Hematology, 2004, No 1, pp 297-317 Myelodysplastic …Hematology 2004 297 Myelodysplastic Syndromes Alan F. List, James Vardiman, Jean-Pierre J. Issa, and Theo M. DeWitte The development

184 American Society of Hematology

Current published data suggest that the 2 drugs as cur-rently studied are equivalently effective in the treatmentof MDS.

The relationship of DNA methyltransferase inhi-bition to the clinical activity of 5AC and DAC in thetreatment of MDS remains a critical question. Lübbertand colleagues reported reversal of p15 methylation,associated with reexpression of p15 protein, in a subsetof patients treated with DAC.24 The frequency of thisreversal of methylation and gene expression cannot beestimated based on current data; nor can the associa-tion of changes in methylation with clinical responsebe ascertained; reversal of p15 methylation did not ap-pear to be required for hematologic response inLübbert’s study.24 Such studies will be important in thefurther development of this class of compounds, aloneand in combination with other drugs.

Histone deacetylase inhibitorsThe transcriptional state of chromatin depends on thetranslation of a complex set of posttranslational modi-fications to histone lysine tails including acetylation,methylation, and phosphorylation, referred to as thehistone code.4,25 Modifications of specific residues onhistones are recognized by specific binding proteins thatimpact chromatin conformation and transcription. Thesespecific modifications are local; that is, histones aremodified in the regions of specific gene promoters, in-ducing local remodeling of chromatin which impactsthe expression of the specific genes. In general, acety-lation of lysine residues on histone tails is associatedwith transcriptionally active chromatin (euchromatin),whereas deacetylated histones are associated with tran-scriptionally inactive chromatin (heterochromatin).Methylation of lysine 9 on histone H3 is associated withheterochromatin, whereas methylation of lysine 4 onthat histone is associated with euchromatin. It is highlylikely that as these epigenetic changes are better char-acterized, many of the enzymes which mediate thesechanges, such as histone methyltransferases, will betargeted in cancer therapeutics.

The earliest described histone modification knownto be associated with transcriptional regulation wasacetylation of specific lysine residues in the tails of hi-stones H2, H3, and H4. Histones are acetylated by en-zymes which contain histone acetyltransferase activ-ity; in contrast, deacetylation is mediated by histonedeacetylases (HDAC). Eleven human HDACs have beenidentified to date (reviewed by de Ruijter et al26). Theseinclude 2 major classes of HDACs (Class I and ClassII); class I HDACs are almost exclusively nuclear, whileclass II HDACs shuttle in and out of the nucleus in re-sponse to specific cellular signals. In addition, a third

class of HDACs known as the SIR2 family includesenzymes with NAD-dependent HDAC activity; unlikethe other 2 classes, these are not inhibited by trichostatinA. HDACs are associated with specific chromatin lociin pairs and triplets.

A variety of HDAC inhibitors (HDACi) are underclinical investigation. Interest in the use of HDACi forthe treatment of myeloid malignancies dates back tothe recognition of the activity of butyrate derivativesand polar planar compounds to induce differentiation.Certain AML fusion genes, such as AML1-ETO, PML-RARα, and PLZF-RARα specifically recruit nuclearcorepression complexes which include HDAC, therebysilencing expression of genes downstream from thepromoters bound by the fusion proteins (reviewed byGore and Carducci27). In such leukemias, HDACi maybe utilized to specifically reverse the transcriptionalrepression induced by the fusion proteins. No such fu-sion proteins have been identified for most cases ofMDS; hence, in this group of diseases, the impact ofHDACi is more likely related to changes in gene ex-pression altered through other epigenetic mechanisms.

Small-chain fatty acids: A variety of small-chainfatty acids inhibit HDAC activity at submillimolar con-centrations. These include sodium and arginine butyrate,sodium phenylbutyrate, and valproic acid. Although bu-tyrate was successfully used to induce terminal differ-entiation in a child with AML28 and has been success-fully used to induce expression of fetal hemoglobin inpatients with sickle cell anemia29 and thalassemia,30 asubsequent Phase II study of butyrate in AML was with-out clinical activity.30 Formal pharmacokinetic/pharma-codynamic studies of butyrate have not been performed.Sodium phenylbutyrate (PB) was selected for develop-ment in part because of the potential for developmentof an oral formulation. PB induces histone acetylation,expression of p21WAF1/CIP1, induction of G

1 cell cycle

arrest, and induction of CD11b expression in myeloidleukemia cells.32 Concentrations that are not clinicallyachievable induce apoptosis. Phase I studies of con-tinuous infusion PB demonstrated that the drug waswell-tolerated; dose-limiting toxicity was a reversibleencephalopathy due to accumulation of phenyl-acetate.33,34 PB was administered for 7/28 days, 7/14days, and 21/28 days. Lineage responses were achievedin several patients with MDS and AML. At the maxi-mum tolerated dose (375 mg/kg/day), sustained plasmaconcentrations ranged from 0.3 to 0.5 mM; in vitro,induction of acetylation of histones in hematopoieticcells requires approximately 0.25 mM.35 Higher peakplasma concentrations of PB have been achieved in solidtumor patients receiving short-term infusions of PB(M.A. Carducci, personal communication). An oral for-

184 American Society of Hematology

Current published data suggest that the 2 drugs as cur-rently studied are equivalently effective in the treatmentof MDS.

The relationship of DNA methyltransferase inhi-bition to the clinical activity of 5AC and DAC in thetreatment of MDS remains a critical question. Lübbertand colleagues reported reversal of p15 methylation,associated with reexpression of p15 protein, in a subsetof patients treated with DAC.24 The frequency of thisreversal of methylation and gene expression cannot beestimated based on current data; nor can the associa-tion of changes in methylation with clinical responsebe ascertained; reversal of p15 methylation did not ap-pear to be required for hematologic response inLübbert’s study.24 Such studies will be important in thefurther development of this class of compounds, aloneand in combination with other drugs.

Histone deacetylase inhibitorsThe transcriptional state of chromatin depends on thetranslation of a complex set of posttranslational modi-fications to histone lysine tails including acetylation,methylation, and phosphorylation, referred to as thehistone code.4,25 Modifications of specific residues onhistones are recognized by specific binding proteins thatimpact chromatin conformation and transcription. Thesespecific modifications are local; that is, histones aremodified in the regions of specific gene promoters, in-ducing local remodeling of chromatin which impactsthe expression of the specific genes. In general, acety-lation of lysine residues on histone tails is associatedwith transcriptionally active chromatin (euchromatin),whereas deacetylated histones are associated with tran-scriptionally inactive chromatin (heterochromatin).Methylation of lysine 9 on histone H3 is associated withheterochromatin, whereas methylation of lysine 4 onthat histone is associated with euchromatin. It is highlylikely that as these epigenetic changes are better char-acterized, many of the enzymes which mediate thesechanges, such as histone methyltransferases, will betargeted in cancer therapeutics.

The earliest described histone modification knownto be associated with transcriptional regulation wasacetylation of specific lysine residues in the tails of hi-stones H2, H3, and H4. Histones are acetylated by en-zymes which contain histone acetyltransferase activ-ity; in contrast, deacetylation is mediated by histonedeacetylases (HDAC). Eleven human HDACs have beenidentified to date (reviewed by de Ruijter et al26). Theseinclude 2 major classes of HDACs (Class I and ClassII); class I HDACs are almost exclusively nuclear, whileclass II HDACs shuttle in and out of the nucleus in re-sponse to specific cellular signals. In addition, a third

class of HDACs known as the SIR2 family includesenzymes with NAD-dependent HDAC activity; unlikethe other 2 classes, these are not inhibited by trichostatinA. HDACs are associated with specific chromatin lociin pairs and triplets.

A variety of HDAC inhibitors (HDACi) are underclinical investigation. Interest in the use of HDACi forthe treatment of myeloid malignancies dates back tothe recognition of the activity of butyrate derivativesand polar planar compounds to induce differentiation.Certain AML fusion genes, such as AML1-ETO, PML-RARα, and PLZF-RARα specifically recruit nuclearcorepression complexes which include HDAC, therebysilencing expression of genes downstream from thepromoters bound by the fusion proteins (reviewed byGore and Carducci27). In such leukemias, HDACi maybe utilized to specifically reverse the transcriptionalrepression induced by the fusion proteins. No such fu-sion proteins have been identified for most cases ofMDS; hence, in this group of diseases, the impact ofHDACi is more likely related to changes in gene ex-pression altered through other epigenetic mechanisms.

Small-chain fatty acids: A variety of small-chainfatty acids inhibit HDAC activity at submillimolar con-centrations. These include sodium and arginine butyrate,sodium phenylbutyrate, and valproic acid. Although bu-tyrate was successfully used to induce terminal differ-entiation in a child with AML28 and has been success-fully used to induce expression of fetal hemoglobin inpatients with sickle cell anemia29 and thalassemia,30 asubsequent Phase II study of butyrate in AML was with-out clinical activity.30 Formal pharmacokinetic/pharma-codynamic studies of butyrate have not been performed.Sodium phenylbutyrate (PB) was selected for develop-ment in part because of the potential for developmentof an oral formulation. PB induces histone acetylation,expression of p21WAF1/CIP1, induction of G

1 cell cycle

arrest, and induction of CD11b expression in myeloidleukemia cells.32 Concentrations that are not clinicallyachievable induce apoptosis. Phase I studies of con-tinuous infusion PB demonstrated that the drug waswell-tolerated; dose-limiting toxicity was a reversibleencephalopathy due to accumulation of phenyl-acetate.33,34 PB was administered for 7/28 days, 7/14days, and 21/28 days. Lineage responses were achievedin several patients with MDS and AML. At the maxi-mum tolerated dose (375 mg/kg/day), sustained plasmaconcentrations ranged from 0.3 to 0.5 mM; in vitro,induction of acetylation of histones in hematopoieticcells requires approximately 0.25 mM.35 Higher peakplasma concentrations of PB have been achieved in solidtumor patients receiving short-term infusions of PB(M.A. Carducci, personal communication). An oral for-

184 American Society of Hematology

Current published data suggest that the 2 drugs as cur-rently studied are equivalently effective in the treatmentof MDS.

