Purpose: Down-regulation of Bcl-2 by the antisense G3139, currently under clinical evaluations, could restore chemosensitivity in otherwise resistant malignant cells. To date, the mechanism of intracellular accumulation of G3139 following in vivo administration remains to be elucidated. This study aimed to assess whether detectable intracellular concentrations of G3139 are achievable in vivo and how these relate to Bcl-2 down-regulation.

Experimental Design: Cellular uptake of G3139 was studied in leukemia myeloid cell lines and blasts collected from treated patients using a newly developed, novel, and highly sensitive ELISA-based assay. Real-time reverse transcription-PCR was used to quantify Bcl-2 mRNA changes in treated cells.

Results: The assay was fully validated and showed a limit of quantification of 50 pmol/L. When exposed to 0.33 to 10 μmol/L G3139, K562 cells exhibited intracellular concentrations in the range of 2.1 to 11.4 pmol/mg protein. When G3139 was delivered with cationic lipids, a 10- to 25-fold increase of the intracellular concentrations was observed. There was an accumulation of G3139 in the nuclei, and the ratio of nucleus to cytoplasm was increased 7-fold by cationic lipids. Intracellular concentrations of G3139 were correlated with Bcl-2 mRNA down-regulation. Robust intracellular concentrations of G3139 were achieved in vivo in bone marrow (range, 3.4-40.6 pmol/mg protein) and peripheral blood mononuclear cells (range, 0.47-19.4 pmol/mg protein) from acute myeloid leukemia patients treated with G3139.

Conclusions: This is the first evidence that measurable intracellular levels of G3139 are achievable in vivo in acute myeloid leukemia patients and that Bcl-2 down-regulation is likely to depend on the achievable intracellular concentrations rather than on plasma concentrations.

The antisense oligodeoxynucleotides are sequences of 16 to 29 bases of ssDNAs that hybridize to specific mRNA by Watson-Crick base pairing (1). Following hybridization, the oligodeoxynucleotide-mRNA duplex becomes a substrate for intracellular RNase H that catalyzes mRNA degradation while allowing the oligodeoxynucleotide to recycle for another base pairing event with the next target mRNA molecule. The net result of this process is a sustained decrease in translation of the target mRNA into the corresponding protein (2, 3).

The use of antisense strategies in cancer is based on the potentials of these oligodeoxynucleotide compounds to down-regulate oncogenic proteins that drive malignant transformation. The phosphorothioate antisense currently available for clinical use contains a sulfur in place of an oxygen in the phosphodiester bond between two subsequent nucleotides and seems to have favorable pharmacologic characteristics, such as nuclease resistance and the ability to recruit and activate RNase H (46), which makes it suitable for in vivo administration (710).

G3139 (Table 1), an 18-mer phosphorothioate antisense oligonucleotide designed to bind to the first six codons of the human Bcl-2 mRNA, is currently being evaluated in several phase I to III clinical trials for both solid tumors and hematologic malignancies (11, 12). As high levels of the antiapoptotic protein Bcl-2 were found to be associated with chemoresistance in malignant cells (1315), it has been hypothesized that by down-regulating Bcl-2 the antisense decreases the apoptosis threshold, thereby restore chemotherapeutic sensitivity in otherwise resistant cells. In our initial OSU 9977 protocol study, G3139 was given as continuous i.v. infusion (CIVI) for 10 days with fludarabine/cytarabine chemotherapy starting on day 6 in patients with refractory or relapsed acute leukemia (16). A response rate of 45% and evidence of target down-regulation in ∼75% of the analyzed patients were observed. Based on these results, we have now incorporated G3139 in the upfront therapy for untreated elderly acute myeloid leukemia (AML; refs. 17, 18).

Table 1.

Oligonucleotide sequence of putative metabolites of G3139 and control oligonucleotides used in the cross-reactivity studies

NameOligonucleotide sequence (5′-3′)Cross-reactivity (%)
  Cell lysate Human plasma 
G3139 TCTCCCAGCGTGCGCCAT 100 100 
3′ N-1 TCTCCCAGCGTGCGCCA 6.1 6.3 
3′ N-2 TCTCCCAGCGTGCGCC 3.0 3.4 
3′ N-3 TCTCCCAGCGTGCGC <0.01 <0.04 
5′ N-2 TCCCAGCGTGCGCCAT 32 41 
Reverse control TACCGCGTGCGACCCTCT 
Mismatch control (G4126) TCTCCCAGCATGTGCCAT 25 26 
NameOligonucleotide sequence (5′-3′)Cross-reactivity (%)
  Cell lysate Human plasma 
G3139 TCTCCCAGCGTGCGCCAT 100 100 
3′ N-1 TCTCCCAGCGTGCGCCA 6.1 6.3 
3′ N-2 TCTCCCAGCGTGCGCC 3.0 3.4 
3′ N-3 TCTCCCAGCGTGCGC <0.01 <0.04 
5′ N-2 TCCCAGCGTGCGCCAT 32 41 
Reverse control TACCGCGTGCGACCCTCT 
Mismatch control (G4126) TCTCCCAGCATGTGCCAT 25 26 

NOTE: All control oligonucleotides are fully phosphorothioate oligonucleotides.

Despite these encouraging clinical results, a recurrent motif in the G3139 trials has been the lack of correlations among plasma drug levels, Bcl-2 down-regulation, and disease response. This problem stems in part from the low sensitivity of the previously reported analytic methods, the uncertainty of whether significant levels of G3139 uptake occur in malignant cells, and how the intracellular distribution of the drug relates to down-regulation of its target. Thus, to address these questions, we developed a novel, highly sensitive, and specific ELISA-based assay for quantification of G3139 in various biological matrices. This assay allowed us to follow plasma drug decay over a much longer period of time than previously reported methods, achieve a more accurate definition of the antisense pharmacokinetics, distinguish the parent compound from its chain-shortened metabolites, and, more importantly, quantify the intracellular drug concentrations in treated patients. Herein, for the first time, we reported that robust levels of G3139 were achievable in blood and bone marrow mononuclear cells from patients with AML treated with the Bcl-2 antisense, and these might determine levels of Bcl-2 down-regulation.

Antisense and reagents. G3139 was supplied by the National Cancer Institute (Bethesda, MD). The putative metabolites shorter of 1, 2, or 3 nucleotides (N-1, N-2, or N-3) were obtained as follows: 3′ N-1 of G3139 was a gift from Dr. William Tong (Memorial Kettering Cancer Center, New York, NY); 3′ N-2, 3′ N-3, 5′ N-2, mismatch control, reverse control (Table 1), and 5′-fluorescein-labeled G3139 (FITC-G3139) were purchased from Integrated DNA Technologies (Coralville, IA). The purity (>95%) and identity of each oligomer was examined by elution sequence of capillary gel electrophoresis and by high-performance liquid chromatography (HPLC)-UV-mass spectrometry (model LCQ, Finnigan Corp., San Jose, CA).

Fluorogenic ELISA assay procedures. The assay principle is illustrated in Fig. 1. Briefly, the capture oligodeoxynucleotide 5′-GAATAGCGAATGGCGCACGCTGGGAGA/biotin-3′ (Integrated DNA Technologies) was first diluted in assay buffer [60 mmol/L phosphate buffer (pH 7.4), 1.0 mol/L NaCl, 5 mmol/L EDTA, 0.3% Tween 20] at a concentration of 200 nmol/L, heated at 95°C for 5 minutes, and mixed with plasma or cell lysates containing G3139. Triton X-100 [final concentration, 0.25% (w/v)] was added into the plasma sample to disrupt nonspecific interaction. The mixture was then incubated at 42°C for 2 hours. The analyte complex was captured by binding to a NeutrAvidin-coated 96-well plate (Pierce Co., Rockford, IL), which was subsequently washed with warm washing buffer (TBS in 0.1% Tween 20) at 30°C. The oligodeoxynucleotide probe (5′-TCGCTATTC-3′ phosphorylated at the 5′ end and digoxigenin modified at the 3′ end, Integrated DNA Technologies) was diluted with ligation buffer [66 mmol/L Tris-HCl (pH 7.6), 10 mmol/L MgCl2, 10 mmol/L DTT, 1 mmol/L ATP] containing 5 units/mL T4 DNA ligase (Amersham Biosciences, Piscataway, NJ). The mixture (150 μL) was dispensed into each well of a 96-well plate and incubated at 18°C for overnight. To remove the excess amount of probe oligodeoxynucleotide bound to capture oligodeoxynucleotide, 30 units of S1 nuclease (Invitrogen, Carlsbad, CA) in 30 mmol/L sodium acetate (pH 4.6), 1 mmol/L zinc acetate, 150 mmol/L NaCl, and 5% glycerol were added into each well for 60 minutes at room temperature, and the plate was then washed five times with washing buffer. Subsequently, anti-digoxigenin-alkaline phosphatase (150 μL) diluted 1:2,500 with bovine serum albumin block buffer in TBS (Roche, Indianapolis, IN) was added into each well. Following 0.5-hour incubation at 37°C, the plate was again washed with washing buffer. Attophos substrate (150 μL, Promega, Madison, WI) in diethanolamine buffer prepared as recommended by the manufacturer was added to each well. Fluorescence intensity was measured at excitation 430/emission 570 (filter = 550 nm) using a Gemini XS plate reader (Molecular Devices, Sunnyvale, CA) following incubation at 25°C for 30 minutes.

