Interaction of deoxycytidine and deoxycytidine analogs in normal and leukemic human myeloid progenitor cells

Leukemia Research Vol. 10, No. 9. pp. 1139-1146, 1986. 0145-2126/86 $3.00 * .00 Printed m Great Britain. Pergamon Journals Ltd.

INTERACTION OF DEOXYCYTIDINE AND DEOXYCYTIDINE ANALOGS IN NORMAL AND

LEUKEMIC HUMAN MYELOID PROGENITOR CELLS*

STEVEN GRANT,'t KAPIL BHALLA and MOSES GLEYZER

Division of Hematology/Oncology, Department of Medicine, Columbia University College of Physicians and Surgeons, New York, NY 10032, U.S.A.

Abstract The inhibitory effects of three deoxycytidine analogs, 1-B-D-arabinofuranosylcytosine (Ara-C), 5-aza-2'-deoxycytidine, (DAZ) and Ara-5-azacytosine (AAC) were compared with respect to the clonogenic behavior of human promyelocytic leukemic cells (HL-60), a deoxycytidine kinase deficient subvariant (HL-60/Ara-C), and normal human myeloid progenitor cells (CFU-GM). When cells were continuously exposed to each agent for 7 days, Ara-C was the most inhibitory, DAZ slightly less effective and AAC the least inhibitory on a molar basis. HL-60/ Ara-C were also highly cross-resistant to both DAZ and AAC. In the absence of deoxycytidine, all three agents were either equally inhibitory or slightly more inhibitory to the growth of CFU- GM than to HL-60, whereas administration of deoxycytidine in ten to one-hundred fold excess protected CFU-GM to a greater extent than HL-60. In contrast, administration of high concentrations of drugs, e.g. 10-~-10 -4 M, in conjunction with excess deoxycytidine exhibited greater toxicity toward CFU-GM than toward HL-60/Ara-C. Coadministration of deoxycytidine in ten- fold excess reduced the total intracellular accumulation and DNA incorporation of each analog in HL-60 cells by approx. 50%, whereas a hundred fold excess was associated with greater than a 90% reduction in these values. These studies demonstrate that deoxycytidine may antagonize the effects of Ara-C, DAZ and AAC in both normal and leukemic human myeloid cells, and that at low drug concentrations the degree of protection may be greater for normal elements. However, regimens employing high drug concentrations in conjunction with deoxycytidine do not appear to exert a selective inhibitory effect toward a highly resistant leukemic subvariant. These data suggest that alternative deoxycytidine/deoxycytidine analog dose relationships and schedules must be sought which are capable of selectively eradicating resistant cells.

Key words: Deoxycytidine, 1-B-D-arabinofuranosylcytosine, 5-aza-2'-deoxycytidine, Ara-5- azacytosine. HL-60 cells. CFU-GM.

I N T R O D U C T I O N

DEOXYCYTIDINE (dCyd) is a naturally occurring nucleoside which is present in low concentrations in the plasma of humans [1]. It is converted intraceilularly to its triphosphate derivative, dCTP, which is felt to play a key role in the initiation and maintenance of DNA synthesis [2]. Because of the potential importance of dCyd in cell replication, deoxycytidine derivatives have been the focus of interest as anti-leukemic agents. For example,

* Supported by NIH grant CA-35601, ACS Research/Clini- cal Investigational Award, ACS Clinical Oncology Career Development Award and the William J. Matheson Foundation. Portions of this work were presented in pre- liminary form at the American Association of Cancer Research. Houston, TX. 1985.

f Scholar of the Leukemia Society of America. Abbreviations: dCyd. deoxycytidine: Ara-C. 1-B-D arab-

inofuranosylcytosine; DAZ, 5-aza-2'-deoxycytidine" AAC. Ara-5-azacytosine.

Correspondence to: Dr Steven Grant at address above.

1-B-D-arabinofuranosylcytosine (Ara-C) is a highly effective agent which is included in virtually all induction regimens in acute myelogenous leukemia [3, 4]. Recent evidence suggests that incorporation of Ara-C into leukemic cell DNA, resulting in premature chain termin- ation, may account for Ara-C mediated cytotoxicity [5]. 5-Aza-2'-deoxycytidine (DAZ) is a newer derivative currently undergoing evaluation in childhood leukemia. It is also known to be incorporated into tumor cell DNA and to interfere with DNA methylation [6]. Arabinosyl- 5-azacytosine (AAC) is a recently synthesized derivative which combines the ring structure of 5-azacytidine with the sugar moiety of Ara-C. Its precise mode of action has not been definitively established, but evidence suggests that it behaves in a manner similar to Ara-C [7]. The structures of dCyd and its related derivatives are illustrated in Fig. 1.

Since dCyd and its derivatives compete at various levels of cellular metabolism including transport, phosphorylation, binding to DNA polymerase, and incorporation into DNA, it might be expected that dCyd would antagonize the cytotoxic effects of its analogs.

