kinetics, cell cycle distribution or global DNA methylation levels and that γ-globin promoter hypomethylation caused by DNMT1 short hairpin RNA was insufficient to induce HbF in primary human erythroblasts [54]. Additional evidence for the conclusion that global hypomethylation is not the primary effect of 5-Aza comes from an in vivo human study with this agent where induction of HbF and γ-globin promoter CpG hypomethylation were not accompanied by hypomethylation of other β-globin locus CpGs [64]. Another problem for this model, dating to early 5-Aza experiments, involves an inability to explain why genes other than γ-globin, whose promoters also become hypomethylated by 5-Aza treatments, are not induced [65].

Whatever the mechanism of the DNMT inhibitors, there is solid clinical evidence to support their status as key lead compounds in development of new HbF-inducing drugs. Important challenges include development of agents that are effective when administered orally, that do not cause significant side effects, including suppression of hematopoiesis, and do not require incorporation into DNA to achieve DNMT inhibition. Even if these obstacles are overcome, concerns over long-term use of drugs that can nonspecifically alter the genome-wide DNA methylation patterns of cycling cells will still need to be considered. The current interest in development of DNMT inhibitors for the treatment of cancer and the hemoglobinopathies should provide new agents that can be evaluated for HbF induction.

Other cytotoxic agents

The hypothesis that normal erythropoiesis was inhibited during 5-Aza administration resulting in a “stress erythropoiesis” response led to the testing of several other cytotoxic agents [62]. As noted previously, this hypothesis led directly to development of HU as an FDA-approved drug for SCD. Most of these agents are best known as cancer chemotherapy drugs. As shown in Table 1, when grouped together, these compounds form the second largest category of HbF-inducing agents. Yet, while they can be grouped together, they induce cytotoxicity through a wide variety of molecular mechanisms, including interfering with DNA, RNA, or protein synthesis, causing DNA damage through many different chemical modifications, and inhibiting of microtubule function. In fact, the great majority of HbF-inducing agents, including the DNMT inhibitors, histone deacetylase (HDAC) inhibitors, most short chain fatty acids (SCFA) and derivatives and the immunomodulatory drugs all cause cytotoxicity.

The hypothesis that altering the kinetics of erythroid differentiation increases HbF production has provided a common mechanistic explanation for in vivo baboon and human experiments, demonstrating the effectiveness of different cytotoxic HbF-inducing agents. However, this model is unable to explain the ability of these agents to induce HbF in vitro, where neither immortalized cell lines nor primary erythroid cultures experience the physiologic stimuli associated with interrupted red cell production that have been proposed to produce altered erythroid differentiation kinetics in vivo. Furthermore, the effects of 5-Aza on erythroid colony formation in marrow samples from several treated subjects found that suppression of colony formation was not necessary for induction of HbF [66]. Indeed, lower doses of 5-Aza caused an increase in colony formation and HbF induction. If the differentiation kinetics model is incorrect, this leaves us with no clear explanation for how most of the large number of agents primarily categorized as cytotoxins induce HbF. Because these drugs include mutagens and carcinogens and all produce clinical toxicities, only HU has been useful for the long-term treatment of β-hemoglobinopathy patients. However, the fact that these agents are effective in inducing fetal globin gene expression makes them important models for understanding the mechanisms underlying HbF induction.

Short-chain fatty acids

Observations that infants of diabetic mothers had extended periods of postnatal HbF production 67, 68 led Perrine and colleagues to perform experiments that identified butyrate as an agent that when administered to sheep, delayed the normal fetal to adult Hb switch [69]. In the same year it was shown that butyrate infusions increased HbF production in baboons [70]. This initial work lead to clinical trials in hemoglobinopathy patients that confirmed the ability of butyrate to increase HbF levels when administered as intravenous arginine butyrate 71, 72 or oral sodium phenylbutyrate 73, 74. Further development of butyrate has been limited by its suppressive effect on erythropoiesis 70, 75, the need to take an intolerably large amounts of the drug and the fact that not all patients responded 71, 72, 76. This has led to the in vitro testing of many other SCFAs and related compounds for their ability to induce HbF (Table 1). While many of these agents stimulate HbF production in model systems only isobutyramide 77, 78 and valproic acid 79, 80 have been evaluated in small human trials.

Based on experiments that predated the identification of butyrate as an inducer of HbF 81, 82, 83, it was initially proposed that this and related SCFAs induced fetal globin gene expression through HDAC inhibition, resulting in global histone hyperacetylation, including nucleosomes at the γ-globin promoters [69]. While increased histone acetylation has been demonstrated following SCFA treatment [84], Boosalis et al. showed that not all SCFA inducers of HbF cause global histone hyperacetylation [85]. In the same article, it was also shown that while butyrate causes p21 induction and cell-cycle arrest, other effective SCFA derivatives do not and actually stimulate erythropoiesis. Recently, Fathalla et al. [86] showed that while butyrate treatment did cause increased acetylation of the γ-globin promoter histones, it also decreased β-globin promoter histone acetylation, again raising the possibility that more specific mechanisms other than global histone hyperacetylation mediate SCFA induction of HbF [86]. Other experiments have shown that butyrate can cause a rapid increase in the association of γ-globin mRNA with ribosomes [55]. As will be discussed in more detail here, other authors have demonstrated activation of p38 mitogen-activated protein kinase (MAPK) and cyclic nucleotide signaling pathways in association with butyrate induction of HbF. Taken together, these studies suggest that global histone hyperacetylation induced by HDAC inhibition is not the primary mechanism underlying SCFA stimulation of HbF.

Histone deacetylase inhibitors

Based on the theory that butyrate induces γ-globin gene expression through HDAC inhibition, McCaffrey et al. [84], tested three known inhibitors of these enzymes (trichostatin A, trapoxin, and HA toxin) and found that each induced γ-globin gene expression in primary erythroid cell cultures [84]. Since that time, many additional HDAC inhibitors have been reported to induce HbF in a variety of experimental systems (Table 1). In several cases, HDAC inhibition and/or histone hyperacetylation have been verified 87, 88, 89. However, here too, other evidence suggests that more specific mechanisms, beyond generalized histone hyperacetylation, may be responsible for HbF induction. Cao et al. [90] studied a wide variety of derivatives of trichostatin, trapoxin, and scriptaid and found a lack of correlation between the potency of HDAC inhibition and γ-globin gene induction [90] and as discussed here, multiple reports have demonstrated that pharmacologic inhibition of p38 MAPK activation prevents induction of HbF by HDAC inhibitors.

So far, none of the HDAC inhibitors have been evaluated in human subjects for their ability to induce HbF, but several are active in human primary erythroid differentiation systems (Table 1). Because the first FDA-approved HDAC inhibitor (Vorinostat or SAHA) has in vitro HbF-inducing activity [87] and because many other HDAC inhibitors, including agents targeting specific HDAC enzymes, are being developed and tested in early clinical trials for cancer [91], this is a promising class of compounds for further development as HbF induction agents. However, if these agents do enter clinical trials, it will be important to consider what long-term side effects are associated with drugs that have the potential to alter global histone acetylation patterns of all exposed cells.

Other inducing agents

We have categorized most of the HbF-inducing agents based on their generally accepted mechanisms of action. While a majority of the drugs fit into these categories, several do not. One example is the so-called immunomodulatory drugs. These include thalidomide and lenalidomide, two drugs that are FDA approved for treatment of myelodysplastic syndrome and multiple myeloma. The mechanisms underlying these anticancer effects remain to be fully elucidated. Within the last year both of these drugs and pomalidomide, another immunomodulatory drug currently undergoing development, have been shown to induce HbF production in K562 or primary human erythroid cultures. One study correlated this activity with locally increased γ-globin promoter histone acetylation [92] while the other implicated reactive oxygen species (ROS) production and p38 MAPK activation [93].

Two reports have suggested that progestins may be active inducers of fetal hemoglobin. Nomegestrol, when administered as an implantable birth-control device to a group of women with SCD, decreased the frequency of painful crises and increased the percentage of F cells and HbF levels compared to an untreated control group [94]. Another report showed that progesterone, when used in a two-phase primary erythroid differentiation system, increased γ-globin steady-state mRNA levels but not HbF production [95].

Erythropoietin was among the first agents to be shown to increase HbF production and has been tested in β-hemoglobinopathy patients [96]. Stem cell factor (kit ligand) and transforming growth factor-β have both been shown to induce HbF production during in vitro differentiation of primary human cells 97, 98, 99, 100. Additional agents are listed in Table 1.