Fig. 1.

Fluorogenic ELISA assay to measure G3139 concentrations in plasma and cell lysate. The assay consists of the following steps: (1) hybridization binding of the analyte to a 3′-biotinylated capture oligodeoxynucleotide (ODN) to form a double-strand complex with a 5′ overhang; (2) attachment of a duplex containing the analyte to the 96-well plate; (3) digoxigenin labeling of the duplex containing the analyte by introducing a secondary probe oligodeoxynucleotide; (4) removal of the excess amount of probe oligodeoxynucleotide bound to template oligodeoxynucleotide by ssDNA endonuclease (S1 nuclease); (5) binding of digoxigenin-labeled product with anti-digoxigenin-alkaline phosphatase conjugate, and (6) fluorometric determination by Attophos substrate. The intensity of the fluorescence signal is directly proportional to the concentration of the G3139 in the clinical samples.

Fig. 1.

Fluorogenic ELISA assay to measure G3139 concentrations in plasma and cell lysate. The assay consists of the following steps: (1) hybridization binding of the analyte to a 3′-biotinylated capture oligodeoxynucleotide (ODN) to form a double-strand complex with a 5′ overhang; (2) attachment of a duplex containing the analyte to the 96-well plate; (3) digoxigenin labeling of the duplex containing the analyte by introducing a secondary probe oligodeoxynucleotide; (4) removal of the excess amount of probe oligodeoxynucleotide bound to template oligodeoxynucleotide by ssDNA endonuclease (S1 nuclease); (5) binding of digoxigenin-labeled product with anti-digoxigenin-alkaline phosphatase conjugate, and (6) fluorometric determination by Attophos substrate. The intensity of the fluorescence signal is directly proportional to the concentration of the G3139 in the clinical samples.

Close modal

Validation studies. Linearity, within-day and between-day assay accuracies, and precision of the method were assessed from blank human plasma concentrations of 50 pmol/L (limit of quantification), 100 pmol/L (low-quality control), 500 pmol/L (medium-quality control), and 2,000 pmol/L (high-quality control). Between-day and within-day validation in cell lysate was also done. The specificity of the assay was evaluated using human plasma from three healthy donors (Red Cross, Columbus, OH) as well as from K562 cell extracts to assess possible interference from endogenous substances and by cross-reactivity studies with putative metabolites (see below).

Cross-reactivity analysis. To evaluate the cross-reactivity of the assay with putative metabolites, various concentrations of 3′ N-1, N-2, and N-3 oligomers from 50 pmol/L to 1,000 nmol/L were added into human blank plasma, and concentration-response curves were constructed. Additionally, different concentrations of 5′ N-2, reverse control, and mismatch control oligomers (Table 1), ranging from 50 pmol/L to 100 nmol/L in cell lysate, were also evaluated for cross-reactivity. All concentration-response curves were fitted to the sigmoid Emax model. The maximal response produced by each compound (Emax) and the concentration that produced 50% of the maximal response (EC50) were obtained by nonlinear regression analysis using WinNonLin version 3.1 (Pharsight Corp., Mountain View, CA). The cross-reactivity was calculated as EC50 of parent compound divided by EC50 of each metabolite or analogue (19).

Cross-validation with the high-performance liquid chromatography-UV method. Samples (n = 45) from AML patients treated on a previous clinical phase I study (16) were analyzed by both the ELISA assay and the established HPLC-UV method.

Cellular uptake of G3139 in cell line and in samples from patients. A K562 cell line was cultured in RPMI 1640 supplemented with l-glutamine (Life Technologies, Carlsbad, CA) and 10% heat-inactivated fetal bovine serum (Life Technologies). About 2 × 106 of cyropreserved mononuclear cells from patient bone marrow and blood samples collected before and after initiation of G3139 on the OSU 9977 protocol were used. Following centrifugation, the cell pellet was incubated with 200 μL of 0.1 μmol/L phosphorothioate 28-mer polycytidine for 2 minutes on ice and washed with PBS to remove membrane-bound oligodeoxynucleotides (20). Following addition of 200 μL lysis buffer [10 mmol/L Tris-HCl (pH 8.0), 0.5 mmol/L EDTA, 1% Triton X-100] and incubation on ice for 10 minutes, the cells were lysed by vortexing and sonication. The homogenate was then centrifuged at 10,000 × g, and the supernatant was transferred to a new tube for the ELISA and protein assays (Bio-Rad protein assay kit, Bio-Rad, Hercules, CA). The intracellular levels of G3139 were measured using the ELISA assay as described above. IC50 of G3139 was expressed as mean ± SD molar concentration using a measured conversion factor of 1 × 106 cells equaling to 1 ± 0.01 μL of cell volume and 70 ± 15 μg protein. Similar procedures were also used to measure G3139 levels in blood and bone marrow mononuclear cells collected from patients with AML. A measured conversion factor of 1 × 106 bone marrow mononuclear cell equaling to 0.6 μL of cell volume and 14.6 μg protein was used for calculation.

In a separate experiment, to assess the intracellular drug distribution between nucleus and cytoplasm, K562 cells were treated with 1 and 3.3 μmol/L G3139 delivered either as free drug or complexed with Oligofectamine. The subcellular fraction was done using the Nuclear Extract kit (Active Motif, Carlsbad, CA) according to the manufacturer's instructions. Briefly, the cells were washed twice with PBS on ice and treated with phosphorothioate 28-mer polycytidine as described above to remove membrane-bound oligonucleotides. The cells were then washed once with hypotonic buffer, resuspended in 500 μL of same, and lysed on ice over a 15-minute period. Following centrifugation for 1 minute at 14,000 × g at 4°C, the supernatant was harvested as cytoplasmic fraction. The resulting pellet was washed once with 500 μL of hypotonic buffer and resuspended in 50 μL of lysis buffer. The suspension was vigorously shaken on a rocking platform at 4°C for 30 minutes to disrupt the nuclear membrane. Following centrifugation at 4°C at 14,000 × g for 10 minutes, the supernatant was saved as the nuclear fraction. Cytosolic lactate dehydrogenase was used as a cytosolic marker, and cross-contamination between cytoplasmic and nuclear fractions was determined using lactate dehydrogenase kit (Roche).

Plasma pharmacokinetics in patients with acute myeloid leukemia. Plasma pharmacokinetics of G3139 in patients whose cellular levels of G3139 in blood and bone marrow mononuclear cells were measured was also studied. Eight patients, five on CIVI at 4 mg/kg and two at 7 mg/kg with available material, were studied during the 10-day infusion and 4 hours following infusion using the newly developed ELISA method. Relevant pharmacokinetic variables were computed following curve fitting to an appropriate model via WinNonLin computer software version 3.1.