1139

1140 S I I . \ } ',, ( } R - \ \ I . KAPII. B I I A I I .\ alTd .~'l()sl s (}I . t "~Zt:R

NH, NH, N H

o o

HO HO OH

NH

o NY HOCH~ I

OH

Deoxycytidine Cytosine 5-Aza-2'-

arabinoside deoxycytidine

Fi(~. 1 The structures ot deoxvcvtidine. I-B-D arabino- furanosvlcvtosinc. 5-aza-2'-dcox~cvtidinc and Ara-5-aza-

cvtosil4e.

Arabmosyl-5- azacytosme

In-vitro studies have in fact demonstrated that Ara-C and D A Z mediated cytotoxicity could be reversed by coadministration of dCyd at the appropriate dose level [8, 9]. Such studies have led to speculation that dCyd might be capable of improving the therapeutic index of an analog such as Ara-C by preferentially protecting host tissues from its lethal effects. For example, in- v ivo investigations in mice have suggested that dCyd infusions might ameliorate Ara-C host toxicity while preserving ant i- tumor effect, thereby enhancing the therapeutic index of the combinat ion [10, 11]. Infor- mation concerning the effect of dCyd on Ara-C effect on human leukemic cells is sparse, although Harris and colleagues demonstra ted that dCvd might be selectively incapable of antagonizing the in-vitro metabolism of this agent in leukemic myeloblasts relatively deficient in dCyd kinase [12]. The aim of the present investigations was to examine the effect of dCyd on the metabolism and cytotoxicity of several dCvd analogs in the continuously cultured human promyelocytic leukemic subline HL-60 [13] as well as a dCyd kinase deficient, highly Ara-C resistant subline (HL-60/Ara- C) recently isolated by our group [14]. An additional objective was to compare the effect of dCyd on dCyd analog mediated cytotoxicity toward leukemic and normal human marrow myeloid progeni tor cells in order to determine whether selectivity might thereby be enhanced.

M A T E R I A L S A N D M E T H O D S

l)eoxycytidine hydrochloride, l-B-D-arabinofuranosylcytosine, and 5-aza-2'-deoxycytidinc were purchased from Sigma C'hemicals, St Louis, MO. Arabinosyl 5-azacytosine were kindly provided by Dr Robert Glazer. Laboratory of Medicinal Chemistry and Pharmacology. National Cancer Institute. Bethesda, MD. [~H]l-B-D-arabinofuranosylcytosine (26 Ci/ mmol) was purchased from Amersham Radiochemicals, It,. [3Hl5-aza-2'-deoxycytidine (18 Ci/mmol) and [3H]arabinosyl- 5-azacytosine (16Ci/mmol) were purchased from Moravck Biochemicals Brea, CA. Drugs were maintained as dry

powders at -20°C. and reconstituted in RPMI medium and filter sterilized immediately prior to use.

Cells

HL-60 cells were derived from the line originally described by Collins and Gallo {13]. The}' are maintained in RPMI medium (GIBCO, Grand Island, NY) containing 10% fetal calf serum, lC,; sodium pyruvate and le;+ nonessential amino acids. (;ells are maintained in a 37°C, 5% CO,, fully humidified incubator, passed twice weekly, and routinely examined for mycoplasma contamination. Logarithmically growing cells (cell density 5 × HI s cells/ml or less) are utilized for all experiments. The characteristics of HL-60/Ara-C have been described in detail previously [ 14}. They are maintained in the same manner as HL-60, except that media contains 10 "M Ara-('. HL-6(I/Ara-C cells are washed thoroughly and grown in drug free medium for three days prior to each experiment.

Intracellular accumulation studies

The total mtracellular accumulation of labeled Ara-C, DAZ, and AAC in I-.IL-6(I and HL-60/Ara-C cells was determined by a rapid centrifugation technique which has been described in detail previously [15]. I,ogarithmically growing cells are placed in 15-ml centrifuge tubes containing 1%. Hepes buffer at a cell density of 5 × 105 cells/ml. Appropriate concentration of labeled Ara-C, AAC, or DAZ are added to the tubes at time 0 along with dCyd at concentrations ranging from 10 7 to 5 × 10 - ' M. The tubes are placed in a 37°C shaking metabolic water bath for 1 h after which aliquots are pipetted into micro- fuge tubes containing I(X) p.I of a silicon and mineral oil mixture and 40 u.l of 5c7c, HCIO~. The tubes are then centrifuged at 10,000 ×g × 60s, frozen, sectioned, and the bottom layer. containing the precipitated cell pellet, placed in scintillation vials containing 12 ml of aqueous scintillation cocktail (Aqua- sol. National Diagnostics). Radioactivity was determined with a Beckman Model LS7()~) scintillation counter and intra- cellularly accumulated drug expressed as pmol dCyd analog./' HI t' cells.