The relationship of DNA methyltransferase inhi-bition to the clinical activity of 5AC and DAC in thetreatment of MDS remains a critical question. Lübbertand colleagues reported reversal of p15 methylation,associated with reexpression of p15 protein, in a subsetof patients treated with DAC.24 The frequency of thisreversal of methylation and gene expression cannot beestimated based on current data; nor can the associa-tion of changes in methylation with clinical responsebe ascertained; reversal of p15 methylation did not ap-pear to be required for hematologic response inLübbert’s study.24 Such studies will be important in thefurther development of this class of compounds, aloneand in combination with other drugs.

Histone deacetylase inhibitorsThe transcriptional state of chromatin depends on thetranslation of a complex set of posttranslational modi-fications to histone lysine tails including acetylation,methylation, and phosphorylation, referred to as thehistone code.4,25 Modifications of specific residues onhistones are recognized by specific binding proteins thatimpact chromatin conformation and transcription. Thesespecific modifications are local; that is, histones aremodified in the regions of specific gene promoters, in-ducing local remodeling of chromatin which impactsthe expression of the specific genes. In general, acety-lation of lysine residues on histone tails is associatedwith transcriptionally active chromatin (euchromatin),whereas deacetylated histones are associated with tran-scriptionally inactive chromatin (heterochromatin).Methylation of lysine 9 on histone H3 is associated withheterochromatin, whereas methylation of lysine 4 onthat histone is associated with euchromatin. It is highlylikely that as these epigenetic changes are better char-acterized, many of the enzymes which mediate thesechanges, such as histone methyltransferases, will betargeted in cancer therapeutics.

The earliest described histone modification knownto be associated with transcriptional regulation wasacetylation of specific lysine residues in the tails of hi-stones H2, H3, and H4. Histones are acetylated by en-zymes which contain histone acetyltransferase activ-ity; in contrast, deacetylation is mediated by histonedeacetylases (HDAC). Eleven human HDACs have beenidentified to date (reviewed by de Ruijter et al26). Theseinclude 2 major classes of HDACs (Class I and ClassII); class I HDACs are almost exclusively nuclear, whileclass II HDACs shuttle in and out of the nucleus in re-sponse to specific cellular signals. In addition, a third

class of HDACs known as the SIR2 family includesenzymes with NAD-dependent HDAC activity; unlikethe other 2 classes, these are not inhibited by trichostatinA. HDACs are associated with specific chromatin lociin pairs and triplets.

A variety of HDAC inhibitors (HDACi) are underclinical investigation. Interest in the use of HDACi forthe treatment of myeloid malignancies dates back tothe recognition of the activity of butyrate derivativesand polar planar compounds to induce differentiation.Certain AML fusion genes, such as AML1-ETO, PML-RARα, and PLZF-RARα specifically recruit nuclearcorepression complexes which include HDAC, therebysilencing expression of genes downstream from thepromoters bound by the fusion proteins (reviewed byGore and Carducci27). In such leukemias, HDACi maybe utilized to specifically reverse the transcriptionalrepression induced by the fusion proteins. No such fu-sion proteins have been identified for most cases ofMDS; hence, in this group of diseases, the impact ofHDACi is more likely related to changes in gene ex-pression altered through other epigenetic mechanisms.

Small-chain fatty acids: A variety of small-chainfatty acids inhibit HDAC activity at submillimolar con-centrations. These include sodium and arginine butyrate,sodium phenylbutyrate, and valproic acid. Although bu-tyrate was successfully used to induce terminal differ-entiation in a child with AML28 and has been success-fully used to induce expression of fetal hemoglobin inpatients with sickle cell anemia29 and thalassemia,30 asubsequent Phase II study of butyrate in AML was with-out clinical activity.30 Formal pharmacokinetic/pharma-codynamic studies of butyrate have not been performed.Sodium phenylbutyrate (PB) was selected for develop-ment in part because of the potential for developmentof an oral formulation. PB induces histone acetylation,expression of p21WAF1/CIP1, induction of G

1 cell cycle

arrest, and induction of CD11b expression in myeloidleukemia cells.32 Concentrations that are not clinicallyachievable induce apoptosis. Phase I studies of con-tinuous infusion PB demonstrated that the drug waswell-tolerated; dose-limiting toxicity was a reversibleencephalopathy due to accumulation of phenyl-acetate.33,34 PB was administered for 7/28 days, 7/14days, and 21/28 days. Lineage responses were achievedin several patients with MDS and AML. At the maxi-mum tolerated dose (375 mg/kg/day), sustained plasmaconcentrations ranged from 0.3 to 0.5 mM; in vitro,induction of acetylation of histones in hematopoieticcells requires approximately 0.25 mM.35 Higher peakplasma concentrations of PB have been achieved in solidtumor patients receiving short-term infusions of PB(M.A. Carducci, personal communication). An oral for-

Hematology 2003 185

mulation has been studied which can supply similarplasma concentrations; however, the daily dosing re-quires as many as 75 pills per day.36 Despite the diffi-culty with delivery of concentrations of PB with HDACinhibitory activity, PB remains extremely well-tolerated.

Recent data demonstrating that valproic acid hassimilar HDAC inhibitory activity has raised hopes thatoral formulations of this drug, which is commonly usedfor neuropsychiatric disorders, could be used to modu-late gene expression.37 Valproate is used clinically atsubmillimolar concentrations; in vitro acetylationchanges have been demonstrated at 0.5 mM. Thepharmacodynamic impact of valproate on leukemiccells is similar to PB. Valproate is currently being in-vestigated in monotherapy and in combination withother drugs in MDS and in AML.

SAHA and FK228: The prototype HDAC inhibitorused for in vitro study, trichostatin, is an hydroxamicacid. To date, the only hydroxamic acid under clinicalinvestigation is suberoylanilide hydroxamic acid(SAHA). SAHA has been developed in both intrave-nous and oral formulations.38 Clinical administrationhas been associated with induction of histone acetyla-tion. No data are yet available detailing the impact ofSAHA on myeloid malignancies.

FK228 (depsipeptide) is a cyclic peptide whoseHDAC inhibitory activity appears to be due in greatpart to a reduced metabolite. FK228 appears to be spe-cific for Class I HDAC. FK228 has been studied inPhase I clinical trials including 1 in AML.39 Tumor ly-sis syndrome has been induced by FK228, indicatingsignificant cytotoxicity to myeloid cells; however, nosustained responses were seen in the Phase I trial. As-thenia appears to be dose limiting.

MS-275: MS-275 is a benzamide HDAC inhibitorundergoing Phase I investigation in MDS and AML.Similar to FK228, fatigue seems to be a major toxicityof MS275. MS275 has a half-life of greater than 24hours; changes in histone acetylation have persisted forseveral weeks following the administration of MS-275.No sustained responses have been reported to date (JEKarp, personal communication).

Integration of HDAC inhibitors intothe treatment of MDSThe recognition that HDAC recruitment accounted atleast in part for the silencing of genes with methylatedpromoters led to the demonstration that optimalreexpression of such genes required sequential expo-sure to a methyltransferase inhibitor followed by anHDACi. While definitive demonstration of a correla-tion between methyltransferase inhibition and clinicalactivity of methyltransferase inhibitors in MDS has yet

to be provided, a great deal of interest remains in aug-menting the clinical activity of 5AC and DAC throughthe addition of an HDACi. To date, data are only avail-able from a dose-finding study of 5AC followed by PB.In this study, a variety of doses and schedules of 5AChave been administered, followed by a 7-day continu-ous infusion of PB.40 The combination has been well-tolerated, and significant sustained clinical responseshave been achieved. Changes in methylation, histoneacetylation, and gene reexpression are being monitored.Similar studies are planned combining DAC plusvalproic acid.

HDACi synergize with retinoids in a variety of sys-tems.35,41 This has led to the concept of potentially com-bining these classes of agents to build on the modestsingle activity of retinoids in MDS as single agents orin combination with growth factors.42 Studies combin-ing all trans-retinoic acid with PB and with valproateare ongoing. Since RARβ is methylated and silencedin a variety of cancers, including myeloid malignan-cies, some rationale exists to study the combination ofa methyltransferase inhibitor, HDACi, and retinoid.

Farnesyl Transferase InhibitionInterest in inhibition of farnesyl transferase in myeloidmalignancies derived from the observation that rasmutations were common in these leukemias. Tipifarnibhas undergone Phase I testing in AML, and a Phase IItrial is ongoing. In the Phase I trial, significant clinicalresponses were seen at all dose levels tested, including2 complete remissions. Interestingly, no patients in thatinitial trial were found to have mutations of ras. Inhibi-tion of farnesyl transferase was demonstrated, begin-ning at the dose level of 300 mg BID; farnesylation oftarget proteins was inhibited at 600 mg BID. In the PhaseII trial, patients with high-risk AML (age > 60, knownadverse cytogenetics, therapy- or MDS-related AML)and high-grade MDS, have been treated with tipifarnib600 mg BID for 21/28 days. An overall response rateof 37% has been observed to date (J.E. Karp, personalcommunication). In a separate Phase I trial of tipifarnibin MDS, Kurzrock and colleagues demonstrated inhi-bition of farnesyl transferase; clinical responses wereseen, including a complete response.43

Another farnsesyl transferase inhibitor, lonafarnib,has undergone early Phase I and II studies in patientswith hematologic malignancies.44,45 Two patients withMDS were treated on the Phase I, and 15 on the PhaseII trial. Five patients with CMML were treated on thePhase I trial, and 12 on the Phase II trial. Of the 5 CMMLpatients on the Phase I trial, 3 developed hematologicimprovement. Of the 15 MDS patients on the Phase IIstudy, 3 developed hematologic improvement, while 6

Hematology 2003 185

mulation has been studied which can supply similarplasma concentrations; however, the daily dosing re-quires as many as 75 pills per day.36 Despite the diffi-culty with delivery of concentrations of PB with HDACinhibitory activity, PB remains extremely well-tolerated.

