Over the last several years, great interest has developed in the potential use of carbon nanostructures (C₆₀ fullerenes and nanotubes) in medicine. In some cases, medical agents derived from these materials have demonstrated greater efficacy than existing clinical agents in many imaging and therapeutic applications. This chapter provides an overall review of the application of these materials in the area of magnetic resonance imaging (MRI), with an emphasis on their future applications in targeted MR molecular imaging for the early detection of cancer and other life-threatening diseases.

Introduction

Widely considered the birth of nanotechnology, the discovery of C₆₀, a new form of elemental carbon, by Curl, Kroto, and Smalley in 1985¹ marshaled in a new wave of research into nanoscale science. These new nanomaterials can be engineered to produce properties unattainable at the larger macro or smaller quantum levels, and they have been the stimulus behind a myriad of proposed nanotechnology applications, including energy storage,^2,3 electronics,^4,5 catalysis,⁶ and space travel.^7,8 Most recently, great strides have been made in the biotechnology sector for nanoscale materials whose proposed uses include: quantum dots as fluorescent-imaging agents,⁹ C₆₀ as drug scaffolds,¹⁰ gold nanoshells as cancer therapeutics,¹¹ and metallofullerenes as radiotracers and radiopharmaceuticals.¹² In this chapter, we discuss one of the more promising new areas of nanobiotechnology, carbon nanomaterials engineered as high-performance magnetic resonance imaging (MRI) contrast agents.

A fundamental aspect of clinical medicine is the imaging and diagnosis of a given malady. Before clinicians can begin treatment, they must first identify the area of concern (cancerous tumor, arterial blockage, ligament tear, etc.) and ascertain the extent of the ailment in order to devise treatment strategies. To this end, a variety of different imaging techniques have been developed, many within the last 10 years. One of the key techniques is MRI, for which the 2003 Nobel Prize in Medicine was awarded to Paul Lauterbur and Peter Mansfield for their development of the technique.

MRI is based on nuclear magnetic resonance, which measures the relaxation rate of water proton spins exposed to a magnetic field. About 35% of the 60 million annual MRI procedures utilize a chemical contrast agent (CA) to enhance the sensitivity of the MR image. Most clinical CAs are Gd³⁺-based due to the ion's 7 unpaired f-electrons (the greatest number of unpaired electrons for any element or ion) and their symmetrical distribution within the ion. Dozens of such CAs have been approved by the FDA, each consisting of a chelated Gd³⁺ compound that circulates in the blood until elimination via the kidney.

Current imaging methods are able to distinguish between neighboring tissues of different composition and anomalies within these tissues. However, when a tumor cannot be detected because of limitations of current imaging technology (limited to the detection of a large mass of tumorigenic cells), they can become metastatic and thereafter much more difficult to treat. To this end, current medical imaging research strives to image the most fundamental unit of life: the single cell. Known as molecular imaging, researchers seek to develop “smart CAs” that are molecularly targeted to diseased cells of interest (tumor cells, arterial plaque cells, etc.) so as to accumulate within targeted cells. This can lead to the detection of disease (via some abnormal cellular biochemistry) at its earliest stage when relatively few cells are present. Such capabilities would allow for earlier diagnosis of cancer or arterial blockage, leading to an increase in successful therapies and the preservation of more lives. Pursuant to these goals, current clinical MRI CAs have important limitations that engineered nanostructures should be able to overcome. Current clinical CAs are generally not targeted to sites of disease, are not readily internalized by cells, and have efficacies much too low to realize molecular imaging.

Engineered carbon nanostructure CAs based on C₆₀ fullerene and shortened single-walled carbon nanotubes (SWNTs), as shown in Figure 1, have the potential to become clinical molecular imaging CAs for the following reasons:

Figure 1

The carbon nanostructure-based MRI contrast agents: (A) Gd@C₆₀ gadofullerene and (B) Gd³⁺_n @US-tube gadonanotube.

• Recent research has demonstrated that this new class of CAs far outperforms conventional CAs, with relaxivities (efficacies) up to 100 times larger than current clinical CAs when internally loaded with Gd³⁺ ions.
• The exterior carbon sheaths of these nanostructures can be functionalized for targeting and biocompatibility without compromising the performance of the sequestered Gd³⁺ ions within. For example, a carbon-based nanostructure CA could be linked to an antibody or peptide that targets a specific cancer cell type.
• Because of their lipophilicities and very small size (1 nm diameters), these carbon nanostructures can readily cross cell membranes and accumulate within cells.

With their extremely high relaxivites and targeting potential, cell-internalized carbon nanostructure CAs are providing one of the first real opportunities to image single cells or small groups of cells, a feat that would revolutionize MRI in medicine.

