Improvement of MRI Probes To Allow Efficient Detection of Gene

Open Access ... Recent Developments in Magnetic Diagnostic Systems .... Chemistry of Materials 2005 17 (20), 5109-5118 ... 4 ) from the hot injection ...
0 downloads 0 Views 187KB Size
Download PDF
Bioconjugate Chem. 2000, 11, 941−946

941

Improvement of MRI Probes To Allow Efficient Detection of Gene Expression Dagmar Ho¨gemann, Lee Josephson, Ralph Weissleder, and James P. Basilion* Center for Molecular Imaging Research, Massachusetts General Hospital, Charlestown, Massachusetts. Received July 7, 2000; Revised Manuscript Received September 18, 2000

Recently, it has been demonstrated that magnetic resonance imaging (MRI) utilizing monocrystalline iron oxide nanoparticles (MIONs) targeted to an engineered transferrin receptor enables imaging of gene expression. However, the relatively high doses of iron oxides used indicated the need for improved MR imaging probes to monitor changes in gene expression in vivo. Using alternative conjugation chemistries to link targeting ligands and iron oxide nanoparticles, we present the development and characterization as well as improved receptor binding and MRI detection of a novel imaging probe. Iron oxide nanoparticles with a cross-linked dextran coat were conjugated to transferrin (Tf) through the linker molecule N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP) to yield Tf-S-S-CLIO. The characteristics of this conjugate were evaluated in comparison to Tf-MION and Tf-CLIO generated by oxidative activation of the dextran-coat with subsequent reduction of Schiff’s base. SPDP conjugation allowed approximately a 4-fold increase in the number of Tf molecules attached per iron oxide nanoparticle and resulted in a more than 10-fold improvement of binding and uptake by cells. This translated into an imaging probe that was 16 times better for imaging gene expression in a cellular MRI assay. This novel probe for MRI may substantially increase the sensitivity for the detection of endogenous or genetically induced transferrin receptor expression in small numbers of cells and may significantly reduce the imaging dose from over 100 mg/kg to doses of iron oxides that are currently used in clinical imaging.

INTRODUCTION

Real-time noninvasive imaging of gene expression in vivo is a major goal in molecular imaging research. The development of such technology would dramatically alter the way studies are performed. This technology would enable temporal monitoring of gene expression in single living animals during the development of disease or during administration of transgenes for gene therapy and eliminate the need for animal-intensive and laborintensive studies. Several different imaging modalities have been exploited to image gene expression in vivo; however, none have the high spatial resolution achievable using magnetic resonance imaging (MRI). We have recently capitalized on the high spatial resolution of this technology demonstrating the utility of MRI for imaging gene expression in vivo and its utility for microscopically mapping transgene expression in tumor sections (1). To image gene expression in vivo, we previously employed an oxidative strategy of conjugating superparamagnetic monocrystalline iron oxide nanoparticles (MION) to transferrin (Tf). This resulted in targeting of the TfMION conjugate to tumors overexpressing engineered human transferrin receptor (ETR, a cell-surface internalizing receptor), cellular accumulation, and detection by MRI (1). However, the high doses of Tf-MION contrast agent needed for these studies emphasized the relatively low sensitivity of MRI for probe detection (1) and demonstrated the need to develop improved MR contrast agents for imaging gene expression. In this report, we conjugate Tf to iron oxide particles using the cross-linking reagent N-succinimidyl 3-(2* To whom correspondence should be addressed. Phone: (617) 726-5788. Fax: (617) 726-5708. E-mail: basilion@ helix.mgh.harvard.edu.

pyridyldithio)propionate (SPDP) to generate a new MR probe and compare its properties with two conjugates prepared by the previously employed oxidative conjugation strategy. We found that the disulfide cross-linked Tf iron oxide conjugates (Tf-S-S-CLIO) had more Tf proteins per iron oxide particle, a higher affinity for the Tf receptor, and altered MR signal intensity in cellular uptake assays at 16-fold lower concentrations. EXPERIMENTAL PROCEDURES