Cellular uptakes of G3139 complexed with cationic liposomes. All transfection experiments were done in Opti-MEM medium (Invitrogen). Stock solutions of Oligofectamine reagent (Invitrogen) and G3139 were prepared using Opti-MEM as the diluent. The appropriate amount of G3139 was diluted in 100 μL Opti-MEM to result in final concentrations of G3139 of 100 nmol/L, 200 nmol/L, 330 nmol/L, 1.0 μmol/L, 3.3 μmol/L, and 10 μmol/L. For 100, 200, and 330 nmol/L G3139, 1.8, 3.6, and 6 μL of Oligofectamine reagent were used. For 1.0, 3.3, and 10 μmol/L G3139, 20 μL of Oligofectamine were used. To avoid cytotoxicity due to Oligofectamine, its final concentration was kept below 20 μL/mL in all transfection experiments. These solutions were incubated at room temperature for 10 to 20 minutes to allow lipid-oligodeoxynucleotide complex to form. Then, 200 μL of each of these complex solutions were overlaid on the cells seeded at a density of 2 × 106 cells per well in 0.8 mL medium on six-well plates for 4 to 5 hours. Another cationic liposome consisting of dimethyldioctadecylammonium bromide (DDAB) and l-α-dioleyl phosphatidylethanolamine (a generous gift by Dr. Robert Lee, The Ohio State University, Columbus, OH) was also used and was prepared as reported previously (21). The mean ± SD particle size of the cationic liposomes was determined to be 94 ± 48 nm. Because each DDAB molecule carries 1 positive charge, whereas one G3139 molecule possesses 17 negative charges, a preliminary cellular uptake study was first carried out to optimize the charge ratio of cationic lipid to G3139 to achieve the highest uptake value. The optimal ratio was found to be 1.43 on K562 cells when 0.33 μmol/L G3139 was used with various amounts of cationic liposomes. Therefore, 8, 24, and 60 μmol/L DDAB/l-α-dioleyl phosphatidylethanolamine were mixed with 0.33, 1, and 3.3 μmol/L G3139, and the DDAB/G3139 complexes were prepared in the similar fashion as for Oligofectamine. Following 4- to 5-hour incubation with Oligofectamine or DDAB, 3 mL of medium containing 10% fetal bovine serum were added to each well and the content was gently mixed. Then, the mixture was incubated for another 20 hours before cell lysis for total RNA isolation and G3139 quantification.

Quantification of Bcl-2 mRNA levels. Quantification of Bcl-2 RNA was done by real-time reverse transcription-PCR as reported previously (22). Briefly, total cellular RNA and cDNAs were prepared as described previously (22). Each cDNA sample was used as a template in a PCR amplification reaction on the ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). The measured Bcl-2 levels were normalized to the internal control of 18S.

Flow cytometry studies. K562 cells (0.5 × 106) were exposed to 0.3 or 0.5 μmol/L FITC-G3139 in the presence or absence of delivery vehicle at 37°C for 24 hours. Then, the cells were harvested, washed thrice by cold PBS/1% fetal bovine serum, and analyzed by flow cytometry on a FACSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ). Data were analyzed and displayed with CellQuest software (Becton Dickinson) as histograms.

Data analyses. Means and SDs were computed for all variables using standard methods. All graphs are plotted as mean ± SD. Two-sided multiple comparison method was done to compare group mean difference with family-wise error at 0.01 (23). Differences were considered statistically significant when P < 0.05. For cross-validation with the HPLC-UV method, Pearson correlation was obtained using S-Plus software (24). Equivalence test was done with the Wilcoxon signed-rank test (25).

Validation of the ELISA-based assay of G3139 in plasma and cell extracts. In human plasma, we determined the limit of detection (defined as 10 times signal-to-background noise; ref. 14) to be 25 pmol/L and the limit of quantification to be 50 pmol/L, equivalent to 0.15 and 0.3 ng/mL, respectively. The assay was linear from 25 to 2,500 pmol/L after log transformation (Fig. 2A). The linear concentration range of the assay could be extended to 500 nmol/L (a 4-log magnitude of dynamic range), as serial dilution from 500 nmol/L yielded a nearly identical calibration curve (data not shown). The mean within-day coefficients of variation (CV) of the assay at 50, 100, 500, and 2,000 pmol/L were found to be 13%, 6%, 6%, and 3% (all n = 6), respectively (Table 2). The corresponding accuracy values were 73%, 103%, 107%, and 82% of the nominal concentrations of the standards. The between-day CVs of the assay were found to be 10%, 5%, and 10% for the 50, 500, and 2,000 pmol/L standards, respectively, with the corresponding accuracy values of 94.2%, 112%, and 94.7% of the nominal concentrations.

Fig. 2.

Validation of the ELISA-based assay of G3139. A, representative standard curve of G3139 in human plasma. B, representative standard curve of G3139 in cell lysate. Each concentration was run in duplicates, and average was used for linear regression analysis. The mean fluorescence signal was plotted against G3139 concentrations (pmol/L). C, concentration-response curves of G3139 and its possible 3′ metabolites. G3139 (•), 3′ N-1 (▿), 3′ N-2 (▪), and 3′N-3 (◊) oligomers at the 3′ end ranging from 0.05 to 1,000 nmol/L were added into human blank plasma, and dose-response curves were constructed as described in Materials and Methods. Points, average of three replicates; bars, SD. The maximal response (fluorescence intensity) produced by each compound (Emax) and the concentration that produced 50% of the maximal response (EC50) was determined by nonlinear regression analysis.

Fig. 2.

Validation of the ELISA-based assay of G3139. A, representative standard curve of G3139 in human plasma. B, representative standard curve of G3139 in cell lysate. Each concentration was run in duplicates, and average was used for linear regression analysis. The mean fluorescence signal was plotted against G3139 concentrations (pmol/L). C, concentration-response curves of G3139 and its possible 3′ metabolites. G3139 (•), 3′ N-1 (▿), 3′ N-2 (▪), and 3′N-3 (◊) oligomers at the 3′ end ranging from 0.05 to 1,000 nmol/L were added into human blank plasma, and dose-response curves were constructed as described in Materials and Methods. Points, average of three replicates; bars, SD. The maximal response (fluorescence intensity) produced by each compound (Emax) and the concentration that produced 50% of the maximal response (EC50) was determined by nonlinear regression analysis.

Close modal
Table 2.

Within-run and between-run precision and accuracy of the ELISA assay of G3139 in human plasma and K562 cell lysate

Nominal concentration (pmol/L)Measured mean (SD), pmol/L *Precision (CV %), n = 6Accuracy (% nominal)Measured mean (SD), pmol/L*Precision (CV %), n = 6Accuracy (% nominal)
 Within-run precision and accuracy in plasma (n = 6)
 
  Between-run precision and accuracy in plasma (n = 5 d)
 
  
50 36.7 (4.8) 13 73 47.1 (4.5) 10 94.2 
100 103.1 (6.1) 103 N/A   
500 532.8 (31.1) 107 562 (29.5) 112.0 
2,000 1,646 (44.4) 82 1,894 (196) 10 94.7 
 Within-run validation in cell lysate (n = 6)
 
  Between-run validation in cell lysate (n = 5 d)
 
  
50 54.6 (4.1) 109 48.4 (6.0) 12 96.7 
100 90.0 (4.2) 90 N/A   
500 515.8 (37.0) 103 533 (30) 106.8 
2,000 1,865 (52.4) 93 2,036 (150) 101.8 
Nominal concentration (pmol/L)Measured mean (SD), pmol/L *Precision (CV %), n = 6Accuracy (% nominal)Measured mean (SD), pmol/L*Precision (CV %), n = 6Accuracy (% nominal)
 Within-run precision and accuracy in plasma (n = 6)
 
  Between-run precision and accuracy in plasma (n = 5 d)
 
  
50 36.7 (4.8) 13 73 47.1 (4.5) 10 94.2 
100 103.1 (6.1) 103 N/A   
500 532.8 (31.1) 107 562 (29.5) 112.0 
2,000 1,646 (44.4) 82 1,894 (196) 10 94.7 
 Within-run validation in cell lysate (n = 6)
 
  Between-run validation in cell lysate (n = 5 d)
 
  
50 54.6 (4.1) 109 48.4 (6.0) 12 96.7 
100 90.0 (4.2) 90 N/A   
500 515.8 (37.0) 103 533 (30) 106.8 
2,000 1,865 (52.4) 93 2,036 (150) 101.8 
*

Concentrations calculated from the linear least-squares regression curve (n = 6).

Expressed as [(mean observed concentration / nominal concentration) ×100].