A Primer in Magnetic Resonance Imaging (MRI)

Derived from the principles of nuclear magnetic resonance (NMR), clinical MRI uses a strong homogenous magnetic field to align the proton spins of water so as to measure the time it takes them to equilibrate, a process known as relaxation.¹³ Differences in relaxation times are perceived in the MRI image as varying shades of white and gray. Although water has a very slow relaxation time and is ubiquitous in living organisms, complex instrumentation and mathematical algorithms are able to filter the data and coarsely differentiate between neighboring tissues which have slightly different relaxation times.

Carbon Nanostructures as a New High-Performance Platform for MR Molecular Imaging 3. Because it is difficult to distinguish between similar tissue (e.g., a tumor within fat tissue), CAs are often used to enhance images. CAs operate by decreasing the relaxation time of the local water protons near the agent, thus allowing for better differentiation between areas of interest. Relaxation occurs via two mechanisms: inner-sphere, in which the water protons are relaxed by unpaired electron spins (spin-lattice relaxation, r₁), and outer-sphere, in which the relaxation is propagated through the lattice field (spin-spin relaxation, r₂). An r1-agent produces a positively-enhanced image (the image brightens), and it is usually a paramagnetic ion (Gd³⁺, Mn²⁺). An r₂-agent produces a negatively-enhanced image (the image darkens), and it is typically a superparamagnetic species (e.g., nanocrystalline Fe₃O₄). All things being equal, clinicians prefer r₁ agents as it is easier to distinguish features in a brightened image opposed to a darkened image.

Relaxivity is relatively simple to measure, although it requires sophisticated instrumentation. About 0.2 ml of sample is placed in an NMR tube and positioned in a relaxometer. Temperatures can be varied to determine characteristics of the system, however, the most useful medical information is acquired at 37°C (body temperature). A uniform magnetic field aligns the proton spins of the sample and a pulse inverts the magnetization vector. The time required for the magnetization vector to randomize is known as the relaxation time (T₁). Fortunately, water protons relax slowly, on the order of several seconds. Upon the addition of a CA, the relaxation time decreases to tens of milliseconds. Because the relaxation time is concentration dependent, mathematical adjustments are made to yield “relaxivity”, a standard by which all contrast agents are compared quantitatively for efficacy. The greater the relaxivity, the more efficacious a given contrast agent; units are mM^-1s^-1. Typical relaxivities for clinical agents at 37°C and clinical-field strengths (˜60 MHz or 1.5 Tesla) are approximately 4 mM^-1s^-1.^14,15

Fullerene(C₆₀)-Based Contrast Agents

Endohedrals are a unique class of fullerenes that have a single metal atom (or in some cases, up to three^16,17) entrapped inside the fullerene carbon cage. These materials are commonly referred to as metallofullerenes.¹⁸ Most of the lanthanide ions have been captured inside fullerene cages of 60, 70, 72, or 82 carbon atoms. These metallofullerenes are generated during fullerene synthesis by using carbon rods soaked in a solution of the appropriate metal salt before carbon-arc ablation. While several lanthanides give appreciable relaxivities, Gd³⁺ endohedrals, as expected, demonstrate the best performance and will be exclusively discussed here as ‘gadofullerenes’.

Gadofullerenes were initially attractive as possible CAs for two reasons: the fullerene cage acts as a perfect chelate, preventing Gd³⁺ leakage in vivo (clinical CAs demetallate in vivo to a minor extent), and initial measurements demonstrated enhanced relaxivity as compared to traditional CAs.^19,20

Initially, the C₆₀-based gadofullerenes were written-off as low-yield, inert polymeric materials, however, technological innovations overcame these limitations and now a wide-variety of C₆₀-functionalized gadofullerenes can be produced in large quantities.²¹ Two of the most studied water-soluble species are Gd@C₆₀[C(COOH)₂]₁₀ and Gd@C₆₀(OH)_x, which have relaxivities ranging from 20 mM^-1s^-1 to 100 mM^-1s^-1 at 60 MHz and 37 °C, respectively, or approximately 5-20 times greater than current clinical agents (˜4 mM^-1s^-1).^19,21

Initial studies exploring the proton relaxation processes of the gadofullerenes revealed that they behave unlike any classical contrast agent, making them especially interesting candidates for study. Because it is entrapped inside a carbon cage, the Gd³⁺ ion relaxes water protons via an outer-sphere mechanism since it has no direct contact with water molecules.¹⁷ O-relaxation studies indicate that the outer-sphere relaxation mechanism is ten times that of chelated Gd³⁺ compounds without inner-sphere water molecules (e.g., [Gd(TETA)]^- where H₄TETA = 1,4,8,11-tetraazacyclotetradecane-1,4,8,11 acetic acid), and it is believed that the C₆₀ carbon cage (formally a C₆₀^3- cage) itself plays an important role in the large relaxivity of the gadofullerenes.²¹ The surface of a gadofullerene is paramagnetic and approximately 200 Å^>2, resulting in a large outer-sphere relaxivity surmised to be caused by the simultaneous relaxation of many water protons on the cage surface.