.%'ucleic actd incorporation ~tudies

Cells arc exposed to labeled d('yd analogs in conjunction with d('yd as in the previous section except that the incubation

Interaction of deoxycytidine and deoxycytidine analogs 1141

period is 4 h. At the end of this time the cells are pelleted, washed, and the cell pellet precipitated with 1 ml of 10% cold trichloroacetic acid. The acid insoluble material is then repetitively washed with 5% trichloroacetic acid until neg- ligible radioactivity remains in the supernatant. DNA is then extracted from the samples by a previously described method 116]. Aliquots of the extracted material arc then placed in scintillation vials containing 12 ml of nonaqueous scintillation fluid (Betafluor, National Diagnostics. Somerset, N J) and radioactivity determined as before. The deoxyribose content of the material is determined by the methylamine method [17] and incorporation of dCyd analogs into DNA expressed as pmol analog/u.g deoxyribose.

Cloning studies ttL-60 and HL-6(I/Ara-C cells are plated utilizing a pre-

viously described soft agar cloning technique [18]. Briefly. logarithmically growing cells are suspended in a mixture consisting of RPMI medium, 20% fetal calf serum, and 0.3% Bacto agar (Difco, Detroit, MI). Half-milliliter aliquots of the mixture arc pipetted into 18-mm 12 well plates (Costar, Cambridge, MA) containing a bottom layer consisting of 0.5 ml of RPMI with 20% fetal calf serum and 0.5% agar. Each well contains 5(X) cells along with the appropriate concentrations of dCyd and each dCyd analog. After gelling, the plates are placed in a 37°C, fully humidified incubator for seven days. At the end of this period, colonies, consisting of groups of 50 or more cells, are scored with an Olympus model CK inverted microscope. The effect of each agent (with or without dCyd) on leukemic cells is expressed as the percentage of colony formation by drug treated cells relative to untreated controls. The I.C.,~ for each agent, defined as the drug concentration associated with a 50% reduction in colony formation, was determined by a previously described method for extrapolating from the dose response curves generated [19].

The effect of dCyd and dCyd analogs in the clonogenicity of normal human mveloid progenitor cells (CFU-GM) is deter- mincd by a similar bilayer agar cloning system [18]. Bone marrow samples are obtained with informed consent from patients undergoing routine diagnostic aspirations for non- malignant hematologic disorders and placed in sterile tubes containing 5(1t) I.U. of preservative free heparin. These studies have been approved by the Human Investigation Committee of the College of Physicians and Surgeons of Columbia Uni- versity. The samples are diluted 1 : 4 with modified McCoy's 5a medium and gently layered over a cushion of lymphocyte separation medium (s.g. 1.077-1.081; Bionetics, Kensington. MD) in sterile 50-ml centrifuge tubes. After centrifugation at 4(111 x g for 38 min at room temperature, the mononuclear laver is extracted with a sterile Pasteur pipette and washed twice with cold McCoy's medium. The cells were then plated in the same manner as HL-60 cells, except that McCoy's 5a medium is used instead of RPMI and 2 x 10 ~ cells are cultured per condition. In addition, 0.1 ml of GCT medium (Gibco) is added to each plate as a source of colony stimulating activity [20]. The plates are then placed in the humidified 5% CO: incubator for 7 days after which colonies, consisting of groups of 50 or more cells with granulocytic or maerophagic-like appearance, are scored as above. Dose response curves and I.C.,,, values for CFU-GM were obtained in the same manner as for 111,-60 cells.

Statistical analysis The statistical significance of differences in clonogenic

responses of cells to various drug regimens was determined utilizing the Student's t-test for paired or unpaired observations.

R E S U L T S

Dose response curves for the three dCyd analogs in each cell type are shown in Fig. 2. In general, Ara-C was the most inhibitory agent on a molar basis and A A C the least, with D A Z exhibiting an intermediate effect. The pattern of response in HL-60 and C F U - G M was similar for all three agents. For example, in both cell types low drug concentrat ions, e.g. 10 .9 M Ara-C and D A Z ; 10-SM A A C , were associated with minimal inhibitory effects, and a steep shoulder in the dose response curve was observed as the drug concentrat ion was increased. In both cases the slope of the dose response curve was steeper for Ara-C and D A Z than for A A C . At Ara-C and D A Z concentrat ions of 10 -7 M there was virtually a 100% inhibition of colony formation for C F U - G M and HL-60 whereas 10 -6 M A A C was required for the same effect. All three agents exhibited similar degrees of cytotoxicity toward CFU- GM and HL-60, although at low concentrations (10 -8 M), Ara -C and D A Z were more inhibitory toward C F U - G M .