Recent data demonstrating that valproic acid hassimilar HDAC inhibitory activity has raised hopes thatoral formulations of this drug, which is commonly usedfor neuropsychiatric disorders, could be used to modu-late gene expression.37 Valproate is used clinically atsubmillimolar concentrations; in vitro acetylationchanges have been demonstrated at 0.5 mM. Thepharmacodynamic impact of valproate on leukemiccells is similar to PB. Valproate is currently being in-vestigated in monotherapy and in combination withother drugs in MDS and in AML.

SAHA and FK228: The prototype HDAC inhibitorused for in vitro study, trichostatin, is an hydroxamicacid. To date, the only hydroxamic acid under clinicalinvestigation is suberoylanilide hydroxamic acid(SAHA). SAHA has been developed in both intrave-nous and oral formulations.38 Clinical administrationhas been associated with induction of histone acetyla-tion. No data are yet available detailing the impact ofSAHA on myeloid malignancies.

FK228 (depsipeptide) is a cyclic peptide whoseHDAC inhibitory activity appears to be due in greatpart to a reduced metabolite. FK228 appears to be spe-cific for Class I HDAC. FK228 has been studied inPhase I clinical trials including 1 in AML.39 Tumor ly-sis syndrome has been induced by FK228, indicatingsignificant cytotoxicity to myeloid cells; however, nosustained responses were seen in the Phase I trial. As-thenia appears to be dose limiting.

MS-275: MS-275 is a benzamide HDAC inhibitorundergoing Phase I investigation in MDS and AML.Similar to FK228, fatigue seems to be a major toxicityof MS275. MS275 has a half-life of greater than 24hours; changes in histone acetylation have persisted forseveral weeks following the administration of MS-275.No sustained responses have been reported to date (JEKarp, personal communication).

Integration of HDAC inhibitors intothe treatment of MDSThe recognition that HDAC recruitment accounted atleast in part for the silencing of genes with methylatedpromoters led to the demonstration that optimalreexpression of such genes required sequential expo-sure to a methyltransferase inhibitor followed by anHDACi. While definitive demonstration of a correla-tion between methyltransferase inhibition and clinicalactivity of methyltransferase inhibitors in MDS has yet

to be provided, a great deal of interest remains in aug-menting the clinical activity of 5AC and DAC throughthe addition of an HDACi. To date, data are only avail-able from a dose-finding study of 5AC followed by PB.In this study, a variety of doses and schedules of 5AChave been administered, followed by a 7-day continu-ous infusion of PB.40 The combination has been well-tolerated, and significant sustained clinical responseshave been achieved. Changes in methylation, histoneacetylation, and gene reexpression are being monitored.Similar studies are planned combining DAC plusvalproic acid.

HDACi synergize with retinoids in a variety of sys-tems.35,41 This has led to the concept of potentially com-bining these classes of agents to build on the modestsingle activity of retinoids in MDS as single agents orin combination with growth factors.42 Studies combin-ing all trans-retinoic acid with PB and with valproateare ongoing. Since RARβ is methylated and silencedin a variety of cancers, including myeloid malignan-cies, some rationale exists to study the combination ofa methyltransferase inhibitor, HDACi, and retinoid.

Farnesyl Transferase InhibitionInterest in inhibition of farnesyl transferase in myeloidmalignancies derived from the observation that rasmutations were common in these leukemias. Tipifarnibhas undergone Phase I testing in AML, and a Phase IItrial is ongoing. In the Phase I trial, significant clinicalresponses were seen at all dose levels tested, including2 complete remissions. Interestingly, no patients in thatinitial trial were found to have mutations of ras. Inhibi-tion of farnesyl transferase was demonstrated, begin-ning at the dose level of 300 mg BID; farnesylation oftarget proteins was inhibited at 600 mg BID. In the PhaseII trial, patients with high-risk AML (age > 60, knownadverse cytogenetics, therapy- or MDS-related AML)and high-grade MDS, have been treated with tipifarnib600 mg BID for 21/28 days. An overall response rateof 37% has been observed to date (J.E. Karp, personalcommunication). In a separate Phase I trial of tipifarnibin MDS, Kurzrock and colleagues demonstrated inhi-bition of farnesyl transferase; clinical responses wereseen, including a complete response.43

Another farnsesyl transferase inhibitor, lonafarnib,has undergone early Phase I and II studies in patientswith hematologic malignancies.44,45 Two patients withMDS were treated on the Phase I, and 15 on the PhaseII trial. Five patients with CMML were treated on thePhase I trial, and 12 on the Phase II trial. Of the 5 CMMLpatients on the Phase I trial, 3 developed hematologicimprovement. Of the 15 MDS patients on the Phase IIstudy, 3 developed hematologic improvement, while 6

Further Analysis of Trials With Azacitidine in Patients WithMyelodysplastic Syndrome: Studies 8421, 8921, and 9221 bythe Cancer and Leukemia Group BLewis R. Silverman, David R. McKenzie, Bercedis L. Peterson, James F. Holland, Jay T. Backstrom, C.L. Beach,and Richard A. Larson

A B S T R A C T

PurposeWithin the last two decades, a new understanding of the biology of myelodysplastic syndrome (MDS)has developed. With this understanding, new classification systems, such as the WHO diagnosticcriteria, and the International Prognostic Scoring System and response criteria guidelines reported bythe International Working Group (IWG) have been developed. We report the combined results of threepreviously reported clinical trials (n � 309) with azacitidine using the WHO classification system forMDS and acute myeloid leukemia (AML) and IWG criteria for response.

Patients and MethodsData from three sequential Cancer and Leukemia Group B trials with azacitidine were recollectedand reanalyzed as part of the New Drug Application process. The trials were conducted with eitherintravenous or subcutaneous azacitidine (75 mg/m2/d for 7 days every 28 days).

ResultsComplete remissions were seen in 10% to 17% of azacitidine-treated patients; partial remissionswere rare; 23% to 36% of patients had hematologic improvement (HI). The median number ofcycles to first response was three, and 90% of responses were seen by cycle 6. Using currentWHO criteria, 103 patients had AML at baseline; 35% to 48% had HI or better responses. Themedian survival time for the 27 AML patients randomly assigned to azacitidine was 19.3 monthscompared with 12.9 months for the 25 patients assigned to observation. Furthermore, azacitidinedid not increase the rate of infection or bleeding above the rate caused by underlying disease.

ConclusionAzacitidine provides important clinical benefits for patients with high-risk MDS.

J Clin Oncol 24:3895-3903. © 2006 by American Society of Clinical Oncology

INTRODUCTION

In 1984, the Cancer and Leukemia Group B(CALGB) began a series of clinical trials with aza-citidine (Vidaza; Pharmion Corporation, Over-land Park, KS) in patients with myelodysplasticsyndrome (MDS).1-4 These studies and other sup-portive data culminated in the 2004 US Food andDrug Administration approval of azacitidine fortreatment of symptomatic patients with MDS.During the intervening two decades, a greater un-derstanding of the biology of myelodysplasia hasevolved, along with a new classification systemdeveloped by WHO that more clearly distin-guishes MDS from acute myeloid leukemia(AML) and from chronic myeloproliferative dis-orders.5,6 In addition, an International WorkingGroup (IWG) sponsored by the National CancerInstitute (NCI) has published new response crite-

ria for evaluation of new treatments for MDS.7,8

As part of the New Drug Application process,Pharmion recollected and reanalyzed the CALGBdata, including expert pathology review of bloodand bone marrow slides. Some of the CALGB datafrom these three trials was previously publishedusing the protocol-specified diagnostic and re-sponse criteria. Here, we report the combinedresults of a reanalysis using the WHO classifica-tion for MDS and AML and the IWG criteria forresponse in MDS.

PATIENTS AND METHODS

Data Collection

For the reanalysis, data were recollected from pa-tients enrolled onto CALGB Protocols 8421, 8921, and9221.1-3 A comprehensive retrospective collection and re-verification of all clinical data in the original protocols

From the Mount Sinai School of Medi-cine, New York, NY; Pharmion Corpora-tion, Overland Park, KS; Cancer andLeukemia Group B Statistical Centerand Duke University, Durham, NC; andthe University of Chicago, Chicago, IL.

Submitted February 2, 2006; acceptedJune 15, 2006.

Supported by federal grants from theUS Food and Drug Administration, byNational Cancer Institute (NCI) GrantNo. CA31946 to the Cancer andLeukemia Group B, and by NCI GrantsNo. CA04457, CA33601, andCA41287. The subsequent recollec-tion and further analyses were spon-sored by Pharmion Corporation.

The content of this article is solely theresponsibility of the authors and doesnot necessarily represent the officialviews of the National Cancer Institute.

Authors’ disclosures of potential con-flicts of interest and author contribu-tions are found at the end of thisarticle.

Address reprint requests to Lewis R.Silverman, MD, Division of Hematology/Oncology, Mount Sinai School of Medi-cine, One Gustave L. Levy Place, Box1129, New York, NY 10029; e-mail:[email protected].