Synthesis of Iron Oxide Particles: MION, CLIO, CLIO-NH2. MION-46 was prepared by reaction of iron salts with ammonium hydroxide in the presence of T10 dextran (3). Also, the preparation of cross-linked iron oxide nanoparticles, now defined as CLIO-10, has been described before (4). Briefly, a dextran-coated monodispersed iron oxide preparation was prepared and then cross-linked with epichlorohydrin and purified (4). To utilize SPDP as the linking reagent, it was necessary to add primary amines to the CLIO. To maintain a chemically defined dextran coat, amidation was performed on CLIO rather than MION particles. To produce iron oxide amine, CLIO-NH2, CLIO was treated with ammonia and then dialyzed as described (4, 5). Synthesis of Tf-MION and Tf-CLIO (Figure 1, reactions A and B). To ensure conjugation of high affinity Tf to iron oxide particles, the integrity of holo-Tf (Sigma, St. Louis, MO) was confirmed photometrically prior to conjugation as described (6). Reaction of MION or CLIO with Tf was accomplished by partial oxidation of dextrans, reaction with Tf, and subsequent reduction of Schiff’s bases as previously described (2) and as modified here. MION or CLIO was dialyzed against sodium acetate buffer (10 mM sodium acetate and 0.15

10.1021/bc000079x CCC: $19.00 © 2000 American Chemical Society Published on Web 11/02/2000

942 Bioconjugate Chem., Vol. 11, No. 6, 2000

Figure 1. Synthetic scheme of the conjugation of Tf-MION, Tf-CLIO, and Tf-S-S-CLIO.

M NaCl, pH 6). Following dialysis, MION or CLIO was oxidized with sodium periodate (4 mg/mg iron) for 40 min, dialyzed against 0.15 M NaCl for 1 h and dialyzed against sodium bicarbonate buffer (20 mM sodium bicarbonate and 0.15 M NaCl, pH 8.7) for another 40 min. To conjugate MION or CLIO with human diferric Tf, Tf was added at a ratio of 1.7 mg/mg iron. After 3 h of incubation, Schiff’s bases were reduced by addition of 15.9 mM sodium cyanoborohydrate and incubation overnight at 4 °C. Magnetic separation columns (Macs separation columns, Miltenyi Biotec, CA) were used to separate nonconjugated Tf from iron oxide particles. To calculate the number of Tf molecules per particle, the protein (BCA, Pierce, Rockford, IL) and iron concentrations of purified material were determined (2, 7). Results were expressed as the number of Tf molecules per particle assuming an average of 2064 Fe atoms/particle (7). The conjugates were stored in phosphate buffered saline (PBS), pH 8 at 4 °C. Synthesis of Tf-S-S-CLIO (Figure 1, reaction C). Derivatization of CLIO with 2-Pyridyl Disulfide. CLIONH2 was used to prepare 2-pyridyl disulfide derivatized CLIO. One milliliter of CLIO-NH2 (26 mg of Fe; in 5 mM citrate buffer), 0.67 mL of PBS (pH 8), and 1.3 mL of 3.4 × 10-2 M (in DMSO) N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP; Molecular Biosciences, Boulder, CO) were incubated for 2 h at room temperature. Low molecular impurities were removed by filtration through Sephadex G25F columns equilibrated with PBS, pH 8. The pooled void volume had an iron concentration of 4.5 mg of Fe/mL. To measure the number of 2-pyridyl disulfide groups attached, 50 µL of the solution were added to 75 µL of DTT (2 × 10-2 M) in PBS, pH 7.4. After 30 min at room temperature, a microconcentrator with a 30 kDa cutoff (Amicon, Beverly, MA) was used to separate the product of the reduction, pyridine-2-thione, from iron. The concentration of pyridine-2-thione was determined using an extinction coefficient of 8100 M-1 cm-1 at 343 nm. On average, 35 2-pyridyl disulfide groups were introduced per CLIO (n ) 3). The compound was kept at 4 °C. Thiolation of Human Diferric Tf (Tf-SH). Equal volumes of diferric Tf (10-3 M ) 80 mg/mL) and iminothiolane (2.5 × 10-3 M in PBS, 2 mM EDTA, pH 8, degassed) were mixed and incubated for 1 h at room temperature in the dark. Unreacted iminothiolane was removed by filtration through a Sephadex G25F column equilibrated with PBS, 2 mM EDTA, pH 8. The absorption at 280 nm with an extinction coefficient of 1.2 × 105 M-1 cm-1 (8) was used to determine the concentration of Tf. The number of reactive sulfhydryl groups was measured at a wavelength of 412 nm after the addition of Ellman’s reagent with an extinction coefficient of 13 600 M-1 cm-1 (9, 10). Under these conditions, a ratio of 0.7:1 sulfhydryl groups/Tf was achieved, and polymerization was not observed.