As might be anticipated, the dCyd kinase deficient HL-60 /Ara -C variant which is highly resistant to Ara- C exhibited cross-resistance to both D A Z and A A C . The dose response curves were shifted dramatically to the right with drug concentrat ions as high as 10 -5 M having relatively minor effects on clonogenicity. At concentrat ions less than 10 -~ M. D A Z was slightly more inhibitory than Ara-C but differences were not significant. In contrast, at higher concentrat ions (10-4 M or greater) Ara -C was significantly more inhibitory than D A Z . At both low and high concentrat ions A A C exhibited less of an effect on clonogenicity than the other dCyd analogs. As with the more sensitive HL-60 and C F U - G M , a steep shoulder in each dose response curve was observed, although at a considerably higher drug concentrat ion, e.g. 10 .5 M.

A similar pattern of response was observed with respect to the 1.C.50 for each agent, or the drug concentration producing a 50% reduction, in soft agar colony formation ~Table 1). Values for each agent were slightly higher for HE-60 than for C F U - G M and more than 3 logs greater for H-60/Ara-C. In each cell type examined. Ara-C was slightly more inhibitory than D A Z on a molar basis, and approx. 1 log more inhibitory than A A C .

q'he ability of dCyd to reverse the inhibitory effects of each analog was examined in Fig. 3. In the upper tier of bar graphs, the ability of dCyd to reverse the inhibitory effects of relatively low drug concentrat ions ( 10 -7 M Ara-C and D A Z : 10 -6 M A A C ) was compared in HL-60 and C F U - G M . When dCyd and analog concentrations were equal, restoration of colony formation ranged from 2% tAra-C) to 42% (AAC) . In each case

1142

¢ ~ - G M

f !

Drug ~ e n t r a t ~ (M)

SII:.XL:N (iR.~NI. KAI'IL BFt,.\II.\ and Mc)SES (II/'~ZER

l, lt. -Ira AI I&-C l, I k . ~

i i ....... Z .......

r , . : - • • , a . . o . , e ,

0¢~9 ~ (k i ) Dn~{i ~ (M)

FI(;. 2. The effect of each deoxycytidinc analog on the ckm- ogenic behavior of CGU-GM. HL-60 and HL-60/Ara-C was compared through soft agar cloning methods, as described in the text. Cells were plated in the continuous prescnce of each analog at thc designated concentrations and colonies. consisting of groups of 50 or more cells, scored at the end of 7 days incubation in a 37~C. 5% CO: incubator. Values for each condition are expressed as the percentage of untreated control cell colony formation and represcnt the mean of at least 5 separatc experiments perlormed in duplicate ± 1 S.D. l'he ordinate and abscissa are both expressed in logarithmic

~cale.

the degree of protection was slightly greater for CFU- GM than for HL-60, but differences were not statistically significant (p = >0.05). When dCyd was present at a ten-fold excess, the degree of protection increased, particularly with AAC, where 62% of control colony formation was observed for CFU-GM. As with the lower dCyd concentration, protection was greater for CFU-GM than for HL-60 and the degree of pro-

I'ABI.k 1

I.C.,, (M) Drug HI.-60 tIL-60/Ara-C CFU-(JM

A r a - C 5.2-+-0.4 x I0 ~ 2 . 0 ± 0 . 2 x I0 5 D A Z 7.5 +-0.a × l() " 4 .5-* 0.3 × I0 A A C 7.6 .':- 0.5 × i0 ~ 2 .4 - * 0.3 × I0 :

3.1 .'- 0.3 x I0 '~ 5.1 :'- 0.3 × I0 " 1.5 ± 0.2 × l() :

The I.C.,,: for each c l ( y d analog was dc te rm incd k)r I l L - 6(1. [ll_-60/Ara-C and CFU-GM. Values represent the drug concentration associatcd with a 5(V~ reduction in colony formation relative to untreated control cells. (;ells were continuously exposed to each agent, and concentrations rcprcsent the mean of at least three separate experiments perk~rmed in duplicate ~ 1 S.D.

tection was not statistically significant (p > 0.05). In the presence of a hundred-fold excess of dCyd, reversal of drug effect in CFU-GM ranged from 68% (Ara-C) to 1()0°~ (AAC). The corresponding values for HL-60 were 50% (DAZ) to 88% (AAC).

When comparisons were made between the effect of dCyd on the response of HL-60/Ara-C and CFU-GM to high drug concentrations, a different pattern emerged (Fig. 3 lower tier). In the presence of a one hundred- fold excess of dCyd, the degree of protection for CFU- GM ranged from 44% (AAC) to 80% (Ara-C). ttowever, substantial protection was also observed in HL-60/Ara-C with equimolar concentrations of dCyd, ranging from 50% (AAC) to 62% (DAZ). At the highest dCyd concentration (100-fold excess), complete reversal of drug mediated growth inhibition was observed.

The effect of varying dCyd concentrations on the total intracellular accumulation and nucleic acid incorporation of the dCyd analogs in HL-60 and HL-60/Ara- C is illustrated in Table 2. In HL-60, administration of equimolar concentrations of dCyd exerted a relatively minor effect on the intracellular accumulation and DNA incorporation of each analog administered at 10 ..7 M.