© 2006 by American Society of ClinicalOncology

0732-183X/06/2424-3895/$20.00

DOI: 10.1200/JCO.2005.05.4346

JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T

VOLUME 24 � NUMBER 24 � AUGUST 20 2006

3895Downloaded from jco.ascopubs.org on April 28, 2012. For personal use only. No other uses without permission.

Copyright © 2006 American Society of Clinical Oncology. All rights reserved.

Further Analysis of Trials With Azacitidine in Patients WithMyelodysplastic Syndrome: Studies 8421, 8921, and 9221 bythe Cancer and Leukemia Group BLewis R. Silverman, David R. McKenzie, Bercedis L. Peterson, James F. Holland, Jay T. Backstrom, C.L. Beach,and Richard A. Larson

A B S T R A C T

PurposeWithin the last two decades, a new understanding of the biology of myelodysplastic syndrome (MDS)has developed. With this understanding, new classification systems, such as the WHO diagnosticcriteria, and the International Prognostic Scoring System and response criteria guidelines reported bythe International Working Group (IWG) have been developed. We report the combined results of threepreviously reported clinical trials (n � 309) with azacitidine using the WHO classification system forMDS and acute myeloid leukemia (AML) and IWG criteria for response.

Patients and MethodsData from three sequential Cancer and Leukemia Group B trials with azacitidine were recollectedand reanalyzed as part of the New Drug Application process. The trials were conducted with eitherintravenous or subcutaneous azacitidine (75 mg/m2/d for 7 days every 28 days).

ResultsComplete remissions were seen in 10% to 17% of azacitidine-treated patients; partial remissionswere rare; 23% to 36% of patients had hematologic improvement (HI). The median number ofcycles to first response was three, and 90% of responses were seen by cycle 6. Using currentWHO criteria, 103 patients had AML at baseline; 35% to 48% had HI or better responses. Themedian survival time for the 27 AML patients randomly assigned to azacitidine was 19.3 monthscompared with 12.9 months for the 25 patients assigned to observation. Furthermore, azacitidinedid not increase the rate of infection or bleeding above the rate caused by underlying disease.

ConclusionAzacitidine provides important clinical benefits for patients with high-risk MDS.

J Clin Oncol 24:3895-3903. © 2006 by American Society of Clinical Oncology

INTRODUCTION

In 1984, the Cancer and Leukemia Group B(CALGB) began a series of clinical trials with aza-citidine (Vidaza; Pharmion Corporation, Over-land Park, KS) in patients with myelodysplasticsyndrome (MDS).1-4 These studies and other sup-portive data culminated in the 2004 US Food andDrug Administration approval of azacitidine fortreatment of symptomatic patients with MDS.During the intervening two decades, a greater un-derstanding of the biology of myelodysplasia hasevolved, along with a new classification systemdeveloped by WHO that more clearly distin-guishes MDS from acute myeloid leukemia(AML) and from chronic myeloproliferative dis-orders.5,6 In addition, an International WorkingGroup (IWG) sponsored by the National CancerInstitute (NCI) has published new response crite-

ria for evaluation of new treatments for MDS.7,8

As part of the New Drug Application process,Pharmion recollected and reanalyzed the CALGBdata, including expert pathology review of bloodand bone marrow slides. Some of the CALGB datafrom these three trials was previously publishedusing the protocol-specified diagnostic and re-sponse criteria. Here, we report the combinedresults of a reanalysis using the WHO classifica-tion for MDS and AML and the IWG criteria forresponse in MDS.

PATIENTS AND METHODS

Data Collection

For the reanalysis, data were recollected from pa-tients enrolled onto CALGB Protocols 8421, 8921, and9221.1-3 A comprehensive retrospective collection and re-verification of all clinical data in the original protocols

From the Mount Sinai School of Medi-cine, New York, NY; Pharmion Corpora-tion, Overland Park, KS; Cancer andLeukemia Group B Statistical Centerand Duke University, Durham, NC; andthe University of Chicago, Chicago, IL.

Submitted February 2, 2006; acceptedJune 15, 2006.

Supported by federal grants from theUS Food and Drug Administration, byNational Cancer Institute (NCI) GrantNo. CA31946 to the Cancer andLeukemia Group B, and by NCI GrantsNo. CA04457, CA33601, andCA41287. The subsequent recollec-tion and further analyses were spon-sored by Pharmion Corporation.

The content of this article is solely theresponsibility of the authors and doesnot necessarily represent the officialviews of the National Cancer Institute.

Authors’ disclosures of potential con-flicts of interest and author contribu-tions are found at the end of thisarticle.

Address reprint requests to Lewis R.Silverman, MD, Division of Hematology/Oncology, Mount Sinai School of Medi-cine, One Gustave L. Levy Place, Box1129, New York, NY 10029; e-mail:[email protected].

© 2006 by American Society of ClinicalOncology

0732-183X/06/2424-3895/$20.00

DOI: 10.1200/JCO.2005.05.4346

JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T

VOLUME 24 � NUMBER 24 � AUGUST 20 2006

3895Downloaded from jco.ascopubs.org on April 28, 2012. For personal use only. No other uses without permission.

Copyright © 2006 American Society of Clinical Oncology. All rights reserved.

relative to standard AML remission induction chemotherapy, theprolongation in survival time to 19.3 months exceeds that typicallyseen with standard induction chemotherapy, suggesting that aza-citidine may alter the natural history of the disease independent ofCR response criteria.10 Furthermore, treatment with azacitidine isassociated with significant reduction in risk of transformation toAML and a significant prolongation of survival in patients withhigh-risk MDS, including RAEB patients, RAEB-T patients � 65years of age, and patients with equivalent intermediate-2 and high-risk disease.11,12 This finding and the data presented in this articlesuggest a paradigm shift, with azacitidine altering the natural his-tory of MDS by modulating the behavior of the MDS clone withoutnecessarily eradicating it. In addition, treatment with azacitidinesignificantly delays the onset of RBC and platelet transfusions inpatients who are transfusion independent at study entry. Thesefindings warrant further studies of azacitidine in patients withsmoldering AML with multilineage dysplasia (ie, patients whowere previously diagnosed as having RAEB-T).

The time to response data indicate that azacitidine can have aneffect at the bone marrow level as early as the first treatment cycle.However, for this effect to translate into an improvement in bone

marrow function leading to clinically significant increases in periph-eral cell counts, the majority of responders can require up to six cyclesof azacitidine. To ensure adequate exposure for patients to demon-strate a clinical response, azacitidine should be administered for aminimum of four cycles. Furthermore, patients will most likely con-tinue to require transfusion support during the first several cycles oftreatment with azacitidine. In patients who were transfusion depen-dent at baseline with a response, azacitidine was associated with amedian of 9 months of transfusion independence.

Given the underlying disease process and the myelotoxicity ofcompounds such as azacitidine and decitabine, an increase in rates ofinfection and bleeding would be expected during treatment. Despitethe potential to exacerbate pre-existing cytopenias early in therapy,azacitidine did not increase the rate of infection or bleeding above therate caused by underlying disease.

This reanalysis demonstrates that azacitidine is effective therapythat directly impacts the disease of MDS rather than just managing thesymptoms. It reconfirms the findings discussed in Silverman et al3 andadds additional data pertaining to safety and more current classifica-tion and response criteria. Furthermore, azacitidine warrants addi-tional studies in patients with AML with dysplasia.

REFERENCES

1. Silverman LR, Holland JF, Weinberg RS, et al:Effects of treatment with 5-azacytidine on the in vivoand in vitro hematopoiesis in patients with myelodys-plastic syndromes. Leukemia 7:21-29, 1993

2. Silverman LR, Holland JF, Demakos EP, et al:Azacitidine (Aza C) in myelodysplastic syndromes(MDS), CALGB studies 8421 and 8921. Ann Hema-tol 68:A12, 1994 (abstr 46)

3. Silverman LR, Demakos EP, Peterson BL, etal: Randomized controlled trial of azacitidine in pa-tients with the myelodysplastic syndrome: A studyof the cancer and leukemia group B. J Clin Oncol20:2429-2440, 2002

4. Kornblith AB, Herndon JE II, Silverman LR, etal: Impact of azacytidine on the quality of life ofpatients with myelodysplastic syndrome treated in arandomized phase III Trial: A Cancer and LeukemiaGroup B Study. J Clin Oncol 20:2441-2452, 2002

5. Harris NL, Jaffe ES, Diebold J, et al: WorldHealth Organization classification of neoplastic dis-eases of the hematopoietic and lymphoid tissues:Report of the Clinical Advisory Committee Meeting–Airlie House, Virginia, November 1997. J Clin Oncol17:3835-3849, 1999

6. Germing U, Gattermann N, Strupp C, et al: Valida-tion of the WHO proposals for a new classification ofprimary myelodysplastic syndromes: A retrospectiveanalysis of 1600 patients. Leuk Res 24:983-992, 2000

Table 9. NCI CTC Grades 1 to 4 Bleeding Rates (patient-years of exposure) in Protocol 9221

Adverse Event†

Azacitidine Patients�

(n � 150)Observation Patients

(n � 92)

Patients With Events PerPatient-Year of Exposure‡

No. ofPatients

Patients With Events PerPatient-Year of Exposure‡

No. ofPatients

Total bleeding§ 0.56 77 0.60 26GI disorders§ 0.26 36 0.25 11

GI hemorrhage 0.03 4 0 0Gingival bleeding 0.13 18 0.09 4Hemorrhoidal hemorrhage 0.04 6 0 0Melena 0.03 4 0.05 2Oral hemorrhage 0.04 5 0.02 1Rectal hemorrhage 0.05 7 0.05 2

Total nervous system disorders§ 0.01 2 0.02 1Intracranial hemorrhage 0.01 2 0 0Subdural hematoma 0 0 0.02 1

Total renal and urinarydisorders,§ hematuria

0.05 7 0.07 3

Total respiratory, thoracic andmediastinal disorders§

0.22 30 0.23 10

Epistaxis 0.18 25 0.21 9Hemoptysis 0.05 7 0.02 1

Abbreviations: NCI CTC, National Cancer Institute Common Toxicity Criteria; NOS, not otherwise specified.�Includes all patients exposed to azacitidine, including patients who crossed over to azacitidine from observation.†Multiple reports of the same adverse event term for a patient are only counted once within each treatment group.‡Total exposure for azacitidine is the cumulative time from the first dose to the end of study (30 days after last dose), and for observation, total exposure is the

cumulative time from random assignment to withdrawal from study or day before cross over.§System organ class using MedDRA, version 5.0 (Northrop Grumman, Los Angeles, CA).

