5-Azacytidine: A New Active Agent for the
`Treatment of Acute Leukemia
`
`By Myron Karon, Lance Sieger, Suzanne Leimbrock, Jerry Z. Finklestein,
`Mark E. Nesbit, and Jerry J. Swaney
`
`Thirty-seven children with acute leukemia
`were treated with 5-azacytidine in 5-day
`courses given every 14 days. Six out of 14
`children with acute myelogenous leukemia
`who were adequately treated achieved an
`M, marrow. Five of
`these subsequently
`developed complete remissions
`lasting
`8 mo. 6 mo. 3 mo. 2 mo. and 2 mo. Of
`22 children with acute lymphocytic leuke-
`mia, one achieved an M, marrow and one
`an M 2 marrow. The former attained a com-
`
`plete remission which lasted 3 mo. The
`maximum tolerated dose is between 150
`to 200 mg/sq m on a daily x 5 schedule
`given every 14 days. The impressive ac-
`tivity of 5-aza-C in patients with acute
`myelogenous leukemia resistant to cyto-
`sine arabinoside indicates that
`this drug
`will become an important addition to the
`therapeutic armamentaria against this type
`of leukemia.
`
`5 -AZACYTIDINE (5-AZA-C) is an analOg. of cytidine first introduced for
`
`the treatment of acute leukemia of childhood in Czechoslovakia.' The
`critical step in the mechanism of drug action is not known, but the activity is
`presumably related to incorporation into DNA and RNA polynucleotides in
`place of cytidine.' The compound is known to interfere with DNA, RNA, and
`protein synthesis.v' The drug shows preferential activity during the S phase of
`the cell cycle in vitro and has shown schedule dependency in animal systems.r"
`This study was undertaken to establish a tolerated dose of 5-aza-C on a daily
`x 5 schedule every 2 wk. During the course of this phase I investigation 5-aza-C
`was found to have impressive activity against acute myelogenous leukemia, and
`some activity against acute lymphocytic leukemia.
`
`From the Division of Hematology, Department of Pediatrics, Childrens Hospital of Los Angeles.
`and the USC School of Medicine; Department of Pediatrics. Harbor General Hospital. and the UCLA
`School of Medicine. Torrance. Calif.; the Department of Pediatrics. University of Minnesota School
`of Medicine. Minneapolis. Minn.; and Childrens Memorial Hospital and the Department of Pedi-
`atrics. Northwestern University. Chicago. 11/.
`Submitted January 22,1973; revised March 12.1973; accepted March 22.1973.
`These studies were performed under the auspices of Childrens Cancer Study Group A in the
`authors' institutions.
`Supported by Grants CA-02649. CA-07306 and CA-07431 from the National Cancer Institute and
`Training Grant H D-00048from the National Institute for Child Health and Human Development,
`NIH, DHEW.
`Myron Karon, M.D.: Professor of Pediatrics. Head, Division of Hematology. Childrens Hospital
`of Los Angeles. Los Angeles, Calif.; Scholar, Leukemia Society of America. Lance Sieger, M.D.:
`Hematology Fellow. Division of Hematology, Childrens Hospital of Los Angeles. Los Angeles. Calif
`90027. Suzanne Leirnbrock, Pharm. D.: Clinical Pharmacist. Division of Hematology. Childrens
`Hospital of Los Angeles. Los Angeles, Calif 90027. Jerry Z. Finklestein, M.D.: Assistant Profes-
`sor of Pediatrics, Head, Pediatric Hematology, Harbor General Hospital. Torrance. Calif Mark E.
`Nesbit, M.D.: Professor of Pediatrics, University of Minnesota School of Medicine, Minneapolis.
`Minn. Jerry J. Swaney, M.D.: Associate in Pediatrics, Childrens Memorial Hospital and the De-
`partment of Pediatrics. Northwestern University. Chicago. 11/.
`© /973 by Grune & Stratton, Inc.
`
`Blood, Vol. 42, No.3 (September), 1973
`
`359
`
`