A powerful tool for studying the factors influencing MRI contrast agent behavior is the nuclear magnetic relaxation dispersion (NMRD) profile, which measures the relaxivity as a function of magnetic field (0.01-400 MHz). Classic Solomon, Bloembergen, and Morgen (SBM) theory of relaxation^22-26 failed to model the NMRD profile of the gadofullerenes. However, using algorithms originally developed by Bertini and coworkers,²⁷ the NMRD behavior of gadofullerenes was successfully modeled as a slowly-tumbling structure in solution.²¹ A typical Gd₃₊ chelate's rotational tumbling time is in the picosecond regime, whereas the gadofullerenes tumble on the nanosecond scale. Cryo-TEM and dynamic light scattering (DLS) studies concluded that the slower-tumbling gadofullerenes exist as large aggregates, estimated to be about 100-200 nm in size, depending on the metallofullerene species in question.²⁸

In addition to determining the relaxivity maxima, evaluating the NMRD profile at several temperatures yields insight into the two separate processes that affect proton relaxivity, namely the proton exchange rate and the molecular reorientation (or tumbling) rate which have opposing temperature dependencies.²⁸ For the most part, the Gd@C₆₀(OH)_x gadofullerenes did not exhibit widely-varying relaxivities as a function of temperature, however the r₁ relaxivity of Gd@C₆₀[C(COOH)₂]₁₀ decreased with increasing temperature, indicating that slow proton exchange is not a limiting factor in relaxivity. This reaffirmed that slowly-tumbling aggregates are the source of the large relaxivities displayed by the gadofullerenes. The influence of pH on r₁ also demonstrated a remarkable dependency, with decreasing pHs resulting in increasing relaxivity. Relaxivites increased up to a factor of three as pH was lowered. Lowering the pH causes the aggregate size to increase, which slows the tumbling time to produce higher relaxivities until the gadofullerenes precipitate from solution around pH ˜3. In the case of Gd@C₆₀[C(COOH)₂]₁₀, DLS data determined the aggregate size to range from 70 to 700 nm as pH was lowered from 9 to 4.

Finally, high salt concentration breaks up gadofullerene aggregates, drastically reducing the relaxivity since the smaller aggregates tumble faster.²⁹ Interestingly, a 10 mM phosphate solution is far more effective than a 150 mM NaCl solution at breaking up the gadofullerene aggregates, indicating that this phenomenon is not solely linked to ionic strength. It is believed that the phosphate species are able to intercalate and hydrogen bond to the malonate or hydroxyl groups facilitating disaggregation of the gadofullerenes. DLS studies supported this disaggregate model.

Another interesting fullerene-based MRI CA of note is the tri-gadolinium nitride endohedral metallofullerene (Gd₃N@C₈₀).³⁰ Synthesized by substituting nitrogen gas for argon during the laser ablation process, three Gd³⁺ ions are bound to a central nitride ion to form a four-atom cluster that is encapsulated by a C₈₀ cage. Further study is underway with these exciting new compounds, but the production of derivatives for study is hampered by the relative poor chemical reactivity of the M₃N@C₈₀ metallofullerenes as compared to their Gd@C₆₀ brethren.³¹

Nanotube-Based Contrast Agents

Gadofullerenes have shown great promise as MRI CAs, but synthetic difficulties (low yields and problematic purification steps) coupled with prohibitively high synthetic cost, reduce the likelihood that such CA materials will become clinical agents. However, a related new Gd³⁺-based nanomaterial which employs ultra-short single-walled carbon nanotubes (US-tubes) has been recently reported (Fig. 1B).³² Not only does this CA platform outperform the gadofullerenes with respect to relaxivity (Fig. 2), but the synthetic limitations that the gadofullerenes present are no longer an issue since Gd³⁺-ion loading occurs post US-tube synthesis. Thus, the synthesis of the Gd³⁺_n@US-tube CAs is limited only by the ability to produce single-walled carbon nanotubes (SWNTs).

Figure 2

Comparative Nuclear Magnetic Resonance Dispersion (NMRD) profiles of the gadonanotubes, Gd@C₆₀(OH)_x, and a common clinical agent, Magnevist™.

