Synthesis and Characterization of a Gd-DOTA- D -Permeation Peptide for Magnetic Resonance Relaxation Enhancement of Intracellular Targets

Abstract

Many MR contrast agents have been developed and proven effective for extracellular nontargeted applications, but exploitation of intracellular MR contrast agents has been elusive due to the permeability barrier of the plasma membrane. Peptide transduction domains can circumvent this permeability barrier and deliver cargo molecules to the cell interior. Based upon enhanced cellular uptake of permeation peptides with ^D-amino acid residues, an all-D Tat basic domain peptide was conjugated to DOTA and chelated to gadolinium. Gd-DOTA-^D-Tat peptide in serum at room temperature showed a relaxivity of 7.94 ± 0.11 mM⁻¹ sec⁻¹ at 4.7 T. The peptide complex displayed no significant binding to serum proteins, was efficiently internalized by human Jurkat leukemia cells resulting in intracellular T1 relaxation enhancement, and in preliminary T1-weighted MRI experiments, significantly enhanced liver, kidney, and mesenteric signals.

Introduction

Contrast agents using paramagnetic metals such as gadolinium are routinely used in magnetic resonance imaging (MRI) to shorten longitudinal relaxation times (T1) of water protons and enhance image contrast [1]. Because gadolinum allows up to nine coordinate sites, chelating moieties are typically octadentate with either acyclic or macrocyclic polyaminopolycarboxylate ligands that form kinetically and thermodynamically stable complexes with gadolinium. The ninth coordination site of Gd (not occupied by the polydentate ligand) allows fast exchanging water molecules to transmit the paramagnetic relaxation effect to bulk solvent. To illustrate the stability of these metal-ligand complexes, the half-life of a (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid)gadolinium(III) [Gd-DOTA] complex is −4000 hr at pH 2 [2].

Clinical contrast agents are confined to extracellular spaces and distribute nonspecifically throughout vascular and interstitial tissues, enabling noninvasive imaging of perfusion and vascular abnormalities [1]. Other agents under development are responsive to environmental changes such as the decreased pH seen in extracellular spaces around tumors [3,4]. However, gene chip microarray analysis and related techniques have recently shown disease-specific gene expression patterns, resulting in biochemical changes long before gross morphological changes are detectable by nonspecific extracellular contrast agents. Thus, targeted MR contrast agents possessing increased sensitivity and specificity have been explored that bind to specific receptors or cell surface antigens [5–7], detect enhanced negative charges on the surface of tumor cells [8,9], or are internalized by receptor-mediated endocytosis [10] or pinocytosis [11,12]. In addition, contrast agents have been developed that demonstrate enhanced relaxivity in response to specific biochemical reactions. This enhanced relaxivity may arise from enzymatic activity either exposing a coordination site for water or creating a binding ligand with higher affinity [13,14], binding a target protein [15,16] or sensing changes in ion concentration [17,18]. However, the compounds employed in these strategies are limited by poor cell permeation properties. Thus, an efficient and generalizable method for delivery of targeted contrast agents into the cell interior would expand the realm of potential applications.

To circumvent the native cell membrane permeability barrier, peptide transduction domains have been explored to carry cargo molecules into cells. The basic domain of HIV-1 Tat protein has been widely used to translocate numerous different cargo molecules across cell membranes, including proteins, peptides, fluorophores, and radiopharmaceuticals [19–24]. As an example of MR contrast agent delivery, Tat basic domain has been used to mediate delivery of cross-linked iron oxides (CLIOs) into the cell interior [25,26]. CLIOs affect the transverse relaxation time (T2) by introducing large susceptibility effects, which alter the local magnetic field and result in signal loss where CLIOs are located. Although both T1 and T2 agents have value, contrast agents that shorten T1 generate signal enhancement and decrease overall imaging time by allowing use of shorter recycle delays. One previous report using DOTA conjugated to a Tat basic domain peptide containing L-amino acids showed the feasibility of rapid delivery of a T1 contrast agent intracellularly to mouse lymphocytes [27]. Recently, we reported that Tat basic domain peptides containing all D-amino acids conjugated to oxotechnetium chelates showed 10-fold enhanced uptake into human Jurkat cells compared with peptides comprising L-amino acid residues [19]. Thus, we sought to exploit the enhanced cellular uptake of D-amino acid sequences by synthesizing nonnative D-Tat basic domain peptides conjugated to DOTA and chelated with Gd for use as improved intracellular MR contrast agents.