Silverman et al

3902 JOURNAL OF CLINICAL ONCOLOGY

Downloaded from jco.ascopubs.org on April 28, 2012. For personal use only. No other uses without permission.Copyright © 2006 American Society of Clinical Oncology. All rights reserved.

Decitabine Improves Patient Outcomes inMyelodysplastic SyndromesResults of a Phase III Randomized Study

Hagop Kantarjian M.D.1

Jean-Pierre J. Issa, M.D.1

Craig S. Rosenfeld, M.D., Ph.D.2

John M. Bennett, M.D.3

Maher Albitar, M.D.4

John DiPersio, M.D.5

Virginia Klimek, M.D.6

James Slack, M.D.7

Carlos de Castro, M.D.8

Farhad Ravandi, M.D.9

Richard Helmer III, M.D.10

Lanlan Shen, M.D.1

Stephen D. Nimer, M.D.6

Richard Leavitt, M.D.11

Azra Raza, M.D.12

Hussain Saba, M.D.13,14

1 Department of Leukemia, University of TexasM. D. Anderson Cancer Center, Houston, Texas.

2 Dallas, Texas.

3 James P. Wilmot Cancer Center, University ofRochester Medical Center, Rochester, New York.

4 Nichols Institute, Quest Diagnostics, San JuanCapistrano, California.

5 Washington University School of Medicine, St.Louis, Missouri.

6 Memorial Sloan-Kettering Cancer Center, NewYork, New York.

7 Roswell Park Cancer Institute, Buffalo, New York.

8 Duke University Medical Center, Durham, NorthCarolina.

9 University of Illinois, Chicago, Illinois.

10 Southwest Regional Cancer Center, Austin,Texas.

11 Pharma Pacific, Inc., Woodside, California.

12 Rush Medical Center, Chicago, Illinois.

13 James A. Haley Veterans Hospital, Tampa, Flor-ida.

14 H. Lee Moffitt Cancer Center, Tampa, Florida.

See related editorial on pages 1650-2, and ac-companying article on pages 1744-50, this issue.

Address for reprints: Hagop Kantarjian, M.D., De-partment of Leukemia, Box 428, The University ofTexas M. D. Anderson Cancer Center, 1515 Hol-combe Blvd., Houston, TX 77030; Fax: (713) 794–4297; E-mail: [email protected]

Dr. Rosenfeld owned SuperGen, Inc. stock and

nonexercised SuperGen options at the time thisarticle was submitted.

Dr. Klimer is a member of the Advisory BoardFaculty for SuperGen, Inc. and has received hon-oraria.

Dr. Leavitt serves as a paid medical and regulatoryaffairs consultant for SuperGen, Inc. and MG/Pharma Inc.

Received October 12, 2005; revision received No-vember 2, 2005; accepted December 2, 2005.

BACKGROUND. Aberrant DNA methylation, which results in leukemogenesis, is

frequent in patients with myelodysplastic syndromes (MDS) and is a potential

target for pharmacologic therapy. Decitabine indirectly depletes methylcytosine

and causes hypomethylation of target gene promoters.

METHODS. A total of 170 patients with MDS were randomized to receive either

decitabine at a dose of 15 mg/m2 given intravenously over 3 hours every 8 hours for

3 days (at a dose of 135 mg/m2 per course) and repeated every 6 weeks, or best

supportive care. Response was assessed using the International Working Group

criteria and required that response criteria be met for at least 8 weeks.

RESULTS. Patients who were treated with decitabine achieved a significantly higher

overall response rate (17%), including 9% complete responses, compared with

supportive care (0%) (P � .001). An additional 12 patients who were treated with

decitabine (13%) achieved hematologic improvement. Responses were durable

(median, 10.3 mos) and were associated with transfusion independence. Patients

treated with decitabine had a trend toward a longer median time to acute myelog-

enous leukemia (AML) progression or death compared with patients who received

supportive care alone (all patients, 12.1 mos vs. 7.8 mos [P � 0.16]; those with

International Prognostic Scoring System intermediate-2/high-risk disease, 12.0

mos vs. 6.8 mos [P � 0.03]; those with de novo disease, 12.6 mos vs. 9.4 mos

[P � 0.04]; and treatment-naive patients, 12.3 mos vs. 7.3 mos [P � 0.08]).

CONCLUSIONS. Decitabine was found to be clinically effective in the treatment of

patients with MDS, provided durable responses, and improved time to AML

transformation or death. The duration of decitabine therapy may improve these

results further. Cancer 2006;106:1794 – 80. © 2006 American Cancer Society.

KEYWORDS: decitabine, azacitidine, myelodysplastic syndrome, acute myelogenousleukemia, hypomethylating.

1794

© 2006 American Cancer SocietyDOI 10.1002/cncr.21792Published online 13 March 2006 in Wiley InterScience (www.interscience.wiley.com).

Decitabine Improves Patient Outcomes inMyelodysplastic SyndromesResults of a Phase III Randomized Study

Hagop Kantarjian M.D.1

Jean-Pierre J. Issa, M.D.1

Craig S. Rosenfeld, M.D., Ph.D.2

John M. Bennett, M.D.3

Maher Albitar, M.D.4

John DiPersio, M.D.5

Virginia Klimek, M.D.6

James Slack, M.D.7

Carlos de Castro, M.D.8

Farhad Ravandi, M.D.9

Richard Helmer III, M.D.10

Lanlan Shen, M.D.1

Stephen D. Nimer, M.D.6

Richard Leavitt, M.D.11

Azra Raza, M.D.12

Hussain Saba, M.D.13,14

1 Department of Leukemia, University of TexasM. D. Anderson Cancer Center, Houston, Texas.

2 Dallas, Texas.

3 James P. Wilmot Cancer Center, University ofRochester Medical Center, Rochester, New York.

4 Nichols Institute, Quest Diagnostics, San JuanCapistrano, California.

5 Washington University School of Medicine, St.Louis, Missouri.

6 Memorial Sloan-Kettering Cancer Center, NewYork, New York.

7 Roswell Park Cancer Institute, Buffalo, New York.

8 Duke University Medical Center, Durham, NorthCarolina.

9 University of Illinois, Chicago, Illinois.

10 Southwest Regional Cancer Center, Austin,Texas.

11 Pharma Pacific, Inc., Woodside, California.

12 Rush Medical Center, Chicago, Illinois.

13 James A. Haley Veterans Hospital, Tampa, Flor-ida.

14 H. Lee Moffitt Cancer Center, Tampa, Florida.

See related editorial on pages 1650-2, and ac-companying article on pages 1744-50, this issue.

Address for reprints: Hagop Kantarjian, M.D., De-partment of Leukemia, Box 428, The University ofTexas M. D. Anderson Cancer Center, 1515 Hol-combe Blvd., Houston, TX 77030; Fax: (713) 794–4297; E-mail: [email protected]

Dr. Rosenfeld owned SuperGen, Inc. stock and

nonexercised SuperGen options at the time thisarticle was submitted.

Dr. Klimer is a member of the Advisory BoardFaculty for SuperGen, Inc. and has received hon-oraria.

Dr. Leavitt serves as a paid medical and regulatoryaffairs consultant for SuperGen, Inc. and MG/Pharma Inc.

Received October 12, 2005; revision received No-vember 2, 2005; accepted December 2, 2005.

BACKGROUND. Aberrant DNA methylation, which results in leukemogenesis, is

frequent in patients with myelodysplastic syndromes (MDS) and is a potential

target for pharmacologic therapy. Decitabine indirectly depletes methylcytosine

and causes hypomethylation of target gene promoters.

METHODS. A total of 170 patients with MDS were randomized to receive either

decitabine at a dose of 15 mg/m2 given intravenously over 3 hours every 8 hours for

3 days (at a dose of 135 mg/m2 per course) and repeated every 6 weeks, or best

supportive care. Response was assessed using the International Working Group

criteria and required that response criteria be met for at least 8 weeks.

RESULTS. Patients who were treated with decitabine achieved a significantly higher

overall response rate (17%), including 9% complete responses, compared with

supportive care (0%) (P � .001). An additional 12 patients who were treated with

decitabine (13%) achieved hematologic improvement. Responses were durable

(median, 10.3 mos) and were associated with transfusion independence. Patients

treated with decitabine had a trend toward a longer median time to acute myelog-

enous leukemia (AML) progression or death compared with patients who received

supportive care alone (all patients, 12.1 mos vs. 7.8 mos [P � 0.16]; those with

International Prognostic Scoring System intermediate-2/high-risk disease, 12.0

mos vs. 6.8 mos [P � 0.03]; those with de novo disease, 12.6 mos vs. 9.4 mos

[P � 0.04]; and treatment-naive patients, 12.3 mos vs. 7.3 mos [P � 0.08]).

