Paramagnetic Oligonucleotides: Contrast Agents for Magnetic

Feb 3, 1999 - COMMUNICATIONS. Paramagnetic ... of Medicine, The Ohio State University, Columbus, Ohio 43210. Received September 2, 1998...
0 downloads 0 Views 113KB Size
MARCH/APRIL 1999 Volume 10, Number 2 © Copyright 1999 by the American Chemical Society

COMMUNICATIONS Paramagnetic Oligonucleotides: Contrast Agents for Magnetic Resonance Imaging with Proton Relaxation Enhancement Effects Jennifer V. Hines,*,† Ghada M. Ammar,† Jill Buss,† and Petra Schmalbrock‡ Division of Medicinal Chemistry & Pharmacognosy, College of Pharmacy, and Department of Radiology, College of Medicine, The Ohio State University, Columbus, Ohio 43210. Received September 2, 1998

An antisense paramagnetic oligonucleotide analogue targeted to a model macromolecular receptor (5S rRNA) was prepared. The paramagnetic agent’s relaxivity (dependence of the relaxation rate on paramagnetic agent concentration) in the presence and absence of the macromolecular receptor was measured at 1.5 and 6.3 T. The relaxivity of the targeted agent increased specifically in the presence of the macromolecular receptor (16% at 6.3 T and 15% at 1.5 T). This effect was specific for a paramagnetic oligonucleotide targeted to the receptor and was larger than the relaxivity enhancement due simply to receptor-induced viscosity differences. Maximizing this relaxivity enhancement of tumor targeted paramagnetic oligonucleotides will aid in contrast agent development for magnetic resonance imaging.

In addition to clinical assessment, cross-sectional diagnostic imaging procedures such as computed tomography (CT) and magnetic resonance imaging (MRI) are currently used to increase the accuracy of cancer diagnosis and staging. Currently approved MRI contrast agents either are tissue-nonspecific (Gd-DTPA, Figure 1) or target normal tissue providing only “cold” spot imaging (superparamagnetic iron oxide) which reduces their value in accurately detecting primary neoplasia or metastases (1-3). Development of cancer-targeted contrast imaging agents for providing “hot” spot MR imaging would improve the sensitivity of MRI for tumor detection over current methods. Biostable oligonucleotide analogues targeted to diseased tissue, either as an antisense agent to overexpressed mRNA (4) or as a ligand (aptamer) to a disease-specific receptor (5), have been investigated for other imaging methods but have not been thoroughly investigated as paramagnetic contrast agents for MRI. To optimize contrast agents for MRI, it is useful to * To whom correspondence should be addressed. Phone: (614) 688-4008. Fax: (614) 292-2435. E-mail: [email protected]. † College of Pharmacy. ‡ College of Medicine.

Figure 1. Gd-DTPA-diethylenetriaminepentaacetic acid.

maximize not only target localization but also the proton relaxation enhancement (PRE) (6) effect upon target binding. Maximizing the PRE effect increases the signal difference between targeted and nontargeted tissues, thus reducing toxicity by reducing the dose necessary for adequate contrast enhancement. We report the preparation of an antisense paramagnetic oligonucleotide targeted to a model macromolecular receptor and the demonstration of a specific PRE effect upon target binding. This information will aid in the development of paramagnetic oligonucleotide contrast agents for MRI. MR contrast imaging relies on the effect a paramagnetic species has on the observed longitudinal and transverse relaxation rates (1/T1 and 1/T2, respectively) of solvent (water) nuclei (6). Differential relaxation rates due to paramagnetic contrast agents can enhance the MR

10.1021/bc980103v CCC: $18.00 © 1999 American Chemical Society Published on Web 02/03/1999

156 Bioconjugate Chem., Vol. 10, No. 2, 1999

Hines et al.