Materials and Methods

Peptide Synthesis

A D-Tat basic domain peptide, Ac-arg-lys-lys-arg-arg-gln-arg-arg-arg-AHA-gly-lys-NH₂, containing residues 49–57 from HIV-1 Tat, was prepared by solid-phase peptide synthesis (Tufts University Peptide Synthesis Core Facility, Boston, MA) on amide resin using Fmoc-6-aminohexanoic acid (AHA) and D N-α-Fmoc-protected amino acids with standard BOP/HOBt coupling chemistry. All amino acids used standard side chain protecting groups, except for the C-terminal lysine residue, which contained a (4,4-dimethyl-2,6-dioxocyclohex-1-ylidene) ethyl (Dde) functionality protecting the e-amino group to allow orthogonal synthesis by selective de-protection of the Dde while the peptide was attached to the resin.

Synthesis of Gd-DOTA-D-Tat Peptide

The Dde protecting group was removed by suspending 20–50 mg of resin in 2% hydrazine in DMF (2 mL; 4 × 10 min), followed by washing with DMF (3 × 4 mL). The resin was stored at 0°C until needed. Thereafter, N-hydroxybenzotriazole (HOBt, 1.5 Eq) was added to a DMF solution (1 mL) of 1,4,7,10-tetraazacyclododecane-1,4,7-tris(acetic acid-t-butyl ester)-10-acetic acid (DOTA-tris(t-butyl ester)) (1.5 Eq; Macrocyclics, Richardson, TX) and stirred for 30 min under argon. The activated DOTA-tris(t-butyl ester) was transferred to a vial containing resin-bound, deprotected D-Tat peptide (1 Eq) suspended in dry DMF (200 μL), and treated with O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU, 1.5 Eq) and N,N-diisopropylethylamine (DIEA, 1.5 Eq). The reaction mixture was stirred for 4 hr under argon. The resin was subsequently washed with DMF (3 × 1 mL), water (1 mL), and acetonitrile (2 × 1 mL). Finally, the peptide was cleaved from resin and protecting groups were removed by stirring the resin in a peptide cleavage cocktail [TFA (2 mL), phenol (0.15 g), thioanisole (0.1 mL), deionized water (0.1 mL), and ethanedithiol (0.05 mL)] at room temperature for 3 hr. The cleaved and deprotected peptide was then filtered through glass wool to remove resin. The filtrate was added to ice-cold ether to generate a white precipitate that was pelleted by centrifugation, and the supernatant was decanted. The pellet was washed three times with ice-cold ether and residual solvent was removed under reduced pressure. Purification of crude product was accomplished using a reversed-phase C-18 column on HPLC at a flow rate of 1 mL/min using an eluent mixture of 0.1% TFA in 5% acetonitrile/95%water (Buffer A) and 0.1% TFA in 90% acetonitrile/10% water (Buffer B) with a linear gradient of 5% to 40% Buffer B over 40 min. The main fraction (retention time: 14.2 min) was collected and found to be DOTA-D-Tat peptide by electrospray ionization mass spectrometry (ESI-MS) (m/z: 2065–0, calcd: C₈5H₁₆₁N3₉O₂₁, 2064.3). For complexation, HPLC fractions containing pure DOTA-D-Tat peptide were lyophilized, dissolved in glycine buffer (50 mM; pH 3.5; 0.5 mL) and added to GdCl₃ dissolved in glycine buffer (5 Eq; 0.5 mL). This solution was heated to 80°C for 3 hr and purified using the HPLC conditions described above yielding pure Gd-DOTA-D-Tat peptide (retention time: 13.9 min; m/z: 2220.0, calcd: C₈₅H₁₅₈N₃₉O₂₁Gd, 2219.2) (Figure 1).

Figure 1.

Synthetic scheme for Gd-DOTA-D-Tat peptide and FITC-D-Tat peptide.

Synthesis of FITC-D-Tat Peptide

The Dde protecting group was removed from the C-terminal lysine residue of peptide bound to resin (30 mg) using 2% hydrazine in DMF as described above. The resin was sequentially washed with DMF (3 × 4 mL) and treated with a solution of DMF (0.5 mL), triethylamine (100 Eq), and fluorescein-5-isothiocyanate (3 Eq). This reaction was stirred at ambient temperature for 4 hr in the dark. The resulting FITC-labeled D-Tat peptide conjugate was deprotected and RP-HPLC purified using the eluent mixture described above. The main fraction (retention time: 30.9 min) was collected and identified (Ac-arg-lys-lys-arg-arg-gln-arg-arg-arg-AHA-gly-lys-(FITC) −NH₂; m/z: 2068.0, calcd: C₉₀H₁₄₆N₃₆O₁₉S, 2067.1) (Figure 1).