To prepare the Gd³⁺_n@US-tube CAs, full-length nanotubes are first cut chemically via fluorination and pyrolysis under inert atmosphere, resulting in US-tubes with an average length of 40 nm.³³ The carbon sheath of the US-tubes can then be functionalized externally for water solubility and biocompatibility, while medically-interesting agents can be loaded internally.

Carbon Nanostructures as a New High-Performance Platform for MR Molecular Imaging 5. One consequence of cutting SWNTs by fluorination is the creation of defects in the sidewalls of US-tubes. This, in turn, serendipitously facilitates filling by Gd3+ ions which accumulate at the defects sites, with the nanotube acting as a template for Gd³⁺ loading. Interestingly, aqueous Gd³⁺ ions do not uniformly accumulate down the length of the US-tube, but rather exist in small 2-5 nm clusters of up to 10 Gd³⁺ ions sitting in the defect site. The Gd³⁺ n@US-tube gadonanotubes can then be suspended in a surfactant solution, 1% sodium dodecylbenzene sulfate (SDBS), for relaxivity measurements and electron microscope imaging.

The Gd³⁺ _n-clusters in the Gd³⁺ _n@US-tubes are clearly distinguished in both cryo-TEM and high-resolution TEM images since the electron-dense Gd³⁺-clusters appear as darkly-contrasted areas (Fig. 3).³² The high-resolution TEM was equipped with EDAX, an elemental analysis attachment, which confirmed that the dark areas tested positive for Gd, while the lighter areas did not (Fig. 3A). It is also worth noting that the cryo-TEM sample (Fig. 3B) was flash-frozen from surfactant-suspended gadonanotubes to give individual-appearing Gd³⁺_n@US-tubes, whereas in the high-resolution TEM sample (Fig. 3A), water (without surfactant) was removed by evaporation, which produced gadonanotube aggregation. For this reason, the cryo-TEM image is a more accurate reflection of the structure of the suspended gadonanotube material in solution.

Figure 3

An HRTEM image of aggregated gadonanotubes (A). The arrows indicate darkcontrast spots which tested positive for Gd³⁺ as measured by EDAX; lightly-contrasted areas tested negative for Gd³⁺. A Cryo-TEM image (B) establishes the presence of Gd³⁺_n clusters (more...)

So what is the functional consequence of the small Gd³⁺_n clusters within the gadonanotubes? By filling the small surface defect sites of US-tubes with only a few ions of Gd³⁺, a superparamagnetic material is created. Superconducting quantum interference device (SQUID) measurements show that not only are the gadonanotubes superparamgnetic, but that the system exists in a spin-glass state.³² This is characteristic of uncommon magnetically-frustrated environments, confirming the existence of nanoscalar superparamagnetic clusters, acting independently of one another. Thus, the gadonanotubes are the first superparamagnetic Gd³⁺-based system which results in never-before-seen performance in an MRI CA.

Comparing the NMRD profile of the gadonanotubes to that of the gadofullerenes, several contrasting features can be seen (Fig. 2). Most striking is the magnitude of the gadonanotube r₁ values, an astounding 174 mM^-1s^-1 per Gd³⁺ at 60 MHz (clinical field strengths), compared to ˜20-40 mM^-1s^-1 for gadofullerenes and 4 mM^-1s^-1 for Magnevist™, a common clinical MRI CA. Also, the shape of the curve for the gadonanotubes is quite different from that of the gadofullerenes and traditional clinical CAs. The shape of the gadonanoube curve cannot be modeled with either SBM theory or with a slowly-tumbling aggregate algorithm. This indicates that there are features of the gadonanotube system that have never been observed for an MRI CA. We currently speculate that this agent has significant inner-sphere relaxation because of the Gd³⁺ ions, as well as a significant outer-sphere contribution from the superparamagnetic character of the Gd³⁺_n-ion cluster. However, much work remains to be done to better understand this unique system.

A major difference between the gadonanotubes and the gadofullerenes is that aggregation does not seem to play as important a role in gadonanotube relaxation, as it does for the gadofullerenes. While pH does strongly affect the relaxivity of gadonanotubes, preliminary DLS data indicates that this is not due to an aggregation effect since the hydrodynamic radius of the gadonanotubes is fairly independent of pH.

Due to their exceptionally high performance and previously unobserved behavior, gadonanotube MRI CAs appear poised to make an important impact on MRI. At current clinical field strengths (20-60 MHz), gadonanotubes are roughly 40 times better CAs than their current clinical counterparts (170 mM^-1s^-1 vs. 4.0 mM^-1s^-1). Furthermore, at higher fields the relaxivity of the gadonanotubes stays fairly constant, in contrast to clinical CAs whose relaxivities decline as the magnetic field is increased. With the trend toward higher-field MRI instrumentation, this behavior is a significant advantage for gadonanotube-based CAs in the future. Perhaps the biggest impact may actually lie in the opposite direction, specifically, millitesla imaging. The astonishing relaxivities of the gadonanotubes at low fields, 635 mM^-1s^-1 for gadonanotubes vs. 7.0 mM^-1s^-1 for Magnevist™ (Fig. 2), will undoubtedly encourage the emerging area of millitesla imaging which requires greater CA relaxivities to overcome instrumentation signal-to-noise issues.