Role of cell signaling in HbF induction

Traditionally, most HbF-inducing agents have been viewed as affecting γ-globin gene expression by altering local promoter chromatin structure (DNMT and HDAC inhibitors including SCFAs) or by altering the kinetics of erythroid differentiation (cytotoxic agents). However, during the past several years, a few groups have begun to view PIFH from a different vantage point—that of cell signaling. Implicated pathways have included those involving cyclic guanosine monophosphate (cGMP) and cyclic adenosine monophosphate (cAMP), nitric oxide (NO), p38 MAPK, ROS, and cytokine signaling. These studies are summarized in Table 2.

Table 2.

Intracellular signaling pathways implicated in γ-globin gene and/or fetal hemoglobin induction


Agent	Class	Implicated pathways	System	References
5-Azacytidine	DNMT inhibitor	cAMP	hc	[109]
Hydroxyurea	Cytotoxic	cAMP	hc	[109]
		NO/cGMP	c, hc	112, 113
		P38 MAPK	c	[115]
Anisomycin	Cytotoxic	P38 MAPK	c	[105]
Butyrate	SCFA	P38 MAPK	c, hc	101, 105
		ROS/p38 MAPK	c, hc	[104]
		cAMP	c	[109]
Valproate	SCFA	P38 MAPK	c	[102]
Apicidin	HDAC	P38 MAPK	c	[89]
Scriptaid	HDAC	P38 MAPK	c, hc	[163]
Trichostatin A	HDAC	ROS/P38 MAPK	c, hc	104, 105
Thalidomide	IMiD	ROS/P38 MAPK	hc	[93]
CysNO	NO donor	NO, cGMP	c, hc	[113]
Rapamycin	mTOR inhibitor	mTOR	c, hc	117, 118

c = immortalized erythroid cell line; cAMP = cyclic adenosine monophosphate; cGMP = cyclic guanosine monophosphate; DNMT = DNA methyltransferase; hc = human primary erythroid cell culture; HDAC = histone deacetylase; IMiD = immunomodulatory drugs; MAPK = mitogen-activated protein kinase; mTOR = mammalian target of rapamycin; NO = nitric oxide; SCFA = short chain fatty acid; ROS = reactive oxygen species.

Witt and colleagues may have provided the first evidence that intracellular signaling is involved in HbF induction by showing that butyrate induction of erythroid differentiation and Hb production in K562 cells was associated with increased p38 MAPK phosphorylation and that pharmacologic inhibition of p38 activity with the drug SB203580 prevented this action [101]. Additional work from this group has demonstrated that HbF induction in K562 cells by the HDAC inhibitor apicidin and the SCFA valproic acid is inhibited by SB203580 89, 102.

The Pace lab has studied signaling events involved with HDAC inhibitors, including butyrate and other SCFAs, trichostatin, and scriptaid. In a series of publications, they provided evidence for a model of HbF induction by HDAC inhibitors in which these agents not only cause histone hyperacetylation, but also lead to production of ROS that, in turn, cause p38 MAPK phosphorylation and downstream activation of the cAMP response element binding protein (CREBP) and activating transcription factor 2 (ATF2) transcriptional activator proteins 103, 104. Evidence in support of this model includes the ability of inhibitors of ROS production to decrease γ-globin induction [104], demonstration of p38 phosphorylation in response to inducing agents and inhibition of induction by SB203580 [105], the ability of constitutively active forms of MKK3 and MKK6 (upstream activators of p38 MAPK) to independently increase γ-globin gene expression [105] and demonstration that CREB and ATF2 bind an element in the γG-globin upstream promoter following HDAC inhibitor treatment [106].

Recently, Aerbajinai et al. [93] reported the ability of thalidomide to increase γ-globin gene expression and the proportion of HbF-containing cells in a human in vitro erythroid differentiation system [93]. They presented evidence for a model similar to that proposed by the Pace lab, in which thalidomide induces production of ROS that in turn cause p38 MAPK phosphorylation and globally increased histone H4 acetylation. Pharmacologic inhibition of either ROS accumulation or p38 MAPK activation lessened the increase in γ-globin mRNA in response to thalidomide.

In 2001, Ikuta and colleagues [107] reported that in K562 cells, activation of the cGMP/soluble guanylate cyclase(sGC)/PKG(cGMP-dependent protein kinase) pathway induced γ-globin gene expression and that inhibition of this pathway prevented butyrate induction of γ-globin expression [107]. Further work in K562 cells led them to propose opposing effects for cGMP (positive) and cAMP (negative) signaling in γ-globin induction [108]. However, subsequent publications by Keefer et al. [109] and the Ikuta group identified cAMP signaling as being of greatest importance for HbF induction by agents including HU, butyrate, and 5-Aza in in vitro human primary erythroid differentiation experiments 109, 110. Most recently, Bailey et al. [111] reported that pharmacologic manipulations that increase cAMP stimulate HbF production in primary cells from β-thalassemia intermedia patients and that intracellular cAMP and phosphorylated CREB levels correlate with HbF in individual patients.

Work from the Schechter laboratory has provided evidence that HU is converted to NO in vivo which then chemically modifies the deoxy-heme form of sGC, increasing its activity and thereby the production of cGMP within differentiating erythroid cells [112]. Pharmacologic inhibition of sGC and the downstream production of cGMP prevented HU induction of HbF, thus providing a mechanistic link between NO, cGMP, and HbF induction [113]. Work by Gladwin et al. [114] demonstrated increased intraerythrocyte levels of NO following HU administration in patients. Other experiments have shown that SB203580 inhibition of p38 MAPK activation prevents HU induced γ-globin gene expression in K562 cells [115].

Additional evidence for involvement of cell signaling in HbF induction includes studies of the protein synthesis inhibitor anisomycin implicating p38 MAPK [105] and of stem cell factor and transforming growth factor-β implicating MEK and SMAD pathways [116] and of inhibitors of the mammalian target of rapamycin (mTOR) pathway 117, 118, 119. Overall, these experiments present a body of evidence that suggests an important role for intracellular signaling in HbF induction.

A unified model of HbF induction

Most current mechanistic models of HbF induction are based on what are thought to be the primary actions of specific classes of inducing agents, e.g., global DNA hypomethylation induced by DNMT inhibitors or global histone hyperacetylation induced by HDAC inhibitors, including SCFA derivatives. While each of these models has strengths, they are unable to account for key experimental results, including the fact that 5-Aza can induce HbF without altering global DNA methylation, that γ-globin promoter hypomethylation is insufficient to induce gene expression and that the ability of HDAC inhibitors to induce HbF does not correlate with potency of HDAC inhibition. An additional problem for current models is that many effective inducing agents are not DNMT or HDAC inhibitors but appear to act through a wide variety of cytotoxicity related mechanisms. Any model of HbF induction must explain how such a functionally diverse group of compounds can all achieve the same result. These considerations have led us to propose a new model of HbF induction that integrates recent cell-signaling data with the stress erythropoiesis model. In the original version of the stress model, cytotoxic drugs were thought to produce a more rapid than normal post-exposure differentiation of erythroid cells, resulting in increased HbF production [62]. In our updated model, we propose that while the emphasis on “stress” was appropriate, the key effect of most HbF-inducing agents is the activation of cell stress signaling pathways that augment γ-globin gene expression and HbF production during adult erythropoiesis. This model is pictured in Figure 3.

View Large Image
Download to PowerPoint

Figure 3 Proposed cell stress signaling model of fetal hemoglobin (HbF) induction. Under this model, many different cellular stresses, including those caused by HbF-inducing drugs activate coordinated stress responses, which include γ-globin gene activation. Potential pathways involved in these responses include the integrated stress response (also known as unfolded protein response), p38 mitogen-activated protein kinase (MAPK) and cyclic adenosine monophosphate (cAMP) signaling pathways. Blue ovals indicate examples of HbF-inducing agents that have been shown to act, at least in part, through the p38 MAPK stress signaling pathway. Solid red ovals indicate pathway members that have been experimentally implicated in HbF induction. Dashed red circles indicate factors that are involved in erythropoiesis but have not been directly linked to HbF induction. See text for details. ATF = activating transcription factor; CHOP = CCAAT/enhancer binding protein homologous protein; CREB = cAMP response element binding protein; EIF2A = eukaryotic translation initiation factor 2α; ELK1 = member of ets oncogene family; ER = endoplasmic reticulum; GADD = growth arrest- and DNA damage-inducible gene; GCN2 = general control nonderepressible-2 (EIF2A kinase 4); HRI = heme-regulated inhibitor (EIF2A kinase 1); HU = hydroxyurea; MAX = MYC-associated factor X; MEF2 = MADS box transcription enhancer factor 2; MKK = mitogen-activated protein kinase kinase; MSK1 = mitogen and stress-activated protein kinase 1; MYC = v-MYC avian myelocytomatosis viral homolog; NRF2 = nuclear factor erythroid 2-like 2; PERK = RNA dependent protein kinase-like ER kinase (EIF2A kinase 3); PKA = protein kinase, cAMP-dependent, regulatory, type I α; PKR = RNA-dependent protein kinase (EIF2A kinase 2); ROS = reactive oxygen species; UV = ultraviolet radiation.