Similar validation studies were also conducted in cell lysate added to known concentrations of G3139. A limit of detection of 25 pmol/L and a limit of quantification of 50 pmol/L equivalent to 2.5 and 5.0 fmol/100 μL cell lysate, respectively, were obtained. A linear calibration curve of the fluorescence signal versus concentration within the range of 25 to 4,000 pmol/L was obtained (Fig. 2B). The mean within-day precision CVs of the assay in cell lysate at 50, 100, 500, and 2,000 pmol/L were 7%, 5%, 7%, and 3%, with corresponding accuracy values of 109%, 90%, 103%, and 93% (Table 2). The between-day CVs of the assay were found to be 12%, 6%, and 7% for the 50, 500, and 2,000 pmol/L samples, respectively, with the corresponding accuracy values of 96.7%, 106.8%, and 101.8%.

Insignificant background was found in human plasma and cell lysate. The specificity of the assay was assessed by measuring the fluorescence signal generated by four putative G3139 metabolites and two control oligodeoxynucleotides. The concentration-response curves of G3139 and the 3′-end metabolites in human plasma from 50 to 1,000 pmol/L are shown in Fig. 2C. The cross-reactivities, calculated as the ratio of EC50 of the parent compound to EC50 of each metabolite, are shown in Table 1. The cross-reactivities of 3′ N-1 and N-2 were estimated to be 6.3% and 3.4%, respectively. The 3′ N-1 metabolite at 5 nmol/L gave a fluorescence signal <10% of that of G3139 at the same concentration. The 3′ N-3 metabolite gave extremely low fluorescence (cross-reactivity <0.04%). In contrast, the cross-reactivity of the 5′ N-2 was similarly evaluated (figure not shown), and the value was estimated to be 41% (Table 1). There was essentially no fluorescence signal when reverse control oligodeoxynucleotide was used (Table 1), but mismatch control oligodeoxynucleotide (figure not shown) gave a cross-reactivity of 26% (Table 1). Similar results were found for G3139 in cell lysate (Table 1).

Comparison of the ELISA-based assay with high-performance liquid chromatography-UV method. Cross-validation of our novel ELISA-based assay with the published HPLC-UV method was done using 45 plasma samples obtained from acute leukemia patients treated on our clinical study (OSU 9977). Because of the low sensitivity of the HPLC-UV method (88 nmol/L or 500 ng/mL), only plasma samples with sufficiently high drug levels were selected. The correlation between the results attained by the two methods was linear with a Pearson correlation coefficient of 0.968 (P < 0.001; Fig. 3). The Wilcoxon signed-rank test also showed that the two methods were equivalent at the 98% confidence level. Importantly, the advantage of the new assay was its detection sensitivity at least 3 orders of magnitude higher than the HPLC-UV method (50 versus 88,000 pmol/L).

Fig. 3.

Correlation curve of G3139 plasma concentrations measured by HPLC-UV and ELISA methods. Y axis, ELISA-based assay; X axis, HPLC-UV assay.

Fig. 3.

Correlation curve of G3139 plasma concentrations measured by HPLC-UV and ELISA methods. Y axis, ELISA-based assay; X axis, HPLC-UV assay.

Close modal

Cellular uptake of noncomplexed and cationic lipid complexed G3139. Intracellular drug levels using our ELISA assay were quantified in K562 cell lysate following incubation with G3139 alone or in the presence of cationic lipid vehicle. Exposure to 0.33 to 10 μmol/L G3139 without delivery vehicle for 24 hours resulted in a concentration-dependent intracellular drug concentration in the range of 2.1 to 11.4 pmol/mg protein (Table 3). The uptake in K562 cells was estimated to be ∼0.2% to 0.6% of the exposed drug. This is the first chemical measurement showing intracellular levels of G3139 when cells were exposed to the free drug. In contrast, exposure to G3139 complexed with cationic lipids, such as Oligofectamine and DDAB/l-α-dioleyl phosphatidylethanolamine, resulted in a 25- and 50-fold increase, respectively, in antisense cellular uptake (Table 3). Notably, at a concentration as high as 10 μmol/L noncomplexed G3139, the cellular uptake of G3139 was even significantly lower than that measured when the cells were exposed to lower concentrations (i.e., 0.33, 1, and 3.3 μmol/L) of the antisense complexed with DDAB or Oligofectamine (P < 0.01). The amount of lipids used was not considered to impart significant toxicity, as IC50 values for Oligofectamine were determined to be 30 μL/mL of the original lipid and 60 μmol/L for DDAB, all using 48-hour exposure.

Table 3.

Comparison of intracellular uptake of G3139 antisense in the absence or presence of liposomal vehicle in K562 leukemic cells

Concentration of G3139 (μmol/L)Free drug, pmol/mg (μmol/L)Complex with Oligofectamine, pmol/mg (μmol/L)Complex with DDAB, pmol/mg (μmol/L)
0.1  7.23 ± 1.35 (0.51± 0.094)  
0.2  15.51± 3.97 (1.09 ± 0.28)  
0.33 2.08 ± 0.47 (0.146 ± 0.33) 50.17 ± 5.34 (3.51 ± 0.37)* 101.67 ± 14.52 (7.12 ±1.01)* 
3.46 ± 0.22 (0.242 ± 0.015) 75.54 ± 6.94 (5.29 ± 0.48)* 143.67 ± 24.03 (10 ± 1.68)* 
3.3 9.16 ± 0.59 (0.64± 0.041) 131.49 ± 25.34 (9.2 ± 1.77)* 161.0 ± 29.82 (11.27± 2.09)* 
8.84 ± 1.28 (0.62 ± 0.09)   
10 11.37 ± 1.97 (0.796 ± 0.138) 261.16 ± 30.11 (18.28 ± 2.10)*  
Concentration of G3139 (μmol/L)Free drug, pmol/mg (μmol/L)Complex with Oligofectamine, pmol/mg (μmol/L)Complex with DDAB, pmol/mg (μmol/L)
0.1  7.23 ± 1.35 (0.51± 0.094)  
0.2  15.51± 3.97 (1.09 ± 0.28)  
0.33 2.08 ± 0.47 (0.146 ± 0.33) 50.17 ± 5.34 (3.51 ± 0.37)* 101.67 ± 14.52 (7.12 ±1.01)* 
3.46 ± 0.22 (0.242 ± 0.015) 75.54 ± 6.94 (5.29 ± 0.48)* 143.67 ± 24.03 (10 ± 1.68)* 
3.3 9.16 ± 0.59 (0.64± 0.041) 131.49 ± 25.34 (9.2 ± 1.77)* 161.0 ± 29.82 (11.27± 2.09)* 
8.84 ± 1.28 (0.62 ± 0.09)   
10 11.37 ± 1.97 (0.796 ± 0.138) 261.16 ± 30.11 (18.28 ± 2.10)*  

NOTE: Intracellular concentration is expressed as pmol/mg protein and molar concentration (in parentheses). A measured conversion factor of 1 × 106 cells equal to 1 ± 0.1 μL of cell volume and 70 ± 15 μg protein was used for calculation. Mean ± SD (n = 3 per group).

*

P < 0.01, significant difference between the groups in comparison with the free drug group.

To further validate our results and confirm that we indeed measured levels of the internalized G3139, we used a fluorescein-labeled G3139, free or complexed with Oligofectamine, to incubate with K562 cells. By flow cytometric analysis, we showed that the mean fluorescence in cells treated with G3139 and cationic lipids was ∼5-fold greater than that in cells treated with G3139 alone, supporting the results obtained with our ELISA-based assay (Fig. 4A). Confocal microscopy indicated that the membrane-bound antisense was negligible compared with the internalized FITC-G3139 (data not shown). As flow cytometry and intracellular concentration determination do not provide information about intracellular distribution, we also examined differential drug distribution of different G3139 formulations in K562 cells by subcellular fractionation. Cell uptake for noncomplexed G3139 was found to be quite low (Fig. 4B), and ∼60% to 80% of the internalized full-length G3139 were found in the cytoplasmic fraction with a nucleus to cytoplasm drug ratio of 0.33 ± 0.053. In contrast, G3139 complexed with Oligofectamine resulted not only in higher intracellular levels but also in a 7-fold higher nucleus to cytoplasm drug ratio (i.e., 0.33 ± 0.053 versus 2.5 ± 0.017; Fig. 4B). To exclude the possibility of cytoplasm to nucleus contamination, the lactate dehydrogenase content in cytoplasm, nuclei wash fraction, and nuclear fraction were measured, and lactate dehydrogenase in nuclei fraction and nuclei wash fraction was found to be <10% of that in cytoplasmic fraction, suggesting minimal cytoplasm to nucleus contamination.