Figure 3. Secondary structure of 5S rRNA and antisense target site. Figure 2. Antisense and aptamer-based targeting strategies.

image obtained and thus lead to better diagnosis. The relaxation rates are linearly dependent on the concentration of the paramagnetic species M where relaxivity (R1) is defined as

(1/T1)obs ) (1/T1)d + R1[M]

(1)

where (1/T1)d is the diamagnetic contribution to the observed relaxation rate (1/T1)obs.1 A similar relationship applies to T2 for R2 relaxivity. The most widely investigated paramagnetic ions for T1 contrast work are chelated Gd(III), Mn(II), and Fe(III). The relaxivity is dependent on many factors, including magnetic field strength, correlation time (τc, tumbling rate in solution), and inner-sphere and outer-sphere water around the paramagnetic ion (6). Of particular importance is the fact that an increase in the correlation time of the paramagnetic species results in increased relaxivity. When the paramagnetic species is targeted to a macromolecular receptor and has a significant change in correlation time upon receptor binding, this leads to a proton relaxation enhancement (PRE) effect or increased relaxivity in the presence of the macromolecular receptor (6). Maximizing the PRE effect enhances the image contrast that can be achieved with a targeted contrast agent beyond that which can be achieved with currently available nonspecific agents. Biopolymers such as paramagnetic antibodies targeted to tumor-specific antigens have been investigated for use as MRI contrast agents. The paramagnetic metal enhances the relaxivity (R1) of bulk water, while the antibody provides tumor localization. Antibodies with only paramagnetic chelates attached, however, are not very effective contrast agents due to their large size, the lack of access of bulk water to the paramagnetic center, and very small correlation time change (and hence small relaxivity change) upon target binding (6). Only the superparamagnetic particle-antibody complexes are useful candidates. Porphyrins (8), liposomes (9), and polysaccharides (10, 11) have also been investigated for tissuespecific delivery of relaxation agents. An alternative to these agents which has not been thoroughly investigated is the use of targeted paramagnetic oligonucleotide analogues. Biostable oligonucleotide analogues can provide tumorspecific delivery of contrast agents (Figure 2) either as an antisense agent to mRNA overexpressed in a cancer cell (4) or as an aptamer to a cell surface (12). The main advantage of an oligonucleotide-based delivery agent over 1 Often, the terms 1/T and R are used interchangeably; 1 1 however, the explicit definition of relaxivity (R1) is that shown in eq 1 (6) and is used accordingly throughout this paper.

an antibody is size. Many oligonucleotide antisense agents and aptamers to complex targets have tight binding specificity contained within a relatively small piece of oligonucleotide (20-40 nucleotides = 7-14 kDa) (13). Consequently, the correlation time of the free oligonucleotide analogue versus the oligonucleotide when bound to the target is different. This difference is sufficient to induce a PRE effect. The extent and lifetime of bound water in nucleic acids (14) can also be exploited to further enhance the PRE effect. Our model system for investigating and ultimately optimizing the PRE effect for an antisense paramagnetic oligonucleotide uses 5S rRNA as the macromolecular target and a complementary 6mer sequence as the antisense agent. The Escherichia coli 5S rRNA was purchased and renatured (15). We chose the sequence 5′CGGCAT-3′ (ON-1) which is complementary to nucleotides 39-44 in 5S rRNA (Figure 3) (15). We chose the sequence 5′-AAAAAA-3′ (A6) as a control to monitor relaxivity changes due to viscosity differences. Since there is no complementary binding site for A6 on 5S rRNA, relaxivity changes in the presence of 5S rRNA will depend only on viscosity changes due to the macromolecule. ON-1 and A6 were purchased with a 5′-amino linker and reacted with DTPA dianhydride (2, Scheme 1). The resulting chelate conjugates were treated with GdCl3 followed by size-exclusion chromatography and dialysis to prepare the paramagnetic-oligonucleotide chelate complexes. HPLC and MALDI-TOF analyses indicated there was pure product with no degradation or dimer formation. DTPA content was verified by spectrophotometric analysis (16). On the basis of the correlation time of 5S rRNA (10 ns) (17) and of a 6mer oligonucleotide (1 ns) (18), there should be a significant change in the correlation time of the 6mer oligonucleotide chelate upon binding to the 5S rRNA target. Such a change in the correlation time should result in a PRE effect as discussed earlier. An electron spin resonance study with a TEMPO-modified single-stranded oligonucleotide showed that there was a change in the correlation time upon binding a complementary region of 5S rRNA (19). The buffered renatured 5S rRNA (15) was used to prepare sample paramagnetic agent:5S rRNA ratios of 0:1, 0.25:1, 0.5:1, 0.75:1, and 1:1. Equivalent samples containing just buffered paramagnetic agent were also prepared. Inversion-recovery studies for measuring the T1 values of the bulk water were carried out at 6.3 and 1.5 T. At 6.3 T, all buffered samples were in 95% D2O/ 5% H2O. Due to the sequential nature of the experiments, a full inversion-recovery experiment for determining each individual T1 was prohibitively long. Consequently, the T1s were estimated from the null spectrum of a 180D2-90-acquire pulse sequence where T1 ) D2/ln 2 (20). At 1.5 T, buffered samples in H2O were placed in a 96-