Ho¨gemann et al.

Synthesis of Tf-S-S-CLIO. Tf-SH (3 × 10-4 M concentration of Tf) and 2Py-S-S-CLIO (1.3 × 10-3 M concentration of 2-pyridyl disulfide groups) were mixed at equal molar ratios of Tf and 2-pyridyl disulfide groups in PBS, pH 8. The mixture was incubated for 18 h at room temperature. Unreacted Tf-SH was separated from Tfconjugates using magnetic columns and the number of Tf molecules per particle was determined as above. The conjugates were stored in PBS, pH 8, either at 4 °C or shock frozen in liquid nitrogen and kept a -20 °C. Cell Culture. Rat 9L gliosarcoma cells (Brain Tumor Research Center, San Francisco, CA) (11) or rat gliosarcoma cells stably overexpressing an altered form of the human transferrin receptor ETR (9L3.9 cells) (1, 2) were grown in Dulbecco’s modified Eagle’s medium (DMEM, Cellgro, Mediatech, Washington, DC) containing 10% fetal bovine serum (FBS, Cellgro). The medium was changed every 3 days. Transfected cells were periodically passaged with G418 to ensure integrity of the plasmid expression (2). Competition Studies. Human diferric Tf (Sigma, St. Louis, MO) was radiolabeled with Na125I in the presence of IodoGen (Pierce, Rockford, IL). Unbound iodide was removed by centrifugation through Bio-Gel P6 spincolumns (Bio-Rad, Hercules, CA). Competition studies were performed for the different conjugates in the presence of [125I]Tf and compared to competition curves generated using different concentrations of unlabeled diferric Tf. Cells (9L wild-type or 9L3.9 cells) were plated at a density of 1 × 105 cells/well onto 24 well plates (Falcon, Becton Dickinson, Lincoln Park, NJ). The next day, different concentrations of unlabeled diferric Tf, thiolated Tf, Tf-MION, Tf-CLIO, or Tf-S-SCLIO (0.25 nM to 2.5 µM in DMEM) were added immediately prior to the addition of [125I]Tf (0.25 nM in DMEM). As a negative control 2Py-S-S-CLIO was mixed with Tf-SH in an amount equal to their respective concentrations in the Tf-S-S-CLIO conjugate and added to cells. After incubation for 1 h at 37 °C the cells were washed three times with Hanks’ Balanced Salt Solution (HBSS, Bio-Whittaker, Walkersville, MD) and detached with trypsin (0.05% trypsin, 0.5 mM ethylenediaminetetraacetic acid). Cell-associated radioactivity was counted in a gamma-counter (1282 Compugamma CS, LKB Wallac, Sweden). SDS-PAGE. Laemmli SDS-PAGE was used to determine the integrity of the different conjugates (12). Samples were mixed with SDS sample loading buffer and analyzed by gel electrophoresis using a 4% stacking gel and a 10% separating gel at a thickness of 0.75 mm. Where reducing conditions were required, 5% β-mercaptoethanol was added to the SDS sample buffer and samples were heated at 95 °C for 5 min. Nonreduced samples were not subjected to heating as controls demonstrated that this had no effect on the integrity of the conjugate (data not shown). Two micrograms of protein was loaded per lane. Gels were run at a constant voltage of 120 V and stained with Coomassie Brilliant Blue R-250. Physical Properties and MR Imaging. The R1 and R2 relaxivities of Tf-MION, Tf-CLIO, and Tf-S-S-CLIO were measured using a 0.47 T tabletop minispec as described (7). For MR imaging, 3.5 × 106 cells (9L wild-type or 3.9 clone) were seeded/10 cm culture dish. The next day, cells were incubated with the conjugates at different concentrations of iron in 2 mL of DMEM. Cells were incubated at 37 °C for 2 h followed by washing with HBSS, trypsinization, washing in DMEM with 10% FBS, and