Fig. 4.

Intracellular distribution of G3139 in K562 cells after incubation with FITC-labeled G3139 in the absence or presence of cationic lipids. A, flow cytometric analysis: histograms were drawn to compare difference in fluorescence intensity between treatments. Y axis, cell counts; X axis, logarithm of fluorescence intensity (FL-1 height). Control, autofluorescence intensity of untreated K562 cells; a, 0.5 μmol/L FITC-G3139 without lipid; b, 0.33 μmol/L G3139 complexed with cationic lipid (6 μL Oligofectamine); c, 0.5 μmol/L G3139 complexed with cationic lipid (6 μL Oligofectamine). B, comparison of G3139 levels in subcellular fractions between free G3139 and G3139 complexed with Oligofectamine at 37°C after 24-hour incubation in K562 cells. Subcellular levels of G3139 are expressed as pmol/1 × 106 cells. White columns, total cell lysate; diagonally hatched columns, cytoplasmic fraction; cross-hatched columns, nuclear fraction. Columns, mean (n = 3 per group); bars, SD.

Fig. 4.

Intracellular distribution of G3139 in K562 cells after incubation with FITC-labeled G3139 in the absence or presence of cationic lipids. A, flow cytometric analysis: histograms were drawn to compare difference in fluorescence intensity between treatments. Y axis, cell counts; X axis, logarithm of fluorescence intensity (FL-1 height). Control, autofluorescence intensity of untreated K562 cells; a, 0.5 μmol/L FITC-G3139 without lipid; b, 0.33 μmol/L G3139 complexed with cationic lipid (6 μL Oligofectamine); c, 0.5 μmol/L G3139 complexed with cationic lipid (6 μL Oligofectamine). B, comparison of G3139 levels in subcellular fractions between free G3139 and G3139 complexed with Oligofectamine at 37°C after 24-hour incubation in K562 cells. Subcellular levels of G3139 are expressed as pmol/1 × 106 cells. White columns, total cell lysate; diagonally hatched columns, cytoplasmic fraction; cross-hatched columns, nuclear fraction. Columns, mean (n = 3 per group); bars, SD.

Close modal

Correlation of G3139 intracellular levels with Bcl-2 down-regulation in acute myeloid leukemia cells in vitro. To determine whether intracellular concentrations of G3139 correlate with down-regulation of its target Bcl-2, K562 cells were again exposed to various concentrations of G3139 alone or to G3139-Oligofectamine complex for 24 hours. As shown in Fig. 5A, at G3139 concentrations between 0.1 and 10 μmol/L in the presence of cationic lipids, down-regulation of Bcl-2 mRNA as measured by real-time reverse transcription-PCR occurred efficiently. Nonlinear regression analysis of the dose-response curve showed that the G3139 concentration that produced 50% of Bcl-2 down-regulation (IC50) was ∼0.29 μmol/L, and maximum Bcl-2 down-regulation (79% decrease relative to the control group) was observed at 10 μmol/L relative to the control group (Fig. 5B). The IC50 of 0.29 μmol/L corresponds to an intracellular G3139 concentration of 37 pmol/mg protein, which was not achievable even at the highest concentration of G3139 (10 μmol/L) when applied alone (Table 3). Further, exposure to 3.3 μmol/L G3139 without lipids failed to result in any significant suppression of Bcl-2 RNA (93% compared with control group). Exposure to 3.3 and 10 μmol/L mismatch control (G4126) and reverse control oligonucleotides complexed with Oligofectamine failed to show any significant Bcl-2 down-regulation, confirming specific target down-regulation (Fig. 5C).

Fig. 5.

Correlation of intracellular G3139 concentration with Bcl-2 down-regulation in K562 cells. A, Bcl-2 mRNA as measured by real-time reverse transcription-PCR (left, Y axis, ▴) and intracellular G3139 levels as quantified by ELISA (right, Y axis, •). Intracellular levels of G3139 were normalized to total cellular protein quantified by Bio-Rad protein assay. B, inhibition of Bcl-2 mRNA by G3139 on K562 cells. Quantification of Bcl-2 mRNA was done by real-time reverse transcription-PCR. Bcl-2 mRNA levels were normalized to 18S mRNA levels and presented as a percentage of lipid-control cells. Bcl-2 mRNA was down-regulated by G3139 in a dose-dependent fashion with IC50 of 0.29 μmol/L. C, Bcl-2 mRNA was down-regulated by G3139 in a dose-dependent fashion, whereas little or no change in Bcl-2 was detected in cells treated with reverse control (RC) oligonucleotides (3.3 and 10 μmol/L) complexed with Oligofectamine, mismatch (MM) antisense (3.3 and 10 μmol/L) complexed with Oligofectamine (20 μL/mL in Opti-MEM), and free drug (3.3 μmol/L).

Fig. 5.

Correlation of intracellular G3139 concentration with Bcl-2 down-regulation in K562 cells. A, Bcl-2 mRNA as measured by real-time reverse transcription-PCR (left, Y axis, ▴) and intracellular G3139 levels as quantified by ELISA (right, Y axis, •). Intracellular levels of G3139 were normalized to total cellular protein quantified by Bio-Rad protein assay. B, inhibition of Bcl-2 mRNA by G3139 on K562 cells. Quantification of Bcl-2 mRNA was done by real-time reverse transcription-PCR. Bcl-2 mRNA levels were normalized to 18S mRNA levels and presented as a percentage of lipid-control cells. Bcl-2 mRNA was down-regulated by G3139 in a dose-dependent fashion with IC50 of 0.29 μmol/L. C, Bcl-2 mRNA was down-regulated by G3139 in a dose-dependent fashion, whereas little or no change in Bcl-2 was detected in cells treated with reverse control (RC) oligonucleotides (3.3 and 10 μmol/L) complexed with Oligofectamine, mismatch (MM) antisense (3.3 and 10 μmol/L) complexed with Oligofectamine (20 μL/mL in Opti-MEM), and free drug (3.3 μmol/L).

Close modal

Intracellular levels of G3139 in acute myeloid leukemia cells in vivo. To probe drug uptake in vivo, we measured levels of G3139 in bone marrow and blood mononuclear cells collected following 72 hours (day 3) and 120 hours (day 5) of G3139 CIVI before FLAG administration in AML patients treated on the protocol OSU 9977 (Tables 4 and 5). Paired bone marrow samples and peripheral mononuclear cells for determination of drug levels and Bcl-2 mRNA were only available from eight patients treated with G3139 (16). G3139 levels ranging from 3.4 to 40.6 pmol/mg protein in bone marrow mononuclear cells and intracellular levels ranging from 0.47 to 19.4 pmol/mg protein in blood mononuclear cells were found. Intracellular drug levels in blood mononuclear cell measured at 120 hours were significantly higher than those measured at 72 hours despite unchanged or even decreased plasma levels, suggesting a slower intracellular clearance of the drug over time. Of note, intracellular levels of G3139 measured following 120 hours of G3139 CIVI were found to be higher in bone marrow mononuclear cells than in blood mononuclear cells, suggesting a site-specific preferential uptake of the drug (Tables 4 and 5).

Table 4.

Plasma concentration and intracellular concentration of G3139 and corresponding Bcl-2 changes in bone marrow cells at 120 hours (day 5) of the G3139 CIVI in acute leukemia patients enrolled on protocol OSU 9977

Unique patient numberG3139 plasma concentration (μmol/L)G3139 intracellular concentration, pmol/mg protein (μmol/L)Bcl-2 mRNA percentage change compared with baseline
10 0.29 40.64 (1.00) ↔ 
17 1.16 11.89 (0.29) ↓ 
16 1.32 11.35 (0.28) ↓ 
18 0.70 9.46 (0.23) ↔ 
12 0.20 5.94 (0.14) ↓ 
14 0.72 5.06 (0.12) ↓ 
0.77 4.18 (0.10) ↑ 
0.42 3.38 (0.08) ↑ 
Unique patient numberG3139 plasma concentration (μmol/L)G3139 intracellular concentration, pmol/mg protein (μmol/L)Bcl-2 mRNA percentage change compared with baseline
10 0.29 40.64 (1.00) ↔ 
17 1.16 11.89 (0.29) ↓ 
16 1.32 11.35 (0.28) ↓ 
18 0.70 9.46 (0.23) ↔ 
12 0.20 5.94 (0.14) ↓ 
14 0.72 5.06 (0.12) ↓ 
0.77 4.18 (0.10) ↑ 
0.42 3.38 (0.08) ↑ 

NOTE: ↓, >20% decrease compared with baseline; ↔, no change; ↑, >20% increase compared with baseline.