Interaction of deoxycytidinc and deoxycytidine analogs 1143

8

© u;

"8

Z

<

O

< z

r-, Z <

0

e~

Z

I - <

<

e¢

:.)

<

Z

<

r,i

e n

<

Y

c

U

~ D D D

~ ' ÷ 1 +1 41

4-~ .

,t~ +1 ÷1 +1

rt3 4-1 +1 4-1

+~ ~ +~

t¢-~ e,3,

4; 41 +1 ~ ¢~1 .,-1- Ox

u 41 4~

~ N U

I I I

I I I

+1 ÷1 +1

I I I

+1 t"i +1

I I I

41 41 +1

o 0 t ' h ' -

;~. ~ =

r - , ~ Z

..e o o - 1

" = Z < ,-r ~ f - - Z

.o ~ . ~ . -

.~o> o

~ ~ - - ' ~ O - = . E ~

t.. ~ t- 0

t~ .E "u .E

--~Zb

, ~ - ~ ~_

c .~ ~ . -

. ~ . . ~

.~.--- ~-~ ~ .E o - ~ >

When dCyd was present in ten-fold excess, the intracellular accumulation of drug was decreased by 50% or more and incorporation into nucleic acid was reduced to an even greater extent. A further increase in the dCyd concentration to 10 -5 M resulted in a 95% reduction in drug accumulation and no detectable incorporation into DNA.

A similar pattern was observed when HL-60/Ara-C cells were exposed to higher concentrations of both dCyd and each analog (Table 2). A dCyd concentration of 10-4M was required to reduce the intracellular accumulations of each analog when the latter were administered at 10 -5 M; a further increase in the dCyd dose resulted in no detectable drug accumulation. It should be noted that even in the absence of dCyd, incorporation of drug into DNA of these highly resistant cells could not be reliably quantitated.

D I S C U S S I O N

The nucleosicle analogs Ara-C, DAZ, and AAC have been shown to be active against a wide variety of animal and human tumor cell lines [21-23]. The present study provides data concerning the relative efficacy of these agents toward sensitive and resistant human leukemic cell sublines, as well as normal human myeloid progenitor cells. Our results suggest that with respect to HL-60 cells as well as normal CFU-GM, Ara-C is slightly more potent than DAZ, and approx. 10-fold more effective than AAC. These results differ from those reported recently by Glazer and colleagues work- ing with the human colon carcinoma line HT-29 [7]. These investigators found that AAC was the most potent agent on a molar basis and that D A Z and Ara- C were marginally less effective. It should be noted that in these studies relatively short drug exposure intervals, e.g. 2 and 24 h, were utilized whereas we examined a continuous 7-day exposure interval. Since 5-azacytidine derivatives such as D A Z and AAC undergo ring scission in aqueous solution [24], it is possible that drug inac- tivation over a long period may have decreased their efficacy relative to Ara-C. Our data also demonstrate that the dCyd kinase deficient, highly Ara-C resistant cell line HL-60/Ara-C also displays a high degree of cross resistance to AAC and DAZ. Since dCyd kinase catalyzes the rate limiting step in clCyd analog metabolism [25], deficiency of this enzyme presumably pre- vents accumulation of sufficient concentrations of the active nucleoside tripHosphate derivatives. Although loss of dCyd kinase is a common mechanism of Ara-C resistance in murine leukemic cells [26], it has not been established to play a consistent role in clinical Ara-C resistance in humans [27]. Nevertheless, if inability of leukemic cells to convert Ara-C to its nucleotide form contributes to lack of efficacy of this agent, refrac- toriness to A A C and D A Z might also be anticipated.

dCyd has previously been shown to reverse the effects of Ara-C in a murine leukemia [11] and D A Z in a hamster fibrosarcoma cell line [9], and has now been demonstrated capable of reversing AAC mediated growth inhibition in normal and leukemic myeloicl pro-

1144

z o I-- <

o i1 >-

. J o o

S I I \ I N ( i R - X N I K . \ I ' I I t 3 t 1 \ I : ~, a n d M O S l s ( ; I . t ' ~ Z l R

• (X} I

4C k-

Ct

D A Z (10 7 M)

10 "0 ' ~:3 . d l . f~q

"0

A R A - C (10 7 M) AAC (10 6 M)

lr ,i l, _ [. "d ' ' ( 1~ ' ",3 1{, •

D A Z (10 "n M) -

lOPr l

~t L L . 1 0 *

20 •-

3L~ . . . . 10 ' 10 ' 10 '

A R A - C (10 s M) J=

_ !l .. 1Q' i(? ' lO '

