Bioconjugate Chem. 1995, 6, 536-540
536
Quantitative Analysis of 3’-Azido-3‘-deoxythymidine Incorporation into DNA in Human Colon Tumor Cells Minoti Sharma,* Rama Jain, Elizabeth Ionescu, a n d J a m e s W. Darnowski’ Department of Biophysics, Roswell Park Cancer Institute, Buffalo, New York 14263. Received March 14, 1995@
We have previously reported that 3’-azido-3’-deoxythymidine(AZT) can possess significant antineoplastic activity in vitro and in vivo when combined with agents which inhibit de novo thymidylate synthesis. Under these conditions cytotoxicity is closely associated with the degree t o which AZT is incorporated into DNA. We now report a fluorescence postlabeling technique by which AZT incorporation into DNA can be quantitated without employing radiolabeled AZT. Cultured human colon tumor (HCT-8) cells were exposed to various concentrations of AZT alone and in combination with 5-fluorouracil (FUra). Control cells received the same amount of medium. DNA was isolated from harvested cell pellets (2 x lo7). Enzymatic digestion of DNA to the mononucleotide level followed by HPLC analysis of the digest showed that the DNA preparation was free of RNA contamination. The DNA digest was conjugated with dansyl chloride in situ via the phosphoramidate derivative with ethylenediamine. HPLC analysis of the postlabeled nucleotides using fluorescence detection detected 105,245, and 479 fmol of 5’-monophosphate of AZT (AZTMP) per pg of DNA from cells exposed to 20, 50, and 100 pM AZT, respectively. FUra (3 pM) doubled the AZT incorporation per pg of DNA in cells exposed to 50 and 100 pM AZT. These findings generally support our previously reported data which quantitated (3H)AZTincorporation into cellular DNA and are discussed in light of the potential clinical utility of this technique in assessing the relationship between AZT incorporation into DNA and therapeutic action.
INTRODUCTION
We have reported, using a human colon tumor model, that AZT in combination with FUra or methotrexate (MTX)exert superior cytotoxic and antineoplastic effects compared to either drug alone (2-3). As a result, clinical analysis of AZT-based regimens for the treatment of cancer have been initiated, and presently phase I1 analysis of AZT FULV is underway. Sommadossi et al. have reported that AZT-induced cytotoxicity in human bone marrow cells is related to AZT incorporation in DNA (4). Our group also has observed that AZT cytotoxicity correlated closely with the size of intracellular pools of di- and triphosphates of AZT and the amount of AZT incorporation into cellular DNA (3). Thus far, all the reports correlating AZT incorporation in DNA and cytotoxicity have been based on studies utilizing radiolabeled AZT. As a result, clinical confirmation of the therapeutic relevance of AZT incorporation into DNA is not practical. Therefore, we now report a technique to analyze AZT incorporation into DNA using nonisotopic detection and demonstrate its capacity to quantitate the incorporation into cellular DNA. These results and the method are discussed in light of their potential utility in ongoing clinical trials.