Because the p38 MAPK signaling pathway has been implicated in the actions of several inducing agents, we have placed it at the center of our model. This pathway is one of three related signaling cascades that are grouped together based on structural similarities of the key MAPK molecules. Besides p38 MAPK, these include the c-Jun N-terminal kinase and extracellular signal-regulated kinase pathways. These kinases integrate multiple upstream signals, including DNA damage, oxidative stress (ROS), heat shock, osmotic shock, NO, and inhibition of protein synthesis and then activate the downstream kinases and transcription factors that mediate coordinated responses to these stimuli (reviewed in 120, 121). Depending on the degree of stress, these pathways can promote cell survival or cell death by apoptosis. In contrast to p38 MAPK signaling, most studies have found that the c-Jun N-terminal kinase and extracellular signal-regulated kinase pathways either do not play a major role in HbF induction 89, 111, or in the case of extracellular signal-regulated kinase signaling, inhibit HbF expression 101, 102, 122, 123. As summarized in Table 2 and Figure 3, evidence has been presented that several inducing agents including HU, SCFAs, HDAC inhibitors, anisomycin (an inhibitor of protein synthesis), and an immunomodulatory drugs induce HbF through p38 MAPK signaling. Several other inducers of HbF, including busulfan [124], cytosine arabinoside [125], HU [126], and NO [126] activate p38 MAPK signaling in other experimental systems. Recent data from our lab indicate that γ-globin mRNA and HbF induction in human primary erythroid cultures by 5-Aza is prevented by the p38 MAPK inhibitor SB203580 (Mabaera et al, unpublished results.) ATF2 and CREB, two factors activated by p38 MAPK signaling, have been implicated in HDAC induction of γ-globin expression in K562 cells [106]. Many of the other transcription factors involved in cell stress signaling also belong to the CREB/ATF family. This is a group of related basic leucine zipper proteins that are activated by phosphorylation and can form homo- and heterodimers and bind sequences similar to the cAMP response element [127]. It is possible that multiple members of this family can increase γ-globin gene expression by binding genetic elements that regulate γ-globin gene expression.

Our model also offers a potential mechanistic link between observations showing that cAMP-mediated signaling in primary erythroid cell cultures can modulate HbF induction, as CREB is activated, not only by the p38 MAPK pathway, but also directly by cAMP-activated protein kinase A (Fig. 3). While the role of cGMP in HbF induction is less clear, studies of HU in human primary cells support a role for this signaling system [113]. cGMP is known to cross-talk with the cAMP pathway by downregulating phosphodiesterase 3, resulting in increased cAMP levels [128]. This interaction has been found to be important in other systems [129] and could offer an explanation for cGMP's role in HbF induction.

Finally, recent publications have demonstrated the importance of what has been termed the unfolded protein response or integrated stress response pathway in erythroid cells [130]. This pathway also responds to a variety of stress stimuli, including viral infection, NO, heat shock, ROS, endoplasmic reticulum stress, ultraviolet irradiation, proteosome inhibition, inadequate nutrients and, in erythroid cells, limiting amounts of heme [131]. Depending on the type of stress, distinct members of a family of four related kinases phosphorylate eukaryotic initiation factor 2 α (EIF2A), which in turn causes a general inhibition of mRNA translation, thus limiting production of misfolded proteins during heat shock or hemoglobin when inadequate heme is available [130]. However, this is not an absolute inhibition of protein production as translation of the transcription factor ATF4 is increased under the same stress conditions, leading to a secondary transcriptional response affecting other transcription factors including ATF3, growth arrest- and DNA damage-inducible gene (GADD) 34, and CCAAT/enhancer binding protein homologous protein, which in turn activate additional downstream genes leading to induction of corrective pathways or apoptosis [131]. GADD34 provides feedback to reactivate global translation. Targeted disruption of the ATF4 gene in mice produces a severe fetal anemia [132], while disruption of the GADD34 gene decreases the efficiency of globin mRNA translation [133]. While there is little direct experimental evidence to suggest that the unfolded protein response/integrated stress response pathway is involved in HbF induction, it is included because it is functionally important in erythroid precursor cells, it is activated by stresses caused by known inducers of HbF, including NO and ROS, because it activates some of the same transcription factors as the p38 MAPK pathway and because it offers a possible explanation for proposed translational effects of HbF inducers.

Implications of a stress signaling model of HbF induction

Our model offers the first unifying theory of HbF induction since the original erythropoietic stress model. It predicts that cell stress caused by a variety of diverse compounds and stimuli will activate similar response genes—including the γ-globin genes. It also provides explanations for several observations not accounted for by previous models. These include the ability of the many functionally diverse inducing agents that are cytotoxic but do not alter global DNA methylation or histone acetylation to increase HbF production; the ability of 5-Aza to induce HbF production without causing global DNA hypomethylation or changes in the kinetics of differentiation [54]; the inability of global γ-globin promoter DNA hypomethylation induced by DNMT1 knock-down to increase fetal globin gene expression or HbF production [54]; the ability of drugs to induce HbF in vitro where no postexposure acceleration of differentiation kinetics occurs; the reported lack of correlation between potency of HDAC inhibition and HbF induction for several HDAC inhibitors [90]; the ability of SCFA derivatives that do not alter global histone acetylation to induce HbF [85]; and the ability of inhibitors of p38 MAPK to block γ-globin and HbF induction by several unrelated compounds (Table 2).

The stress signaling model offers several testable predictions. These include the prediction that most inducing agents will activate cell stress signaling and that this signaling is required for their action. We have recently observed, e.g., that 5-Aza induction of γ-globin gene expression and HbF production in human primary erythroid cells is associated with p38 MAPK phosphorylation and that induction is inhibited by the p38 MAPK inhibitor SB203580 (R.M., C.L., unpublished results). Another prediction is that several members of stress signaling pathways, from the sensors of each stress to the activated transcription factors that bind to the γ-globin promoters, are required for γ-globin gene and HbF induction. The requirement for specific factors can be tested through pharmacologic manipulation of protein production or function. The model also predicts that cellular stresses beyond chemical agents, such as ultraviolet or x-ray irradiation or protein misfolding will also induce HbF if these pathways are intact in differentiating adult erythroid precursors. Another prediction is that γ-globin–inducing agents or stresses other than DNMT inhibitors and HDAC inhibitors should also cause γ-globin promoter hypomethylation and histone hyperacetylation. The predicted promoter demethylation has recently been demonstrated for butyrate [86].

Our model also offers an explanation for clinical observations that the blood counts of patients receiving inducing agents including 5-Aza, decitabine, HU, and butyrate must be closely monitored to prevent the development of low blood counts. Under the model, the dose of inducing agents must be high enough to activate stress signaling but not so high as to induce levels of cell-cycle arrest or apoptosis in hematopoietic precursor cells that cause dangerously low blood counts. Our model also provides alternative explanations for observations that have been used to support previous models of HbF induction. For example, the γ-globin promoter hypomethylation seen with nucleoside analogue DNMT inhibitors can be viewed, not as a result of DNMT inhibition and global DNA hypomethylation, but as being caused by the binding of stress-activated transcription factors that protect the promoter from CpG methylation. Similarly, the increased γ-globin promoter acetylation seen with HDAC inhibitors and SCFAs can be explained by the binding of stress-induced transcriptional activating complexes that contain histone acetyl transferases and/or displace HDACs [134]. Finally, for the original stress erythropoiesis model, while there is strong evidence to support an association between suppression of erythropoiesis, in vivo alteration in differentiation kinetics and HbF induction, it may be that this association is due to activation of stress signaling which can induce γ-globin gene expression and, at higher doses, cell-cycle arrest or apoptosis.