Transport Assays

Suspensions of human Jurkat leukemia cells were maintained in RPMI supplemented with L-glutamine, penicillin/streptomycin, and heat-inactivated fetal bovine serum at 37°C in an atmosphere of 5% CO₂ [22]. Assays were performed with a minor modification to a previously described procedure [22]. Briefly, transport assays were carried out with Jurkat cells incubated in modified Earl's balanced salt solution (MEBSS) containing (mM): 145 Na⁺, 5.4 K⁺, 1.2 Ca²⁺, 0.8 Mg²⁺, 152 Cl⁻, 0.8 H₂PO₄⁻, 0.8 SO₄^2–, 5.6 dextrose, 4.0 HEPES, and 1% fetal bovine serum (v/v), pH 7.4 ± 0.05. Experiments were performed with an MEBSS suspension of 10⁷ cells in siliconized microfuge tubes each containing either Gd-DOTA-D-Tat peptide (100 μM), Omniscan (gadolinium complex of diethylenetriamine pentaacetic acid bismethylamide, Amersham Health) (100 μM), or MEBSS alone in a total volume of 750 μL. The tubes were incubated at 37°C for 60 min with occasional mixing and the reaction terminated by pelleting the cells for 10 sec with a microcentrifuge. The cells were resuspended in 750 μL of 4°C MEBSS buffer, followed immediately by another rapid spin, and then the cell pellet was suspended in PBS (10 μL; pH 7.4) prior to dilution in a 0.6-mL microcentrifuge tube to 1% agarose using a stock solution of 1.3% agarose in PBS. This tube was placed inside a 1.5-mL microcentrifuge tube containing 1.3% agarose in PBS and used for MR spectroscopy and MR imaging experiments.

MR Spectroscopy and MR Imaging

All MR data were collected at room temperature in an Oxford Instruments 4.7-T magnet (33 cm clear bore) equipped with 16 cm inner-diameter, actively shielded, high-performance gradient coils (18 G/cm, 100 μsec rise time). The magnet/gradient coils are interfaced with a Varian (Palo Alto, CA) INOVA console. Proton spectroscopy and in vitro imaging experiments were performed with a 2-cm ID, linear RF Litz coil obtained from Doty Scientific (Columbia, SC), while in vivo imaging data were collected using a home-built 3-cm ID birdcage coil. T1 data for relaxivity determinations were obtained using a standard inversion recovery pulse sequence with 20–25 delay times, τ, ranging from 1 μsec to 5 × T1. In vitro single-slice T1-weighted spin-echo images (500/15, TR/TE) were acquired with a 1-mm slice thickness and a 2-cm FOV, while in vivo multislice gradient-echo images (100/2.2, TR/TE) were collected with a 1-mm slice thickness and 3 cm FOV.

Omniscan samples (0.1, 1.0, and 3–0 mM) were prepared in deionized water by serial dilution of the commercially available 500 mM stock solution. Samples in fetal calf serum were prepared by lyophilizing aliquots of these water-based samples and subsequent mixing with serum. A Gd-DOTA-D-Tat peptide stock solution (3 mM) was prepared by dissolving the appropriate weight of Gd-DOTA-D-Tat peptide in deionized water. Aliquots of appropriate stock solution were transferred to separate microcentrifuge tubes, lyophilized, and the dried peptide was then dissolved in either deionized water or fetal calf serum. Sample concentrations in water were independently determined to be 0.06, 0.56, and 1.65 mM by amino acid analysis. Individual T1 values were estimated by fitting signal intensity versus τ using t1s, a software program (Varian) that performs exponential curve fitting. Relaxivities of Omniscan and Gd-DOTA-D-Tat peptide in water and fetal calf serum were determined by fitting 1/T1 versus contrast agent concentration using standard linear regression methods.

Animal studies were conducted with approval of the Washington University Animal Studies Committee. Prior to in vivo imaging experiments, ICR mice (Harlan; n = 2) were anesthetized with isoflurane/O₂ and maintained on isoflurane/O₂ (2% v/v) throughout. An initial set of precontrast, multislice gradient-echo images were collected. Subsequently, Gd-DOTA-D-Tat peptide (0.06 mmol/kg) was administered through an intraperitoneal catheter and postcontrast, multislice gradient-echo images were collected every 10 min for 90 min.