Communications

Bioconjugate Chem., Vol. 10, No. 2, 1999 157

Scheme 1. Synthesis of Paramagnetic Oligonucleotides

Table 1. R1 Relaxivity of Paramagnetic Oligonucleotides at 6.3 Ta ON-1-DTPA-Gd ON-1-DTPA-Gd and 5S rRNAd A6-DTPA-Gd A6-DTPA-Gd and 5S rRNAd ON-1-DTPA-Gd and dextrane

R1b (mM-1 s-1)

% PREc

8.84 10.24 13.49 13.90 9.13

16 3 3

ON-1-DTPA-Gd ON-1-DTPA-Gd and 5S rRNAd A6-DTPA-Gd A6-DTPA-Gd and 5S rRNAd

a

In 10 mM Tris-HCl (pH 6.9), 10 mM MgCl2, 20 mM Na2B4O7 buffer. Relaxation studies were performed at 300 K on a Bruker 6.3 T 270 MHz NMR spectrometer. b See the text for R1 determination. Error estimated to be less than 5% based on replicate T1 determinations.

well polypropylene tube array and surrounded by buffer to minimize magnetic susceptibility problems. To investigate the possibility of the nucleic acid binding to the well walls, which could potentially affect the relaxation measurements, duplicates of the samples were placed in wells previously treated with a siliconizing agent and then rinsed with 95% ethanol and air-dried. To avoid difficulties with phase variation across the image, MR systems typically generate magnitude images. Thus, only the absolute value of the MR signal was measured. Images were analyzed using IDL (Research Systems Inc.), and the image intensities of each sample well (7 mm × 7 mm, 14 pixel × 14 pixel image) were measured over an area of 121 pixels. The intensities at different TIs were analyzed using SigmaPlot (Jandell Scientific). A nonlinear curve fit of the form

Iτ ) |I∞(1 - Ce-τ/T1)|

Table 2. R1 Relaxivity of Paramagnetic Oligonucleotides at 1.5 Ta

(2)

was applied where τ is the delay time TI, T1 is the spinlattice relaxation time, Iτ is the image intensity at delay time τ ()TI), I∞ is the image intensity when τ ) ∞, and C ≈ 2. I∞, C, and T1 were treated as adjustable parameters. The resulting 1/T1 values were then plotted versus the paramagnetic agent concentrations. The relaxivity (R1) was calculated for each experiment using a linear regression fit to the equation (1/T1)obs ) R1x + b, where (1/T1)obs is measured in s-1 and x (the paramagnetic agent concentration) in millimolar. Using our experimental conditions for sample preparation, data acquisition, and analysis, the relaxivity measured for Gd-DTPA (R1 ) 4.43 mM-1 s-1 at 1.5 T) compares favorably with the literature value (21, 22). The antisense paramagnetic oligonucleotide targeted to 5S rRNA produced a 16% PRE effect at 6.3 T (Table 1) and a 15% PRE effect at 1.5 T (clinical MR field strength, Table 2). The nonspecific effect that viscosity changes would have on the observed relaxivity of the paramagnetic oligonucleotides due to the presence of a macromolecule was estimated two different ways. The R1 of ON-1-DTPA-Gd in the presence of dextran (MW