+
MATERIALS AND METHODS
Chemicals and Reagents. The standard deoxynucleotides, 1-methylimidazole, 1-(3,3-bis(methylamino)propyl)-3-ethylcarbodiimide hydrochloride (CDI), 5-(dimethy1amino)naphthalene 1-sulfonylchloride(Dansyl chloride), protected mononucleotide phosphodiester, the enzyme nuclease P1, and FUra were purchased from Sigma
* To whom all correspondence should be addressed. Phone: (716) 845-8296. Fax: (716) 845-8899. ’ Department of Medicine, Brown University and Roger Williams Hospital, Providence, Rhode Island 02908. Abstract published in Advance ACS Abstracts, July 15, 1995. @
1043-1802/95/2906-0536$09.00/0
(St. Louis, MO). Ethylenediamine (EDA), triisopropylbenzenesulfonyl chloride (TPSCI), anhydrous pyridine, and 2-cyanoethyl phosphate (barium salt dihydrate) were obtained from Aldrich Chemical Co. AZT was the generous gift of the Burroughs Wellcome Co. (Research Triangle Park, NC), and RPMI 1640 medium and fetal bovine serum (FBS) were purchased from Gibco (Grand Island, NY). HPLC grade solvents, chemical, and disposable tissue culture supplies were obtained from Fisher Scientific (Medford, MA). Cell Line. Continuous cultures of HCT-8 human colon adenocarcinoma cells, obtained from the American Type Culture Collection, were used in these studies. The biochemical and histological characterization of this cell line has been reported (5). Cells were cultured in sterile plastic tissue culture flasks as monolayers in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and passed twice weekly. Cell cultures were maintained in a humidified incubator a t 37 “C in an atmosphere of 5% COZ. Under these conditions their doubling time was 20 h and cells in logarithmic growth were used in all studies. In Vitro Evaluation of Cytotoxicity. HCT-8 cells (1 x lo5) were added to 10 mL of RPMI 1640 media containing 10% FBS in 25 mL culture flasks. AZT and FUra previously dissolved in media were added a t concentrations of 20,50, and 100 pM AZT and 3 pM FUra either alone or in various noted combinations. Control cultures received the same amount of media without drug. After 5 days, cells were harvested and growth inhibition was determined as described previously ( 2 , 2 ) . Each experiment was performed in duplicate and repeated a minimum of four times. For experiments to quantitate AZT incorporation into DNA and the effect of FUra on this parameter, approximately 3 x lo6 HCT-8 cells were added to 150 mL flasks containing 50 mL of RPMI 1640 media, 10%FBS, and the various noted concentrations of either FUra and/ 0 1995 American Chemical Society
Bioconjugafe Chem., Vol. 6,No. 5, 1995 537
AZT Incorporation into DNA
.“-Of OR I
HO{
i
O
T
I11
1 +
HO-/-O{N3 OH
N,
O
A
*
HO-!;toH-
T
HO-[;tO-[;$N,
v
I1
O
IV
Figure 1. Scheme for the chemical synthesis of the 5’-monophosphate of AZT and its dinucleotide d(pApAZT).
0 I1 -O-P-OCH~
-b
~
‘coj
COI 1 MI0AZOLE BUFFER, pH6
N3
0
4
N3
CD I = 1 -ethyl-3,3-dimethylominopropylcorbodiimide
Figure 2. Scheme for the synthesis of the fluorescence-labeled 5’-monophosphate of AZT
or AZT. After 5 days the cells (approximately 2 x lo7) were harvested following reported procedure (3). Preparation of 5‘-Monophosphate of AZT (AZTMP) and Its Dinucleotide d(pApAZT1. The scheme for the synthesis of AZTMP and its dinucleotide d(pApAZT)is shown in Figure 1. AZT was phosphorylated with 2-cyanoethyl phosphate following a previously reported procedure (6). The fully protected dinucleoside monophosphate (111) was synthesized in the solution phase by a modified phosphotriester approach (7). The product was purified by column chromatography on silica using 0-5% methanol in dichloromethane. The pure product was phosphorylated with 2-cyanoethyl phosphate as in the case of AZT. After being deblocked with ammonia and pyridine (9:l v/v), the product (IV) was isolated by C18 reversed phase chromatography (5 ym, 10 mm x 25 cm) using a 30 min linear gradient of 0-20% acetonitrile in 0.1 M ammonium acetate buffer and desalted on the same system using a linear gradient of 0-100% methanol in water. AZTMP (11) was also isolated by reversed phase HPLC under similar conditions. The chemical shifts in ppm with reference to TSP a t 1.93 (d, 3H, CHd, 2.49-2.52 (m, 2H, CH-2’ and CH2’7, 4.06-4.23 (m, 3H, CH-5’, C H - 5 , CH-4’1, 4.50-4.53 (m, l H , CH-3’), 6.26-6.30 (t, l H , CH-l’), and 7.7 (s, l H , CH-6) were in agreement with the structure of AZTMP.