Novel Probe for MRI of Gene Expression

Bioconjugate Chem., Vol. 11, No. 6, 2000 943

Table 1. Physical Properties of Superparamagnetic Iron Oxide Conjugates

compd Tf-MION Tf-CLIO Tf-S-S-CLIO

average Tf-molecule R2 R1 -1 -1 (mM × s) (mM × s) per particle 20 26 32

62 114 146

0.6 1.2 4

linker chemistry nonreversible nonreversible reversible

centrifugation. After centrifugation, the cell pellets were resuspended in culture medium and transferred into 250 µL tubes. To form uniform pellets the cells were again centrifuged at 500 rpm for 2 min (Sorvall 7 RT, Kendro Laboratory Products, Newtown, CT). To prevent drying and susceptibility artifacts, the supernatant was not removed. The tubes were placed into a water bath at RT for MR imaging. The phantom allowed the simultaneous imaging of up to 50 different samples. MR imaging was performed with a clinical 1.5 T superconducting magnet (Signa 5.0; GE Medical Systems, Milwaukee, WI) using a 5-in. surface coil. The imaging protocol consisted of a T2 weighted spin-echo sequence (SE, TR 3000 ms, variable TE 16-100 ms). The 1.9 mm imaging slice was carefully placed to avoid partial volume effects. At a field of view of 8 cm2 and a 256 × 256 imaging matrix each voxel has a size of 0.186 mm3. Accordingly, in each tube a volume of 9.5 µL accounting for approximately 1.5 × 106 cells was imaged. After MR imaging, the viability of the cells was confirmed by trypan blue exclusion and was always greater than 95%. Signal intensity (SI) measurements of cell pellets were obtained by manual placement of regions of interest (ROI) on the cell pellets. ROIs were also placed on the surrounding water to determine the overall homogeneity of the image. Signal intensities plotted against the 10 different TEs showed the differences in T2 decay of each pellet. Curve fitting for exponential decay allowed the calculation of T2 relaxation times (SI ) Ae(-TE/T2) + B, where SI is the signal intensity, TE the echo time, A the amplitude, and B the offset for each incubation concentration of conjugate. RESULTS

Synthesis of Tf-Iron Oxide Conjugates. Previous work demonstrated that the conjugation of Tf to periodate-treated MION decreased the affinity of the conjugated Tf for its receptor (2). We hypothesized that the contrast agent could be significantly improved if the affinity of conjugated Tf for transferrin receptor could be increased. To minimize the effects of conjugation on Tftransferrin receptor binding, we utilized a chemical crosslinking reagent, N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP), which has previously been shown to have excellent cross-linking efficiency while maintaining the biological activity of Tf (10). The synthesis schemes for all three types of iron oxide conjugates are shown in Figure 1. The use of a chemical cross-linking agent significantly improved the number of Tf molecules conjugated per iron oxide particle. Approximately 4 Tf molecules were bound/iron crystal in Tf-S-S-CLIO compared to 0.6-1.2 Tf/particle in Tf-MION and Tf-CLIO. Physical properties of the different conjugates are summarized in Table 1. To determine if the conjugation and purification strategies above resulted in successful binding and adequate purity, SDS-PAGE was used to analyze the three different conjugates (Figure 2). A brown color, indicative of iron oxide nanoparticles before staining (upper panel), and a blue color after Coomassie Blue staining (lower panel) showed the co-localization of iron oxide nanopar-

Figure 2. Nonreducing and reducing SDS-PAGE of Tf-MION, Tf-CLIO, and Tf-S-S-CLIO. Two micrograms of nonconjugated Tf (lane 1) or of Tf in each of the conjugates (lanes 2-9) were incubated in SDS-sample buffer with or without β-mercaptoethanol (β-ME), heat treated as indicated in the text and electrophoresed. Upper panel: Gel prior to staining for protein. Lower panel: Gel after Coomassie Blue staining. Arrow indicates migration of authentic Tf standard.