Table 5.

G3139 concentrations in plasma and blood mononuclear cells at 72 hours (day 3) and 120 hours (day 5) following G3139 CIVI in acute leukemia patients enrolled on protocol OSU 9977

Unique patient numberG3139 intracellular concentration, pmol/mg protein (μmol/L)G3139 plasma concentration (μmol/L)Time point (h)
18 8.25 (0.20) 1.16 72 
18 19.44 (0.47) 0.62 120 
16 3.0 (0.73) 1.32 72 
16 6.3 (0.15) 1.32 120 
0.47 (0.11) 0.13 72 
8.77 (0.21) 0.14 120 
17 3.18 (0.078) 1.09 72 
17 4.05 (0.10) 1.16 120 
Unique patient numberG3139 intracellular concentration, pmol/mg protein (μmol/L)G3139 plasma concentration (μmol/L)Time point (h)
18 8.25 (0.20) 1.16 72 
18 19.44 (0.47) 0.62 120 
16 3.0 (0.73) 1.32 72 
16 6.3 (0.15) 1.32 120 
0.47 (0.11) 0.13 72 
8.77 (0.21) 0.14 120 
17 3.18 (0.078) 1.09 72 
17 4.05 (0.10) 1.16 120 

Because only few pairs of plasma and viable cell samples were available, a statistical analysis between plasma and intracellular drug concentration could not be made at this time. However, a direct correlation between drug plasma concentrations and cell uptakes did not seem to occur, as higher concentrations of G3139 were achieved in patients in whom lower plasma concentrations of the antisense were measured. Similarly, no linear correlation was found between intracellular levels of G3139 and Bcl-2 down-regulation, although it appeared that decrease of the Bcl-2 mRNA was more likely with a measured intracellular concentration of G3139 >5.0 pmol/mg protein (Table 4).

Plasma pharmacokinetics of G3139 in acute myeloid leukemia patients.Figure 6 shows the profiles of eight patients with AML whose blood and bone marrow intracellular drug levels were monitored. Plasma G3139 levels reached steady-state concentration from 24 hours and declined biexponentially when infusion was stopped. Because the protocol was designed based on the lower assay sensitivity of the previous HPLC-UV method, plasma samples to only 4 hours postinfusion were collected, and the 4-hour levels showed a range of 22.8 to 84 nmol/L, readily measurable by the ELISA method. Based on these data and curve fitting to a two-compartment model, the relevant pharmacokinetic variables were computed and shown in Table 6. Mean steady-state concentration and area under the curve values were proportional to the infusion dose. The total clearance and half-lives are similar to those reported previously (16).

Fig. 6.

Logarithmic plasma concentration versus time profiles during and after infusion of two different doses of G3139 (4 and 7 mg/kg). Unique patient numbers (UPN) 17 and 18 were given 7 mg/kg G3139 CIVI and the rest 4 mg/kg. Symbols represent the measured concentrations.

Fig. 6.

Logarithmic plasma concentration versus time profiles during and after infusion of two different doses of G3139 (4 and 7 mg/kg). Unique patient numbers (UPN) 17 and 18 were given 7 mg/kg G3139 CIVI and the rest 4 mg/kg. Symbols represent the measured concentrations.

Close modal
Table 6.

Relevant pharmacokinetic variables of G3139 in plasma in acute leukemia patients treated on two different doses of G3139 CIVI (4 and 7 mg/kg)

Variables (units)/dose4 mg/kg, mean ± SD or mean (range)7 mg/kg, mean ± difference
AUC0-∞ (μmol/L h) 124.2 ± 32.5 204.1 ± 21.5 
Css (μmol/L) 0.5 ± 0.11 0.88 ± 0.14 
CL (L/h) 4.45 ± 0.84 3.97 ± 0.85 
t
\(\frac{1}{2}\)
α
(h) 
0.3 (0.2-0.5) 0.33 ± 0.1 
t
\(\frac{1}{2}\)
β
(h) 
3.8 (1.29-7.90) 3.2 ± 0.1 
Vss (L) 6.68 ± 3.88 5.42 ± 1.44 
Variables (units)/dose4 mg/kg, mean ± SD or mean (range)7 mg/kg, mean ± difference
AUC0-∞ (μmol/L h) 124.2 ± 32.5 204.1 ± 21.5 
Css (μmol/L) 0.5 ± 0.11 0.88 ± 0.14 
CL (L/h) 4.45 ± 0.84 3.97 ± 0.85 
t
\(\frac{1}{2}\)
α
(h) 
0.3 (0.2-0.5) 0.33 ± 0.1 
t
\(\frac{1}{2}\)
β
(h) 
3.8 (1.29-7.90) 3.2 ± 0.1 
Vss (L) 6.68 ± 3.88 5.42 ± 1.44 

NOTE: All variables were estimated by fitting to a two-compartment method. Five patients at 4 mg/kg G3139 CIVI and two patients at 7 mg/kg were included in the calculation. AUC0-∞, area under the plasma concentration-time curve calculated by linear trapezoid rule; Css, plasma steady-state concentration; CL, total body clearance; Vss, steady state volume of distribution.

One of the limitations in assessing the clinical activity of G3139 or other antisense therapeutics has been the inability to obtain information on the fate of the drug following in vivo administration. Specifically, it is unknown whether detectable intracellular concentration can be achieved and how these relate to the drug plasma levels and if any stoichiometric relationship between intracellular levels of the antisense and baseline Bcl-2 levels is necessary to attain a clinically significant down-regulation of the target. To date, none of the assay methods available for G3139 were either specific or sensitive enough to measure intracellular levels of the drug (12, 16, 18, 2629). To overcome these problems and to quantify Bcl-2 antisense in different biological matrices, we developed and validated a sensitive and specific ELISA assay for G3139 and used it to measure the intracellular levels of the drug in cell extracts of blood and bone marrow samples from in vivo–treated patients. Although this method used an approach conceptually similar to those published for other antisense oligodeoxynucleotides (3032), it also presents substantial differences. The addition of S1 nuclease specific for ssDNA following the second hybridization step seems to decrease the excess amount of probe oligodeoxynucleotide bound to capture oligodeoxynucleotide, thereby reducing the background fluorescence in biological samples and enhancing the linearity of the assay. Further, the addition of Triton X-100 at low concentration plays a role in disrupting the nonspecific interaction between antisense and cellular or plasma protein and seems quite effective at improving assay precision.

Herein, we showed that our assay is highly sensitive and specific toward the parent compound and its metabolites, as it detects only the 3′-end intact sequence of G3139 at a single nucleotide resolution, with little or no cross-reactivity with the putative 3′-end metabolites (33, 34). In contrast, the assay was not selective toward 5′ metabolites; however, 5′-end metabolism has not been considered a major pathway (3436). Preliminary metabolism data in our laboratory (not shown) supported this contention. The design of the capture oligodeoxynucleotide was based on the sequence of G3139 in a way that its 3′ terminal sequence is complementary to the G3139 and the 5′ terminal overhang is complementary to the sequence of the probe oligodeoxynucleotide. The probe oligodeoxynucleotide had a predetermined random sequence with no similar match from BLAST database (http://www.ncbi.nlm.nih.gov/BLAST). The length and sequence of probe oligodeoxynucleotide were not critical in terms of assay specificity or sensitivity, and it is likely that probe oligodeoxynucleotides with different designs might work equally well as long as the sequence of the capture oligodeoxynucleotide is complementary to the probe oligodeoxynucleotide at the 5′ terminal sequence.