Viewing PIFH as a cell signaling phenomenon also has important implications for development of targeted HbF-inducing agents. Most targeted drug therapies have been based on designing compounds that alter the function of cell signaling molecules. Thus, if our model is correct, it should be amenable to targeted drug development using standard approaches. There is, however, one exception to this point. In most cases of targeted drug development, agents are designed that inhibit specific signaling pathways. In the case of HbF induction, what may be required are agents that activate stress signaling in the absence of actual DNA damage or other cell stresses. One way to achieve this might be through the activation of specific stress sensors, not by the stress they are design to detect, but by pharmacologic agonists. Another approach would be to activate distal portions of the signaling pathways that include γ-globin gene activation but not other portions of cell stress responses. It would also be desirable that such activation not involve DNA damage, global epigenomic changes, or suppression of erythropoiesis. While this is likely to represent a challenging problem in drug development, the Perrine group has identified active SCFA derivatives that induce HbF without HDAC inhibition or suppression of erythropoiesis 85, 134.

While our model offers potential explanations for several previously unexplained results, it remains to be critically tested. Verification of the model will require demonstration of a complete series of molecular events leading from a primary drug effect to the activation of γ-globin gene expression and increased HbF production. Much of the data that led to our hypothesis were produced using immortalized cell lines. Confirmatory experiments will need to be performed in human primary erythroid cells. Similarly, experiments involving the use of pharmacologic inhibitors of pathways should be verified by more specific experiments. For example, SB203580 is a potent inhibitor of the α and β isoforms of p38 MAPK that has been used in thousands of publications, including several implicating p38 MAPK signaling in HbF induction. As for many other kinase inhibitors, recent studies suggest that this compound also inhibits other signaling molecules [135]. Pathway and target verification can be pursued through the use of strategies, including knockdown of specific factors by small interfering RNA or short hairpin RNA or by the expression of dominant negative molecules or mutant signaling proteins that are resistant to drug inhibition [135].

Summary

We have reviewed the pharmacologic agents reported to induce human HbF production and their proposed mechanisms of action. We find that for many agents complete mechanisms have not been proposed and that for other agents, experimental evidence leads us to question mechanisms of action involving global changes in DNA methylation or histone acetylation or alterations in erythroid differentiation kinetics. The fact that most HbF-inducing agents are cytotoxic and many activate the p38 MAPK cell stress signaling pathway has led us to propose that this and other stress-related pathways may be the keys to understanding HbF induction. It is our hope that this review will stimulate discussion, experimentation, and development of improved agents for HbF induction so that erythroid precursors of hemoglobinopathy patients will not have to be stressed “to make their red cells stronger.”

Acknowledgments

This work was supported by National Institutes of Health grants HL52243 and HL73442 to C.H.L. and by funding from the Knights of the York Cross of Honour. The authors thank Drs. David Bodine, Steven Fiering, and Alan Eastman for critical reading of this manuscript. We have attempted to provide a comprehensive review of pharmacologic induction of fetal hemoglobin. We apologize to any investigators whose work was inadvertently not included.

References

1. 1Weatherall DJ, JGGL , Warrell DA. In: Weatherall DJ, Ledingharm JGG, Warrell DA editor. Oxford textbook of medicine. 3rd ed. Oxford: Oxford University Press; 1996;.

2. 2Nagel RL, Roth EF. Malaria and red cell genetic defects. Blood. 1989;74:1213–1221. MEDLINE

3. 3Aidoo M, Terlouw DJ, Kolczak MS, et al. Protective effects of the sickle cell gene against malaria morbidity and mortality. Lancet. 2002;359:1311–1312. Abstract | Full Text | Full-Text PDF (68 KB) | CrossRef

4. 4Ingram VM. Gene mutations in human haemoglobin: the chemical difference between normal and sickle cell haemoglobin. Nature. 1957;180:326–328. MEDLINE | CrossRef

5. 5Frenette PS, Atweh GF. Sickle cell disease: old discoveries, new concepts, and future promise. J Clin Invest. 2007;117:850–858. MEDLINE | CrossRef

6. 6Platt OS, Brambilla DJ, Rosse WF, et al. Mortality in sicskle cell disease. Life expectancy and risk factors for early death. N Engl J Med. 1994;330:1639–1644. MEDLINE | CrossRef

7. 7Sickle-cell anaemia. Report by the Secretariat Fifty-Ninth World Health Assembly. Geneva: World Health Organization; 2006;p. 1–5.

8. 8Patrinos GP, Giardine B, Riemer C, et al. Improvements in the HbVar database of human hemoglobin variants and thalassemia mutations for population and sequence variation studies. Nucleic Acids Res. 2004;32:D537–D541.

9. 9Schrier SL. Pathophysiology of thalassemia. Curr Opin Hematol. 2002;9:123–126. MEDLINE | CrossRef

10. 10Gardenghi S, Marongiu MF, Ramos P, et al. Ineffective erythropoiesis in beta-thalassemia is characterized by increased iron absorption mediated by down-regulation of hepcidin and up-regulation of ferroportin. Blood. 2007;109:5027–5035. MEDLINE | CrossRef

11. 11Delea TE, Edelsberg J, Sofrygin O, et al. Consequences and costs of noncompliance with iron chelation therapy in patients with transfusion-dependent thalassemia: a literature review. Transfusion. 2007;47:1919–1929. CrossRef

12. 12Bhatia M, Walters MC. Hematopoietic cell transplantation for thalassemia and sickle cell disease: past, present and future. Bone Marrow Transplant. 2008;41:109–117. CrossRef

13. 13Shenoy S. Has stem cell transplantation come of age in the treatment of sickle cell disease?. Bone Marrow Transplant. 2007;40:813–821. CrossRef

14. 14Lucarelli G, Gaziev J. Advances in the allogeneic transplantation for thalassemia. Blood Rev. 2008;22:53–63. | CrossRef

15. 15Vermylen C, Cornu G, Ferster A, et al. Haematopoietic stem cell transplantation for sickle cell anaemia: the first 50 patients transplanted in Belgium. Bone Marrow Transplant. 1998;22:1–6. MEDLINE

16. 16Bernaudin F, Socie G, Kuentz M, et al. Long-term results of related myeloablative stem-cell transplantation to cure sickle cell disease. Blood. 2007;110:2749–2756. CrossRef

17. 17Walters MC, Storb R, Patience M, et al. Impact of bone marrow transplantation for symptomatic sickle cell disease: an interim report. Multicenter investigation of bone marrow transplantation for sickle cell disease. Blood. 2000;95:1918–1924. MEDLINE

18. 18Walters MC, Patience M, Leisenring W, et al. Barriers to bone marrow transplantation for sickle cell anemia. Biol Blood Marrow Transplant. 1996;2:100–104. MEDLINE

19. 19Sadelain M. Recent advances in globin gene transfer for the treatment of beta-thalassemia and sickle cell anemia. Curr Opin Hematol. 2006;13:142–148. MEDLINE

20. 20Bank A, Dorazio R, Leboulch P. A phase I/II clinical trial of beta-globin gene therapy for beta-thalassemia. Ann N Y Acad Sci. 2005;1054:308–316. MEDLINE | CrossRef

21. 21Dunbar CE. The yin and yang of stem cell gene therapy: insights into hematopoiesis, leukemogenesis, and gene therapy safety. Hematology Am Soc Hematol Educ Program. 2007;2007:460–465.

22. 22Hargrove PW, Kepes S, Hanawa H, et al. Globin lentiviral vector insertions can perturb the expression of endogenous genes in beta-thalassemic hematopoietic cells. Mol Ther. 2008;16:525–533. CrossRef

23. 23Wood WG. Haemoglobin synthesis during human fetal development. Br Med Bull. 1976;32:282–287. MEDLINE

24. 24Conley CL, Weatherall DJ, Richardson SN, Shepard MK, Charache S. Hereditary persistence of fetal hemoglobin: a study of 79 affected persons in 15 Negro families in Baltimore. Blood. 1963;21:261–281. MEDLINE

25. 25Edington GM, Lehmann H. Expression of the sickle-cell gene in Africa. Br Med J. 1955;1:1308–1311. MEDLINE

26. 26Goldberg MA, Husson MA, Bunn HF. Participation of hemoglobins A and F in polymerization of sickle hemoglobin. J Biol Chem. 1977;252:3414–3421. MEDLINE

27. 27Nathan DG, Gunn RB. Thalassemia: the consequences of unbalanced hemoglobin synthesis. Am J Med. 1966;41:815–830. Full-Text PDF (6459 KB) | CrossRef

28. 28Papayannopoulou TH, Brice M, Stamatoyannopoulos G. Stimulation of fetal hemoglobin synthesis in bone marrow cultures from adult individuals. Proc Natl Acad Sci U S A. 1976;73:2033–2037. MEDLINE | CrossRef