ticles and protein in the gel (lanes 3, 5, and 7). The TfMION (lane 3) and to a lesser extent also the Tf-CLIO conjugate (lane 5) contained compounds that migrated through the stacking gel presumably due to Tf bound to iron oxide particles of smaller size. The lack of detection of free Tf or protein degradation products showed that magnetic separation efficiently removed nonconjugated Tf from Tf-MION, Tf-CLIO, and Tf-S-S-CLIO preparations. Boiling of Tf-S-S-CLIO in sample buffer containing β-mercaptoethanol, as expected, cleaved the disulfide bond, and no change in the integrity of Tf-S-S-CLIO was observed after prolonged freezing (Figure 2, lanes 6 and 8). In contrast, boiling of Tf-MION and Tf-CLIO in reducing SDS-sample buffer resulted in no release of Coomassie Blue detectable Tf. Binding and Uptake of Tf, Tf-MION, Tf-CLIO, and Tf-S-S-CLIO. To compare the binding and uptake of the three compounds cells overexpressing a known amount of transferrin receptor, 9L3.9 cells (1, 2) were incubated with a constant amount of [125I]diferric Tf (0.25 nM) in the presence of increasing concentrations of either unlabeled diferric Tf, Tf-S-S-CLIO, Tf-CLIO, Tf-MION, unreacted Tf-SH and 2Py-S-S-CLIO, or 2Py-S-S-CLIO alone (Figure 3A). The concentration for which 50% of the cell associated radioactivity was displaced occurred at 18.7 ( 3.2 nM for diferric Tf and the IC50 for Tf-S-SCLIO was 306.8 ( 21.2 nM. Although it was not possible to generate high enough concentrations of the other conjugates to obtain complete displacement curves, the binding and uptake of Tf-S-S-CLIO was significantly better (greater than 10-fold improved) than either the Tf-MION or the Tf-CLIO conjugates (Figure 3A). MR Imaging of Tf-MION, Tf-CLIO, and Tf-S-SCLIO. The next step in these studies was to determine

944 Bioconjugate Chem., Vol. 11, No. 6, 2000

Ho¨gemann et al.

Figure 3. (A) Transferrin receptor competitive binding assay. To compare the relative affinity of the different conjugates and synthesis intermediates for the transferrin receptor a constant amount of [125I]transferrin and increasing amounts of unlabeled diferric Tf, Tf-S-S-CLIO, Tf-CLIO, Tf-MION, Tf-SH freshly mixed with 2Py-S-S-CLIO, and 2Py-S-S-CLIO were co-incubated with cells and total cell associated radioactivity was measured. To account for differences in Tf molecules/particle for the different conjugates the amount of added conjugate was normalized for Tf concentration. 2Py-S-S-CLIO was added in an amount equal to its respective concentration in the Tf-S-S-CLIO conjugate. (B) Transverse relaxivity versus iron concentrations. MR imaging was performed at 10 different echo times (TE 16-100 ms) on the 9L3.9 cell phantoms to allow the calculation of T2 relaxation times. T2 relaxation times were plotted as a function of the concentration of Tf-iron oxide conjugates in the incubation media. The generated curves were used to compare the relative detection of the different imaging probes by MRI.

if changes in Tf-iron oxide binding and uptake would be translated into detectable changes in MR signal intensities. To quantify the changes in signal intensities, we calculated the changes as transverse relaxation times (T2) for each contrast probe at different iron concentrations (see Experimental Procedures). T2 values were plotted versus the iron concentrations in the incubation media of the cells and used to compare the potency of the different conjugates resulting from binding and intracellular accumulation (Figure 3B). A 50% lower T2 relaxation time was found after incubation of cells with 16fold lower concentration of iron for Tf-S-S-CLIO than for Tf-CLIO or Tf-MION (Figure 3B). The changes in signal intensity measured at a single TE for the Tf-MION, TfCLIO, and Tf-S-S-CLIO imaging probes are displayed in Figure 4. Signal intensities were dependent on binding and uptake of the conjugates and highly specific for the transferrin receptor. Little to no signal changes were observed after incubation with 250 µg/mL nonconjugated MION or CLIO (data not shown) and a 1000-fold excess of nonconjugated diferric Tf completely inhibited a signal change in comparison to untreated controls. MR Imaging of Different Levels of Receptor Expression in 9L and 3.9 Cells. Finally, we determined if the improved MRI contrast probe, Tf-S-S-CLIO, was capable of being used to image differences in receptor levels. To assess this, 9L rat glioma cells (9L wild-type) and a derivation of that cell line engineered to overexpress ETR (9L3.9 cells) were incubated with different concentrations of Tf-S-S-CLIO and imaged by MRI. A decrease in signal intensity was observed at much lower

