|Home | About | Journals | Submit | Contact Us | Français|
While antisense oligonucleotide (AS-ODN) technology holds promise for the treatment of cancer, to date there have been no clinical successes. Unfortunately, current assays are not sufficiently sensitive to measure tissue ODN levels. Hence it has not been possible to ascertain whether treatment failures result from failure of drug delivery. To investigate the relationship between drug uptake and therapeutic effect, we developed an ultrasensitive noncompetitive hybridization-ligation enzyme-linked immunosorbent assay (NCHL-ELISA) to quantify Kinase Suppressor of Ras1 (KSR1) AS-ODN drug uptake in plasma and tumor tissues. In mice harboring PANC-1 pancreatic cancer xenografts and continuously infused with AS-ODN, our ELISA detects plasma and tumor KSR1 AS-ODN levels over an extended range, from 0.05 nM to 20 nM. Using this sensitive assay, we demonstrate that KSR1 repression in pancreatic cancer xenografts correlates highly with AS-ODN uptake into tumor tissues. In contrast, plasma drug levels do not correlate with tumor drug content or target downregulation. These studies indicate the efficacy of our ELISA, and suggest that tumor biopsy material will need to be procured to estimate the potential of this antisense technology.
Kinase Suppressor of Ras1 (KSR1) was originally identified in D. melanogaster and C. elegans as a positive modulator of Ras/MAPK signaling either upstream of or parallel to Raf.1-4 In C. elegans, KSR-1 is dispensable for normal development and adult worm function, but necessary for transmission of increased signaling resulting from gain-of-function (gf) mutation of RAS-1, or LET-23 the epidermal growth factor receptor (EGFR) homolog.2,3 Isolation of murine and human KSR1 orthologs with high sequence identity4 suggested KSR1 signaling is evolutionarily conserved. As in C. elegans, KSR1 inactivation has little, if any, impact on Ras function in normal mammalian development or physiology, rather KSR1 appears specifically engaged when the demand for Ras function is increased, such as via activating point mutations or enhanced upstream signaling through hyperactivated tyrosine kinase receptors.5 Consistent with this observation, mice lacking KSR1 develop and reproduce normally, and manifest a normal lifespan. However, ksr1−/− mice develop an unusual disorganized hair follicle phenotype also manifest in EGFR knockout mice, indicating that as in C. elegans, EGFR, Ras and KSR1 are on the same signaling pathway.5 Moreover, papilloma formation in Tg.AC mice, resulting from skin-specific v-Ha-ras expression, was abrogated in a ksr1 knockout background, indicating v-Ha-ras-mediated skin cancer, signaled through the Raf-1/MAPK cascade, requires KSR1.5 These results suggest KSR1 may represent a therapeutic target for Ras/MAPK signaling of human tumorigenesis.
Based on the selectivity of KSR1 for involvement in gf Ras phenotypes, we designed a KSR1 phosphorothioate antisense oligonucleotide (AS-ODN) to inactivate KSR1 pharmacologically. KSR1 is comprised of five distinct domains, termed CA1-CA5, the first four of which are N-terminal.6 CA1 is unique to KSR1, and conserved in most KSR1 orthologs. Its function awaits characterization. CA2 is a Src homology 3 (SH3) recognition site and CA3 a Cys-rich domain similar to the lipid-binding domain of protein kinase C. CA4 is a Ser/Thr-rich region resembling the CR2 domain of c-Raf-1. The C-terminal CA5 region contains the 11 conserved kinase subdomains found in all kinases. KSR1 AS-ODN was designed against the unique CA1 KSR1 domain (n214−231 of human KSR1 cDNA) conserved between mouse and human.6 Continuous infusion of AS-ODN into nude mice harboring xenografts of PANC-1 pancreas or non-small lung cancers, which are driven through mutant K-Ras, resulted in significant anti-tumor activity.7 As overall 5-year survival for all patients with pancreatic cancer is below 5% and mean survival is usually less than 1 year,8 we are pursuing development of our KSR1 AS-ODN for eventual use in human clinical trials.