29. 29DeSimone J, Heller P, Hall L, Zwiers D. 5-Azacytidine stimulates fetal hemoglobin synthesis in anemic baboons. Proc Natl Acad Sci U S A. 1982;79:4428–4431. MEDLINE | CrossRef

30. 30Ley TJ, DeSimone J, Anagnou NP, et al. 5-azacytidine selectively increases gamma-globin synthesis in a patient with beta+ thalassemia. N Engl J Med. 1982;307:1469–1475. MEDLINE | CrossRef

31. 31Ley TJ, Anagnou NP, Noguchi CT, et al. DNA methylation and globin gene expression in patients treated with 5-azacytidine. Prog Clin Biol Res. 1983;134:457–474. MEDLINE

32. 32Dover GJ, Charache SH, Boyer SH, Talbot CC, Smith KD. 5-Azacytidine increases fetal hemoglobin production in a patient with sickle cell disease. Prog Clin Biol Res. 1983;134:475–488. MEDLINE

33. 33Koshy M, Dorn L, Bressler L, et al. 2-deoxy 5-azacytidine and fetal hemoglobin induction in sickle cell anemia. Blood. 2000;96:2379–2384. MEDLINE

34. 34Saunthararajah Y, Molokie R, Saraf S, et al. Clinical effectiveness of decitabine in severe sickle cell disease. Br J Haematol. 2008;141:126–129. CrossRef

35. 35Lowrey CH, Nienhuis AW. Brief report: treatment with azacitidine of patients with end-stage beta-thalassemia. [see comments] N Engl J Med. 1993;329:845–848. MEDLINE | CrossRef

36. 36Saunthararajah Y, Lavelle D, DeSimone J. DNA hypo-methylating agents and sickle cell disease. Br J Haematol. 2004;126:629–636. MEDLINE | CrossRef

37. 37Fathallah H, Atweh GF. Induction of fetal hemoglobin in the treatment of sickle cell disease. Hematolpgy Am Soc Hematol Educ Program. 2006;58–62.

38. 38Steinberg MH, Lu ZH, Barton FB, Terrin ML, Charache S, Dover GJ. Fetal hemoglobin in sickle cell anemia: determinants of response to hydroxyurea. Multicenter Study of Hydroxyurea. Blood. 1997;89:1078–1088. MEDLINE

39. 39Fucharoen S, Siritanaratkul N, Winichagoon P, et al. Hydroxyurea increases hemoglobin F levels and improves the effectiveness of erythropoiesis in beta-thalassemia/hemoglobin E disease. Blood. 1996;87:887–892. MEDLINE

40. 40Zeng YT, Huang SZ, Ren ZR, et al. Hydroxyurea therapy in beta-thalassaemia intermedia: improvement in haematological parameters due to enhanced beta-globin synthesis. Br J Haematol. 1995;90:557–563. MEDLINE | CrossRef

41. 41Hajjar FM, Pearson HA. Pharmacologic treatment of thalassemia intermedia with hydroxyurea. J Pediatr. 1994;125:490–492. Abstract | Full-Text PDF (244 KB) | CrossRef

42. 42Charache S, Terrin ML, Moore RD, et al. Effect of hydroxyurea on the frequency of painful crises in sickle cell anemia. Investigators of the Multicenter Study of Hydroxyurea in Sickle Cell Anemia. N Engl J Med. 1995;332:1317–1322. MEDLINE | CrossRef

43. 43Schultz WH, Ware RE. Malignancy in patients with sickle cell disease. Am J Hematol. 2003;74:249–253. MEDLINE | CrossRef

44. 44Steinberg MH, Barton F, Castro O, et al. Effect of hydroxyurea on mortality and morbidity in adult sickle cell anemia: risks and benefits up to 9 years of treatment. JAMA. 2003;289:1645–1651. MEDLINE | CrossRef

45. 45Debaun MR, Field JJ. Limitations of clinical trials in sickle cell disease: a case study of the Multi-Center Study of Hydroxyurea (MSH) Trial and the Stroke Prevention (STOP) trial. Hematology Am Soc Hematol Educ Program. 2007;2007:482–488.

46. 46Lanzkron S, Haywood C, Segal JB, Dover GJ. Hospitalization rates and costs of care of patients with sickle-cell anemia in the state of Maryland in the era of hydroxyurea. Am J Hematol. 2006;81:927–932. MEDLINE | CrossRef

47. 47Ritho JN, Mayhew DY, Hartzema AG, Liu H, Lottenberg R. Hydroxyurea use in sickle cell disease patients in a Florida Medicaid population. [abstract] Blood. 2007;110:Abstract 79.

48. 48Aliyu ZY, Babadok A. Hydroxyurea utilization in Nigeria, a lesson in public health. [abstract] Blood. 2007;110:Abstract 80.

49. 49Chaulk CP, Kazandjian VA. Directly observed therapy for treatment completion of pulmonary tuberculosis: Consensus Statement of the Public Health Tuberculosis Guidelines Panel. JAMA. 1998;279:943–948. MEDLINE | CrossRef

50. 50Laurent C, Diakhate N, Gueye NF, et al. The Senegalese government's highly active antiretroviral therapy initiative: an 18-month follow-up study. AIDS. 2002;16:1363–1370. MEDLINE | CrossRef

51. 51Wood WG, Stamatoyannopoulos G, Lim G, Nute PE. F-cells in the adult: normal values and levels in individuals with hereditary and acquired elevations of Hb F. Blood. 1975;46:671–682. MEDLINE

52. 52Papayannopoulou T, Kalmantis T, Stamatoyannopoulos G. Cellular regulation of hemoglobin switching: evidence for inverse relationship between fetal hemoglobin synthesis and degree of maturity of human erythroid cells. Proc Natl Acad Sci U S A. 1979;76:6420–6424. MEDLINE | CrossRef

53. 53Farquhar MN, Turner PA, Papayannopoulou T, Brice M, Nienhuis AW, Stamatoyannopoulos G. The asynchrony of gamma- and beta-chain synthesis during human erythroid cell maturation. III. gamma- and beta-mRNA in immature and mature erythroid clones. Dev Biol. 1981;85:403–408. MEDLINE | CrossRef

54. 54Mabaera R, Greene MR, Richardson CA, Conine SJ, Kozul CD, Lowrey CH. Neither DNA hypomethylation nor changes in the kinetics of erythroid differentiation explain 5-azacytidine's ability to induce human fetal hemoglobin. Blood. 2008;111:411–420. CrossRef

55. 55Weinberg RS, Ji X, Sutton M, et al. Butyrate increases the efficiency of translation of gamma-globin mRNA. Blood. 2005;105:1807–1809. MEDLINE | CrossRef

56. 56Dover GJ, Boyer SH. Fetal hemoglobin-containing cells have the same mean corpuscular hemoglobin as cells without fetal hemoglobin: a reciprocal relationship between gamma- and beta-globin gene expression in normal subjects and in those with high fetal hemoglobin production. Blood. 1987;69:1109–1113. MEDLINE

57. 57Chakalova L, Osborne CS, Dai YF, et al. The Corfu deltabeta thalassemia deletion disrupts gamma-globin gene silencing and reveals post-transcriptional regulation of HbF expression. Blood. 2005;105:2154–2160. MEDLINE | CrossRef

58. 58Clegg JB, Weatherall DJ, Bodmer WF. 5-azacytidine for beta-thalassemia?. Lancet. 1983;1:536. MEDLINE

59. 59Stamatoyannopoulos G. Control of globin gene expression during development and erythroid differentiation. Exp Hematol. 2005;33:259–271. | CrossRef

60. 60Jane SM, Gumucio DL, Ney PA, Cunningham JM, Nienhuis AW. Methylation-enhanced binding of Sp1 to the stage selector element of the human gamma-globin gene promoter may regulate development specificity of expression. Mol Cell Biol. 1993;13:3272–3281. MEDLINE

61. 61Sengupta PK, Lavelle D, DeSimone J. Increased binding of Sp1 to the gamma-globin gene promoter upon site-specific cytosine methylation. Am J Hematol. 1994;46:169–172. MEDLINE | CrossRef

62. 62Stamatoyannopoulos G, Veith R, Galanello R, Papayannopoulou T. Hb F production in stressed erythropoiesis: observations and kinetic models. Ann N Y Acad Sci. 1985;445:188–197. MEDLINE | CrossRef

63. 63Veith R, Galanello R, Papayannopoulou T, Stamatoyannopoulos G. Stimulation of F-cell production in patients with sickle-cell anemia treated with cytarabine or hydroxyurea. N Engl J Med. 1985;313:1571–1575. MEDLINE | CrossRef

64. 64Charache S, Dover G, Smith K, Talbot CC, Moyer M, Boyer S. Treatment of sickle cell anemia with 5-azacytidine results in increased fetal hemoglobin production and is associated with nonrandom hypomethylation of DNA around the gamma-delta-beta-globin gene complex. Proc Natl Acad Sci U S A. 1983;80:4842–4846. MEDLINE | CrossRef

65. 65Ley TJ, Chiang YL, Haidaris D, Anagnou NP, Wilson VL, Anderson WF. DNA methylation and regulation of the human beta-globin-like genes in mouse erythroleukemia cells containing human chromosome 11. Proc Natl Acad Sci U S A. 1984;81:6618–6622. MEDLINE | CrossRef

66. 66Humphries RK, Dover G, Young NS, et al. 5-Azacytidine acts directly on both erythroid precursors and progenitors to increase production of fetal hemoglobin. J Clin Invest. 1985;75:547–557. MEDLINE | CrossRef

67. 67Bard H, Prosmanne J. Relative rates of fetal hemoglobin and adult hemoglobin synthesis in cord blood of infants of insulin-dependent diabetic mothers. Pediatrics. 1985;75:1143–1147.