concentrations of Tf-S-S-CLIO in 9L3.9 cells compared to wild-type 9L cells (Figure 5). Since receptor-mediated binding for human diferric Tf is identical for both cell types (data not shown), the change in signal intensity most likely resulted from increased binding and uptake of the probe due to an increase in the number of receptors and possibly from changes in relaxivity. Endogenous expression of transferrin receptor could be detected in 9L cells at the highest concentration of contrast probe. DISCUSSION

We have recently demonstrated that uptake of superparamagnetic monocrystalline iron oxide nanoparticles via the transferrin receptor can be imaged by MRI and correlated to receptor expression (1, 2), while other groups have employed similar strategies to probe for endogenously overexpressed transferrin receptor (13) or to load cells with iron oxides (14). However, in their current form the targeted iron oxides used to assess transferrin receptor expression in vivo require administration of relatively high doses of Tf-MION (1). To improve probe binding and presumably reduce the dose of targeted iron oxide necessary to image transferrin receptor expression we utilized the chemical cross-linking reagent SPDP to generate Tf-S-S-CLIO. The studies presented here show that Tf-S-S-CLIO is an imaging probe that is 16-fold better for MR imaging of transferrin receptor expression and suitable to measure differences in endogenous receptor expression at much lower concentrations of probe.

Novel Probe for MRI of Gene Expression

Bioconjugate Chem., Vol. 11, No. 6, 2000 945

Figure 4. MR imaging of 9L3.9 cell phantoms. 9L3.9 cells, which overexpress TfR, were incubated with the indicated concentrations of Tf-iron oxide conjugates. After the incubation the cells were harvested, washed, pelleted into phantoms and imaged with a T2weighted MR sequence (GE 1.5 T, Signa 5.0; Spin-Echo Sequence, TR 3000 ms/TE 100 ms). Note that the addition of excess diferric Tf resulted in no change in MR signal compared to untreated controls.

Figure 5. MR imaging of TfR expression. Wild-type 9L cells, expressing relatively low levels of TfR, or 9L3.9 cells, expressing relatively high levels of TfR, were incubated with the indicated concentrations of Tf-S-S-CLIO and imaged as phantoms. The relative difference in TfR expression is clearly visible by T2-weighted MRI (GE 1.5 T, Signa 5.0; Spin-Echo Sequence, TR 3000 ms/TE 50 ms).

The oxidative conjugation strategy that has been used in previous studies produces aldehydes on the dextran coat of the iron oxides. This method can result in multiple direct attachment points between Tf and the polysaccharide of the iron oxides. Therefore, a substantial loss in the biological activity of the protein may result from (1) conjugation to sites that are important for receptor binding or (2) particle interference with receptor binding. To minimize these effects we, sought to bind Tf and iron oxide particles via a linker molecule. We first identified reaction conditions that would produce biologically active Tf bearing a defined number of sulfhydryls (0.7-1.0/Tf) and have determined conditions that yield Tf-SH with binding characteristics indistinguishable from unreacted Tf (Figure 3A and data not shown). The reaction of activated iron oxide nanoparticles, 2Py-S-S-CLIO, with Tf-SH yielded a conjugate, Tf-S-S-CLIO, with approximately 4 Tf- molecules/particle compared to 0.6-1.2 for the oxidative conjugates (Tf-MION and Tf-CLIO). Even after adjustment for the increased number of Tf molecules per particle, the competitive binding assay revealed that Tf-S-S-CLIO has a 10-fold higher affinity for the transferrin receptor than Tf-MION and Tf-CLIO (Figure 3A). Since the difference in biological activity