Phosphorothioate oligodeoxynucleotides have been investigated extensively in recent years as potential therapeutic agents for viral infections,9 inflammatory diseases,10 and various types of cancer.11 One reason that AS-ODN drugs are particularly attractive for cancer therapy is that unlike the nonspecific killing by radiation or chemotherapy, AS-ODNs can be used to target specific molecules involved in cell proliferation and cell death in cancer cells.12 Further, AS-ODNs display a favorable toxicity profile.13,14 However, AS-ODNs have not yet proven reliable therapeutics for human diseases. The biggest success was the approval of Formivirsen as a drug to treat CMV retinitis, the first AS-ODN drug to enter the market.11 On the other hand, the recent lack of success in large randomized phase III trials testing AS-ODNs against Bcl-2 for multiple myeloma and PKCα for non small cell lung carcinoma has reduced enthusiasm for AS-ODN therapeutics.15 While the reasons of these clinical failures are unknown, lack of adequate drug uptake in tumor tissues, particularly solid tumors, might be one of them. Prior clinical trials rarely defined tumor drug levels due to the fact that in clinical trials, tumor specimen availability is often restricted. Furthermore the amount of tumor tissue when available is often scarce, limiting determination of AS-ODN levels by conventional approaches such as ion pair reversed-phase or strong anion-exchange high-performance liquid chromatography,16,17 capillary gel electrophoresis with ultraviolet detection18 or laser-induced fluorescence detection,19 mass spectrometry20 or competitive hybridization assays.21-23 To measure drug uptake in body fluids and tumor tissues, a highly sensitive assay is needed. Recently, two non-competitive hybridization-ligation enzyme-linked immunosorbent assays (NCHL-ELISAs) were reported with a lower limit of quantitation in the low picomolar range.24,25 In the present study, we adapted the Ultrasensitive NCHL-ELISA assay for detection of KSR1 AS-ODN. We found that KSR1 expression in pancreatic cancer PANC-1 xenografts correlates well with tumor tissue AS-ODN uptake, whereas plasma drug levels do not reflect AS-ODN levels in tumor tissue.
To begin to explore the pharmacodynamics of KSR1 AS-ODN treatment as applied to KSR1 expression, tumor xenografts were established by injecting 10 × 106 PANC-1 cells subcutaneously into the flank of male athymic nude mice, and used for experiments between passage 5−20. AS-ODN, control sense ODN or PBS vehicle control were infused subcutaneously via Alzet osmotic minipump when tumors reached a volume of approximately 300 mm3. After 5 days of treatment, tumors were harvested, homogenized in NP-40 lysis buffer and tumor lysates subjected to western blot. Expression of the housekeeping gene GAPDH was used as loading control. KSR1 and GAPDH bands were quantified by densitometry and percent KSR1 expression in arbitrary densitometry units for each ODN-treated tumor was calculated relative to the mean of the PBS-treated control tumors. A representative experiment is shown in Figure 1. After 5 days of treatment with 40 mg/kg/d AS-ODN or PBS control, endogenous expression of KSR1 in five PANC-1 xenografts was reduced by 77.8% (range 64−85% downregulation) compared to the mean of the controls, indicating our KSR1 AS-ODN is capable of significantly inactivating KSR1 expression in vivo. In contrast, 40 mg/kg/d sense ODN failed to reduce KSR1 levels (data not shown).
Subsequent studies investigated the relationship between KSR1 AS-ODN drug levels and KSR1 target inactivation by adapting a NCHL-ELISA to measure KSR1 AS-ODN levels in blood and tumor specimens. The principle of this ultrasensitive ELISA is shown in Figure 2A. Briefly, this assay is based on hybridization between KSR1 AS-ODN, a complementary template, and a signal probe. The template contains nine extra nucleotides at the 5'-end and biotin at the 3'-end. The signal probe is a 9-mer ODN complementary to the template nine-nucleotide overhang and contains a phosphate group at 5' end and digoxigenin at 3' end. To separate KSR1 AS-ODN from irrelevant nucleic acids, template probe is added to biologic samples, allowed to hybridize with KSR1 AS-ODN, and subsequently the template probe/KSR1 AS ODN complex is bound to a streptavidin-coated plate via 3' biotin-streptavidin interaction. Signal probe is then hybridized to the template probe/KSR1 AS ODN complex, ligated therein by T4 DNA ligase, and the entire complex is detected with an anti-digoxigenin antibody conjugated to alkaline phosphatase (AP). AP catalyzes formation of fluorescent AttoPhos, quantified using a microtiter plate reader.
Figure 2B shows a representative standard curve using our ultrasensitive KSR1 AS-ODN ELISA assay. Different concentrations of commercially-synthesized KSR1 AS-ODN were used as standard. Detection was linear within the range of 0.05 nM to 20 nM with a correlation coefficient of 0.9949. This extended range suggests that this ELISA might have potential for detecting AS-ODN levels in clinically-relevant samples.