The downfield chemical shifts of the dinucleotide shown at 8.52 (s, l H , AH-8), 8.27 (s, l H , AH-2), 7.5 (s, l H , AZT H-6), 6.46-6.49 (t, lH, AH-1’1, and 6.13-6.16 (t, lH, AZT H-1’) with respect to TSP supported the structure of d(pApAZT). Nuclease P1 digestion of IV afforded U P (V) and AZTMP (11) as shown in Figure 1. ‘H NMR measurements were done on a Bruker WP 200 spectrometer. Figure 2 shows the scheme used for the synthesis of dansylated nucleotide. The phosphate group of AZTMP (1)reacts with water-soluble l-(3,3-bis(methylamino)propyl)-3-ethylcarbodiimide (CDI) in l-methylimidazole buffer a t pH 6 to generate the phosphorimidazolidate 2 which when exposed to ethylenediamine (EDA) results in the formation of a stable 5’-phosphoramidate derivative 3. The free amino group of the phosphoramidate reacts readily with dansyl chloride in 50 mM borate buffer, pH 9.5, to yield the fluorescentlabeled nucleotide 4. Fluorescence Postlabeling Assay of d(pApAZT1. The dinucleotide (1 OD) was digested with nuclease P1 following reported procedure (8). The digest was filtered on an ultrafree microunit with 10 000 NMWL polysulfone membrane and labeled with dansyl chloride as described below. The filtrate was adjusted to pH 6 with 0.1 M NaOH and lyophilized. A cocktail (50 yL) prepared by
538 Bioconjugate Chem., Vol. 6,No. 5, 1995 mixing 40 mg of CDI, 15 pL of EDA, and 8 pL of 1-methylimidazole in 1mL of water, pH 6, was added to the lyophilized material. The reaction mixture was left a t room temperature overnight to prepare the phosphoramidate derivative of the digested nucleotides. The pH of the reaction mixture was then adjusted to 9.5 with 0.2 M NaZC03, and 50 p L of dansyl chloride (1g/10 mL acetone) was added. After being stirred a t room temperature in the dark for 1h, the dansyl-labeled reaction mixture was filtered on a microcentrifuge to remove any particulate material, and the filtrate was frozen until ready for HPLC analysis. A sample of AZTMP (1 OD) was also dansylated following the same procedure. Postlabeling and HPLC Analysis of Postlabeled DNA Digests. Typically, DNA was isolated from a pellet containing 2.5 x lo7 cells. The pellet was lysed by a guanidinethiocyanate procedure (9). The protocol was modified by adding 0.75 vol of ethanol to the lysate. After 2 h a t -22 "C, the mixture was centrifuged a t l O O O O g for 20 min a t 4 "C, and DNA was isolated from the supernatant by ethanol precipitation, proteinase K treatment of the precipitate, and organic extraction following the standard procedure. DNA was digested enzymatically to the nucleotide level, and the nucleotide profile of the digest was routinely monitored by HPLC to determine the purity of the preparation. Typically, 100 pg of DNA (2 OD) was digested with 2 pL of DNAse-1 (2000 KU/lOO pL) and 2 pL of nuclease P l ( 1 mg/mL) in 40 pL of Tris (10 mM), EDTA (0.1 mm), MgClz (4 mM) buffer, pH 7.5 containing 2 pL of ZnS04 (10 mM) and 4 p L of NaOAc (38 mM, pH 5.0) at 37 "C overnight. The digest was derivatized with EDA cocktail following the procedure described for the DNA model study. However, in order to enrich the modified nucleotide from the normal nucleotides in the DNA digest, the phosphoramidate reaction mixture was fractionated by HPLC. A fraction corresponding to the retention time of phosphoramidate derivative of AZTMP was collected, lyophilized, and labeled by adding 25 pL of dansyl chloride in 100 pL of 0.1 M carbonate bicarbonate buffer, pH 9.5. An aliquot of the labeled, enriched fraction was analyzed by HPLC using fluorescence detection as shown in Figure 5. Alternatively, the EDA reaction mixture was reacted with dansyl chloride following the procedure described in the model study, and the postlabeled digest was fractionated by HPLC using fluorescence detection. A fraction corresponding to the retention time of dansylated AZTMP was collected isocratically with 23% acetonitrile in 0.1 M ammonium acetate. Reanalysis of an aliquot of the enriched fraction using the high-sensitive detector accessory under the same elution conditions detected the peak of interest. HPLC Separation of Nucleotides of Normal and Modified Bases. Prior to labeling, the nucleotides were separated on a Radial-Pak LC cartridge 8MBC18 (10 pm, 8 mm x 10 cm) using a 30 min linear gradient of 0-20% acetonitrile in 0.1 M ammonium acetate, pH 7. The profiles were monitored a t 254 nm using a Beckman variable wavelength detector with an Altex spectrophotometer cell. The postlabeled nucleotides were analyzed on a microsorb C18 column (5 pm, 4.6 mm x 25 cm) using a McPherson detector model 750B equipped with a highly sensitive detection accessory to monitor fluorescence. The elution conditions are described in the figure legends. RESULTS AND DISCUSSION
