Paramagnetic Clusters of Mn₃(O₂CCH₃)₆(Bpy)₂ in Polyacrylamide Nanobeads as a New Design Approach to a T₁−T₂ Multimodal Magnetic Resonance Imaging Contrast Agent

Abstract

There is an increasing need for gadolinium-free magnetic resonance imaging (MRI) contrast agents, particularly for patients suffering from chronic kidney disease. Using a cluster–nanocarrier combination, we have identified a novel approach to the design of biomedical nanomaterials and report here the criteria for the cluster and the nanocarrier and the advantages of this combination. We have investigated the relaxivity of the following manganese oxo clusters: the parent cluster Mn₃(O₂CCH₃)₆(Bpy)₂ (1) where Bpy = 2,2′-bipyridine and three new analogs, Mn₃(O₂CC₆H₄CH=CH₂)₆(Bpy)₂ (2), Mn₃(O₂CC(CH₃)=CH₂)₆(Bpy)₂ (3), and Mn₃O(O₂CCH₃)₆(Pyr)₂ (4) where Pyr = pyridine. The parent cluster, Mn₃(O₂CCH₃)₆(Bpy)₂ (1), had impressive relaxivity (r₁ = 6.9 mM⁻¹ s⁻¹, r₂ = 125 mM⁻¹ s⁻¹) and was found to be the most amenable for the synthesis of cluster-nanocarrier nanobeads. Using the inverse miniemulsion polymerization technique (1) in combination with the hydrophilic monomer acrylamide, we synthesized nanobeads (~125 nm diameter) with homogeneously dispersed clusters within the polyacrylamide matrix (termed Mn₃Bpy-PAm). The nanobeads were surface-modified by co-polymerization with an amine-functionalized monomer. This enabled various postsynthetic modifications, for example, to attach a near-IR dye, Cyanine7, as well as a targeting agent. When evaluated as a potential multimodal MRI contrast agent, high relaxivity and contrast were observed with r₁ = 54.4 mM⁻¹ s⁻¹ and r₂ = 144 mM⁻¹ s⁻¹, surpassing T₁ relaxivity of clinically used Gd-DTPA chelates as well as comparable T₂ relaxivity to iron oxide microspheres. Physicochemical properties, cellular uptake, and impacts on cell viability were also investigated.

Graphical Abstract

INTRODUCTION

Gadolinium chelate contrast agents for magnetic resonance imaging (MRI), used in approximately 35% of MRI scans, are established in both the clinic and the research community as the “gold standard” of contrast agents.¹ However, there is a growing need for non-gadolinium-based contrast agents for patient safety, particularly for those with compromised renal function.² These patients are excluded from the administration of gadolinium agents because of the high risk of nephrogenic systemic fibrosis, a potentially lethal fibrosis of skin and organs.³ The statistics from the National Kidney Foundation suggests that kidney disease affects over 26 million adults in the United States.⁴ Chronic kidney disease and the gradual loss of kidney function are often associated with diabetes, a rapidly growing health problem, thus these numbers are anticipated to rise. In addition, there is evidence that some patients accumulate gadolinium despite having normal renal function,⁵ with both linear and to some extent macrocyclic gadolinium chelate agents linked to metal accumulation.⁶ Finally, a future technological limitation is that gadolinium has reduced brightness at higher scanning fields,⁷ which are growing in clinical use for their increased spatial resolution and decreased acquisition times.

Manganese is a promising alternative to gadolinium chelates because of its strong T₁ relaxivity.⁸ For patients with compromised kidney function, manganese has the advantage that it is cleared primarily via hepatobiliary excretion rather than exclusively through glomerular filtration.⁹ Despite these advantages chelates such as [Mn(DPDP)]³⁻, (commercially known as Teslascan), have been removed from the market because of the lability of the metal,¹⁰ and high levels of free manganese can cause neurological disorders.¹¹ Successful approaches to developing thermodynamically and kinetically stable manganese-based chelates include both linear and macrocyclic ligands,¹² which have led to reports of several promising chelates.^13–15 The relaxivity values are competitive with gadolinium chelates, and they are likely to have similar pharmacokinetic and pharmacodynamic behaviors. However, the in vivo stability of these novel chelates is yet to be established.