Molecular Targeting of Carbon Nanostructures

While gadofullerenes and gadonanotubes clearly demonstrate superior relaxivity properties, they still must be solubilized (without the aid of surfactant) and targeted to specific areas of interest in the body if they are to realize their full potential. An obvious advantage of these engineered carbon-based nanostructures over their inorganic nanostructure counterparts (e.g., iron oxide nanocrystals) is the potential for the facile derivatization of their carbon exterior. Gadofullerenes and gadonanotubes can be functionalized for biocompatibility, water-solubility, and molecular targeting. Such functionalization can link the carbon nanostructure and its cargo to a variety of different targeting moieties that may include peptides, steroids, antibodies, or viral vectors.

While seemingly a simple process in design, the strong innate bundling of nanotubes to one another is a property often overlooked. Due to the π-conjugated electronic structure of carbon nanotubes, the π-π interaction between different tubes results in aggregated bundles with van der Waals energies of approximately 0.5 eV per nanometer of tube contact.³⁴ A single, full-length nanotube can be isolated with ultrasonication and polymer or surfactant coating.^35,36 It is believed that the full-length tubes are long enough to permit structural bending, eventually allowing for them to be peeled free from the bundle. However, this strategy is not possible with US-tubes because their small length:width ratio (relative to full-length SWNTs) results in rigid-rod behavior, preventing separation.

In order to obtain single-molecule nanotube “capsules” derived from US-tubes, the tubes must first be separated from one another and quickly derivatized to prevent reaggregation. The π-π system responsible for their extreme bundling can also be exploited to overcome the aggregation. The π-conjugated system of a nanotube has been shown to accept and stabilize a large number of electrons (1 electron per 10 carbon atoms).^37,38 Taking advantage of this behavior, we have chemically reduced US-tubes by an amended Birch reduction method (Na°/THF reduction), creating single US-tubes salt complexes in tetrahydrofuran, which are stable for well over 10 days. The extent of exfoliation can be determined by spin-coating the samples onto a mica surface and measuring their height, and therefore, bundle thickness, by atomic force microscopy (AFM). AFM measurements of the reduced US-tubes yield z-heights (tube diameters) of 0.5-1.5 nm, far smaller than for bundled US-tubes which have z-heights in excess of 7 nm. Since a typical HiPco-produced, single-wall nanotube has an average diameter of 1.0 nm (although it may vary from 0.5 nm to 2.0 nm³⁹) an AFM height measurement of ˜1.0 nm indicates that complete exfoliation of the nanotubes is achieved by this method.

Individualized US-tubes are then easily functionalized via the Bingel reaction.^40,41 Briefly, a bromomalonate is deprotonated by a strong base, creating a nucleophile that readily attacks the fullerene or nanotube conjugated system, resulting in formation of a cyclopropyl group on the surface. The Bingel group is both chemically and biologically stable and easily allows for subsequent derivatization of esters or amides to produce any desired targeting moiety as shown schematically in Fig. 4 for the case of the RGD targeting peptide, which specifically targets the α_vβ₃ integrin receptor upregulated in breast cancers.⁴² Several analytical techniques, including NMR can confirm covalent attachment of the malonate group, and x-ray photoelectron spectroscopy (XPS), an elemental analysis technique, has been used to establish that approximately 5 malonate groups per nanometer of US-tube can be attached. AFM indicated that once functionalized in this manner, US-tubes do not flocculate and reform bundles, but exist as individual functionalized US-tubes. The attainment of “single-molecule” nanotube building block of the nature shown in Fig. 4 is an important milestone for the continued development of carbon nanotube biotechnology.

Figure 4

A reaction scheme for attaching the RGD peptide to a gadonanotube.