CONCLUSIONS. Decitabine was found to be clinically effective in the treatment of

patients with MDS, provided durable responses, and improved time to AML

transformation or death. The duration of decitabine therapy may improve these

results further. Cancer 2006;106:1794 – 80. © 2006 American Cancer Society.

KEYWORDS: decitabine, azacitidine, myelodysplastic syndrome, acute myelogenousleukemia, hypomethylating.

1794

© 2006 American Cancer SocietyDOI 10.1002/cncr.21792Published online 13 March 2006 in Wiley InterScience (www.interscience.wiley.com).

Chromosome 5q deletion and epigenetic suppressionof the gene encoding a-catenin (CTNNA1) in myeloidcell transformationTing Xi Liu1,2,8, Michael W Becker3,8, Jaroslav Jelinek4, Wen-Shu Wu2, Min Deng1,2, Natallia Mikhalkevich3,Karl Hsu2, Clara D Bloomfield5, Richard M Stone6, Daniel J DeAngelo6, Ilene A Galinsky6, Jean-Pierre Issa4,Michael F Clarke7 & A Thomas Look2

Interstitial loss of all or part of the long arm of chromosome 5,

or del(5q), is a frequent clonal chromosomal abnormality

in human myelodysplastic syndrome (MDS, a preleukemic

disorder) and acute myeloid leukemia (AML)1, and is thought to

contribute to the pathogenesis of these diseases by deleting one

or more tumor-suppressor genes2. Although a major commonly

deleted region (CDR) has been delineated on chromosome band

5q31.1 (refs. 3–7), attempts to identify tumor suppressors

within this band have been unsuccessful. We focused our

analysis of gene expression on RNA from primitive leukemia-

initiating cells, which harbor 5q deletions8,9, and analyzed

12 genes within the CDR that are expressed by normal

hematopoietic stem cells. Here we show that the gene encoding

a-catenin (CTNNA1) is expressed at a much lower level in

leukemia-initiating stem cells from individuals with AML or

MDS with a 5q deletion than in individuals with MDS or AML

lacking a 5q deletion or in normal hematopoietic stem cells.

Analysis of HL-60 cells, a myeloid leukemia line with deletion

of the 5q31 region10,11, showed that the CTNNA1 promoter

of the retained allele is suppressed by both methylation and

histone deacetylation. Restoration of CTNNA1 expression in

HL-60 cells resulted in reduced proliferation and apoptotic

cell death. Thus, loss of expression of the a-catenin tumor

suppressor in hematopoietic stem cells may provide a growth

advantage that contributes to human MDS or AML with del(5q).

Leukemia-initiating cells (L-ICs), with a phenotype ofCD34+CD38–CD123+Lin–, are a minor population within bothMDS and AML but can generate leukemia after transplantation intononobese diabetic/severe combined immunodeficient (NOD/SCID)mice12, a property not shared by the larger subset of more matureCD34+CD38+ or CD34–leukemic blast cells8,13,14. This distinction,

together with detection of the del(5q) abnormality in CD34+CD38–

cells from both MDS and AML individuals9, provided a compellingrationale to search for tumor-suppressor genes within the 5q31.1 CDR(refs. 3–7) of L-ICs isolated from MDS and AML cases with thedel(5q). First, we collected cell samples from normal healthy donors,12 MDS/AML individuals with either del(5q) (n ¼ 10) or loss of asingle copy of chromosome 5 (–5; n ¼ 2), and 10 MDS/AMLindividuals without abnormalities of chromosome 5 (SupplementaryTable 1 online). The percentage of morphologically identifiableleukemic blast cells in the bone marrow or peripheral blood of theseindividuals ranged from 3% to 91%, with variable blast cell cyto-genetic findings (Supplementary Table 1). Normal hematopoieticstem cells (HSCs) (CD34+CD38–CD90+Lin–, Fig. 1a) from normaldonors and L-ICs (CD34+CD38–CD123+Lin–) from individuals withMDS/AML with or without chromosome 5 abnormalities (Fig. 1a)were enriched by flow-activated cell sorting (FACS) to a puritydetermined to be greater than 90%.

Previous studies9 have demonstrated the involvement ofCD34+CD38– cells by the del(5q) clone in individuals with AML/MDS; however, as this population of cells is highly heterogenous andcontains many non–L-ICs, we used dual-probe in situ hybridization todetermine the frequency of involvement of the L-IC by the malignantdel (5q) clone in highly enriched cell populations from five MDS/AMLindividuals. As a control, we performed similar analyses in HL-60 andKG-1 human leukemic cell lines, both of which harbor a deletion of5q11-q31 (refs. 10,11,15). As the probe for del(5q), we used a bacterialartificial chromosome that contains the EGR1, ETF1 and HSPA9Bgenomic loci, ending within 164 kb of the CTNNA1 locus, and thathas previously been used to demonstrate the 5q deletion in subsets ofFACS-sorted primary bone marrow cells9. The results show that 90%of HL-60 cells and 99% of KG1 cells carried the 5q deletion (Fig. 1band Supplementary Table 2 online). In contrast, a probe specific for

Received 25 July; accepted 3 November; published online 10 December 2006; doi:10.1038/nm1512

1Laboratory of Development and Diseases, Institute of Health Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences and Shanghai JiaoTong University School of Medicine, Shanghai 200025, China. 2Department of Pediatric Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston,Massachusetts 02115, USA. 3Division of Hematology/Oncology, University of Rochester, Rochester, New York 14642, USA. 4University of Texas M.D. Anderson CancerCenter, Houston, Texas 77030, USA. 5Ohio State University Comprehensive Cancer Center, Columbus, Ohio 43210, and Cancer and Leukemia Group, University ofChicago, Chicago, Illinois 60606, USA. 6Department of Adult Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115, USA.7Institute of Stem Cell Biology and Regenerative Medicine, and Division of Hematology/Oncology, Internal Medicine, Stanford University, Palo Alto, California 94304,USA. 8These authors contributed equally to this work. Correspondence should be addressed to A.T.L. ([email protected]).

78 VOLUME 13 [ NUMBER 1 [ JANUARY 2007 NATURE MEDICINE

L E T T ERS

©20

07 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

Chromosome 5q deletion and epigenetic suppressionof the gene encoding a-catenin (CTNNA1) in myeloidcell transformationTing Xi Liu1,2,8, Michael W Becker3,8, Jaroslav Jelinek4, Wen-Shu Wu2, Min Deng1,2, Natallia Mikhalkevich3,Karl Hsu2, Clara D Bloomfield5, Richard M Stone6, Daniel J DeAngelo6, Ilene A Galinsky6, Jean-Pierre Issa4,Michael F Clarke7 & A Thomas Look2

Interstitial loss of all or part of the long arm of chromosome 5,

or del(5q), is a frequent clonal chromosomal abnormality

in human myelodysplastic syndrome (MDS, a preleukemic

disorder) and acute myeloid leukemia (AML)1, and is thought to

contribute to the pathogenesis of these diseases by deleting one

or more tumor-suppressor genes2. Although a major commonly

deleted region (CDR) has been delineated on chromosome band

5q31.1 (refs. 3–7), attempts to identify tumor suppressors

within this band have been unsuccessful. We focused our

analysis of gene expression on RNA from primitive leukemia-

initiating cells, which harbor 5q deletions8,9, and analyzed

12 genes within the CDR that are expressed by normal

hematopoietic stem cells. Here we show that the gene encoding

a-catenin (CTNNA1) is expressed at a much lower level in

leukemia-initiating stem cells from individuals with AML or

MDS with a 5q deletion than in individuals with MDS or AML

lacking a 5q deletion or in normal hematopoietic stem cells.

Analysis of HL-60 cells, a myeloid leukemia line with deletion

of the 5q31 region10,11, showed that the CTNNA1 promoter

of the retained allele is suppressed by both methylation and

histone deacetylation. Restoration of CTNNA1 expression in

HL-60 cells resulted in reduced proliferation and apoptotic

cell death. Thus, loss of expression of the a-catenin tumor

suppressor in hematopoietic stem cells may provide a growth

advantage that contributes to human MDS or AML with del(5q).

Leukemia-initiating cells (L-ICs), with a phenotype ofCD34+CD38–CD123+Lin–, are a minor population within bothMDS and AML but can generate leukemia after transplantation intononobese diabetic/severe combined immunodeficient (NOD/SCID)mice12, a property not shared by the larger subset of more matureCD34+CD38+ or CD34–leukemic blast cells8,13,14. This distinction,

together with detection of the del(5q) abnormality in CD34+CD38–

cells from both MDS and AML individuals9, provided a compellingrationale to search for tumor-suppressor genes within the 5q31.1 CDR(refs. 3–7) of L-ICs isolated from MDS and AML cases with thedel(5q). First, we collected cell samples from normal healthy donors,12 MDS/AML individuals with either del(5q) (n ¼ 10) or loss of asingle copy of chromosome 5 (–5; n ¼ 2), and 10 MDS/AMLindividuals without abnormalities of chromosome 5 (SupplementaryTable 1 online). The percentage of morphologically identifiableleukemic blast cells in the bone marrow or peripheral blood of theseindividuals ranged from 3% to 91%, with variable blast cell cyto-genetic findings (Supplementary Table 1). Normal hematopoieticstem cells (HSCs) (CD34+CD38–CD90+Lin–, Fig. 1a) from normaldonors and L-ICs (CD34+CD38–CD123+Lin–) from individuals withMDS/AML with or without chromosome 5 abnormalities (Fig. 1a)were enriched by flow-activated cell sorting (FACS) to a puritydetermined to be greater than 90%.