Chelates in general have limitations as contrast agents as they lack sufficient contrast capabilities and because of the long circulation times required in applications such as oncological imaging.¹⁶ There has been a surge in studies of nanomaterials, which are capable of delivering a multimetal agent, with higher contrast and longer circulation times, as well as providing a large surface area for conjugation of targeting molecules.¹⁷ Where chelates equilibrate between the intravascular and extravascular spaces leading to nonspecific biodistribution and rapid renal excretion, size restricts nanoparticles (>5 nm) from doing so.¹⁸ For example, ferumoxytol (iron oxide nanoparticles) a T₂ contrast agent and therapeutic iron supplement has a long intravascular half-life (14–15 h).¹⁹ This increased circulation makes it feasible for nanoparticles to act as blood-pooling agents, as well as opportunities to enhance their tumor accumulation. Nanoparticles present advantages for multifunctional approaches including multimodal imaging, drug, and gene delivery.²⁰ In addition, nanoparticles have been used successfully for cell tracking, which may have a significant impact on preclinical studies.²¹

Nanoparticles have fundamentally different pharmacokinetics from chelates, and as a result, they have drawn interest for imaging tumor vascularization and angiogenesis.²² One of the most studied classes of nanoparticles for these applications, because of their low toxicity, are iron oxide nanomaterials for T₂ contrast. Several groups have also targeted ultrasmall iron oxide nanoparticles for T₁ contrast because of the well-established size effects on the relaxivity of iron oxide nanoparticles,²⁰ where the T₂ decreases more rapidly with size than T₁ does. While it is possible to bring the r₂/r₁ within the range for T₁-weighted imaging (~10), unfortunately, as the size decreases the relaxivity values are also notably diminished.²³ One of the challenges shared by other inorganic core nanoparticle materials is that the coatings used to reduce metal leaching and to stabilize the core also prevent the aqueous proton interaction required for efficient T₁ relaxivity.²⁴

Our approach to developing a nanoparticle platform for MRI is distinctively centered on a “cluster-nanocarrier” design. Clusters are underexplored largely because of their tendency to speciate, as observed for the high relaxing iron oxo cluster Fe₈, which has consequences on the relaxivity measured.²⁵ Nanocarriers can stabilize clusters preventing both speciation and metal leaching. The clusters we have investigated are high-spin, paramagnetic manganese, and iron metal-oxo clusters with carboxylate ligands. The role of the nanocarrier is similar to that of the chelate in that it controls the local environment of the paramagnetic center and interfaces with the biological environment. We have focused on polymeric nanocarriers because of the extensive literature on the design of polymer nanoparticles for biomedical applications.^26,27

Here, we report a new cluster-nanocarrier system with ultra-bright T₁ contrast as well as appreciable T₂ contrast, and a modifiable surface chemistry, as a potential MRI contrast agent. We characterized the relaxivity of four clusters: Mn₃(O₂CCH₃)₆(Bpy)₂ (Bpy = bipyridyl) (1) or Mn₃Bpy, Mn₃(O₂CC₆H₄CH=CH₂)₆(Bpy)₂ (2) or Mn₃BpyVBA, Mn₃(O₂CC(CH₃)=CH₂)₆(Bpy)₂ (3) or Mn₃BpyMA and Mn₃O(O₂CCH₃)₆(Pyr)₂ (4) or Mn₃Pyr, and we structurally characterized (2) and (3). Despite the lack of polymerizable ligands, only the cluster (1), simply called Mn₃Bpy, was used in the cluster-nanocarrier synthesis because of the weaker solubility of the other three clusters (2–4) in the monomer. The miniemulsion polymerization method requires precise matching of the cluster solubility with the monomer of the nanocarrier. This cluster in combination with the hydrophilic polymer, polyacrylamide, forms a new cluster-nanocarrier, termed Mn₃Bpy-PAm. The Mn₃Bpy-PAm nanobeads prepared by inverse miniemulsion polymerization were monodispersed, 125 nm in diameter, and exhibited ultrahigh relaxivity. We describe the synthesis and characterization of the clusters and the physicochemical properties of the cluster–nanocarrier. The cytotoxicity of the cluster-nanocarrier was measured, hemolytic properties studied, and because of the importance of metal leaching for toxicity of manganese, several approaches to confirm the chemical stability of the nanobeads were also investigated. Finally, we also demonstrate the modularity to form multifunctional nanobeads through surface modification of amine groups for labeling with a near-IR dye, and for attaching the targeting and therapeutic agent methotrexate.