cannot be explained by the increase in particle associated Tf molecules, the improvement in binding of Tf-S-S-CLIO compared to either Tf-MION or Tf-CLIO suggests that the oxidative conjugation chemistry significantly interferes with the binding of the conjugates to the receptor. This may be due to chemical alteration of Tf, cross-linking to a site of Tf directly involved in receptor binding, significant particle-dependent sterical inhibition of Tftransferrin receptor binding, or a combination of these. While none of these possibilities can be excluded, the binding characteristics of Tf-S-S-CLIO to the transferrin receptor suggest it is likely that sterical inhibition of TfMION and Tf-CLIO binding plays a role. Although TfS-S-CLIO binds the transferrin receptor much better than Tf-CLIO or Tf-MION, when compared to nonconjugated diferric transferrin Tf-S-S-CLIO is a significantly less effective competitor for [125I]Tf (Figure 3A). Since, there was no detectable difference in the ability of Tf and Tf-SH to compete for [125I]Tf (Figure 3A and data not shown), the difference between Tf and Tf-S-S-CLIO to compete in the displacement assay is likely due to steric hindrance and the overall size of the conjugate rather than being due to thiolation of Tf (i.e., chemical damage or attachment site). These data strongly suggest that

946 Bioconjugate Chem., Vol. 11, No. 6, 2000

changing the proximity of the iron oxide nanoparticle and Tf can have a significant effect on the interaction of the conjugate with receptor. Furthermore, these data suggest that by lengthening the linker between Tf and the iron oxide it may be possible to synthesize a MR probe with still higher affinity for the transferrin receptor. It has been reported that the type of disulfide bond formed by conjugation of thiolated Tf and 2Py-S-S-CLIO is relatively stable in plasma with a half-life of 8 h (15, 16) making this a useful coupling chemistry for in vivo probes. The use of this linker chemistry to provide a chemically cleavable linkage between the Tf and CLIO was initially instituted to allow monitoring of the integrity of conjugated Tf following the coupling reaction (Figure 2). However, it is interesting to note that several studies where disulfide bonds have been used to target immunotoxins (17-19), opioid peptides (20) and antibody fragments (21) to cells have demonstrated that a labile disulfide bond was necessary for the expression of full activity of these agents. Whether these findings can be transferred to Tf-S-S-CLIO or whether a stable linker is preferable remains to be shown. Further studies to improve this probe in terms of stability, pharmacological profile, bioavailability, relaxation and number of Tf molecules per particle as well as the attachment of other targeting molecules are currently underway. ACKNOWLEDGMENT

We would like to thank Dr. Anna Moore for her expert advice and help in performing the studies. The work was supported by NIH Grant RO1 CA 85240. Dr. Dagmar Ho¨gemann is a recipient of a fellowship from the Deutsche Forschungsgemeinschaft, Germany. LITERATURE CITED (1) Weissleder, R., Moore, A., Mahmood, U., Bhorade, R., Benveniste, H., Chiocca E. A., and Basilion, J. P. (2000) In vivo MR imaging of transgene expression. Nat. Med. 6, (3), 351-355. (2) Moore, A., Basilion, J. P., Chiocca, E. A., and Weissleder, R. (1998) Measuring transferrin receptor gene expression by NMR imaging. Biochim. Biophys. Acta 1402, 239-249. (3) Moore, A., Marecos, E., Bogdanov, A., and Weissleder, R., (2000) Tumoral distribution of long-circulating dextran-coated iron oxide nanoparticles in a rodent model. Radiology 214, 568-574. (4) Palmacci, S., and Josephson, L. (1993) Synthesis of polysaccharide covered superparamagnetic oxide colloids. U.S. Patent 5,262,176. (5) Josephson, L., Tung, C.-H., Moore, A., and Weissleder, R. (1999) High-efficiency intracellular magnetic labeling with novel superparamagnetic-tat peptide conjugates. Bioconjugate Chem. 10 (2), 186-191. (6) Klausner, R. D., Van Renswoude, J., Ashwell, G., Kempf, C., Schechter, A. N., Dean, A., and Bridges, K. R. (1983) Receptor-mediated endocytosis of transferrin in K562 cells J. Biol. Chem. 258 (8), 4715-4724. (7) Shen, T., Weissleder, R., Papisov, M., Bogdanov, A., and Brady, T. J. (1993) Monocrystalline iron oxide nanocom-