The higher sensitivity and selectivity of the ELISA assay permitted monitoring of drug decay for a longer time period and revealed a second slower declined phase with a mean half-life of over 3 hours; however, due to the design of the current protocol, only samples to 4 hours postinfusion were obtained. A better pharmacokinetic characterization for this drug in another protocol with longer sampling time will be obtained. More importantly, the high sensitivity of our assay allowed us to measure intracellular concentrations of G3139. This is a critical point, because assessment of intracellular drug levels could provide us with currently unavailable information on cell uptake and the fate of the antisense, once it has been internalized following in vivo administration. To date, the process through which oligodeoxynucleotide cellular uptake occurs remains to be elucidated, although adsorptive and fluid-phase endocytosis seems involved (3739). Once internalized, oligodeoxynucleotides are sequestered in the endosomal-lysosomal compartment, and only small proportion of oligodeoxynucleotides could escape from the degradation vesicles and reached the intended targets either in the cytosol or in the nucleus. However, pharmacologic activity of antisense could be limited if insufficient concentrations are attained. Using our sensitive assay, we were able to show that indeed low uptake and no antisense activity occurred in K562 cells exposed to 3.3 μmol/L G3139 in the absence of cationic lipids. This is consistent with previously published data showing that an excess amount of antisense without cationic lipids was needed to achieve the desired target down-regulation (40, 41). In contrast, using cationic lipids (Oligofectamine and DDAB/l-α-dioleyl phosphatidylethanolamine) as delivery vehicles, marked concentration-dependent intracellular G3139 levels were observed with a more effective target down-regulation. The increase in intracellular availability was 10- to 25-fold using Oligofectamine and 20- to 50-fold by DDAB/l-α-dioleyl phosphatidylethanolamine compared with free G3139. This difference may be due to the intrinsic difference in uptake behaviors between noncomplexed G3139 and G3139 complexed with cationic lipids generally considered for antisense oligodeoxynucleotide (42, 43). Alternatively, the lower uptake of noncomplexed G3139 in cells may be due to its high binding with proteins in the cell culture medium. Of note, using FITC-labeled G3139, we showed only a 5-fold increase of cellular uptake of oligodeoxynucleotide by Oligofectamine compared with free antisense. The discrepancy may be related to quenching due to the protein binding of fluorescence-labeled oligonucleotide, a small alteration in uptake behavior of labeled G3139, or difference in methodology.

Cationic lipids not only enhanced the rate and amount of G3139 uptake into K562 cells but also might alter the intracellular distribution of G3139 as reported previously. It is widely accepted that cationic liposomes deliver oligodeoxynucleotide into cells through an endocytotic pathway (44) followed by dissociation between the oligodeoxynucleotides and the cationic lipids (45). Here, we showed that the presence of cationic lipids enhanced nuclear accumulation and concentration-dependent down-regulation of Bcl-2 RNA of the antisense in K562 cells. Our results for the intracellular localization of G3139 in K562 cells are consistent with the previous finding in which >70% of the radiolabeled phosphorothioate oligonucleotides were found to be associated with the cytoplasmic fraction, and various nucleus to cytoplasm ratios ranging from 0.146 to 0.34 were found for different sequences (46). Because the cytoplasmic fraction obtained with hypotonic lysis comprised membranes, cytosol, and endosome/lysosome, except the nuclei, the level of G3139 in cytoplasmic fraction might still be somewhat overestimated. Nevertheless, the distinct accumulation of G3139 in the nuclei by cationic liposome suggests that intranuclear content of G3139 may correlate with our observed concentration-dependent Bcl-2 down-regulation, because RNase H is enriched in the nuclei (47).

Interestingly, whereas in vitro antisense activity required the use of cationic liposomes, G3139 in aqueous saline solution showed pharmacologic activity in vivo as shown by target down-regulation. Thus, it seems that cationic liposomes are not required to achieve adequate intracellular levels of antisense in vivo. In fact, for the first time, we showed that a significant cellular uptake of G3139 occurs in mononuclear cells from patients' bone marrow and blood mononuclear cell samples collected following 72 to 120 hours of CIVI of the antisense. Of eight patients evaluated, four of six patients who had intracellular drug concentrations >5 pmol/mg proteins showed down-regulation of Bcl-2 mRNA in bone marrow. This result may suggest a threshold concentration for the pharmacologic effect. However, given the small size of the sample population, a larger-sized study needs to be conducted to validate these preliminary results. To our surprise, there was no correlation between plasma steady-state concentration of G3139 and intracellular levels of the drug or Bcl-2 down-regulation. Further, although G3139 was infused in patients in aqueous solution without any delivery vehicle, the cellular uptakes of G3139 in bone marrow or blood samples were significantly higher than those observed in leukemia cell lines treated in vitro in the absence of cationic lipids. These results, therefore, suggest that additional unidentified factors or conditions in vivo might be responsible for an efficient internalization of the antisense into mononuclear cells. Future studies to recognize such factors or conditions are important to optimize G3139 uptake in vivo.

In conclusion, using a novel, highly sensitive hybridization-based ELISA method developed in our laboratory, we have found for the first time evidence that measurable intracellular levels of Bcl-2 antisense G3139 are achievable in vivo in AML patients when a noncomplexed form of the drug was given and that Bcl-2 down-regulation is likely to depend on the achievable intracellular concentration rather than on plasma concentrations.