To investigate the practical value of our ultrasensitive ELISA for in vivo samples, we assayed plasma samples from mice infused with different doses of KSR1 AS-ODN. A five-day infusion with 20, 30 or 40 mg/kg/d KSR1 AS-ODN, sense ODN or PBS via subcutaneous Alzet osmotic minipump was used, consistent with our tumor treatment protocol. As shown in Figure 3, no signal was detected in plasma samples from mice treated with PBS vehicle control, nor was signal detected using sense ODN (data not shown). However, there was a linear relationship between plasma drug levels and treatment doses from 20−40 mg/kg/d suggesting our KSR1 AS-ODN ELISA assay is both sensitive and specific.
Sufficient free drug within tumor tissues to inactivate the cognate target is the ultimate determining factor to achieve local tumor control. While plasma drug levels, which are relatively easy to obtain, are commonly used to evaluate drug uptake, plasma drug levels may not faithfully reflect tissue levels, especially for drugs that are highly protein bound in the circulation, as is the case for AS-ODNs. In fact, phosphorothioate AS-ODNs can be more than 97% bound in to plasma proteins.26 Here we investigated the value of our ultrasensitive ELISA for detecting drug levels in plasma and tumor tissues, and compared these results to tumor KSR1 downregulation. For these studies, we treated animals for 5 days with 10−75 mg/kg/d KSR1 AS-ODN or PBS via subcutaneous Alzet osmotic minipump. As shown Figure 4A, there was a very high correlation between tumor drug uptake and KSR1 target downregulation. In contrast, KSR1 AS-ODN levels in plasma did not correlate with drug levels in PANC-1 tumor xenografts (Fig. 4B), nor KSR1 downregulation. Preliminary data performed using human MiaPaCa pancreas cancer xenografts yielded similar correlations (data not shown).
In summary, we have developed an ultrasensitive ELISA assay for quantifying KSR1 AS ODN levels in both plasma and tumor tissues with great sensitivity and specificity. The detection was linear within a wide range of drug levels from 0.05 nM to 20 nM. With this ultrasensitive assay, we found that drug levels in pancreatic xenografts correlate highly with target protein downregulation. In contrast, drug levels in plasma did not correlate with tumor drug uptake or target downregulation. Our results suggest that direct measurement of drug uptake and target downregulation in tumor biopsies may be necessary to evaluate the efficacy of AS ODNs in cancer therapy.
PANC-1 cells were purchased from ATCC (Rockville, MD) and grown at 37°C in a 5% CO2 incubator in DME medium supplemented with glucose (4.5 g/L), 10% fetal bovine serum (Life Technologies, Inc., Grand Island, NY), 100 IU/ml penicillin, and 100 IU/ml streptomycin. Six to eight week-old male athymic NCRnu/nu (National Cancer Institute, Frederick, MD) were housed at the animal core facility of Memorial Sloan-Kettering Cancer Center. This facility is maintained in accordance with the regulations and standards of the United States Department of Agriculture and the Department of Health and Human Services, NIH. To establish PANC-1 xenografts, mice were injected with 10 × 106 PANC-1 cells in 100 μl Matrigel (BD Biosciences, San Jose, CA) subcutaneously into the flank. Mice were examined every other day for tumor formation and volume, based on caliper measurements. Once established, tumors were maintained by successive in vivo passage without intermittent growth in vitro. For passaging, tumors were surgically removed, separated of stroma and necrotic areas, and then passed through the Stainless Steel Homogenizer Cellector Tissue Sieve (Bellco Glass Inc., Vineland, NJ). Collected tumor cells were washed once with ice-cold Dulbecco PBS (0.2 gm/L KCl, 0.2 gm/L KH2PO4, 8 gm/L NaCl, 1.15 gm/L Na2HPO4, pH 7.0), and 100 mg tumor tissue equivalents in PBS were injected subcutaneously into the flank. Tumors of passage 5−20 were used for experiments.
Human KSR1 AS-ODN (5'-CTT TGC CTC TAG GGT CCG-3') and KSR1 sense ODN (5'-CGG ACC CTA GAG GCA AAG-3') were generated as phosphorothioate derivatives against nucleotides 214 to 231 of the unique CA1 domain (AAs 42−82) of KSR1 by Genelink Inc., (Hawthorne, NY). AS-ODN and sense ODN were resuspended in PBS and infused subcutaneously as per manufacturer's instructions into experimental animals harboring 250−350 mm3 PANC-1 tumor xenografts via Alzet osmotic minipump at doses of 10−75 mg/kg/d. PBS was used as vehicle control.