Previous studies9 have demonstrated the involvement ofCD34+CD38– cells by the del(5q) clone in individuals with AML/MDS; however, as this population of cells is highly heterogenous andcontains many non–L-ICs, we used dual-probe in situ hybridization todetermine the frequency of involvement of the L-IC by the malignantdel (5q) clone in highly enriched cell populations from five MDS/AMLindividuals. As a control, we performed similar analyses in HL-60 andKG-1 human leukemic cell lines, both of which harbor a deletion of5q11-q31 (refs. 10,11,15). As the probe for del(5q), we used a bacterialartificial chromosome that contains the EGR1, ETF1 and HSPA9Bgenomic loci, ending within 164 kb of the CTNNA1 locus, and thathas previously been used to demonstrate the 5q deletion in subsets ofFACS-sorted primary bone marrow cells9. The results show that 90%of HL-60 cells and 99% of KG1 cells carried the 5q deletion (Fig. 1band Supplementary Table 2 online). In contrast, a probe specific for

Received 25 July; accepted 3 November; published online 10 December 2006; doi:10.1038/nm1512

1Laboratory of Development and Diseases, Institute of Health Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences and Shanghai JiaoTong University School of Medicine, Shanghai 200025, China. 2Department of Pediatric Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston,Massachusetts 02115, USA. 3Division of Hematology/Oncology, University of Rochester, Rochester, New York 14642, USA. 4University of Texas M.D. Anderson CancerCenter, Houston, Texas 77030, USA. 5Ohio State University Comprehensive Cancer Center, Columbus, Ohio 43210, and Cancer and Leukemia Group, University ofChicago, Chicago, Illinois 60606, USA. 6Department of Adult Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115, USA.7Institute of Stem Cell Biology and Regenerative Medicine, and Division of Hematology/Oncology, Internal Medicine, Stanford University, Palo Alto, California 94304,USA. 8These authors contributed equally to this work. Correspondence should be addressed to A.T.L. ([email protected]).

78 VOLUME 13 [ NUMBER 1 [ JANUARY 2007 NATURE MEDICINE

L E T T ERS

©20

07 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

REVIEW

Nature Clinical Practice Oncology (2005) 2, S4-S11doi:10.1038/ncponc0354Received 16 August 2005 | Accepted 30 August 2005

DNA methylation and gene silencing in cancer

Stephen B Baylin

Correspondence Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Departments of Oncology and Medicine, The Johns Hopkins Medical Institutions, 1650 East Orleans Street, Baltimore, MD 21231, USA

Email [email protected]

SUMMARY

Epigenetic changes such as DNA methylation act to regulate gene expression in normal mammalian development. However, promoter hypermethylation also plays a major role in cancer through transcriptional silencing of critical growth regulators such as tumor suppressor genes. Other chromatin modifications, such as histone deacetylation and chromatin-binding proteins, affect local chromatin structure and, in concert with DNA methylation, regulate gene transcription. The DNA methylation inhibitors azacitidine and decitabine can induce functional re-expression of aberrantly silenced genes in cancer, causing growth arrest and apoptosis in tumor cells. These agents, along with inhibitors of histone deacetylation, have shown clinical activity in the treatment of certain hematologic malignancies where gene hypermethylation occurs. This review examines alteration in DNA methylation in cancer, effects on gene expression, and implications for the use of hypomethylating agents in the treatment of cancer.

Keywords:

chromatin, DNA methylation, gene silencing, histone modifications

Top of page

INTRODUCTION

Epigenetic events play a significant role in the development and progression of cancer. Mutations occurring in oncogenes frequently result in a gain of function, while mutations or deletions associated with tumor suppressor genes cause a loss or inactivation of negative regulators. Loss of function, however, can also occur through epigenetic changes such as DNA methylation. 'Epigenetics' refers to heritable changes in gene expression that do not result from alterations in the gene nucleotide sequence. When DNA is methylated in the promoter region of genes, where transcription is initiated, genes are inactivated and silenced.1, 2This process is often dysregulated in tumor cells. In cancer, EPIGENETIC SILENCING through methylation occurs at least as frequently as mutations or deletions and leads to aberrant silencing of normal tumor-suppressor function.2, 3There are a number of tumor types, particularly hematopoietic malignancies such as myelodysplastic syndromes (MDSs), in which hypermethylation occurs and results in leukemic transformation.4, 5, 6, 7, 8

Current research is aimed at providing a better understanding of the molecular mechanisms underlying methylation-related gene silencing in normal and tumor cells and its effects on chromatin structure and function. Since epigenetic changes are reversible, inhibitors of DNA methylation such as azacitidine (5-azacytidine; Vidaza®, Pharmion Corp., Boulder, CO, USA) and decitabine (Dacogen™, SuperGen Inc., Dublin, CA, USA,

and MGI Pharma Inc., Minneapolis, MN, USA) can derepress silenced tumor suppressor genes and restore their normal function. The therapeutic application of these methylation inhibitors is providing new and effective options for patients with leukemias and related diseases. This article summarizes the molecular biology underlying DNA methylation, its effects on gene transcription and chromatin structure, alterations in DNA methylation in cancer, and the rationale for treatment with inhibitors of DNA methylation to reactivate inappropriately silenced genes in cancer.Top of page

DNA METHYLATION

DNA methylation occurs by covalent addition of a methyl group at the 5' carbon of the cytosine ring, resulting in 5-methylcytosine.1These methyl groups project into the major groove of DNA and effectively inhibit transcription. In mammalian DNA, 5-methylcytosine is found in approximately 4% of genomic DNA, primarily at cytosine–guanosine dinucleotides (CpGs). Such CpG sites occur at lower than expected frequencies throughout the human genome but are found more frequently at small stretches of DNA called CpG islands. These islands are typically found in or near promoter regions of genes, where transcription is initiated.3In contrast to the bulk of genomic DNA, in which most CpG sites are heavily methylated, CpG islands in germ-line tissue and promoters of normal somatic cells remain unmethylated, allowing gene expression to occur.1

DNA methylation helps to maintain transcriptional silence in nonexpressed or noncoding regions of the genome. For example, pericentromeric HETEROCHROMATIN, which is condensed and transcriptionally inactive, is heavily methylated. Hypermethylation thus ensures this DNA is late-replicating and transcriptionally quiescent, and suppresses the expression of any potentially harmful viral sequences or transposons that may have integrated into such sites containing highly repetitive sequences.1, 2, 9By contrast, these sites are generally unmethylated in promoter regions of EUCHROMATIN, regardless of the transcriptional state of the gene. Exceptions to this rule, however, can be found in mammalian cells where these regions are methylated to maintain transcriptional inactivation. Thus, CpG islands in promoters of genes located on the inactivated X chromosome of females are methylated, as are certain imprinted genes in which only the maternal or paternal allele is expressed.1, 3

Epigenetic effects such as hypermethylation can also induce inevitable alterations in gene expression. Methylation of the DNA repair genes MLH1 and MGMTcan lead to their inactivation, resulting in microsatellite instability and increased frequency of mutations, respectively.10, 11Methylation can also promote spontaneous deamination, enhance DNA binding of carcinogens, and increase ultraviolet absorption by DNA, all of which serve to increase the rate of mutations and DNA adduct formation and subsequent gene inactivation.Top of page

REGULATION OF DNA METHYLATION

DNA methylation is controlled at several different levels in normal and tumor cells. The addition of methyl groups is carried out by a family of enzymes, DNA methyltransferases (DNMTs). Chromatin structure in the vicinity of gene promoters also affects DNA methylation and transcriptional activity. These are in turn regulated by various factors such as nucleosome spacing and histone acetylases, which affect access to transcriptional factors.

DNA METHYLTRANSFERASESDNMTs are enzymes that catalyze the addition of methyl groups to cytosine residues in DNA. DNMTs found in mammalian cells include DNMT1, DNMT3a, and DNMT3b.9, 12, 13In mouse development, DNMT1 appears to be responsible for maintenance of established patterns of DNA methylation, while DNMTs 3a and 3b seem to mediate establishment of new, or de novo, DNA methylation patterns.9, 12, 13In this regard, and potentially important for translational purposes, cancer cells may be different in that DNMT1 alone is

REGULATION OF THE HUMAN XANTHINEOXIDOREDUCTASE GENE

Eeva Martelin

Hospital for Children and AdolescentsUniversity of Helsinki

Finland

and

Program for Developmental and Reproductive BiologyBiomedicum HelsinkiUniversity of Helsinki

Finland

ACADEMIC DISSERTATION

Helsinki University Biomedical Dissertations No. 42

To be publicly discussed with the permission of the Medical Faculty of the University of Helsinki, in the Niilo Hallman Auditorium

of the Hospital for Children and Adolescents,on January 16th, 2004, at 12 noon.

Helsinki 2004

Review of the Literature

23

Regulation of gene expression

Levels of regulation

The purpose of the regulation of gene expression is to assure temporally and spatially

accurate expression of proteins which enables cells, on one hand, to accommodate to the

changing requirements of the environment and, on the other hand, to undergo normal cell

growth and differentiation. Differences between organisms are mainly based on the evolution

of the regulatory networks that control gene expression, not on genes themselves, which are

often conserved between species (Hood and Galas 2003). To accomplish these goals, the

regulation of gene expression takes place at several levels (Orphanides and Reinberg 2002)

(Figure 5).