RESULTS AND DISCUSSION

In our first effort to prepare a cluster-nanocarrier for MRI, we investigated the single-molecule magnet, Mn₁₂O₁₂(O₂CCH₃)₁₆·4H₂O,²⁸ and the improved iron-substituted version, Mn₈Fe₄O₁₂(O₂CCH₃)₁₆·4H₂O,²⁹ or Mn₈Fe₄, which has a higher spin state and a greater aqueous stability. The Mn₈Fe₄ cluster has strong T₂ relaxivity properties (r₁ = 2.37 mM⁻¹ s⁻¹, r₂ = 27.74 mM⁻¹ s⁻¹), and the ability to undergo ligand exchange with polymerizable carboxylates such as vinyl benzoic acid (VBA), while maintaining the core structure. Using direct miniemulsion polymerization conditions, we synthesized homogenous 50–80 nm beads of Mn₈Fe₄(VBA)₁₆ co-polymerized with styrene, to form Mn₈Fe_4-co-PS.³⁰ Polystyrene, a hydrophobic polymer, is often used as a noncovalent coating for nanobeads. Importantly, the r₁ for Mn₈Fe₄-co-PS increased to 3.37 mM⁻¹ s⁻¹ while the r₂ decreased to 11.2 mM⁻¹ s⁻¹, making it possible for the nanobeads to support T₁ weighted imaging. While the relaxivity changes observed are of interest for academic reasons, the ability to act as a positive contrast agent is highly preferred over negative contrast agents for clinical applications.

At first consideration, the relaxivity of the cluster alone, Mn₈Fe₄, is not particularly impressive, the r₁ is less than gadolinium complexes. However, Mn₈Fe₄ does not have the optimum spin state (S = 5/2) for manganese. Despite the large molecular spin state, the core (Mn^IV₄Mn^III₄Fe^III₄), has four manganese atoms with S = 3/2 (Mn^IV), and four atoms with S = 2 (Mn^III), while only the four Fe^III (S = 5/2) have the maximum transition-metal spin state, d⁵. Typically, both Mn(III) and Fe(III) have significantly lower relaxivity versus Mn(II), so it is interesting that this cluster has such a large r₁ (per metal) value. Building on the insights gained from the Mn₈Fe₄ prototype system to identify new high-relaxivity clusters, we searched for reported clusters with five or fewer metals (to avoid solution speciation), focusing on Mn(II) (S = 5/2), and carboxylate ligands for ligand exchange. Using these criteria, we identified the molecule, Mn₃(O₂CCH₃)₆(Bpy)₂ (Bpy = bipyridyl) (1).³¹ This cluster has a linear trinuclear set of Mn(II) atoms bridged by acetate groups and, based on magnetic susceptibility and electron paramagnetic resonance, has weakly antiferromagnetically coupled metals, as seen in Mn₈Fe₄. We used ligand substitution for the acetate groups with a polymerizable carboxylate for both clusters to generate (2) and (3). A second type of cluster was identified, Mn₃O(O₂CCH₃)₆(Pyr)₂ where Pyr = pyridine (4) or Mn₃Pyr, was also trinuclear but oxo-centered with a triangular arrangement of Mn with mixed oxidation states.³²