R1b (mM-1 s-1)

% PREc

85 98 156 168

15 7.7

a In 10 mM Tri- HCl (pH 6.9), 10 mM MgCl , 20 mM Na B O 2 2 4 7 buffer. Relaxation studies were performed at room temperature on a GE/Signa 1.5 T MR imager with the wrist coil as the receiver coil and the large volume body coil as the radio frequency transmitter. A single slice that was 4 mm thick was acquired corresponding to 40% of the sample height. The field of view was 12 cm × 9 cm and the matrix size 256 × 256 points. An inversion recovery fast spin-echo (IR-FSE) program was used with TIs (delay time for monitoring inversion recovery) of 50, 200, 300, 500, 700, 900, 1100, 1500, 1800, 2100, 2500, 3200, and 4000 ms with a constant TE (echo time) and TR (repetition time) of 16 and 8000 ms, respectively. b See the text for R1 determination. Error estimated to be 5% based on replicate R1 determinations. c Proton relaxation enhancement (% R1 enhancement in the presence of a macromolecule); see the text. d At 0.006 mM in 5S rRNA.

) 40 kDa = 5S rRNA) was measured. The other viscosity control involved measuring the R1 of A6-DTPA-Gd in the presence of 5S rRNA. Both methods for estimating viscosity effects indicated the same enhancement of R1 (3%, Table 1) in the presence of a macromolecule. On the basis of these controls, the enhancements simply due to viscosity were smaller than the PRE effects due to the targeted paramagnetic oligonucleotide binding to the 5S rRNA target. This indicates that a PRE effect can be observed when paramagnetic oligonucleotide analogues bind specifically to a macromolecular receptor. Maximizing this PRE effect will be useful in adding hot spot contrast to MR imaging with diseased tissue-targeted paramagnetic oligonucleotide analogues. As mentioned above, in the 96-well array used for the MRI experiment we were concerned there might be some nonspecific attachment to the well sides. However, no significant difference in R1 values was seen between the treated and the nontreated wells. An additional control was run to determine if there was any nonspecific attraction between Gd-DTPA and 5S rRNA. No significant difference in the R1 values for Gd-DTPA in the presence (5.22 mM-1 s-1 at 6.3 T) or absence (5.20 mM-1 s-1 at 6.3 T) of 5S rRNA was detected. While the PRE for A6-DTPA-Gd at 1.5 T was significantly smaller than that for ON-1-DTPA-Gd, we were somewhat surprised that it was larger than that at 6.3 T. It is possible that A6-DTPA-Gd hydration differences for samples with and without 5S rRNA are affecting the PRE at 1.5 T more than at 6.3 T. Since unchelated lanthanide ions and certain lanthanide complexes can lead to RNA degradation (23), we investigated whether any 5S rRNA degradation was occurring due to the paramagnetic oligonucleotides. UV