`

`360
`
`KARON ET AL.
`
`MATERIALS AND METHODS
`
`least one of the parents of each
`Informed consent for the use of 5-aza-C was obtained from at
`child prior to treatment. This request was based on the potential efficacy of 5-aza-C and the fact
`that all other agents of known activity had been used. There were no parents who refused treat-
`ment for their child.
`Thirty-seven patients ranging in age from 2 to 17 yr with acute leukemia refractory to standard
`chemotherapeutic agents were treated with 5-aza-C at a starting dose of 2 mg/sq m at
`the authors'
`institutions. The drug was administered as either an intravenous "push" or a 15-min "fast drip,"
`daily for 5 days and repeated every 14 days, depending upon response and/or toxicity. Generally
`in the absence of significant
`toxicity,
`the dosage was increased by 50~o increments (Fig.
`I).
`In
`some instances the drug was administered in divided doses either every 8 or 12 hr to forestall
`nausea and vomiting.
`included physical examination, a complete blood
`Observations at
`the beginning of treatment
`count, a bone marrow examination, platelet count, uric acid, SGOT, alkaline phosphatase, BUN,
`and urinalysis. During therapy, complete blood counts were repeated at
`least weekly and a bone
`marrow aspiration performed at
`least every 28 days. Criteria for evaluation were those in use
`by the Childrens Cancer Study Group A. 7 An M I marrow contains 5% or less of blasts or other
`abnormal cells.
`5-aza-C was supplied in 50-mg vials by the Clinical Drug Evaluation Branch of the National
`Cancer Institute. The drug was dissolved in 5 ml of distilled water and administered within 15
`min of reconstitution since the drug begins to decompose in aqueous solution within I hr.
`
`RESULTS
`Determination of Maximally Tolerated Dose
`The pattern of dose escalation is illustrated in Fig. I. Dosage increments were
`logarithmic below 100 mg/sq m. The rate of initial escalation was based on the
`lack of response and toxicity at the lower doses. The broken curve represents a
`modified Fibonachi numerical progression recently proposed by Dr. Oleg Sel-
`
`1 0 0 0 . - - - - - - - - - - - - - -....
`
`100
`
`Fig. 1. Dose escalation of 5-azacytidine.
`Dose escalation was logarithmic until a dose of
`100 mg/sq m when there was a decrease in
`this rate based on the occurrence of response
`and/or toxicity. Each point
`represents a dose
`increment actually used one or more times. The
`broken curve represents a modified Fibonachi
`search procedure designed to permit
`rapid
`escalation of the dosage initially while diminish-
`ing the rate of
`increase near
`the maximum
`tolerated dose in order to avoid excessive tox-
`icity. The modified Fibonachi series is as fol-
`lows: 2. 0.7. 0.5. 0.3.... These numbers are
`used as coefficients to predict dose increments.
`If the starting dose based on 0.1 LD,o in ani-
`mals was 2 mg/sq m,
`then dosage increments
`would be 2. 4. 6.8. 10.2. 13.5 mg/sq m ....
`Blind adherence to such a scheme would un-
`necessarily prolong a phase I trial if the starting
`dose was too low.
`
`2
`
`3
`
`8
`7
`6
`5
`4
`INCREMENTS
`
`9
`
`10 II
`
`12
`
`(\J
`
`E<,
`0'E
`/)oo
`
`w(
`
`