68. 68Perrine SP, Greene MF, Faller DV. Delay in the fetal globin switch in infants of diabetic mothers. N Engl J Med. 1985;312:334–338. MEDLINE | CrossRef

69. 69Perrine SP, Rudolph A, Faller DV, et al. Butyrate infusions in the ovine fetus delay the biologic clock for globin gene switching. Proc Natl Acad Sci U S A. 1988;85:8540–8542. MEDLINE | CrossRef

70. 70Constantoulakis P, Papayannopoulou T, Stamatoyannopoulos G. Alpha-amino-N-butyric acid stimulates fetal hemoglobin in the adult. Blood. 1988;72:1961–1967. MEDLINE

71. 71Atweh GF, Sutton M, Nassif I, et al. Sustained induction of fetal hemoglobin by pulse butyrate therapy in sickle cell disease. Blood. 1999;93:1790–1797. MEDLINE

72. 72Perrine SP, Ginder GD, Faller DV, et al. A short-term trial of butyrate to stimulate fetal-globin-gene expression in the beta-globin disorders. N Engl J Med. 1993;328:81–86. MEDLINE | CrossRef

73. 73Collins AF, Pearson HA, Giardina P, McDonagh KT, Brusilow SW, Dover GJ. Oral sodium phenylbutyrate therapy in homozygous beta thalassemia: a clinical trial. Blood. 1995;85:43–49. MEDLINE

74. 74Dover GJ, Brusilow S, Charache S. Induction of fetal hemoglobin production in subjects with sickle cell anemia by oral sodium phenylbutyrate. Blood. 1994;84:339–343. MEDLINE

75. 75Fibach E, Prasanna P, Rodgers GP, Samid D. Enhanced fetal hemoglobin production by phenylacetate and 4-phenylbutyrate in erythroid precursors derived from normal donors and patients with sickle cell anemia and beta-thalassemia. Blood. 1993;82:2203–2209. MEDLINE

76. 76Sher GD, Ginder GD, Little J, Yang S, Dover GJ, Olivieri NF. Extended therapy with intravenous arginine butyrate in patients with beta-hemoglobinopathies. N Engl J Med. 1995;332:1606–1610. MEDLINE | CrossRef

77. 77Reich S, Buhrer C, Henze G, et al. Oral isobutyramide reduces transfusion requirements in some patients with homozygous beta-thalassemia. Blood. 2000;96:3357–3363. MEDLINE

78. 78Domenica Cappellini M, Graziadei G, Ciceri L, et al. Oral isobutyramide therapy in patients with thalassemia intermedia: results of a phase II open study. Blood Cells Mol Dis. 2000;26:105–111. MEDLINE | CrossRef

79. 79Kieslich M, Schwabe D, Cinatl J, Driever PH. Increase of fetal hemoglobin synthesis indicating differentiation induction in children receiving valproic acid. Pediatr Hematol Oncol. 2003;20:15–22. MEDLINE | CrossRef

80. 80Collins AF, Dover GJ, Luban NL. Increased fetal hemoglobin production in patients receiving valproic acid for epilepsy. Blood. 1994;84:1690–1691. MEDLINE

81. 81Simpson RT. Structure of chromatin containing extensively acetylated H3 and H4. Cell. 1978;13:691–699. MEDLINE | CrossRef

82. 82Candido EP, Reeves R, Davie JR. Sodium butyrate inhibits histone deacetylation in cultured cells. Cell. 1978;14:105–113. MEDLINE | CrossRef

83. 83Sealy L, Chalkley R. The effect of sodium butyrate on histone modification. Cell. 1978;14:115–121. MEDLINE | CrossRef

84. 84McCaffrey PG, Newsome DA, Fibach E, Yoshida M, Su MS. Induction of gamma-globin by histone deacetylase inhibitors. Blood. 1997;90:2075–2083. MEDLINE

85. 85Boosalis MS, Bandyopadhyay R, Bresnick EH, et al. Short-chain fatty acid derivatives stimulate cell proliferation and induce STAT-5 activation. Blood. 2001;97:3259–3267. MEDLINE | CrossRef

86. 86Fathallah H, Weinberg RS, Galperin Y, Sutton M, Atweh GF. Role of epigenetic modifications in normal globin gene regulation and butyrate-mediated induction of fetal hemoglobin. Blood. 2007;110:3391–3397. CrossRef

87. 87Skarpidi E, Cao H, Heltweg B, et al. Hydroxamide derivatives of short-chain fatty acids are potent inducers of human fetal globin gene expression. Exp Hematol. 2003;31:197–203. | CrossRef

88. 88Wei GH, Zhao GW, Song W, et al. Mechanisms of human gamma-globin transcriptional induction by apicidin involves p38 signaling to chromatin. Biochem Biophys Res Commun. 2007;363:889–894. CrossRef

89. 89Witt O, Monkemeyer S, Ronndahl G, et al. Induction of fetal hemoglobin expression by the histone deacetylase inhibitor apicidin. Blood. 2003;101:2001–2007. MEDLINE | CrossRef

90. 90Cao H, Stamatoyannopoulos G, Jung M. Induction of human gamma globin gene expression by histone deacetylase inhibitors. Blood. 2004;103:701–709. MEDLINE | CrossRef

91. 91Pan LN, Lu J, Huang B. HDAC inhibitors: a potential new category of anti-tumor agents. Cell Mol Immunol. 2007;4:337–343.

92. 92Moutouh-de Parseval LA, Verhelle D, Glezer E, et al. Pomalidomide and lenalidomide regulate erythropoiesis and fetal hemoglobin production in human CD34+ cells. J Clin Invest. 2008;118:248–258. CrossRef

93. 93Aerbajinai W, Zhu J, Gao Z, Chin K, Rodgers GP. Thalidomide induces gamma-globin gene expression through increased reactive oxygen species-mediated p38 MAPK signaling and histone H4 acetylation in adult erythropoiesis. Blood. 2007;110:2864–2871. CrossRef

94. 94Nascimento Mde L, Ladipo OA, Coutinho EM. Nomegestrol acetate contraceptive implant use by women with sickle cell disease. Clin Pharmacol Ther. 1998;64:433–438. MEDLINE | CrossRef

95. 95da Silva Santos Duarte A, Sales TS, Mengel JO, Costa FF, Saad ST. Progesterone upregulates GATA-1 on erythroid progenitors cells in liquid culture. Blood Cells Mol Dis. 2002;29:213–224. MEDLINE | CrossRef

96. 96Stamatoyannopoulos G, Veith R, al-Khatti A, Papayannopoulou T. Induction of fetal hemoglobin by cell-cycle-specific drugs and recombinant erythropoietin. Am J Pediatr Hematol Oncol. 1990;12:21–26. MEDLINE

97. 97Miller BA, Perrine SP, Bernstein A, et al. Influence of steel factor on hemoglobin synthesis in sickle cell disease. Blood. 1992;79:1861–1868. MEDLINE

98. 98Wojda U, Leigh KR, Njoroge JM, et al. Fetal hemoglobin modulation during human erythropoiesis: stem cell factor has “late” effects related to the expression pattern of CD117. Blood. 2003;101:492–497. MEDLINE | CrossRef

99. 99Gabbianelli M, Morsilli O, Massa A, et al. Effective erythropoiesis and HbF reactivation induced by kit ligand in beta-thalassemia. Blood. 2008;111:421–429. CrossRef