Fluorescence Microscopy

Just prior to the study, FITC-D-Tat peptide or FITC solutions (100 μM) (freshly made or from stock solutions stored at –20°C) were diluted in RPMI supplemented with L-glutamine, penicillin/streptomycin, and heat-inactivated fetal bovine serum at 37°C to obtain 10 μM solutions. For cell uptake experiments, Jurkat cells at a concentration of 10⁵ cells/90 μL RPMI were directly added to 10 μL of the 10 μM stock solutions for a final FITC concentration of 1 μM. Following a 20-min incubation at 37°C, cells were cytospun onto slides, fixed for 10 min with paraformaldehyde (4%), and washed three times with PBS. All slides were then mounted with anti-fading mounting medium according to recommended procedures of the manufacturer (Vector) and analyzed by epifluorescent microscopy on a Zeiss microscope coupled to a Nikon camera equipped with a CCD interfaced to a PC [28,29].

Statistics

To determine the statistical significance of differences between intensity ratios obtained from single-slice T1-weighted spin-echo MRI experiments, an ANOVA analysis was performed followed by a t test with a Bonferroni adjusted alpha value for post hoc analysis.

Results

The dependence of 1/T1 on contrast agent concentration for Omniscan and Gd-DOTA-D-Tat peptide in water and fetal calf serum is shown in Figure 2. T1 relaxivity ± standard error of Omniscan increased from 4.01 ± 0.03 in water to 4.78 ± 0.09 mM⁻¹ sec⁻¹ in serum, while the relaxivity of Gd-DOTA-D-Tat peptide increased from 6.81 ± 0.02 to 7.94 ± 0.11 mM⁻¹ sec⁻¹, respectively. Parameters that most significantly impact relaxivity of a compound are the molecular correlation time (τ_c) and the number of coordinated water molecules (q) [30]. The relaxivity of a compound increases with either an increase in τ_c or q. However, increasing the q value leads to a corresponding decrease in chelate stability, so τ_c is the parameter most often modified when synthesizing contrast agents. Upon coordination, chelation of Gd involves the amide oxygen functionality in Gd-DOTA-D-Tat peptide, which replaces the carboxylate oxygen of the parent DOTA as a coordinating ligand, thereby leaving q unchanged [1,3,31–35]. Thus, the higher relaxivity of Gd-DOTA-D-Tat peptide relative to Omniscan likely results from an increase in τ_c Both compounds showed a similar percent increase in relaxivity values between water and calf serum (19% for Omniscan and 17% for Gd-DOTA-D-Tat peptide) and the relaxation rates remained a linear function of the contrast agent concentration. Because Omniscan does not bind to serum proteins [36–38], the increase in relaxivity of both compounds in serum over the value observed in water could be best attributed to the enhanced viscosity of calf serum increasing τ_c by slowing molecular rotation, not tight binding between Gd-DOTA-D-Tat peptide and serum proteins.

Figure 2.

Relaxivity of Omniscan and Gd-DOTA-D-Tat peptide in deionized water and fetal calf serum. The mean 1/T1 ± standard deviation (when larger than the symbol) of triplicate determinations is shown and relaxivity (mM⁻¹ sec⁻¹) was determined by linear regression: 4.01 ± 0.02 (Omniscan/water: •), 4.78 ± 0.09 (Omniscan fetal calf serum: ˆ), 6.81 ± 0.02 (Gd-DOTA-D-Tat peptide/water: ▪), and 7.94 ±0.11 (Gd-DOTA-D-Tat peptide/fetal calf serum: □).

To determine the presence of intracellular relaxivity changes, the behavior of Jurkat cells incubated with either Omniscan or Gd-DOTA-D-Tat peptide was monitored by MRI. A T1-weighted spin-echo imaging sequence was used to obtain a single slice through the various samples shown in Figure 3. Square regions of interest were manually drawn in both the inner circle and the outer ring, the mean voxel signal intensities were determined and the ratios of the mean signal intensity of the inner circle to the outer ring were calculated (I^in/out). The means and standard deviations (n = 3–4) were: (I^in/out)_agar = 1.07 ± 0.02, (I^in/out)_{Jurkat cells} = 1.05 ± 0.03, (I^in/out)_{Jurkat cells + Omniscan} = 1.03 ± 0.04, and (I^in/out)_{Jurkat cells + peptide} = 1.28 ± 0.10. There was no significant difference in the intensity ratios between the agar, Jurkat cells or Jurkat cells plus Omniscan samples, but (I^in/out)_{Jurkat cells + peptide} was significantly different from (I^in/out)_{Jurkat cells + omniscan} (p < .006).