`

`5-AZACYTIDINE
`
`361
`
`awry for predicting dose increments." The shape of this theoretical curve is
`quite similar to that which was obtained by increasing the dose until either re-
`sponse or toxicity occurred.
`The maximum tolerated dose was obtained by determining the number of
`patients given a particular dose of 5-aza-C per course. As the dose of drug was
`increased, the number of patients at a given dose increased to a maximum and
`then declined abruptly between 150 and 200 mg/sq m. The plateau value of
`such a curve was used to estimate the lower limit of drug toleration.
`
`Therapeutic Efficacy
`The over-all results of treatment are tabulated in Table I. Six out of 15 pa-
`tients with acute myelogenous leukemia achieved an M I bone marrow status.
`Five of these children obtained a complete remission which lasted 8 mo,
`6 mo, 3 mo, 2 mo, 2 mo, and 2 mo. One patient with an M, marrow died
`within 3 wk in partial remission. The over-all remission rate for AM L, there-
`fore, is 6/15 (40%). If one excludes one patient with acute myelocytic leukemia
`who received only one-fiftieth of the maximum tolerated dose early in the study,
`then the remission rate becomes 6/14 (42%).
`Of the 22 patients with acute lymphocytic leukemia (ALL), one patient
`achieved an M I marrow and another an M 2 marrow. The child with the M I
`marrow obtained a complete remission which lasted for 3 mo. The child with
`the M 2 marrow (9% lymphoblasts) died of infection before achieving a com-
`plete remission. The difference in remission rates (M I) between AML and ALL
`is significant, p < 0.008 (Fishers exact test).
`A summary of the clinical data on the eight patients who achieved an objec-
`tive response is given in Table 2. Of these patients, six had white blood cell
`counts below 50,000 cells/cu mm at the time of diagnosis and two had a white
`blood cell count above 100,000 cells/cu mm. With the exception of I patient
`who received a lower dose of 5-aza-C 5 days each week, the best response oc-
`curred after two courses of drug in five patients and after four courses in two
`patients. At a dose of 150 to 200 mg/sq m per day x 5 the nadir in the white
`blood cell count occurred between the 10th and the 14th day.
`There was no uniform approach to maintenance therapy for those patients
`who achieved complete remission. Patient CB, who stayed in remission for 8
`mo, received 90 mg/sq m on 2 consecutive days each week. Patient DC (3 mo)
`
`Table 1. Treatment Results
`
`Diagnosis·
`
`Patients
`
`Adequate Treatmentt
`
`AML
`ALL
`Total
`
`15
`22
`37
`
`14
`22
`36
`
`Marrow
`
`Over-all Status!
`
`M-'
`
`6
`1
`7
`
`M-2
`
`M-3
`
`CR
`
`PR
`
`0
`1
`1
`
`9
`20
`29
`
`5
`1
`6
`
`1
`1
`2
`
`NR
`
`9
`20
`29
`
`• AM L acute myelogenous leukemia as a generic name for all subcategories of non-ALL ALL acute
`lymphocytic leukemia. including acute undifferentiated leukemia. AU L
`t Adequate treatment is at least one course at a dose of 60 mg/sq m or greater daily x 5. The one
`inadeouatelv treated patient had less than 10 mg/sq m daily x 5.
`t CR. complete remission. PRo partial remission. NR. no response or progressive disease.
`
`

`

`...
`
`>r
`m.
`
`::XJ
`
`~
`
`z
`0
`
`thi-
`
`2mo
`
`CR
`
`M-1
`
`4mo
`
`CR
`
`3wks
`
`3wks
`
`PR
`
`PR
`
`M-l
`
`M-2
`
`M-1
`
`6mo
`
`CR
`
`M-l
`
`3mo
`
`CR
`
`3mo
`
`CR
`
`3mo
`
`CR
`
`tion
`Dura-
`
`Status
`Over-all
`
`Response
`
`n
`we
`
`f',J
`
`M-l
`
`M-l
`
`M-l
`
`8
`
`2
`
`4
`
`4
`
`2
`
`2
`
`2
`
`2
`
`4
`
`1500
`
`150
`
`5
`
`9
`
`8
`
`6
`
`4
`
`4
`
`1500
`
`150
`
`1944
`
`222
`
`1395
`
`93
`
`93
`
`3411
`
`268
`
`2431
`
`150
`
`80
`
`1180
`
`160
`
`15
`
`1597
`
`90
`
`2
`
`Marrow
`
`Courses
`
`Wks
`
`Total
`
`Final
`
`InductionPhase
`
`Durationof
`
`inInduction
`
`Dosage(mg/sqm)t
`
`tVCR.vincristine;ara-C.cytosinearabinoside;MP.6-mercaptopurine;MTX,methotrexate;CTX.cyclophosphamide;ASP,L-asparaginase;P.prednisone;TG.
`•AML.acutemyelocyticleukemia;AMMLacutemono-myelocyticleukemia;ALLacutelymphocyticleukemia.
`
`tDailyx5every14days.exceptforCB(seetext).
`oguanine;DBD.dibromodulcitol.DNM.daunomycin.
`
`140
`
`0.6
`
`0.4
`
`0.9
`
`16.6
`
`174
`
`5.9
`
`3.4
`
`3.9
`
`113
`
`Initial
`
`Nadir
`
`Diagnosismentwith5-aza-C
`
`StartofTreat-
`
`WhiteBloodCellCountx103
`
`Table2.ClinicalCharacteristicsofRespondingPatients
`
`VCR.ara-C
`
`AMML
`
`5M
`
`VCR.ara-C
`TG.P
`ara-C.CTX
`
`DBD
`ara-C.MP
`Asp
`
`VCR.MTX
`VCR.MP
`
`ALL
`
`ara-C.
`
`AMML
`
`5M
`
`3M
`
`crx.MP
`VCR.ara-C
`
`Treatment
`Previoust
`
`AML
`
`F
`
`14
`
`DiagnOSIs·
`
`Age/Sex
`
`Patients
`
`K.K.
`
`C.K.
`
`R.R.
`
`J.B.
`
`D.C.
`
`C.B.
`
`150
`
`0.7
`
`1.0
`
`15.6
`
`150
`
`166
`
`4.0
`
`0.7
`
`0.6
`
`10.9
`
`0.7
`
`4.3
`
`20.2
`
`21.5
`
`21.6
`
`P.MP
`ara-C.TG
`
`VCR.CTX
`
`MP
`CTX.TG
`ara-C.
`VCR.P.
`
`MTX.Asp
`
`P.VCR.
`
`CTX.DNM
`
`AML
`
`10M
`
`P.J.
`
`AML
`
`2F
`
`ALL
`
`8M
`
`125
`
`0.6
`
`2.0
`
`25.7
`
`VCR.
`
`AMML
`
`7F
`
`N.S.
`
`