Fullerene-Antibody Conjugates

Using similar Bingel functionalization strategies, the first fullerene(C₆₀)-antibody immunoconjugates have been recently produced.⁴³ Antibodies are a particularly promising class of cellular-targeting agents.^44,45 They are large, specialized proteins (MW >100,000 Daltons) that contain antigen binding sites which preferentially bind to a specific antigen that is over-expressed on a given cell type, allowing for cell-specific targeting. Recently, our group engineered the first fullerene(C₆₀)-antibody conjugate using a murine antibody that specifically targets melanoma cancer cells.⁴³

To produce the fullerene(C₆₀)-antibody immunoconjugate, two unique Bingel-derivatized fullerenes were investigated—one that can form a covalent disulfide bond with the antibody (C₆₀-SPDP) and one that cannot form covalent bonds with the antibody (C₆₀-Ser).⁴³ These C₆₀ derivatives are shown in Fig. 5. Amazingly, the results of this study indicated that fullerene(C₆₀)-antibody covalent bond conjugation may not be requisite for immunoconjugate formation. In fact, the C₆₀-Ser derivative loaded into the antibody to a greater extent (on average, 42 C₆₀'s derivatives per antibody) than the C₆₀-SPDP derivative (on average, 15 C₆₀'s derivatives per antibody). These C₆₀:antibody ratios were obtained from the UV-Vis spectroscopic signatures of the C₆₀ derivatives and Biorad protein assays for the antibody. TEM images of the C₆₀-Ser immunoconjugates and the antibody-only control are shown in Fig. 6. From the image, a contrast between the unloaded antibody (Fig. 6A) with the C₆₀-SPDP-loaded antibody (Fig. 6B) is observed, where the loaded antibody appears to have tripled in size and exhibits small dark clusters (˜5 nm) believed to be small fullerene aggregates inside the antibody. Enzyme-linked immunoabsorbant assay (ELISA) determined that the C₆₀-Ser-antibody is still 70 times more specific than a generic murine antibody. For the C₆₀-SPDP -antibody derivative, there is essentially no loss in specificity and binding efficiency as compared to an unconjugated antibody. This data suggests that the C₆₀ moieties are internalized by the antibody and that the antibody can be highly loaded with a payload of ˜15 fullerene nanostructures and still retain its biological function. It is believed that no chemical attachment is needed to link the fullerene to the antibody since the water-soluble, lipophillic fullerene derivative, C₆₀-Ser, is readily absorbed into the antibody interior.

Figure 5

The water-soluble, Bingel-derivatized fullerene (C₆₀) derivatives. The C₆₀-SPDP derivative (A) was designed to form a covalent attachment through the disulfide linker, whereas the C₆₀-Ser derivative (B) cannot form a covalent attachment with the antibody. (more...)

Figure 6

TEM images of an unloaded melanoma antibody (A) and a C₆₀-Ser loaded melanoma antibody (B), at identical magnification (scale bar = 20 nm). The fullerene derivative is readily taken up by the antibody, resulting in increased size as can easily be seen. (more...)

Since gadofullerenes are very similar materials to C₆₀, it is logical to assume that they will also load into an antibody. Experiments are currently underway in our laboratory to determine proton relaxivity properties of gadofullerene-antibody conjugates and to attempt targeted, MR imaging of melanoma cells dispersed in a field of normal cells. Perhaps the greatest advantage of carbon-based nanostructures for molecular targeting is the ability of these materials to cross cell membranes,^46,47 resulting in the accumulation of deliverable payloads inside cells, whether for molecular imaging or guided therapy.

Closing Remarks

Clearly the future of carbon-based nanostructures in MRI is bright. In only a few short years, these materials have been shown to dramatically outperform classical MRI CAs in vitro, and they have the potential to revolutionize the field of medical MR imaging. They could compliment (or even replace) current MRI CAs because of their superior relaxivities or they may encourage the development of new techniques of the future, such as millitesla imaging. Clearly, the most exciting frontier for these carbon nanostructures is their potential for overcoming many of the limitations of classical MRI CAs and thus, to produce new agents for molecular imaging. In particular, they are high-performance nanoscalar MRI probes which are intrinsically intracellular and chemically very versatile. A worthy long-term goal is their development as a new universal platform, based on single-molecule US-tubes as disease-targeted carbon nanocapsules, for the containment and delivery of an array of diagnostic and therapeutic agents in medicine.

Acknowledgements

The research at Rice University was supported by the Robert A. Welch Foundation (Grant C-0627), the U.S. National Institutes of Health (Grant NIH R01 EB000703), the NASA Johnson Space Center and University of Texas Health Science Center (Grant NNJ05HE75A). KBH is supported by a Rice University Center for Biological and Environmental Nanotechnology Fellowship (EEC-0118007).

References

1.

Kroto HW, Heath JR, O'Brien SC. et al. C₆₀: Buckminsterfullerene. Nature. 1985;318(6042):162–163.

2.

Arico AS, Bruce P, Scrosati B. et al. Nanostructured materials for advanced energy conversion and storage devices. Nature Materials. 2005;4(5):366–377. [PubMed: 15867920]

3.