HL.6O CF ~,-'LiI'¢ 'tL 60 AI4A.C

Fl(i. 3. The ahil i tv o t~ar ious d( 'wt concentrat ions to reverse the inhib i tory cffcct~ o f l ixed concentrat ions of Ara-C. D A Z and A A C were compared in II1.-60. CFL?-GM, and HL-60., A ra -C uti l izing a soft agar cloning mcthod as described in the text. ('ells were plated in the continuous presence of each analog at the designated concentration in c,mjunction with increasing concentrations of dCvd. lhe height of each bar graph represents the percentage ot colony formation for each condition relative to untreated controls. In the upper tier ol graphs, the responses ot CFU-GM and HL-6(} arc compared: in the lower tier. comparisons arc made bctv.een the grovdh of CFU-GM and l]l,-61)Ara-(" Values tor each condition represent the mean for 5 separate cxpcrimcnt~, performed in

duplicate -" S.D

A A C (10 4 M) .[.,

10 'C ' 1(I '

genitor cells as well. It is possible that dCyd may compete with its analogs at a variety of steps in the pyrim- idine salvage biosynthetic pathway. For example, the affinity of calf thymus dCyd kinase for dCyd is higher than for Ara-C, e.g. K m l . 4 x 10-SM for dCyd: 4 × 10 5 M for Ara-C [28]. Similarly, dCyd may compete with its derivatives at the level of nucleoside transport, phosphorylation by the mono and diphosphate kinase, binding to DNA polymerase, and incorporation into DNA. Although controversy exists, recent evidence suggests that incorporation of dCyd analogs into tumor cell DNA may be a prime determinant of drug mediated cytotoxicity. Kufe and co-workers have shown that leukemic cell incorporation into DNA most closely correlates with cytotoxic effects [29] and Momparler has found that incorporation of DAZ into DNA of EMT~ and L1210 tumor ceils may be directly related to lethality [301. The recent studies by Glazer in the HT-29 colon tumor cell line demonstrate that while AAC, like Ara- C, is a potent inhibitor of DNA synthesis, it is also incorporated into DNA [7]. In our studies, a dCyd concentration in excess by ten-fold or more was necessary to antagonize the effect of Ara-C, AAC, or DAZ in HL-60 cells, and these dCyd levels were associated

with a significant reduction in nucleic acid incorporation of drug. However, since coadministration of dCyd also reduced the total intracellular accumulation of drug at the same relative concentration, it cannot be concluded from our data that inhibition of DNA incorporation was solely responsiblc for reversal of drug effect.

A major aim of this study was to determine whether altering exogenous dCyd concentrations might ameliorate analog mediated toxicity toward normal myeloid elements while preserving activity towards their leukemic counterparts. Previous in-v ivo studies involving mice innoculated with the murine leukemia L1210 demonstrated that coadministration of dCyd might improve the therapeutic index of Ara-C [31]. However. our in-vitro data suggest that a continuous coadministration of dCyd for 7 days has only a modest effect on improving the selectivity of dcoxycytidine analogs. For example, Ara-C, AAC, and DAZ were slightly more toxic to CFU-GM than to HL-60 when administered separately, but when dCyd was added, the com- bined regimens proved somewhat more toxic to HL-60. In all cases the improvcmcnt in selectivity was relativeh, minor, and it remains questionable whether these dil:- fcrcnces could be exploited in an in-v ivo setting. It is

Interaction of deoxycytidine and deoxycytidine analogs 1145

possible that al ternative dCyd/dCyd analog dose relationships and schedules might display superior selectivity, but these remain to be uncovered. Our data utilizing the highly Ara-C resistant leukemic subline HL-60 /Ara -C differ from results reported by Harris et al. who studied Ara-C metabol ism in patient derived human leukemic myeloblasts. These workers found that dCyd was selectively incapable of antagonizing Ara-C metabolism in cells displaying low dCyd kinase activity [12]. In contrast , we found that when Ara-C, A A C , and D A Z were administered at the high concentrat ions necessary for anti-proliferative effect, e.g. 10 -~ M or greater, addition of dCyd in excess was capable of antagonizing both intracellular drug accumulation as well as cytotoxicity. Moreover , the ability of dCyd to protect C F U - G M from nucleoside analogs administered at high concentrat ions was in general less than that observed for HL-60 /Ara -C . It should be noted that in contrast to the studies by Harris et al. in which a dCyd concentrat ion of 10 -~'M was used, we examined the effect of dCyd concentrat ion considerably higher than physiologic levels. In addition, the HL-60 /Ara -C subline represents a relatively homogeneous populat ion of dCyd kinase deficient cells which display very high degrees of resistance to dCyd analogs. It is likely that patient derived leukemic myeloblasts are considerably more heterogenous in nature, and may develop resistance through a variety of mechanisms.