In the present study the ICE,, of AZT in HCT-8 cells after a 5-d exposure was approximately 67.5 pM. Under these conditions the ICsoof FUra was 2.3 pM. Analysis of combined effect of AZT and FUra revealed that these
Sharma et al. Table 1. Analysis of the Combined Effect of FUra and AZT on IC50 of FUra or AZT in HCT-8 Cellsa
incubation condition control 0.5 pM FUra 1.5p M FUra 20 pM AZT 50 pM AZT
of FUra 2.3 i 0.2
of AZT 67.5 i 4.8 49.4 i 6.2 32.0 i 4.6
1.8 6 0.3 1.4 f 0.4
Twenty-five mL tissue culture flasks containing 10 mL of RPMI 1640 media 10% FBS, 1 x lo5 cells and various concentrations of AZT alone, FUra alone, or their noted combinations, were incubated at 37 "C. After 5 days, the cells were harvested and the cell number was determined. Percent growth inhibition was quantitated using cells incubated without AZT or FUra as control. Each value represents the mean f SE from four determinations.
+
agents exerted additive cytotoxic effects (Table 1). Upon closer examination, however, it was apparent that while FUra was able to increase the cytotoxic activity of AZT, AZT exerted less of a n effect on the cytotoxicity of FUra. Previous results using radiolabeled AZT have suggested that, in this model, AZT-induced cytotoxicity was closely associated with the degree to which AZT was incorporated into DNA (1-3). Reflecting both the expense of these isotopic studies and their lack of utility in the clinical setting we have therefore attempted to develop a method to assess AZT incorporation into DNA by directly quantitating the AZT nucleotide content of DNA in cells exposed to this thymidine analogue. To this end, the 5'-monophosphate of AZT and its dinucleotide derivative were synthesized and labeled as described in the methods (Figures 1and 2). The labeling reactions, though shown in three steps in Figure 2, were carried out in a single pot. The overall yield was 90%. Thereafter, chromatographic conditions were devised to resolve the nucleotides of normal and modified bases both before and after labeling. Figure 3 shows the reversed phase HPLC profiles of (a) the dinucleotide d(pApAZT1, (b) nuclease P1 digest of the dinucleotide, and (c) the normal nucleotides and AZTMP. It has been reported that nuclease P1 releases the normal nucleotides from modified DNA as 5'-monophosphate (pN) whereas the modified nucleotides, depending on the nature of modification, are excised as dinucleotide (pNpX, X = modified base) (10). A DNA model study shows that nuclease P1 excised AZT as 5'-monophosphate (pX). The retention times of the two peaks in profile b matched the retention times of authentic dAMP and AZTMP shown in profile c. This was further confirmed by cochromatography (results not shown). HPLC analysis of the postlabeled digest also supported such an observation (Figure 4b). Cochromatography (profile 4c) of the postlabeled digest of d(pApAZT) with the postlabeled dinucleotide demonstrated clearly that the peak a t 21 min in the profile 4b is from labeled dAMP and not from the labeled dinucleotide. The labeled dinucleotide is eluted a t 22.5 min as shown in the profile 4a. HPLC analysis of the postlabeled nucleotides of known concentrations also revealed that the labeling yields of AZTMP and its dinucleotide derivative were quantitative. Figure 5 shows fluorescence postlabeling assay from control (profile c) and AZT exposed (profile b) HCT-8 cells. Unlike the DNA model study, enrichment of modified nucleotide is essential in order to detect the peak of interest in modified, cellular DNA by postlabeling technique from the huge background of normal nucleotides (10).As described in the methodology, the enrichment of the modified nucleotides from the normal nucleotides
Bioconjugate Chem., Vol. 6,No. 5, 1995 539
AZT Incorporation into DNA
A Z T mD
Oonsyi Sulfonic Acid
10
20
(0
20
30
40
10
20
30
40
AZTmp
c,
4 AZTmp
10
20
Time (min)
30
40
c
1,
d I
10
20
Figure 3. HPLC profiles of (a) d(pApAZT), (b) nuclease P1 digest of d(pApAZT), and (c) 5’-monophosphates of C, T, G, A, and AZT eluted with a 30 min linear gradient of 0-20% acetonitrile in 0.1 M ammonium acetate buffer, pH 7.