After 5 days of KSR1 AS-ODN treatment, mice were anesthetized and 0.7−1.0 ml blood was extracted from the left ventricle in the presence of the anticoagulant EDTA (BD Microtainer K2E, Franklin NJ). Blood samples were centrifuged for 15 min at 800 xg at room temperature. Subsequently plasma was centrifuged for 5 min at 15,000 xg to remove remaining cells and platelets, and then subjected to ELISA to quantify AS-ODN levels or immediately frozen at −80°C.
After blood collection, tumors were harvested, weighed, pared of connective tissue, and minced into small pieces. 0.2−0.5 g tumor tissue was Dounce homogenized in 4 vol NP-40 lysis buffer (20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 2 mM EDTA, 10% Glycerol, 1% Nonidet P-40 plus protease and phosphatase inhibitors), centrifuged at 10,000 xg for 10 min and tumor lysate protein concentration was determined using the Bio-Rad protein assay.
250 μg tumor lysate was mixed with Laemmli SDS-loading buffer and resolved by 7% SDS-PAGE. Separated proteins were then transferred to a Hybond-ECL nitrocellulose membrane (Amersham Pharmacia Biotech, Piscataway, NJ). KSR1 was detected using a mouse monoclonal anti-KSR-1 antibody (BD biosciences, Franklin Lakes, NJ; 1:500 dilution), and HRP-conjugated sheep anti-mouse secondary antibody (Amersham Pharmacia Biotech, 1:1000 dilution). Autoradiography was performed as described by Immobilon Western (Millipore, Bedford, MA). Expression of the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), detected by a rabbit anti-GAPDH monoclonal antibody (Cell Signaling Technology, Inc., Beverly, MA; 1:10,000 dilution) and HRP-conjugated donkey anti-rabbit secondary antibody (Amersham Pharmacia Biotech, 1:15000 dilution) was used as loading control. KSR1 and GAPDH bands were quantified by densitometry using Imagej 1.37v (National Institutes of Health, Bethesda, MD).
KSR AS-ODN, template probe (5'-GAA TAG CGA CGG ACC CTA GAG GCA AAG-Biotin-3') and signal probe (5'-Phosphate-TCG CTA TTC-digoxigenin-3') were synthesized by Genelink Inc. Reacti-Bind NeutrAvidin-coated polystyrene strip plates were purchased from Pierce (Rockford, IL). Template probe solution (0.05 μM) was prepared in hybridization buffer containing 60 mM Na2HPO4 (pH 7.4), 0.9 M NaCl, 0.24% Tween 20. Ligation probe (0.067 μM) was prepared in 1xOne-Phor-All Buffer Plus (Amersham Pharmacia Biotech) with 1.33 units/mL of T4 DNA ligase (Amersham Pharmacia Biotech, Piscataway, NJ) and 0.05 mM ATP (Amersham Pharmacia Biotech). Washing buffer contains 25 mM Tris-HCl (pH 7.2), 0.15 M NaCl, and 0.1% Tween. Anti-digoxigenin-AP (conjugated with alkaline phosphatase, Fab fragments) was obtained from Roche Diagnostics Corporation (Indianapolis, IN). Alkaline phosphatase fluorescent substrate AttoPhos and its reconstitution solution were purchased from Promega Life Science (Madison, WI).
The assay was performed as follows: 100 μl template probe solution was added to 25−100 μg tumor lysate or 5−20 μl plasma in 100 μl lysate buffer at 37°C in a polypropylene 96-well plate. After 1 h hybridization, 150 μl of solution was transferred to a NeutrAvidin coated 96-well plate and incubated at 37°C for 30 min to allow binding of biotin to streptavidin-coated wells. After washing 4x with washing buffer, ligation was performed by adding 150 μl ligation probe solution to the plate followed by 2 h incubation at room temperature. The plate was then washed 2x with washing buffer followed by 3x with deionized water. Subsequently, 150 μl of a 1:2000 dilution of anti-digoxigenin-AP was added at room temperature for 30 min. After washing, AttoPhos was added to the plate. After 30 min, fluorescence intensity was determined using a Cytofluor microtiter plate reader (Perkin Elmer Biosystems, Framingham, MA; excitation 450/50 and emission 580/50).
Pearson and Spearman correlation coefficients and corresponding p values were computed to illustrate the strength of association between drug uptake and protein expression. All computations were done in R 2.5.0 (http://cran.r-project.org/).
The present study was supported by NIH grant RO1 CA42385 to R.K.