Gomez-Romero P, Cuentas-Gallegos K, Lira-Cantu M. et al. Hybrid nanocomposite materials for energy storage and conversion applications. J Mat Sci. 2005;40(6):1423–1428.

4.

Seminario JM. Molecular electronics. Approaching reality. Nature Materials. 2005;4(2):111–113. [PubMed: 15689945]

5.

Balandin AA. Nanophononics: Phonon engineering in nanostructures and nanodevices. J Nanosci Nanotechnol. 2005;5(7):1015–1022. [PubMed: 16108421]

6.

Johnson BFG. Nanoparticles in Catalysis. Top Catalysis. 2003;24(1-4):147–159.

7.

Arepalli S, Fireman H, Huffman C. et al. Carbon-nanotube-based electrochemical double-layer capacitor technologies for spaceflight applications. J Mat. 2005;57(12):26–31.

8.

Sharma SK, Kumar R, Dolia SN. et al. Magnetic nanoparticles for space applications. Materials Research Society Symposium Proceedings. 2005;851:481–486.

9.

Alexson D, Chen H, Cho M. et al. Semiconductor nanostructures in biological applications. J Phys: Condensed Matter. 2005;17(26):R637–R656.

10.

Zakharian TY, Seryshev A, Sitharaman B. et al. A Fullerene-paclitaxel chemotherapeutic: Synthesis, characterization, and study of biological activity in tissue culture. J Am Chem Soc. 2005;127(36):12508–12509. [PubMed: 16144396]

11.

Hirsch LR, Stafford RJ, Bankson JA. et al. Nanoshell-mediated near-infrared thermal therapy of tumors under magnetic resonance guidance. PNAS. 2003;100(23):13549–13554. [PMC free article: PMC263851] [PubMed: 14597719]

12.

Cagle DW, Kennel SJ, Mirzadeh S. et al. In vivo studies of fullerene-based materials using endohedral metallofullerene radiotracers. PNAS. 1999;96(9):5182–5187. [PMC free article: PMC21838] [PubMed: 10220440]

13.

In: Merbach AE, Toth E,eds. The Chemistry of Contrast Agents in Medical Magnetic Resonance Imaging. Chichester: John Wiley and Sons. 2001

14.

Muller RN, Vander Elst L, Rinck PA. et al. The importance of nuclear magnetic relaxation dispersion (NMRD) profiles in MRI contrast media development. Inv Rad. 1988;23(Suppl 1):S229–231. [PubMed: 3198350]

15.

Lauffer RB. Paramagnetic metal complexes as water proton relaxation agents for NMR imaging: Theory and design. Chem Rev. 1987;87:901–927.

16.

Stevenson S, Fowler PW, Heine T. et al. A stable nonclassical metallofullerene family. Nature. 2000;408(6811):427–428. [PubMed: 11100715]

17.

Stevenson S, Rice G, Glass T. et al. Small-bandgap endohedral metallofullerenes in high yield and purity. Nature. 1999;401(6748):55–57.

18.

Cioslowski J, Fleischmann ED. Endohedral complexes: Atoms and ions inside the carbon sixty-atom molecule (C₆₀.

19.

Bolskar RD, Benedetto AF, Husebo LO. et al. First soluble M@C₆₀ derivatives provide enhanced access to metallofullerenes and permit in vivo evaluation of Gd@C₆₀[C(COOH)₂]₁₀ as a MRI contrast agent. J Am Chem Soc. 2003;125(18):5471–5478. [PubMed: 12720461]

20.

Mikawa M, Kato H, Okumura M. et al. Paramagnetic water-soluble metallofullerenes having the highest relaxivity for MRI contrast agents. Bioconj Chem. 2001;12(4):510–514. [PubMed: 11459454]

21.

Toth E, Bolskar RD, Borel A. et al. Water-soluble gadofullerenes: Toward high-relaxivity, pH-responsive MRI contrast agents. J Am Chem Soc. 2005;127(2):799–805. [PubMed: 15643906]

22.

Bloembergen N, Morgan LO. Proton relaxation times in paramagnetic solutions effects of electron spin relaxation. J Chem Phys. 1961;34:842–850.

23.

Bloembergen N. Proton relaxation times in paramagnetic solutions. J Chem Phys. 1957;27:572–573.

24.

Solomon I, Bloembergen N. Nuclear magnetic interactions in the HF molecule. J Chem Phys. 1956;25:261–266.

25.

Solomon I. Relaxation processes in a system of two spins. Phys Rev. 1955;99:559–565.

26.

Bloembergen N, Purcell EM, Pound RV. Relaxation effects in nuclear magnetic resonance absorption. Phys Rev. 1948;73:679–712.

27.

Bertini I, Galas O, Luchinat C. et al. A computer program for the calculation of paramagnetic enhancements of nuclear-relaxation rates in slowly rotating systems. J Mag Res A. 1995;113(2):151–158.