`

`5-AZACYTIDINE
`
`363
`
`received 150 rng/sq m q.d. for 2 consecutive days every week. Patient NS is
`receiving 230 mg/sq m q.d. x 5 every 3 wk. Patient JB has received 150
`mg/sq m q.d. x 5 every 5-6 wk. This more prolonged interval was necessary
`because J8's white blood cell count continued to decrease for approximately
`4 wk following therapy. Patients KK and PJ were maintained on 5-aza-C 100
`mg/sq m q.d. x 5 once per month.
`
`Toxicity
`
`The most disabling toxicity involves the gastrointestinal tract. At doses of
`150 mg/sq m or above, 5-aza-C causes profound nausea and vomiting and diar-
`rhea in all patients. The severity of these toxic manifestations, especially the
`former, can be reduced by giving the drug in divided doses or in some in-
`stances by the use of a 15 min intravenous fast drip. Antiemetics such as
`chlorpromazine can be of use when given in relatively large doses at
`least
`24-48 hr prior to beginning the course of 5-aza-C. A pruritic, follicular skin
`rash occurred in 50% of the patients, but was usually transient and did not re-
`quire drug dosage modification.
`The drug is myelosuppressive. Although recovery usually occurs by day 14
`at a dose of 150 tug] sq m daily x 5, myelosuppression may be more pro-
`longed, especially at 200 mg /sq m or above. Myelosuppression is not neces-
`sarily untoward and accompanied response in every instance except one (Ta-
`ble 2).
`
`DISCUSSION
`
`These data indicate that 5-aza-C is an active drug for the treatment of acute
`leukemia in children. The compound has particularly impressive activity
`against acute myelogenous leukemia. Activity of 5-aza-C against adult
`leu-
`kemia has recently been demonstrated by McCredie and co-workers" and
`against acute lymphocytic leukemia in newly diagnosed patients who were also
`receiving prednisone.'
`The fact that the activity of 5-aza-C could be demonstrated in children with
`advanced refractory leukemia has important implications for the evaluation of
`other new drugs. Agents of uncertain potency do not need to be tested early
`in the course of a child's disease to demonstrate activity. Indeed, the activity
`of vincristine, ara-C, and L-asparaginase, all effective antileukemic agents, was
`demonstrated under similar circumstances. Since 5-aza-C is not cross-resistant
`to ara-C, the use of these two agents in combination for the treatment of AM L
`should be promising.
`Although the original plan was to adhere to a modified Fibonachi search
`for dose escalation, this approach was abandoned. The Fibonachi search in-
`volves a rapid initial dose escalation followed by decreasing dosage increments
`which ·are predetermined and aimed at reducing the chance of grossly over-
`shooting the maximum tolerated dose. The original plan for escalation is shown
`on Fig. 1 as an interrupted line, and is based on the series 2, 0.7, 0.5, 0.3 ....
`The initial starting dose was 2 mg/M 2
`l~
`the LD lO determined in preclinical
`,
`pharmacology. Because this dose proved to be 70-IOO-fold less than the maxi-
`mum tolerated dose, strict adherence to Fibonachi's escalation program would
`
`