158 Bioconjugate Chem., Vol. 10, No. 2, 1999

melting (24) of 1:1 ON-1-DTPA-Gd:5S rRNA and 1:1 A6-DTPA-Gd:5S rRNA each showed a slight decrease in Tm (∆Tm ) 5 °C) compared to that of 5S rRNA alone (approximate Tm ) 75 °C). The predicted Tm for the ON1-DTPA-Gd:5S rRNA duplex is 20 °C; however, the expected hyperchromicity upon melting would be very small relative to the 5S rRNA hyperchromicity and, thus, was not detected. Denaturing polyacrylimide gel studies for detecting full-length 5S rRNA and any degradation products were also carried out. Under the ratio and incubation time conditions used in the NMR and MRI studies, no degradation was observed. However, some degradation was observed at a ratio of 2:1. While 5S rRNA degradation effects could affect the observed R1 values, there was no difference in the degradation effects of ON-1-DTPA-Gd versus A6DTPA-Gd on 5S rRNA. Consequently, degradation does not appear to contribute to the observed difference in PRE between the targeted ON-1-DTPA-Gd and the viscosity control A6-DTPA-Gd. The degradation is most likely due to gadolinium having one less ligand in the oligonucleotide-DTPA-Gd complexes than in Gd-DTPA, leading to an increased likelihood of free Gd3+ and nonspecific degradation. Since Gd-DTPA alone does not degrade 5S rRNA under the conditions used in the NMR and MRI studies or at a 2:1 ratio of paramagnetic agent: 5S rRNA as determined by UV melting and gel studies, we believe that utilizing a more stable chelate (DTPA isothiocyanate) (25) will reduce the degradation effects we are observing. This analogue allows full chelation of Gd3+ (stability constant equivalent to that of Gd-DTPA) with an added tether for conjugation to a macromolecule. In summary, these studies indicate that paramagnetic oligonucleotides targeted to a macromolecule produce a PRE effect which is specific and greater than simple viscosity differences in the presence of the macromolecular receptor. Maximizing this PRE effect will lead to enhanced MR image contrast with paramagnetic oligonucleotides targeted to tumors since it enhances relaxivity without increasing the paramagnetic agent concentration. We are currently investigating the optimal biophysical features necessary to maximize the PRE effect of tumor-targeted biostable paramagnetic oligonucleotide analogues. Results from these experiments will be presented in due course. ACKNOWLEDGMENT

This work was supported by The American Cancer Society of Ohio. We thank Profs. E. Wisner and I. Tinoco, Jr., for helpful discussions. LITERATURE CITED (1) Sabel, M., and Aichinger, H. (1996) Recent Developments in Breast Imaging. Phys. Med. Biol. 41, 315-368. (2) Greenfield, G. B., Arrington, J. A., and Kudryk, B. T. (1993) MRI of Soft Tissue Tumors. Skeletal Radiol. 22, 77-84. (3) Tanimoto, A., Satoh, Y., Uasa, Y., Jinzaki, M., and Hiramatsu, K. (1997) Performance of Gd-EOB-DTPA and Superparamagnetic Iron Oxide Particles in the Detection of Primary Liver Cancer: A Comparative Study by Alternative Free-Response Receiver Operating Characteristic Analysis. J. Magn. Reson. Imaging 7, 120-124. (4) Dewanjee, M. K., Ghafouripur, A. K., Kapadvanjwala, M., Dwanjee, S., Serafini, A. N., Lopez, D. M., and Sfakianakis, G. N. (1994) Noninvasive Imaging of C-myc Oncogene Messenger RNA with Indium-111-antisense Probes in a Mammary Tumor-Bearing Mouse Model. J. Nucl. Med. 35, 10541063.