Chromatin is a complex of DNA and proteins that in dividing cells is packaged into

chromosomes. In non-dividing cells, chromatin is distributed diffusely throughout the nucleus

and appears as condensed heterochromatin or more open euchromatin. Chromatin structure is

related to gene expression and it is critically regulated by histones, the principal proteins of

chromatin, which can either promote or repress gene activation (Weintraub and Groudine

1976). Modifications of histones and their higher-order structures, nucleosomes, by chromatin

remodelling complexes determine whether a specific area of chromatin is active or inactive at

certain time point (Dillon and Festenstein 2002; Robertson 2002; Felsenfeld and Groudine

2003). Distinctive chromatin remodelling complexes either activate or suppress gene

transcription and may associate with coregulator proteins, which interact with proteins

binding directly to the regulatory elements of genes. Histone modifications include

acetylation, methylation, phosphorylation, and ubiquitination of amino acids. Reflecting the

complexity of the regulation of chromatin structure, several histone acetylases and

deacetylases have been identified (Neely and Workman 2002; de Ruijter et al. 2003).

Transcriptional regulation of gene expression depends on cis-acting DNA elements

interacting with trans-acting regulatory proteins, i.e. transcription factors. Enhancers,

silencers, and regulatory elements located in introns may further modulate transcription.

Ultimately, the transcriptional activation of a gene is determined by cross-talk between

chromatin remodeling enzymes and transcription factors.

“Epigenetic therapy “in MDS

Pierre FENAUX

AP-HP,Hôpital Avicenne,

Université Paris 13

Groupe francophone des myélodysplasies

(GFM)

Tel Aviv July 2007

Epigenetic modifications

Potentially reversible DNA and chromatin modifications transmitted from a cell to its progeny, able to induce altered gene

expression without changing DNA sequence and without any “new” genetic

information

Epigenetic modifications

Potentially reversible DNA and chromatin modifications transmitted from a cell to its progeny, able to induce altered gene

expression without changing DNA sequence and without any “new” genetic

information

HistonesHistones

• Histone proteins organize DNA into nucleosomes

• Nucleosomes consist of DNA wrapped around a core histone

• A histone core comprises two copies each of four different core proteins (H2A, H2B, H3, H4)

• Histones can be acetylated and methylated at N-terminal lysines

• Histone proteins organize DNA into nucleosomes

• Nucleosomes consist of DNA wrapped around a core histone

• A histone core comprises two copies each of four different core proteins (H2A, H2B, H3, H4)

• Histones can be acetylated and methylated at N-terminal lysines

H3

H4

H2A

H2B

Published online 17 August 2007 Nucleic Acids Research, 2007, Vol. 35, No. 16 e106doi:10.1093/nar/gkm560

Choreography for nucleosomes: the conformationalfreedom of the nucleosomal filament and itslimitationsMogens Engelhardt*

Department of Cellular and Molecular Medicine, Faculty of Health Sciences, University of Copenhagen,Blegdamsvej 3, DK-2200 Copenhagen N, Denmark

Received May 3, 2007; Revised July 7, 2007; Accepted July 9, 2007

ABSTRACT

Eukaryotic DNA is organized into nucleosomes bycoiling around core particles of histones, forminga nucleosomal filament. The significance for theconformation of the filament of the DNA entry/exitangle (a) at the nucleosome, the angle of rotation (b)of nucleosomes around their interconnecting DNA(linker DNA) and the length of the linker DNA, hasbeen studied by means of wire models with straightlinkers. It is shown that variations in a and b endowthe filament with an outstanding conformationalfreedom when a is increased beyond 60–90u, owingto the ability of the filament to change betweenforward right-handed and backward left-handedcoiling. A wealth of different helical and loopedconformations are formed in response to repeated bsequences, and helical conformations are shownto be able to contract to a high density and toassociate pairwise into different types of doublefibers. Filaments with random b sequences arecharacterized by relatively stable loop clustersconnected by segments of higher flexibility.Displacement of core particles along the DNA insuch fibers, combined with limited twisting of thelinkers, can generate the b sequence necessary forcompaction into a regular helix, thus providing amodel for heterochromatinization.

INTRODUCTION

DNA in the cell nucleus is organized in nucleosomes byrepetitive coiling �1.8 times around core particles consist-ing of two copies of each of four different core histones(H2A, H2B, H3 and H4). In higher eukaryotes a fifthhistone, H1, is bound to the majority of the nucleosomes,bridging the DNA at the entry and exit of its coilingaround the core particle. On average, one nucleosome

is present for every 160–240 bp of DNA [the nucleosomalrepeat length (NRL)]. The structure of the nucleosome isknown in detail (1), and several models have beenproposed for folding of the nucleosomal filaments intofibers with a diameter of �30 nm (2), which have beenobserved by electron microscopy. These models are basedon experimental evidence obtained by a number ofdifferent techniques, using nuclear preparations, isolatedchromatin and reconstituted oligonucleosomes. One of themain differences between the models is whether the DNA,which connects the nucleosomes (linker DNA), is straightor bent. Studies of chromatin in isolated nuclei haveprovided evidence for an organization of the filament intoglobular assemblies of nucleosomes (supranucleosomesor superbeads) (3) as well as for interdigitation of 30 nmfibers (4,5). Supranucleosomes have been proposed to becaused by dislocations in a cross-linker helical structure(6), but no models have been developed to explain thegeometry of double fibers.The conformational freedom of the nucleosomal

filament, i.e. the number of different conformations itcan attain, is limited by collisions between the nucleo-somes, and is in first instance determined by the basicgeometry of the filament, which is therefore of importancefor the structure and function of the chromatin in the cell.One function concerns its role in transcription and trans-criptional regulation, in which its dynamic nature (7) andstructural plasticity (8) must play an important role.Another function concerns the putative architectural roleof the chromatin in the nucleus, of which not so muchis known. A long-standing question is to which extentthe chromatin is a self-organizing polymer and to whichextent it is being organized by other structures in thenucleus (9,10). It is not known whether the basic geometryof the filament by itself makes possible the formationof stable fiber associations.Provided that the linker DNA is straight—and there

is evidence to suggest that this is so in the cell (11)—conformational variations of the filament are mainlydetermined by two angles: the change in direction of the

*To whom correspondence should be addressed. Tel: +45 35327768; Fax: +45 35327732; Email: [email protected]

� 2007 The Author(s)

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/

by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Published online 17 August 2007 Nucleic Acids Research, 2007, Vol. 35, No. 16 e106doi:10.1093/nar/gkm560

Choreography for nucleosomes: the conformationalfreedom of the nucleosomal filament and itslimitationsMogens Engelhardt*

Department of Cellular and Molecular Medicine, Faculty of Health Sciences, University of Copenhagen,Blegdamsvej 3, DK-2200 Copenhagen N, Denmark

Received May 3, 2007; Revised July 7, 2007; Accepted July 9, 2007

ABSTRACT

Eukaryotic DNA is organized into nucleosomes bycoiling around core particles of histones, forminga nucleosomal filament. The significance for theconformation of the filament of the DNA entry/exitangle (a) at the nucleosome, the angle of rotation (b)of nucleosomes around their interconnecting DNA(linker DNA) and the length of the linker DNA, hasbeen studied by means of wire models with straightlinkers. It is shown that variations in a and b endowthe filament with an outstanding conformationalfreedom when a is increased beyond 60–90u, owingto the ability of the filament to change betweenforward right-handed and backward left-handedcoiling. A wealth of different helical and loopedconformations are formed in response to repeated bsequences, and helical conformations are shownto be able to contract to a high density and toassociate pairwise into different types of doublefibers. Filaments with random b sequences arecharacterized by relatively stable loop clustersconnected by segments of higher flexibility.Displacement of core particles along the DNA insuch fibers, combined with limited twisting of thelinkers, can generate the b sequence necessary forcompaction into a regular helix, thus providing amodel for heterochromatinization.

INTRODUCTION

DNA in the cell nucleus is organized in nucleosomes byrepetitive coiling �1.8 times around core particles consist-ing of two copies of each of four different core histones(H2A, H2B, H3 and H4). In higher eukaryotes a fifthhistone, H1, is bound to the majority of the nucleosomes,bridging the DNA at the entry and exit of its coilingaround the core particle. On average, one nucleosome

is present for every 160–240 bp of DNA [the nucleosomalrepeat length (NRL)]. The structure of the nucleosome isknown in detail (1), and several models have beenproposed for folding of the nucleosomal filaments intofibers with a diameter of �30 nm (2), which have beenobserved by electron microscopy. These models are basedon experimental evidence obtained by a number ofdifferent techniques, using nuclear preparations, isolatedchromatin and reconstituted oligonucleosomes. One of themain differences between the models is whether the DNA,which connects the nucleosomes (linker DNA), is straightor bent. Studies of chromatin in isolated nuclei haveprovided evidence for an organization of the filament intoglobular assemblies of nucleosomes (supranucleosomesor superbeads) (3) as well as for interdigitation of 30 nmfibers (4,5). Supranucleosomes have been proposed to becaused by dislocations in a cross-linker helical structure(6), but no models have been developed to explain thegeometry of double fibers.The conformational freedom of the nucleosomal

filament, i.e. the number of different conformations itcan attain, is limited by collisions between the nucleo-somes, and is in first instance determined by the basicgeometry of the filament, which is therefore of importancefor the structure and function of the chromatin in the cell.One function concerns its role in transcription and trans-criptional regulation, in which its dynamic nature (7) andstructural plasticity (8) must play an important role.Another function concerns the putative architectural roleof the chromatin in the nucleus, of which not so muchis known. A long-standing question is to which extentthe chromatin is a self-organizing polymer and to whichextent it is being organized by other structures in thenucleus (9,10). It is not known whether the basic geometryof the filament by itself makes possible the formationof stable fiber associations.Provided that the linker DNA is straight—and there

is evidence to suggest that this is so in the cell (11)—conformational variations of the filament are mainlydetermined by two angles: the change in direction of the

*To whom correspondence should be addressed. Tel: +45 35327768; Fax: +45 35327732; Email: [email protected]

� 2007 The Author(s)

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/

by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.