`

`364
`
`KARON ET AL
`
`Frequency distribu-
`Fig. 2.
`tion of patients receiving a
`particular dose of 6-aza-C per
`course. The
`plateau of
`this
`curve determines the maxi-
`mum tolerated dose of 6-aza-C
`since courses which do not
`produce toxicity or therapeutic
`effect or which produce exces-
`sive toxicity are usually not
`given to a large number of
`patients.
`while
`tolerated
`courses that affect the disease
`favorably are used in an in-
`creasing number of patient. All
`patients receiving the drug are
`included as long as they were
`in the induction phase. A pa-
`tient can be represented only
`once for each dose level.
`
`40
`
`80
`
`120
`
`160
`
`200
`
`240
`
`280
`
`320
`
`DOSE PER COURSE (mg/M2)
`
`14
`
`13
`12
`
`"1O
`
`9
`8
`7
`
`6
`5
`4
`
`3 2 °
`
`.
`
`~
`z
`
`w~
`
`a:
`~
`a:
`~
`~
`:::>z
`
`have been completely unworkable. The shape of the actual dose escalation
`curve, however, is similar to the Fibonachi modification, indicating that there
`is a need to red uce the rate of dosage increment once there is either a thera-
`peutic effect or some toxic manifestations, but that this mathematical series per
`se has no particular "magic."
`The maximum tolerated dose was estimated by plotting the number of pa-
`tients receiving a given dose per course. Clearly such a curve would have a posi-
`tive slope at doses which were well tolerated and, therefore, used in more pa-
`tients and a negative slope when the tolerated dose was exceeded. This proved
`to be the case (Fig. 2) and was a useful way of predicting drug dosage. The main
`danger of a more rapid escalation scheme is the development of cumulative
`toxicity. This can be obviated by escalating the dosage every second course as
`the maximum tolerated dose is approached.
`The achievement of complete remission in two patients whose initial white
`counts were in excess of 100,000 indicates that such high white blood cell
`counts may not necessarily predict for a poor prognosis, especially in advanced
`disease. This may be the result of a selection since patients with acute leukemia
`who live long enough to receive phase I agents have usually responded to other
`agents. The practice in some phase II studies designed to estimate the remis-
`sion rate in advanced disease to exclude certain patients because of the height
`of their initial white blood cell count may not necessarily be justified.
`
`REFERENCES
`I. Hrodek 0, Vesely J: 5-azacytidine in child-
`hood leukemia. Neoplasm 18:493, 1971
`2. Juvovcik M, Raska K Jr, Sovmova F,
`Sorm F: Anabolic transformation of a novel
`antimetabolite, 5-azacytidine and evidence for
`its incorporation into ribonucleic acid. Coli
`Czech Chern Commun 30:3370, 1965
`3. Li LH, Olin J, Boskivk HH, Reineke
`
`LM: Cytotoxicity and mode of action of 5-
`azacytidine on LI210 leukemia. Cancer Res
`30:2760, 1970
`4. Raska KJ Jr, Juvovcik M, Fucik V, Tykva
`R, Sovmova Z, Sorm F: Metabolic effects of 5-
`azacytidine and 5-aza-2 1-deoxycytidine in mice.
`Coli Czech Chern Commun 31:2803,1966
`5. Li LH, Olin EJ, Fvosev TJ, Bhuyan BK:
`
`

`

`5-AZACYTIDINE
`
`365
`
`Phase specificity of 5-azacytidine against mam-
`malian cells in tissue culture. Cancer Res 30:-
`2770, 1970
`6. Fucik V, Michaelis A, Reigev R: On the
`induction of segment extension and chromatid
`structural changes in vicia faba chromosomes
`after treatment with 5-azacytidine and 5-azade-
`oxycytidine. M utat Res 9:599, 1970
`7. Leikin SL, Brubaker C, Hartmann JR,
`Murphy ML, Wolff JA, Perrin E: Varying
`prednisone dosage in remission induction of
`
`previously untreated childhood leukemia. Can-
`cer 21:346, 1968
`8. Selawry as: Considerations for
`initial
`clinical trial of anti-neoplastic agents. Sympo-
`sium on Statistical Aspects of Protocol Design,
`December 9-10, 1970
`
`9. McCredie KB, Bodey GP, Burgess MA,
`Rodriguez V, Sullivan M P, Freireich EJ: The
`treatment of acute leukemia with 5-azacytidine.
`Blood 40:975, 1972
`
`