Figure 3.

Representative T1-weighted MRI images. In all four images, the outer ring is 1.3% agarose and the inner circle is: (A) 1% agarose, (B) 10⁷ Jurkat cells in 1% agarose, (C) 10⁷ Jurkat cells in 1% agarose after incubation in buffer containing 100 μM Omniscan followed by a wash, and (D) 10⁷ Jurkat cells in 1% agarose after incubation in buffer containing 100 μM Gd-DOTA-D-Tat peptide followed by a wash.

The observed (I^in/out)_{Jurkat> cells + peptide} ratio was consistent with Tat basic domain peptide translocating Gd-chelate cargo molecules across cell membranes. To further evaluate permeability and intracellular localization of Gd-DOTA-D-Tat peptide, FITC was substituted for DOTA to create a fluorescent probe. While a control uptake experiment using FITC alone showed no fluorescence (Figure 4A), the FITC-D-Tat peptide conjugate was internalized and showed cytosolic and punctate nucleolar localization (Figure 4B) characteristic of other small Tat-peptide conjugates [22].

Figure 4.

Fluorescence microscopy showing the intracellular localization of FITC-D-Tat peptide. Jurkat cells were incubated in RPMI containing FITC (1 μM) alone (A) or FITC-D-Tat peptide (1 μM) (B) at 37°C for 20 min and then fixed. Magnification: ×40.

To demonstrate contrast enhancement in vivo, pilot MRI studies were performed in mice following intraperitoneal injection of Gd-DOTA-D-Tat peptide (0.06 mmol/kg). Consistent with previous biodistribution studies performed with ^99mTc-Tat-peptide conjugates [22] and Tat-peptide conjugates dual labeled with ^99mTc and fluorescein [20], enhancement of liver and kidney tissues was readily observed on T1-weighted gradient-echo images (Figure 5). However, at the concentration tested, both mice died within 90 min, an event not previously observed at the significantly lower concentrations required for radiotracer or optical imaging studies with similar ^99mTc or fluorescein tagged Tat-peptide conjugates.

Figure 5.

Single transaxial images from a multislice T1-weighted gradient-echo acquisition before (A) and I hour after intraperitoneal injection of 0.06 mmol/kg Gd-DOTA-D-Tat peptide (B). The postcontrast image shows signal enhancement in the kidneys (arrows), collecting systems and mesenteric structures.

Discussion

There exists significant interest in development of targeted MR contrast agents that show enhanced relaxivity in response to specific biochemical interactions. One of the best demonstrations of a biochemically responsive signal arising from an intracellular MR contrast agent is with the β-galactosidase activatible (1-(2-(β-galactopyranosyloxy)-propyl)-4,7,10-tris(carboxymethyl)-1,4,7,10-tetraazacyclododecane) gadolinium (III) complex (EgadMe) [13]. Gene expression of β-galactosidase was interrogated after microinjection of β-galactosidase mRNA and EgadMe into a Xenopus laevis embryo at the two-cell stage. Descendants of the injected cell showed increased relaxivity arising from enzymatic activity of β-galactosidase removing a galactopyranose group from EgadMe, thereby exposing a coordination site for water and increasing agent relaxivity. However, microinjection delivery limits the general applicability of this compound for studies in vivo. Thus, an efficient method for delivery of targeted MR contrast agents into the cell interior is desired. Herein, we characterize a nonnative D-Tat basic domain peptide conjugated to DOTA and chelated with Gd as a potential scaffold for membrane permeant MR contrast agents.

The relaxivity of the orthogonally labeled Gd-DOTA-D-Tat peptide was compared with the relaxivity of Omniscan at 4.7 T and room temperature. The relaxivity of Gd-DOTA-D-Tat peptide in water reported herein, 6.81 ± 0.02 mM⁻¹ sec⁻¹, differs from the value of 4.1 mM⁻¹ sec⁻¹ (1.5 T, 20°C) previously reported for a similar DOTA conjugated Tat peptide compound [27]. At high fields, the molecular correlation time (τ_c) is dominated by the rotational correlation time (τ_R); relaxivity for similar species is, therefore, a linear function of τ_R which is proportional to complex size and molecular weight [33,39,40]. Thus, similar relaxivities are expected for values determined at various high fields, as demonstrated by the similarity between the T1 relaxivity values for Omniscan that we observed and those reported (3–91 mM⁻¹ sec⁻¹, 1.2 T, 25°C [41]). However, regarding Gd-DOTA-D-Tat peptide, it is known that peptides obtained by lyophilization from HPLC buffers containing TFA are in the form of trifluoroacetic acid salts [42,43]. Assuming that the three lysines, six arginines, and the amidated C-terminus were ion paired with TFA, the actual mass of the Gd-DOTA-D-Tat peptide salt would be 45% more than nominally expected from the peptide sequence alone, which was consistent with the peptide concentrations that we determined by amino acid analysis being 40–45% lower than those calculated by weight. (Indeed, using peptide concentration calculated by weight and ignoring the TFA salt, the relaxivity of Gd-DOTA-D-Tat peptide in this study would have been measured as 3.8 mM⁻¹ sec⁻¹.) Overall, we conclude that the Gd-DOTA-D-Tat peptide had a higher relaxivity than Omniscan, consistent with an increase in rotational correlation time due to increased molecular weight and nonspecific interactions with serum proteins.