28.

Sitharaman B, Bolskar RD, Rusakova I. et al. Gd@C₆₀[C(COOH)₂]₁₀ and Gd@C₆₀(OH)_x: Nanoscale aggregation studies of two metallofullerene MRI contrast agents in aqueous solution. Nano Lett. 2004;4(12):2373–2378.

29.

Laus S, Sitharaman B, Tóth É. et al. Destroying gadofullerene aggregates by salt addition in aqueous solution of Gd@C₆₀(OH)_x and Gd@C₆₀[C(COOH₂)]₁₀. J Am Chem Soc. 2005;127(26):9368–9369. [PMC free article: PMC2597542] [PubMed: 15984854]

30.

Stevenson S, Phillips JP, Reid JE. et al. Pyramidalization of Gd₃N inside a C₈₀ cage. Chem Commun. 2004;(24):2814–2815. [PubMed: 15599420]

31.

Stevenson S, Stephen RR, Amos TM. et al. Synthesis and purification of a metallic nitride fullerene bisadduct: Exploring the reactivity of Gd₃N@C₈₀. J Am Chem Soc. 2005;127(37):12776–12777. [PubMed: 16159252]

32.

Sitharaman B, Kissell KR, Hartman KB. et al. Superparamagnetic gadonanotubes are high-performance MRI contrast agents. Chem Commun. 2005;(31):3915–3917. [PubMed: 16075070]

33.

Gu Z, Peng H, Hauge RH. et al. Cutting single-wall carbon nanotubes through fluorination. Nano Lett. 2002;2(9):1009–1013.

34.

Girifalco LA, Hodak M, Lee RS. Carbon nanotubes, buckyballs, ropes, and a universal graphitic potential. Phys Rev B. 2000;62(19):13104–13110.

35.

Bachilo SM, Strano MS, Kittrell C. et al. Structure-assigned optical spectra of single-walled carbon nanotubes. Science. 2002;298(5602):2361–2366. [PubMed: 12459549]

36.

Moore VC, Strano MS, Haroz EH. et al. Individually suspended single-walled carbon nanotubes in various surfactants. Nano Lett. 2003;3(10):1379–1382.

37.

Penicaud A, Poulin P, Derre A. et al. Spontaneous dissolution of a single-wall carbon nanotube salt. J Am Chem Soc. 2005;127(1):8–9. [PubMed: 15631422]

38.

Sauvajol JL, Bendiab N, Anglaret E. et al. Phonons in alkali-doped single-wall carbon nanotube bundles. Comptes Rendus Physique. 2003;4(9):1035–1045.

39.

Kim UJ, Gutierrez HR, Kim JP. et al. Effect of the tube diameter distribution on the high-temperature structural modification of bundled single-walled carbon nanotubes. J Phys Chem B. 2005;109(49):23358–23365. [PubMed: 16375307]

40.

Worsley KA, Moonoosawmy KR, Kruse P. Long-range periodicity in carbon nanotube sidewall functionalization. Nano Lett. 2004;4(8):1541–1546.

41.

Coleman KS, Bailey SR, Fogden S. et al. Functionalization of single-walled carbon nanotubes via the bingel reaction. J Am Chem Soc. 2003;125(29):8722–8723. [PubMed: 12862456]

42.

Chen X, Conti PS, Moats RA. In vivo near-infrared fluorescence imaging of integrin α_vβ₃ in brain tumor xenografts. Can Res. 2004;64(21):8009–8014. [PubMed: 15520209]

43.

Ashcroft JM, Tsyboulski DA, Hartman KB. et al. Fullerene (C₆₀) immunoconjugates: Interaction of water-soluble C₆₀ derivatives with the murine anti-gp240 melanoma antibody. Chem Commun. 2006;3004-3006 [PubMed: 16832518]

44.

Goodwin DA. Strategies for antibody targeting. Antibody, Immunoconjugates, and Radiopharmaceuticals. 1991;(4):4. 427–434.

45.

Courtenay-Luck NS, Epenetos AA. Targeting of monoclonal antibodies to tumors. Curr Opin Immunol. 1990;2(6):880–883. [PubMed: 2486568]

46.

Pantarotto D, Briand JP, Prato M. et al. Translocation of bioactive peptides across cell membranes by carbon nanotubes. Chem Commun. 2004;(1):16–17. [PubMed: 14737310]

47.

Kam NWS, Jessop TC, Wender PA. et al. Nanotube molecular transporters: Internalization of carbon nanotube-protein conjugates into mammalian cells. J Am Chem Soc. 2004;126(22):6850–6851. [PubMed: 15174838]