can be achieved by HPLC fractionation of the DNA digest both before and after labeling. The latter approach offers the advantage of enriching the modified nucleotide not only from the normal nucleotides but also from the excess labeling reagent. However, the successful application of this procedure depends on the chemical nature of the modified nucleotide. We observed that AZTMP can be resolved from the normal nucleotides both before and after labeling. The peak a t 19 min in profile b of Figure 5 was identified as AZTMP by cochromatography of the postlabeled DNA digest with authentic, dansylated AZTMP (profile a). The percent digestion of DNA was calculated by HPLC, prior to labeling, from the integrated area of excised dAMP peak and the response factor of standard dAMP. The efficiency of digestion of modified DNA was 90% of control. The profiles in Figure 5 represent analysis of approximately 1pg of DNA. The results shown in Table 2 reveal the relationship between the media concentration of AZT and the degree of AZT incorporation into cellular DNA detected by fluorescence postlabeling technique. Each number in this table is a n average of three independent measurements. Having demonstrated that the fluorescence postlabeling technique could be used to quantitate AZT incorporation into cellular DNA, we next assessed the effect of exposing cells to both FUra and AZT on the degree to which AZT was incorporated into cellular DNA in this model using nonisotopic detection. The results of these studies indicated that FUra (3 pM) coexposure doubled the incorporation of AZT into DNA from cells exposed to 50 pM ( I C d and 100 pM (Table 2). Biochemical analysis of acid (PCA)-insoluble material from HCT-8 cells exposed to 5 pM FUra increased the incorporation of radiolabeled AZT into the nucleic acid fraction by 52% and also decreased the IC50 of U T considerably. Furthermore, there appeared to be a FUra-related dose dependency to this
Time (min) Figure 4. HPLC profiles of dansylated (a) d(pApAZT), (b) nuclease P1 digest of d(pApAZT), and (c) cochromatography of d(pApAZT) and its nuclease P1 digest eluted with a 30 min linear gradient of 18-30% acetonitrile in 0.1 M ammonium acetate buffer, pH 7. A Z T Monoplimphate
E
(a1 AZT Expored DNA Digest
E
Authentic AZT Monophosphate
ON ln
t
W
I
0
_.
.o X
w
Is,
N
cz
( I d AZT Exposed DNA Digest
v)
i!! z
I
I
.3
n
W
u z w
c
&I
xw
a 0
(c) Coritrol DNA Digest
3
U
w
L t-
4
W
a TIME (inin)
Figure 5. HPLC profiles of postlabeled DNA digests of HCT-8 cells (c) control, (b) AZT exposed, and (a) AZT exposed cochromatographed with dansylated AZTMP, eluted with 25% acetonitrile in 0.1 M ammonium acetate buffer, pH 7.
effect, perhaps reflecting more thymidylate synthatase inhibition at higher FUra concentration (1).
Sharma et al.