The relevance of comparing the effects of dCyd/ nucleoside analog containing regimens in normal myeloid progeni tor cells and continuously cultured human leukemic cell lines will depend in part upon the degree to which the latter reflect primary leukemic cell behavior. Primary cultures of human leukemic mveloblasts, unlike continuously cultured cell lines, grow irregularly in short- term culture and cannot be propa- gated indefinitely. These biologic differences may reflect fundamental biochemical characteristics unique to continuously cultured lines. Recent studies bv McCulloch and co-workers have suggested that the secondary plating efficiency of leukemic mveloblasts following exposure to various agents mav be the most relevant biological correlate of clinical response [32]. The group has also reported that dCyd analog such as D A Z may be highly inhibitory to the secondary plating efficiency of leukemic myelol~lasts in culture [33]. In light of these findings, it would be of interest to compare the effect of dCvd on analog mediated cytotoxicity in normal vs patient derived leukemic myeloid progeni tor cells in order to identify dose relationships and schedules selectively toxic to leukemic elements. Such studies are currently underway in our laboratory.

R E F E R E N C E S

I. Danhauser L. L. & Rustum Y. M. (198(I) Effect of th~- midine on the toxicity, anti-tumor effect and metabolism of 1-B-D arabinofuranosylcytosin¢ in rats bearing a chemi- cally induced colon carcinoma. Cancer Res. 40, 1274.

2. ,Moore E. ('. & Hurlbert R. B. (1966) Regulation of

mammalian deoxyribonucleotide biosynthesis by nucleo- tides as activators and inhibitors. J. biol. Chem. 241, 4802.

3. Ellison R. R., Holland J. F., Well. M., Jacquillat C.. Boison M., Bernard J., Sawitsky A., Rosner F., Gussof B.. Silver R. T., Karanas A.. Cuthner J., Spurr C. L.. Haves D. M., Biota J., Leone L. A., Haurani F.. Kyle R., Hutchison J. L., Forcier R. J. & Moon J H. (1968) Arabinosyl cytosine: a useful agent in the treatment of acute leukemia in adults. Blood 32, 507.

4. Clarkson B. D.. Dowling M. D. 8̀ : Gee T. S., Cunningham I. B. & Burchenal J. H. (1975) Treatment of acute leukemia in adults. Cancer 36, 775.

5. Major P. P.. Eagan E. M., Beardsley G. P.. Minden M. & Kufe D. W. (1981) Lethality of human myeloblasts correlates with the incorporation of Ara-C into DNA. Proc. ham. A c a d Sci. U.S.A.78, 3235.

6. Bouchard J. 8,: Momparler R. L. (1983)Incorporation of 5-aza-2'-deoxycytidine-5'-triphosphate into DNA. Mol. Pharmac. 24, 109.

7. Glazer R. I. & Knode M. C. (1984) 1-B-D arabinosyl-5- azacytosine cytocidal activity and effects on the synthesis and methvlation of DNA in human colon carcinoma cells. Mol. Pharmac. 26, 381.

8. Biro V. & Goldenberg D. M. (1971) Hematological toxicity of 1-B-D-arabinofuranosylcytosine and its prevention by deoxycytidine in the mouse. Chemother. 16, 29.

9. Momparler R. L. 8̀ : Goodman J. (1977) ln-vltro cytotoxic and biochemical effects of 5-aza-2-deoxycytidine. Cancer Res. 37, 1636.

10. Buchman V. M.. Svet-Moldavskv G. J., Lichenister M. R. 8̀ : Mkheidzc D. M. (1977) Selective chemotherapy for advanced murine L1210 leukemia bv combination of 2'- deoxycytidine and lethal doses of cytosine arabinoside. Biomed. 27, 179.

11. Buchman V. M.. Belyanchikova N. K. & Mkheidze D. M. (1979) 2'-Dcoxycytidinc chloride protection of mice against the lethal toxicity of cytosine arabinoside. Cancer Chemother. Pharmac. 3, 229.

12. llarris A. I,. 8̀ : Grahamc-Smith D. G. (1981) Cytosine arabinoside triphosphate production in human leukemic mveloblasts: Interaction with deoxycytidine. Cancer ('hemother. Pharmac. 5, 185.

13. Gallcgher R.. Collins S.. TrujiHo J.. McCredie M.. Ahearn M.. Tsai S.. Metzgar R., Aulakh G.. Ting F. Rusceni F. & Gallo R. (1979) Characterization of the continuous differentiating mvcloid cell line (HL-60) from a patient with acute prom.~eloc.vtic leukemia. Blood 54, 713.

14. Bhalla K.. Navak R. 8̀ : Grant S. (1984) Isolation and characterization of a dcoxvcvtidine kinasc-deticient. human promyclocytic leukemic cell highly resistant to 1- B-D arabinofuranosylc.vtosinc. Cancer Res. 44, 5029.

15. (;rant S.. I,chman C & Cadman E. ( 1981 ) Enhancement of l-B-D-arabinofuranosvlcvtosine accumulation with L1210 cells and increased cvtotoxicitv follo~ing thvmidine exposure. Cancer Res. 40, 1525.