The success of the cluster-nanocarrier depends on incorporating the highest relaxing cluster possible, but also on the design of the nanocarrier. The nanocarrier has an important role in the biocompatibility and preventing metal release. Our previous data suggested that a cross-linked polymer is an effective tool for preventing metal leaching, a concern when using any metal-based agent and significant problem for manganese chelates. Because positive relaxivity (T₁ agents) depends on an inner-sphere aqueous interaction with the metal, the nanocarrier should be hydrophilic. Therefore, polyacrylamide, a hydrophilic polymer, was selected because of its biocompatibility and low toxicity.³³ Polyacrylamide nanobeads have been used for a wide array of biomedical applications ranging from imaging and MRI contrast to photodynamic therapy because they are amenable to the addition of fluorophores,³⁴ photosensitizers,³⁵ or to encapsulate magnetic iron oxide nanoparticles.³⁶

Synthesis and Structure of Manganese Clusters.

The parent linear trimetallic cluster, Mn₃(O₂CCH₃)₆(Bpy)₂ (1), was prepared as reported,³¹ to form a crystalline product whose structure is illustrated in Figure 1. This core trimetallic cluster was then modified using ligand exchange for the carboxylate groups, and we were able to grow single crystals of these analogs. We report the structure of two new manganese clusters, Mn₃(L)₆(Bpy)₂ for L = vinyl-benzoic acid, and methacrylic acid or Mn₃BpyVBA (2) and Mn₃BpyMA (3), respectively. The single-crystal X-ray diffraction structures for the two substituted compounds are provided in Supporting Information (S1), along with a brief summary of crystallographic data (Supporting Information S1). The overall bonding is directly analogous to the parent compound, with three bridged Mn atoms in a linear configuration. In the parent acetate structure, the three Mn atoms are all linked by two bridging acetates. The third acetate bridges the central Mn (1) and terminal Mn (2) through one oxygen only, while the other acetate oxygen form a longer bond to the terminal Mn (2) (see Figure 1). Thus, in all three complexes, the metal geometry is distorted octahedral. In Mn₃Bpy, the Mn–Mn separation is 3.61 Å for the parent complex, whereas the analogs have slightly shorter separations of 3.497(2) for Mn₃BpyVBA (2) and 3.461(6) for Mn₃BpyMA (3). The final cluster, Mn₃Pyr (4) was mixed valent Mn^III₂Mn^II; however, the metals were reported to be crystallographically equivalent.

Figure 1.

Structure of Mn₃(O₂CCH₃)₆(Bpy)₂. The manganese atoms are in teal, the carbon in black, nitrogen in blue, and oxygen in red.

The purpose of the carboxylate ligand exchange is to introduce a polymerizable moiety to the cluster to create a covalent link with the polymer nanocarrier. A direct synthesis of the clusters with the polymerizable ligand fails because the redox chemistry used to form the metal oxo core initiates polymerization of the ligand, leading to phase separation. With the core intact, this postsubstitution reaction proceeds cleanly; however, it also alters the solubility of the cluster. For example, the analogs (2) and (3) had significantly diminished aqueous solubility compared to the parent acetate cluster (1). In prior studies, we observe similar behavior for other clusters we have ligand substituted, such as Mn₈Fe₄O₁₂(O₂CCH₃)₁₆ (hydrophilic) and Mn₈Fe₄O₁₂(VBA)₁₆ (hydrophobic). Unfortunately, the analogs of Mn₃Bpy, clusters (2 and 3) are not hydrophobic enough for solubility in styrene for direct miniemulsion polymerization. This was surprisingly given that the vinyl benzoic acid substitution was quite effective for enhancing the styrene solubility of the Mn₈Fe₄(VBA)₁₆ cluster. Although we tested the new substituted clusters (2, 3 as well as 4) in inverse miniemulsion polymerization, the difference in solubility of these clusters with acrylamide was problematic.

Cluster Relaxivity.