540 Bioconjugate Chem., Vol. 6,No. 5, 1995 Table 2. Effect of Various Concentrations of AZT Alone and AZT plus FUra on AZT Incorporation into DNA of Exposed HCT-8Cells
drug and concn
AZTMP (fmoyun DNA)
none AZT, 20 pM AZT, 50 ,uM AZT, 100 pM AZT, 20 pM + FUra, 3 pM AZT, 50 pM + FUra, 3 pM AZT, 100 ,uM FUra, 3 pM
109 f 5 245 f 7 479 f 11 114 & 5 476 i 9 980 i 10
+
An AZT concentration-dependent relationship between AZT incorporation into cellular DNA and cytotoxicity has been reported by quantitating the tritium content of PCA insoluble material obtained from HCT-8 cells after exposure to (3H)AZT( I 1. Fluorescence postlabeling assay of DNA in this model also shows a similar trend (Table 2) although the absolute amount of AZT incorporated into DNA as determined using radiolabeled AZT is 1order of magnitude higher than those observed by nonisotopic detection. The labeling efficiency of AZTMP, as monitored by HPLC analysis of each reaction step shown in Figure 2, was 90-95%. It is unlikely that the procedural loss can account for the observed result reported by radiolabeling study ( I ) . This difference may reflect, therefore, several differences between these methods. In the fluorescence postlabeling assay, prior to labeling, the nucleotide profile of DNA digest is routinely assessed by HPLC analysis to check the purity of DNA preparation. No such information was available for PCA insoluble material used to quantitate the radiolabeled AZT. In addition, it is difficult to make PCA insoluble material completely free of unwanted radiolabel, especially when material of relatively high specific activity is used. AZT is presently used extensively in the treatment of AIDS and ARC, and significant research is directed toward identifying the mechanism(s) responsible for both its therapeutic activity and toxicity. In addition, several groups have reported that AZT can possess antineoplastic activity in combination with drugs which disrupt DNA synthesis and chemical evaluation of this potential is underway. Our present findings utilizing the fluorescence postlabeling technique to quantitate AZT incorporation into DNA suggest the important potential of this technique in studies to clinically monitor the fate of AZT without using radiolabeled drug. Clearly, further in vitro studies are essential in order to evaluate the therapeutic
relevance of the two-drug regimen in cancer chemotherapy under clinical settings. ACKNOWLEDGMENT
Supported in part by National Cancer Institute grants CA46896 and CA55358. LITERATURE CITED (1) Brunetti, I., Falcone, A., Calabresi, P., Goulette, F. A., and
Darnowski, J. W. (1990) 5-Fluorouracil enhances azidothymidine cytotoxicity: In vitro, in vivo, and biochemical studies. Cancer Res. 50, 4026-4031. (2) Tosi, P., Calabresi, P., Goulette, F. A., Renoud, C. A., and Darnowski, J. W. (1992)Azidothymidine-induced cytotoxicity and incorporation into DNA in the human colon tumor cell line HCT-8 is enhanced by methotrexate in vitro and in vivo. Cancer Res. 52, 4069-4073. (3) Darnowski, J. W., and Goulette, F. A. (1994) 3’-Azido-3’deoxythymidine cytotoxicity and metabolism in the human colon tumor cell line HCT-8. Mol. Pharmacol. 48, 1797-1805. (4) Sommadossi, J. P., Carlisle, R., and Zhou, Z. (1989) Cellular pharmacology of 3’-azido-3’deoxythymidinewith evidence of incorporation into DNA of human bone marrow cells. Mol. Pharmacol. 36, 9-14. (5) Tompkins, W. A. F., Watrach, A. M., Schmale, J. D., Schultz, R. M., and Harris, J. A. (1974) Cultural and antigenic properties of newly established cell strains derived from adenocarcinoma of the human colon and rectum. J . Natl. Cancer Inst. 52, 1101-1110. (6) Kelman, D. J., Lilga, K. L., and Sharma, M. (1988) Synthesis and application of fluorescent labeled nucleotides to assay DNA damage. Chem.-Biol. Interact. 66, 85- 100. (7) Sharma, M., and Box, H. C. (1985) Synthesis, modification with N-acetoxy-2-acetylaminofluoreneand physicochemical studies of DNA model compound d(TACGTA). Chem.-Biol. Interact. 56, 73-88. ( 8 ) Sharma, M., Jain, R., and h a c , T. V. (1991) A novel technique to assay adducts of DNA induced by anticancer agent cis-diamminedichloroplatinum (11).Bioconjugate Chem. 2 , 403-406. (9) Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutler, W. J. (1979) Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18, 5294-5299. (10) Randerath, K., Randerath, E., Danna, T. F., Van Golen, K. L., and Putnam, K. L. (1989) A new sensitive 32Ppostlabeling assay based on the specific enzymatic conversion of bulky DNA lesions t o radiolabeled dinucleotides and nucleoside monophosphates. Carcinogenesis 10, 1231- 1239. BC950042Q