16. Grant S. 8̀ : Cadman E. (1980) Augmentation of 5-azacv- tidmc incorporation and effect by 3-dcazauridine in LSITSY cells and human leukemic mveloblasts. Cancer R('.~. 40, 4000.

17. Burton K. A. (1956) A study of the conditions and mechanism of the dephen.vlamine reaction for the colorimctric estimation ,.'ff deoxvribonucleic acid. l~io('hem. J. 62, 315.

1146 St~-',t~ Gray t . K.-,,PJt. BHa~ t..a and Most!s (h.t:'~zLr

18. Grant S.. Rauseher F. & Cadman E. (1982) Differential effect of N-phosphonacetyl-t.-aspartatc on I-B-D arabinofuranosylcytosine metabolism and eytotoxicity in human leukemia and normal bone marro~ progenitors. Cancer Res. 42, 4007.

19. Chou T-C. Burchenal J. H , Schmid F. A.. Braun T. S.. Su T-L.. Watanabe K. A.. Fox J. J. & Phillips F. S. (1982) Biochemical effects of 2'-fluoro-5-methvl-l-B-D arab- inofuranosvl-uracil and 2'-fluoro-5-iodo-l-B-D arabino- furanos.vlc.vtosinc. Cancer Res. 42, 3957

20. Dipersio J. F., Brennan J. K., Lichtman M. A. et al. (1980) The fractionation, characterization and subcellular localization of colony stimulating activities released by the human monocyte-like cell line GCT. Blood .5~, 716.

21. Momparler R. L. & Gonzales F. A. (1978) Effect of intravenous 5-aza-2'-deoxycytidine on survival of mice with L1210 leukemia. Cancer Res. 38, 2673.

22. Beisler J. A.. Abbasi M. M. & Driscoll J. S. (1977) Synthesis and anti-tumor activity of arabinosyl-5-azac~- tosine. Biochem. Pharrnac. 26, 2469.

23. Furth J. J. & Cohen S. S. (1968) Inhibition of mammalian DNA polymerase by the 5'-triphosphate of 9-B-D-arabinofuranosylcytosine and the 5'-triphosphate of 9-1)-arab- inofuranosyl-adenine. Cancer Res. 28, 2061.

24. Notari R. E. & De Young J. L. (1975) Kinetics and mechanism of degradation of the anti-leukemic agent 5- azacytidine in aqueous solution. J. Pharm. Sci. 64, 1 I48.

25. Plagemann P. G. W., Marz R. & Wohlheutcr R. M. (1978) Transport and metabolism of deox.vcytidine and 1-B-I)- arabinofuranosylcytosine into cultured Novikoff hepatoma cells, relationship to phosphorylation and regulation of triphosphate synthesis. Cancer Res. 38, 978.

26. Chu M. Y. & Fisher G. A. (1965) Comparative studies of leukemic cells sensitive and resistant to cytosine arab- moside. Biochem. Pharmac. 14, 333.

27. Yattersall M. N.. Ganeshagaru K. & Hoffbrand A. V. (1974) Mechanisms of resistance of human acute leukemia cells to cytosine arabinoside. Br. J. Haemat. 27, 39.

28. Momparler R. 1.. & Fischer G. A. (1968) Mammalian deoxynucleoside kinases. 1. Deoxycytidine kinase: purific- ,ltion. properties and kinetic studies with cytosine arab- inosidc. J. biol. Chem. 243, 4298.

29. Major P. P.. Eagan E. M., Beardsley G. P., Minden M. & Kufe D. W. (1981) Lethality of human myeloblasts correlates with the incorporation of Ara-C into DNA. Proc. natn. A c a d Sci. 78, 3235.

30. Momparler R. 1... Vesely J.. Momparler L. F. & Rivard G E. (1979) Synergistic action of 5-aza-2'-deoxycytidine and 3-deazauridine on L1210 leukemic cells and EMT~ tumor cells. Cancer Res. 39, 3822.

31. Buchman V. M.. Svct-Moldavsky G. J., Lichinitscr M. R. & Mkheidze D. M. (1977) Selective chemotherapy for advanced murine LI210 leukemia by combination of 2'- deoxycytidine and lethal doses of cytosine arabinoside. Biomedicine 27, 179.

32. Curtis J. E., Messner H. A., Hasselback R., Elhakim T. & McCulloch E. A. (1984) Contributions of host and disease related attributes to the outcome of patients with acute myelogenous leukemia. J. clin. Oncol. 4, 253.

33. Motoji T.. Hoang T., Tritchler D. & McCulloch E. A. (1985) The effect of 5-azacytidine and its analogues on blast cell renewal in acutc myelogenous leukemia. Blood 65, N94.

Documents

Interaction of deoxycytidine and deoxycytidine analogs in normal and leukemic human myeloid progenitor cells