100. 100Bohmer RM. Reactivation of fetal hemoglobin in adult stem cell erythropoiesis by transforming growth factor-beta. J Hematother Stem Cell Res. 2003;12:499–504. MEDLINE

101. 101Witt O, Sand K, Pekrun A. Butyrate-induced erythroid differentiation of human K562 leukemia cells involves inhibition of ERK and activation of p38 MAP kinase pathways. Blood. 2000;95:2391–2396. MEDLINE

102. 102Witt O, Monkemeyer S, Kanbach K, Pekrun A. Induction of fetal hemoglobin synthesis by valproate: modulation of MAP kinase pathways. Am J Hematol. 2002;71:45–46. MEDLINE | CrossRef

103. 103Pace BS, Zein S. Understanding mechanisms of gamma-globin gene regulation to develop strategies for pharmacological fetal hemoglobin induction. Dev Dyn. 2006;235:1727–1737. MEDLINE | CrossRef

104. 104Hsiao CH, Li W, Lou TF, Baliga BS, Pace BS. Fetal hemoglobin induction by histone deacetylase inhibitors involves generation of reactive oxygen species. Exp Hematol. 2006;34:264–273. | CrossRef

105. 105Pace BS, Qian XH, Sangerman J, et al. p38 MAP kinase activation mediates gamma-globin gene induction in erythroid progenitors. Exp Hematol. 2003;31:1089–1096. | CrossRef

106. 106Sangerman J, Lee MS, Yao X, et al. Mechanism for fetal hemoglobin induction by histone deacetylase inhibitors involves gamma-globin activation by CREB1 and ATF-2. Blood. 2006;108:3590–3599. MEDLINE | CrossRef

107. 107Ikuta T, Ausenda S, Cappellini MD. Mechanism for fetal globin gene expression: role of the soluble guanylate cyclase-cGMP-dependent protein kinase pathway. Proc Natl Acad Sci U S A. 2001;98:1847–1852. MEDLINE | CrossRef

108. 108Inoue A, Kuroyanagi Y, Terui K, Moi P, Ikuta T. Negative regulation of gamma-globin gene expression by cyclic AMP-dependent pathway in erythroid cells. Exp Hematol. 2004;32:244–253. | CrossRef

109. 109Keefer JR, Schneidereith TA, Mays A, Purvis SH, Dover GJ, Smith KD. Role of cyclic nucleotides in fetal hemoglobin induction in cultured CD34+ cells. Exp Hematol. 2006;34:1151–1161.

110. 110Kuroyanagi Y, Kaneko Y, Muta K, et al. cAMP differentially regulates gamma-globin gene expression in erythroleukemic cells and primary erythroblasts through c-Myb expression. Biochem Biophys Res Commun. 2006;344:1038–1047. CrossRef

111. 111Bailey L, Kuroyanagi Y, Franco-Penteado CF, et al. Expression of the gamma-globin gene is sustained by the cAMP-dependent pathway in beta-thalassaemia. Br J Haematol. 2007;138:382–395. CrossRef

112. 112Cokic VP, Andric SA, Stojilkovic SS, Noguchi CT, Schechter AN. Hydroxyurea nitrosylates and activates soluble guanylyl cyclase in human erythroid cells. Blood. 2008;111:1117–1123. CrossRef

113. 113Cokic VP, Smith RD, Beleslin-Cokic BB, et al. Hydroxyurea induces fetal hemoglobin by the nitric oxide-dependent activation of soluble guanylyl cyclase. J Clin Invest. 2003;111:231–239. MEDLINE | CrossRef

114. 114Gladwin MT, Shelhamer JH, Ognibene FP, et al. Nitric oxide donor properties of hydroxyurea in patients with sickle cell disease. Br J Haematol. 2002;116:436–444. MEDLINE | CrossRef

115. 115Park JI, Choi HS, Jeong JS, Han JY, Kim IH. Involvement of p38 kinase in hydroxyurea-induced differentiation of K562 cells. Cell Growth Differ. 2001;12:481–486. MEDLINE

116. 116Bhanu NV, Trice TA, Lee YT, et al. A sustained and pancellular reversal of gamma-globin gene silencing in adult human erythroid precursor cells. Blood. 2005;105:387–393. MEDLINE | CrossRef

117. 117Fibach E, Bianchi N, Borgatti M, et al. Effects of rapamycin on accumulation of alpha-, beta- and gamma-globin mRNAs in erythroid precursor cells from beta-thalassaemia patients. Eur J Haematol. 2006;77:437–441. MEDLINE | CrossRef

118. 118Mischiati C, Sereni A, Lampronti I, et al. Rapamycin-mediated induction of gamma-globin mRNA accumulation in human erythroid cells. Br J Haematol. 2004;126:612–621. MEDLINE | CrossRef

119. 119Zuccato C, Bianchi N, Borgatti M, et al. Everolimus is a potent inducer of erythroid differentiation and gamma-globin gene expression in human erythroid cells. Acta Haematol. 2007;117:168–176. CrossRef

120. 120Obata T, Brown GE, Yaffe MB. MAP kinase pathways activated by stress: the p38 MAPK pathway. Crit Care Med. 2000;28:N67–N77. MEDLINE | CrossRef

121. 121Kyriakis JM, Avruch J. Mammalian mitogen-activated protein kinase signal transduction pathways activated by stress and inflammation. Physiol Rev. 2001;81:807–869. MEDLINE

122. 122Tang DC, Zhu J, Liu W, et al. The hydroxyurea-induced small GTP-binding protein SAR modulates gamma-globin gene expression in human erythroid cells. Blood. 2005;106:3256–3263. MEDLINE | CrossRef

123. 123McElveen RL, Lou TF, Reese K, Xia S, Baliga BS, Pace BS. Erk pathway inhibitor U0126 induces gamma-globin expression in erythroid cells. Cell Mol Biol (Noisy-le-grand). 2005;51:215–227.

124. 124Probin V, Wang Y, Zhou D. Busulfan-induced senescence is dependent on ROS production upstream of the MAPK pathway. Free Radic Biol Med. 2007;42:1858–1865. MEDLINE | CrossRef

125. 125Sanchez-Arevalo Lobo VJ, Aceves Luquero CI, Alvarez-Vallina L, et al. Modulation of the p38 MAPK (mitogen-activated protein kinase) pathway through Bcr/Abl: implications in the cellular response to Ara-C. Biochem J. 2005;387:231–238. CrossRef

126. 126Raman M, Earnest S, Zhang K, Zhao Y, Cobb MH. TAO kinases mediate activation of p38 in response to DNA damage. EMBO J. 2007;26:2005–2014. MEDLINE | CrossRef

127. 127Persengiev SP, Green MR. The role of ATF/CREB family members in cell growth, survival and apoptosis. Apoptosis. 2003;8:225–228. MEDLINE | CrossRef

128. 128Maurice DH. Cyclic nucleotide phosphodiesterase-mediated integration of cGMP and cAMP signaling in cells of the cardiovascular system. Front Biosci. 2005;10:1221–1228. CrossRef

129. 129Zaccolo M, Movsesian MA. cAMP and cGMP signaling cross-talk: role of phosphodiesterases and implications for cardiac pathophysiology. Circ Res. 2007;100:1569–1578. CrossRef

130. 130Chen JJ. Regulation of protein synthesis by the heme-regulated eIF2alpha kinase: relevance to anemias. Blood. 2007;109:2693–2699. MEDLINE

131. 131Wek RC, Jiang HY, Anthony TG. Coping with stress: eIF2 kinases and translational control. Biochem Soc Trans. 2006;34:7–11. MEDLINE | CrossRef

132. 132Masuoka HC, Townes TM. Targeted disruption of the activating transcription factor 4 gene results in severe fetal anemia in mice. Blood. 2002;99:736–745. MEDLINE | CrossRef

133. 133Patterson AD, Hollander MC, Miller GF, Fornace AJ. Gadd34 requirement for normal hemoglobin synthesis. Mol Cell Biol. 2006;26:1644–1653. MEDLINE | CrossRef

134. 134Mankidy R, Faller DV, Mabaera R, et al. Short-chain fatty acids induce gamma-globin gene expression by displacement of a HDAC3-NCoR repressor complex. Blood. 2006;108:3179–3186. MEDLINE | CrossRef

135. 135Bain J, Plater L, Elliott M, et al. The selectivity of protein kinase inhibitors: a further update. Biochem J. 2007;408:297–315. CrossRef

136. 136Saunthararajah Y, Hillery CA, Lavelle D, et al. Effects of 5-aza-2'-deoxycytidine on fetal hemoglobin levels, red cell adhesion, and hematopoietic differentiation in patients with sickle cell disease. Blood. 2003;102:3865–3870. MEDLINE | CrossRef

137. 137Carr BI, Rahbar S, Doroshow JH, et al. Fetal hemoglobin gene activation in a phase II study of 5,6-dihydro-5-azacytidine for bronchogenic carcinoma. Cancer Res. 1987;47:4199–4201. MEDLINE

138. 138Lavelle D, Vaitkus K, Banzon V, et al. Fetal hemoglobin induction in baboons (P. Anubis) following administration of a novel decitabine dinucleotide (S110) compound. [abstract] Blood. 2007;110:Abstract 571.