Hines et al. (5) Charlton, J., Sennello, J., and Smith, D. (1997) In Vivo Imaging of Inflammation using an Aptamer Inhibitor of Human Neutrophil Elastase. Chem. Biol. 4, 809-816. (6) Lauffer, R. B. (1987) Paramagnetic Metal Complexes as Water Proton Relaxation Agents for NMR Imaging: Theory and Design. Chem. Rev. 87, 901-927. (7) Teifenauer, L. X., Kuhne, G., and Andres, R. Y. (1993) Antibody-Magnetite Nanoparticles: In Vitro Characterization of a Potential Tumor-Specific Contrast Agent for Magnetic Resonance Imaging. Bioconjugate Chem. 4, 347-352. (8) van Zijl, P. C., Place, D. A., Cohen, J. S., Faustino, P. J., Lyon, R. C., and Patronas, N. J. (1990) Metalloporphyrin Magnetic Resonance Contrast Agents. Feasibility of TumorSpecific Magnetic Resonance Imaging. Acta Radiol. Suppl. 374, 75-79. (9) Storrs, R. W., Tropper, F. D., Li, H. Y., Song, C. K., Kuniyoshi, J. K., Sipkins, D. A., Li, K. C. P., and Bednarski, M. D. (1995) Paramagnetic Polymerized Liposomes: Synthesis, Characterization, and Applications for Magnetic Resonance Imaging. J. Am. Chem. Soc. 117, 7301-7306. (10) Rongved, P., and Laveness, J. (1991) Water-Soluble Polysaccharides as Carriers of Paramagnetic Contrast Agents for Magnetic Resonance Imaging: Synthesis and Relaxation Properties. Carbohydrate Res. 214, 315-323. (11) Rebizak, R., Schaefer, M., and Dellacherie, E. (1997) Polymeric Conjugates of Gd3+-Deithylenetriaminepentaacetic Acid and Dextran. 1. Synthesis, Characterization, and Paramagnetic Properties. Bioconjugate Chem. 8, 605-611. (12) Schering, A.-G. (1996) Ger. Offen. DE 4424923; Chem. Abstr. 124, 283703. (13) Wang, Y., and Rando, R. R. (1995) Specific Binding of Aminoglycoside Antibiotics to RNA. Chem. Biol. 2, 281-290. (14) Johannesson, H., and Halle, B. (1998) Minor Groove Hydration of DNA in Solution: Dependence on Base Composition and Sequence. J. Am. Chem. Soc. 120, 6859-6870. (15) Goringer, H. U., and Wagner, R. (1988) 5S RNA Structure and Function. Methods Enzymol. 164, 721-737. (16) Pippin, C. G., Parker, T. A., McMurry, T. J., and Brechbiel, M. W. (1992) Spectrophotometric Method for the Determination of a Bifunctional DTPA Ligand in DTPA-Monoclonal Antibody Conjugates. Bioconjugate Chem. 3, 342-345. (17) Marshall, G., and Smith, J. L. (1980) Raman and 19F(1H) Nuclear Overhauser Evidence for a Rigid Solution Conformation of Escherichia coli 5-Fluorouracil 5S Ribonucleic Acid. Biochemistry 19, 5955-5959. (18) Cantor, C. R., and Schimmel, P. R. (1980) Biophysical Chemistry II: Techniques for the Study of Biological Structure and Function, p 461, Freeman & Co., San Francisco. (19) Murakami, A., Nagahara, S., Nakaura, M., Uematsu, H., Mukae, M., and Makino, K. (1990) Structural Analysis of Nucleic Acids by Labeled Oligonucleotides. Nucleic Acids Res. Symp. Ser. 22, 27-28. (20) Derome, A. E. (1987) Modern NMR Techniques for Chemistry Research, pp 86-89, Pergamon, Oxford, U.K. (21) Steen, R. G., Gronemeyer, S. A., Kingsley, P. B., Reddick, W. E., Langston, J. S., and Taylor, J. S. (1994) Precise and Accurate Measurement of Proton T1 in Human Brain in vivo: Validation and Preliminary Clinical Application. J. Magn. Reson. Imaging 4, 681-691. (22) Donahue, K. M., Weisskoff, R. M., and Burstein, D. (1997) Water Diffusion and Exchange as They Influence Contrast Enhancement. J. Magn. Reson. Imaging 7, 102-110. (23) Kazunari, M., and Komiyama, M. (1997) Enormously fast RNA hydrolysis by Lanthanide(III) Ions Under Physiological Conditions: Eminent Candidates for Novel Tools of Biotechnology. Jpn. J. Biochem. 122, 387-394. (24) Puglisi, J. D., and Tinoco, I., Jr. (1989) Absorbance Melting Curves of RNA. Methods Enzymol. 180, 304-325. (25) Brechbiel, M. W., Gansow, O. A., Atcher, R. W., Schlom, J., Esteban, J., Simpson, D. E., and Colcher, D. (1986) Synthesis of 1-(p-Isothiocyanatobenzyl) Derivatives of DTPA and EDTA. Antibody Labeling and Tumor-Imaging Studies. Inorg. Chem. 25, 2772-2781.

BC980103V