Ho¨gemann et al. pounds (MION): physicochemical properties. Magn. Reson. Med. 29, 599-604. (8) Wellho¨ner, H. H., Neville, D. M., Jr., Srinivasachar, K., and Erdmann, G. (1991) Uptake and concentration of bioactive macromolecules by K562 cells via the transferrin cycle utilizing an acid-labile transferrin conjugate. J. Biol. Chem. 266 (7), 4309-4314. (9) Ellman, G. L. (1959) Tissue sulphydryl groups. Arch. Biochem. Biophys. 82, 70-77. (10) Kratz, F., Beyer, U., Roth, T., Tarasova, N., Collery, P., Lechenault, F., Cayabat, A., Schumacher, P., Unger, C., and Falken, U. (1998) Transferrin conjugates of doxorubicin: synthesis, characterization, cellular uptake, and in vitro efficacy. J. Pharm. Sci. 87 (3), 338-346. (11) Jacobs, R., and Fraser, S. (1994) Magnetic resonance microscopy of embryonic cell lineages and movements. Science 263, 681-684. (12) Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. (13) Kresse, M., Wagner, S., Pfefferer, D., Lawaczeck, R., Elste, V., and Semmler, W. (1998) Targeting of ultrasmall superparamagnetic iron oxide (USPIO) particles to tumor cells in vivo by using transferrin receptor pathways. Magn. Reson. Med. 40, 236-242. (14) Bulte, J. W. M., Zhang, S.-C., van Gelderen, P., Herynek, V., Jordan, E. K., Duncan, I. D., and Frank, J. A. (1999) Neurotransplantation of magnetically labeled oligodendrocyte progenitors: magnetic resonance tracking of cell migration and myelination. Proc. Natl. Acad. Sci. U.S.A. 96 (26), 1525615261. (15) Winkler, M. A., Price J. O., Foglesong, P. D., and West, W. H. (1990) Biodistribution and plasma survival in mice of anti-melanoma monoclonal antibody cross-linked to OKT-3. Cancer Immunol., Immunother. 31 (5), 278-284. (16) Thorpe, P. E., Wallace, P. M., Knowles, P. P., Relf, M. G., Brown A. N. F., Watson, G. J., Knyba, R. E., Wawrzynczak, E. J., and Blakey, D. C. (1987) New coupling agents for the synthesis of immunotoxins containing a hindered disulfide bond with improved stability in vivo. Cancer Res. 47, 59245931. (17) Ramakrishnan, S., Fryxell, D., Mohonraj, D., Olson, M., and Li, B.-Y. (1992) Cytotoxic conjugates containing translational inhibitory proteins. Annu. Rev. Pharmacol. Toxicol. 32, 579-621. (18) Vitetta, E. S., Thorpe, P. E., and Uhr, J. W. (1993) Immunotoxins: magic bullets or misguided missiles? Immunol. Today 14, 252-259. (19) Masuho, Y., Kishida, K., Saito, M., Umemoto, N., and Hara, T. (1982) Importance of the antigen-binding valency and the nature of the cross-linking bond in ricin A-chain conjugates with antibody. J. Biochem. 91, 1583-1591. (20) Bickel, U., Kang, Y.-S., and Pardridge, W. M. (1995) In vivo cleavability of a disulfide-based chimeric opioid peptide in rat brain. Bioconjugate Chem. 6, 211-218. (21) Stein, S., Weiss, A., Adermann, K., Lazarovici, P., Hochman, J., and Wellho¨ner, H. (1999) A disulfide conjugate between anti-tetanus antibodies and HIV (37-72) Tat neutralizes tetanus toxin inside chromaffin cells. FEBS Lett. 458 (3), 383-386.

BC000079X