Grant support: NIH/National Cancer Institute grants R21 CA 94552, UO1-CA 76576, and P30CA16058.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Stein CA, Cheng YC. Antisense oligonucleotides as therapeutic agents—is the bullet really magical?
Science
1993
;
261
:
1004
–12.
2
Cotter FE. Antisense therapy for lymphomas.
Hematol Oncol
1997
;
15
:
3
–11.
3
Furdon PJ, Dominski Z, Kole R. RNase H cleavage of RNA hybridized to oligonucleotides containing methylphosphonate, phosphorothioate and phosphodiester bonds.
Nucleic Acids Res
1989
;
17
:
9193
–204.
4
Chiang MY, Chan H, Zounes MA, et al. Antisense oligonucleotides inhibit intercellular adhesion molecule 1 expression by two distinct mechanisms.
J Biol Chem
1991
;
266
:
18162
–71.
5
Skorski T, Nieborowska-Skorska M, Nicolaides NC, et al. Suppression of Philadelphia1 leukemia cell growth in mice by BCR-ABL antisense oligodeoxynucleotide.
Proc Natl Acad Sci U S A
1994
;
91
:
4504
–8.
6
Mullen P, McPhillips F, MacLeod K, et al. Antisense oligonucleotide targeting of Raf-1: importance of raf-1 mRNA expression levels and raf-1-dependent signaling in determining growth response in ovarian cancer.
Clin Cancer Res
2004
;
10
:
2100
–8.
7
Tamm I, Dorken B, Hartmann G. Antisense therapy in oncology: new hope for an old idea?
Lancet
2001
;
358
:
489
–97.
8
Klasa RJ, Gillum AM, Klem RE, Frankel SR. Oblimersen Bcl-2 antisense: facilitating apoptosis in anticancer treatment.
Antisense Nucleic Acid Drug Dev
2002
;
12
:
193
–213.
9
Lee Y, Vassilakos A, Feng N, et al. GTI-2040, an antisense agent targeting the small subunit component (R2) of human ribonucleotide reductase, shows potent antitumor activity against a variety of tumors.
Cancer Res
2003
;
63
:
2802
–11.
10
Tolcher AW, Reyno L, Venner PM, et al. A randomized phase II and pharmacokinetic study of the antisense oligonucleotides ISIS 3521 and ISIS 5132 in patients with hormone-refractory prostate cancer.
Clin Cancer Res
2002
;
8
:
2530
–5.
11
Jansen B, Wacheck V, Heere-Ress E, et al. Chemosensitisation of malignant melanoma by BCL2 antisense therapy.
Lancet
2000
;
356
:
1728
–33.
12
Chi KN, Gleave ME, Klasa R, et al. A phase I dose-finding study of combined treatment with an antisense Bcl-2 oligonucleotide (Genasense) and mitoxantrone in patients with metastatic hormone-refractory prostate cancer.
Clin Cancer Res
2001
;
7
:
3920
–7.
13
Gross A, McDonnell JM, Korsmeyer SJ. BCL-2 family members and the mitochondria in apoptosis.
Genes Dev
1999
;
13
:
1899
–911.
14
Campos L, Rouault JP, Sabido O, et al. High expression of bcl-2 protein in acute myeloid leukemia cells is associated with poor response to chemotherapy.
Blood
1993
;
81
:
3091
–6.
15
Reed JC. Bcl-2 family proteins: regulators of apoptosis and chemoresistance in hematologic malignancies.
Semin Hematol
1997
;
34
:
9
–19.
16
Marcucci G, Byrd JC, Dai G, et al. Phase 1 and pharmacodynamic studies of G3139, a Bcl-2 antisense oligonucleotide, in combination with chemotherapy in refractory or relapsed acute leukemia.
Blood
2003
;
101
:
425
–32.
17
Marcucci G, Stock W, Zwiebel J, et al. Clinical activity of Genasense (GNS, oblimersen sodium), a pro-apoptotic to Bcl-2 oligonucleotide, in combination with daunorubicin and cytarabine: a phase I study in previously untreated elderly acute myeloid leukemia (AML).
Blood
2003
;
102
:
385a
.
18
Marcucci G, Dai G, Klisovic MI, et al. Pharmacologic and biologic assessment of Genasense (GNS, oblimersen sodium), a pro-apoptotic to Bcl-2 oligonucleotide, in combination with daunorubicin and cytarabine: a phase I study in previously untreated elderly acute myeloid leukemia (AML).
Blood
2003
;
102
:
874a
.
19
Inoue T, Matsuura E, Nagata A, et al. Enzyme-linked immunosorbent assay for human pulmonary surfactant protein D.
J Immunol Methods
1994
;
173
:
157
–64.
20
Tonkinson JL, Stein CA. Patterns of intracellular compartmentalization, trafficking and acidification of 5′-fluorescein labeled phosphodiester and phosphorothioate oligodeoxynucleotides in HL60 cells.
Nucleic Acids Res
1994
;
22
:
4268
–75.
21
Rose JK, Buonocore L, Whitt MA. A new cationic liposome reagent mediating nearly quantitative transfection of animal cells.
Biotechniques
1991
;
10
:
520
–5.
22
Marcucci G, Livak KJ, Bi W, et al. Detection of minimal residual disease in patients with AML1/ETO-associated acute myeloid leukemia using a novel quantitative reverse transcription polymerase chain reaction assay.
Leukemia
1998
;
12
:
1482
–9.
23
Hsu JC. Multiple comparisons with a control. In: Multiple comparisons: theory and methods. 1st ed. CRC Press LLC, FL; 1996. p. 43–78.
24
Neter J. Applied linear regression models. 3rd ed. Chicago (IL): Irwin; 1996. p. xv, 720.
25
Wolfe DA. HM. the two sample location problem. In: Nonparametric statistical methods. 2nd ed. John Wiley & Sons, Inc., New York, NY; 1999. p. 106–10.
26
Raynaud FI, Orr RM, Goddard PM, et al. Pharmacokinetics of G3139, a phosphorothioate oligodeoxynucleotide antisense to bcl-2, after intravenous administration or continuous subcutaneous infusion to mice.
J Pharmacol Exp Ther
1997
;
281
:
420
–7.
27
Lopes D, Mayer LD. Pharmacokinetics of Bcl-2 antisense oligonucleotide (G3139) combined with doxorubicin in SCID mice bearing human breast cancer solid tumor xenografts.
Cancer Chemother Pharmacol
2002
;
49
:
57
–68.
28
Waters JS, Webb A, Cunningham D, et al. Phase I clinical and pharmacokinetic study of bcl-2 antisense oligonucleotide therapy in patients with non-Hodgkin's lymphoma.
J Clin Oncol
2000
;
18
:
1812
–23.
29
Morris MJ, Tong WP, Cordon-Cardo C, et al. Phase I trial of BCL-2 antisense oligonucleotide (G3139) administered by continuous intravenous infusion in patients with advanced cancer.
Clin Cancer Res
2002
;
8
:
679
–83.
30
Boutet V, Delaunay V, De Oliveira MC, et al. Real-time monitoring of the hybridization reaction: application to the quantification of oligonucleotides in biological samples.
Biochem Biophys Res Commun
2000
;
268
:
92
–8.
31
Deverre JR, Boutet V, Boquet D, et al. A competitive enzyme hybridization assay for plasma determination of phosphodiester and phosphorothioate antisense oligonucleotides.
Nucleic Acids Res
1997
;
25
:
3584
–9.
32
Yu RZ, Baker B, Chappell A, et al. Development of an ultrasensitive noncompetitive hybridization-ligation enzyme-linked immunosorbent assay for the determination of phosphorothioate oligodeoxynucleotide in plasma.
Anal Biochem
2002
;
304
:
19
–25.
33
Gaus HJ, Owens SR, Winniman M, Cooper S, Cummins LL. On-line HPLC electrospray mass spectrometry of phosphorothioate oligonucleotide metabolites.
Anal Chem
1997
;
69
:
313
–9.
34
Phillips JA, Craig SJ, Bayley D, et al. Pharmacokinetics, metabolism, and elimination of a 20-mer phosphorothioate oligodeoxynucleotide (CGP 69846A) after intravenous and subcutaneous administration.
Biochem Pharmacol
1997
;
54
:
657
–68.
35
Griffey RH, Greig MJ, Gaus HJ, et al. Characterization of oligonucleotide metabolism in vivo via liquid chromatography/electrospray tandem mass spectrometry with a quadrupole ion trap mass spectrometer.
J Mass Spectrom
1997
;
32
:
305
–13.
36
Cohen AS, Bourque AJ, Wang BH, Smisek DL, Belenky A. A nonradioisotope approach to study the in vivo metabolism of phosphorothioate oligonucleotides.
Antisense Nucleic Acid Drug Dev
1997
;
7
:
13
–22.
37
Loke SL, Stein CA, Zhang XH, et al. Characterization of oligonucleotide transport into living cells.
Proc Natl Acad Sci U S A
1989
;
86
:
3474
–8.
38
Benimetskaya L, Loike JD, Khaled Z, et al. Mac-1 (CD11b/CD18) is an oligodeoxynucleotide-binding protein.
Nat Med
1997
;
3
:
414
–20.
39
Gray GD, Basu S, Wickstrom E. Transformed and immortalized cellular uptake of oligodeoxynucleoside phosphorothioates, 3′-alkylamino oligodeoxynucleotides, 2′-O-methyl oligoribonucleotides, oligodeoxynucleoside methylphosphonates, and peptide nucleic acids.
Biochem Pharmacol
1997
;
53
:
1465
–76.
40
Campos L, Sabido O, Rouault JP, Guyotat D. Effects of BCL-2 antisense oligodeoxynucleotides on in vitro proliferation and survival of normal marrow progenitors and leukemic cells.
Blood
1994
;
84
:
595
–600.
41
Kitada S, Miyashita T, Tanaka S, Reed JC. Investigations of antisense oligonucleotides targeted against bcl-2 RNAs.
Antisense Res Dev
1993
;
3
:
157
–69.
42
Crooke ST, Bennett CF. Progress in antisense oligonucleotide therapeutics.
Annu Rev Pharmacol Toxicol
1996
;
36
:
107
–29.
43
Thierry AR, Vives E, Richard JP, et al. Cellular uptake and intracellular fate of antisense oligonucleotides.
Curr Opin Mol Ther
2003
;
5
:
133
–8.
44
Beltinger C, Saragovi HU, Smith RM, et al. Binding, uptake, and intracellular trafficking of phosphorothioate-modified oligodeoxynucleotides.
J Clin Invest
1995
;
95
:
1814
–23.
45
Zelphati O, Szoka FC Jr. Mechanism of oligonucleotide release from cationic liposomes.
Proc Natl Acad Sci U S A
1996
;
93
:
11493
–8.
46
Crooke RM, Graham MJ, Cooke ME, Crooke ST. In vitro pharmacokinetics of phosphorothioate antisense oligonucleotides.
J Pharmacol Exp Ther
1995
;
275
:
462
–73.
47
Walder RY, Walder JA. Role of RNase H in hybrid-arrested translation by antisense oligonucleotides.
Proc Natl Acad Sci U S A
1988
;
85
:
5011
–5.