Cell-associated relaxivity enhancement in Jurkat cells using Gd-DOTA-D-Tat peptide was seen by MRI and the membrane transduction potential of an analogous orthogonally labeled fluorescent conjugate was directly demonstrated by fluorescence microscopy. A more elegant and direct demonstration of intracellular localization would be the use of a D-Tat peptide double labeled with FITC and Gd-DOTA [20]. However, one recently elucidated limitation on Tat peptide-mediated cargo delivery is the presence of a permeability barrier provided by differentiated epithelial cells [21]. Nonetheless, this study illustrated the feasibility of using a peptide transduction domain to overcome the permeablility barrier presented by most cellular membranes for the intracellular delivery of MR contrast agents into human cells.

It is interesting to note that although acute toxicity has not been previously observed with these peptides in rodent models of optical or radiotracer applications [20,22], a potential disadvantage for intracellular MR applications with this permeation peptide relates to the potential toxicity of the relaxivity agent in mice at the tested dose. Detection of MR relaxivity in vivo requires ˜10 to 100 times higher concentrations of conjugate than that required for fluorescence detection and 10⁶ times higher concentrations than that required for no-carrier-added radiotracers [22], and thus, the MR diagnostic window would appear narrow for this agent. These D-peptides show 10-fold enhanced uptake compared to L-peptides in cell culture [19] and thus, conventional whole-body dosing may not be predictive of the localized concentration of these MR contrast agents within cells, further complicating pharmacokinetic and dosing estimates. It is unclear from these pilot studies whether the apparent toxicity is related to high intracellular concentrations of the permeation peptide per se or to decomposition (transmetallation) reactions of the Gd-DOTA moiety within subcellular compartments (organelles) of critical organs in vivo. Although all preparations of Gd-DOTA-D-Tat peptide were HPLC purified prior to injection into animals, toxicities from trace metals in the injectate cannot be absolutely excluded [44] and the stability of these Gd-complexes within highly reducing or acidic compartments (lysosomes) of cells remains to be investigated.

As shown previously, a large fraction of cell-associated Tat peptide both nonspecifically accumulates and washes out of most cells rapidly [22], which may limit the time scale and content accessible for contrast-enhanced MRI. Indeed, under the washing conditions used with Jurkat cells in the present protocol, the residual compartment of retained Tat peptide (˜20% [20,22]) was the component likely detected within the cells by MRI. A logical future extension of this conjugate would be the introduction of a target-specific sequence between the Tat basic domain and the lysine residue derivatized with DOTA to form a selective membrane-permeable intracellular contrast agent. Various target sequences can be envisioned that could either bind to an intracellular protein of interest, be modified to reduce reverse membrane permeability, or be cleaved by an intracellular protease liberating the MR contrast moiety from the membrane-permeable protein sequence. These cell targeting or activating mechanisms would provide specificity, while concentrating the contrast agent in cells of interest, thus enhancing both their relaxivity and ease of detection. In addition, the trapped contrast moiety would show decreased washout kinetics from cells, allowing for the reduction of background nontarget contrast enhancement by allowing the extracellular pool of contrast agent to be cleared from the system by renal or hepatobiliary excretion.

In addition, once Gd-DOTA-D-Tat peptide has been modified with an appropriate targeting functionality to provide high target selectivity, the compound could potentially be used as a dual agent for MRI and gadolinium neutron capture therapy [45,46]. Therefore, biodistribution could be monitored using T1-weighted imaging sequences and the highly ionizing Auger electrons and γ-rays generated by neutron capture with ¹⁵⁷Gd (highly effective over short distances) would be capable of destroying labeled cells.