139. 139Liu DP, Liang CC, Ao ZH, et al. Treatment of severe beta-thalassemia (patients) with Myleran. Am J Hematol. 1990;33:50–55. MEDLINE | CrossRef

140. 140Bianchi N, Ongaro F, Chiarabelli C, et al. Induction of erythroid differentiation of human K562 cells by cisplatin analogs. Biochem Pharmacol. 2000;60:31–40. MEDLINE | CrossRef

141. 141Iyamu WE, Adunyah SE, Fasold H, et al. Enhancement of hemoglobin and F-cell production by targeting growth inhibition and differentiation of K562 cells with ribonucleotide reductase inhibitors (Didox and Trimidox) in combination with streptozotocin. Am J Hematol. 2000;63:176–183. MEDLINE | CrossRef

142. 142Yu J, Lemas V, Page T, Connor JD, Yu AL. Induction of erythroid differentiation in K562 cells by inhibitors of inosine monophosphate dehydrogenase. Cancer Res. 1989;49:5555–5560. MEDLINE

143. 143Pace BS, Elford HL, Stamatoyannopoulos G. Transgenic mouse model of pharmacologic induction of fetal hemoglobin: studies using a new ribonucleotide reductase inhibitor. Didox. Am J Hematol. 1994;45:136–141.

144. 144Rodrigue CM, Arous N, Bachir D, et al. Resveratrol, a natural dietary phytoalexin, possesses similar properties to hydroxyurea towards erythroid differentiation. Br J Haematol. 2001;113:500–507. MEDLINE | CrossRef

145. 145Delgado-Canedo A, Chies JA, Nardi NB. Induction of fetal haemoglobin expression in erythroid cells—a model based on iron availability signalling. Med Hypotheses. 2005;65:932–936. Abstract | Full Text | Full-Text PDF (86 KB) | CrossRef

146. 146Bianchi N, Osti F, Rutigliano C, et al. The DNA-binding drugs mithramycin and chromomycin are powerful inducers of erythroid differentiation of human K562 cells. Br J Haematol. 1999;104:258–265. MEDLINE | CrossRef

147. 147Bianchi N, Chiarabelli C, Borgatti M, Mischiati C, Fibach E, Gambari R. Accumulation of gamma-globin mRNA and induction of erythroid differentiation after treatment of human leukaemic K562 cells with tallimustine. Br J Haematol. 2001;113:951–961. MEDLINE | CrossRef

148. 148Lampronti I, Bianchi N, Borgatti M, Fibach E, Prus E, Gambari R. Accumulation of gamma-globin mRNA in human erythroid cells treated with angelicin. Eur J Haematol. 2003;71:189–195. MEDLINE | CrossRef

149. 149Viola G, Vedaldi D, Dall'Acqua F, et al. Induction of gamma-globin mRNA, erythroid differentiation and apoptosis in UVA-irradiated human erythroid cells in the presence of furocumarin derivatives. Biochem Pharmacol. 2008;75:810–825. CrossRef

150. 150Veith R, Dautenhahn AG, Roth RC. Methotrexate stimulates fetal hemoglobin production in anemic baboons. Prog Clin Biol Res. 1989;316B:363–370. MEDLINE

151. 151Veith R, Papayannopoulou T, Kurachi S, Stamatoyannopoulos G. Treatment of baboon with vinblastine: insights into the mechanisms of pharmacologic stimulation of Hb F in the adult. Blood. 1985;66:456–459. MEDLINE

152. 152Letvin NL, Linch DC, Beardsley GP, McIntyre KW, Miller BA, Nathan DG. Influence of cell cycle phase-specific agents on simian fetal hemoglobin synthesis. J Clin Invest. 1985;75:1999–2005. MEDLINE | CrossRef

153. 153Pace BS, White GL, Dover GJ, Boosalis MS, Faller DV, Perrine SP. Short-chain fatty acid derivatives induce fetal globin expression and erythropoiesis in vivo. Blood. 2002;100:4640–4648. MEDLINE | CrossRef

154. 154Witt O, Schmejkal S, Pekrun A. Tributyrin plus all-trans-retinoic acid efficiently induces fetal hemoglobin expression in human erythroleukemia cells. Am J Hematol. 2000;64:319–321. MEDLINE | CrossRef

155. 155Stamatoyannopoulos G, Blau CA, Nakamoto B, et al. Fetal hemoglobin induction by acetate, a product of butyrate catabolism. Blood. 1994;84:3198–3204. MEDLINE

156. 156Torkelson S, White B, Faller DV, Phipps K, Pantazis C, Perrine SP. Erythroid progenitor proliferation is stimulated by phenoxyacetic and phenylalkyl acids. Blood Cells Mol Dis. 1996;22:150–158. MEDLINE | CrossRef

157. 157Cao H, Jung M, Stamatoyannopoulos G. Hydroxamide derivatives of short-chain fatty acid have erythropoietic activity and induce gamma gene expression in vivo. Exp Hematol. 2005;33:1443–1449. | CrossRef

158. 158Safaya S, Ibrahim A, Rieder RF. Augmentation of gamma-globin gene promoter activity by carboxylic acids and components of the human beta-globin locus control region. Blood. 1994;84:3929–3935. MEDLINE

159. 159Liakopoulou E, Blau CA, Li Q, et al. Stimulation of fetal hemoglobin production by short chain fatty acids. Blood. 1995;86:3227–3235. MEDLINE

160. 160Bohacek R, Boosalis MS, McMartin C, Faller DV, Perrine SP. Identification of novel small-molecule inducers of fetal hemoglobin using pharmacophore and ‘PSEUDO’ receptor models. Chem Biol Drug Des. 2006;67:318–328. MEDLINE | CrossRef

161. 161Mai A, Jelicic K, Rotili D, et al. Identification of two new synthetic histone deacetylase inhibitors that modulate globin gene expression in erythroid cells from healthy donors and patients with thalassemia. Mol Pharmacol. 2007;72:1111–1123. CrossRef

162. 162Cao H, Stamatoyannopoulos G. Histone deacetylase inhibitor FK228 is a potent inducer of human fetal hemoglobin. Am J Hematol. 2006;81:981–983. MEDLINE | CrossRef

163. 163Johnson J, Hunter R, McElveen R, Qian XH, Baliga BS, Pace BS. Fetal hemoglobin induction by the histone deacetylase inhibitor, scriptaid. Cell Mol Biol (Noisy-le-grand). 2005;51:229–238.

164. 164Smith RD, Li J, Noguchi CT, Schechter AN. Quantitative PCR analysis of HbF inducers in primary human adult erythroid cells. Blood. 2000;95:863–869. MEDLINE

165. 165Haynes J, Baliga BS, Obiako B, Ofori-Acquah S, Pace B. Zileuton induces hemoglobin F synthesis in erythroid progenitors: role of the L-arginine-nitric oxide signaling pathway. Blood. 2004;103:3945–3950. MEDLINE | CrossRef

166. 166Amoyal I, Prus E, Fibach E. Vanadate elevates fetal hemoglobin in human erythroid precursors by inhibiting cell maturation. Exp Biol Med (Maywood). 2007;232:654–661. MEDLINE

167. 167Hsieh MM, Linde NS, Wynter A, et al. HIF prolyl hydroxylase inhibition results in endogenous erythropoietin induction, erythrocytosis, and modest fetal hemoglobin expression in rhesus macaques. Blood. 2007;110:2140–2147. CrossRef

Departments of Medicine and Pharmacology and Toxicology and the Norris Cotton Cancer Center, Dartmouth Medical School, Lebanon, NH., USA

Offprint requests to: Christopher H. Lowrey, M.D., Dartmouth Medical School/Dartmouth-Hitchcock Medical Center, Department of Medicine, One Medical Center Drive, Lebanon, NH 03756

PII: S0301-472X(08)00332-9

doi:10.1016/j.exphem.2008.06.014

2 of 18