The relaxivities of the clusters (Table 1) were determined using NMR (400 MHz, room temperature, pH = 7 and 2). The relaxivity of the linear trinuclear clusters (1–3) were quite similar ranging from r₁ = 6.9–7.3 and r₂ = 96.15–124 at pH of 7. These are approximately twice that of reported MnDPDP values, and four times the value of Gd-DTPA at the same field (see Supporting Information S2). Due to solubility issues, the relaxivity of the Mn₃Pyr was measured at pH of 2 and gave an r₁ of 8.72 ±0.21, close to the relaxivity of clusters (1–3) at neutral pH. Interestingly, the Mn₃(O₂CCH₃)₆(Bpy)₂ (1) was pH sensitive and under the same conditions as the Mn₃Pyr cluster, attained a relaxivity value as high as 18.06 mM⁻¹ s⁻¹, and r₂ = 174 mM⁻¹ s⁻¹. Although relaxivity is a complex interplay between structural features, dynamic interactions, and electronic characteristics, it is likely that the increased diameter and rigidity of the cluster contributes to the high relaxivity value observed compared to chelates.^37–39

Table 1.

Comparison of Relaxivity Values of Manganese Oxo Clusters (1–4)

composition	r1 (mM−1 s−1)	r2 (mM−1 s−1)	r2/r1
(1) Mn3Bpy or [Mn3(O2CCH3)6(Bpy)2], pH 2	18.06 ± 0.21	174.72 ± 8.9	18
Mn3Bpy or [Mn3(O2CCH3)6(Bpy)2], pH 7	6.9 ± 0.25	124.7 ± 5.00	10
(2) Mn3BpyVBA or [Mn3(O2CC6H4CH=CH2)6(Bpy)2], pH 7	7.3 ± 0.1	95.0 ± 3.5	13
(3) Mn3BpyMA or [Mn3(O2CC(CH3)=CH2)6(Bpy)2], pH 7	7.0 ± 0.1	95.9 ± 3.3	14
(4) Mn3Pyr or [Mn3O(O2CCH3)6(Pyr)2], pH 2	8.72 ± 0.12	96.15 ± 1.67	11
Mn3Bpy-PAm nanobeads, pH 7	54.4 ± 1.02	144 ± 2.88	3

Generally, Mn(II)-based complexes have r₁ in the range of 2.44–3.7 mM⁻¹ s⁻¹,⁴⁰ so the relaxivity of the clusters reported here are comparatively higher. The solid-state structure suggests an absence of inner sphere water, which typically correlates with r₁; however, we have not yet done solution studies to determine whether the ligands remain in the same bonding arrangement. Beyond, rotational correlation (τ_r), strategies to attain high relaxivity involve increasing the hydration (q) and the water exchange rate (τ_m), although appreciable relaxivity can be observed for q = 0 complexes.⁴⁰ Finally, the third factor, metal electron relaxivity (T_1e), which is magnetic field dependent, is one potential way clusters may differ from chelates. The current understanding of homo- and hetero-metal dimers argues against this, as magnetic coupling is thought to enhance spin relaxation, causing a decrease of T_1e, which lowers relaxivity (the opposite trend).⁴¹ The relaxivity of the clusters are interesting in their own right, and we intend to further investigate the relaxivity to identify the mechanism of relaxivity in future work. Unfortunately, while providing strong support for the enhanced relaxivity of clusters, the problem of solubility was an impediment to solubility was an impediment to incorporating the clusters (2–4) into a polymer nanocarrier via miniemulsion techniques.

Synthesis of Cluster-Nanocarriers.

Generally, the cluster-nanocarrier nanobead synthesis begins with the dissolution of the metal cluster in the monomer, followed by miniemulsion polymerization. Forming the miniemulsion involves high-intensity shearing of the monomer solution and a poor solvent using ultrasonication to form nanometer-sized droplets, which are then polymerized to form polymer beads. The miniemulsion polymerization approach has many advantages in terms of nanobead uniformity and size-control and has been adapted to a variety of polymers.^42,47 Although ligand substitution of the clusters was used to incorporate a polymerizable group, as described previously, it also changed the solubility which affected the miniemulsion synthesis. We have found that the cluster solubility must closely match that of the monomer to form homogeneous inorganic–organic hybrid polymers. We believe this is because the solubility affects the stability of the initial nanodroplet, a critical step in miniemulsion polymerization. Thus, our efforts to prepare polyacrylamide copolymer nanobeads were challenged by the low solubility. With the MA-substituted cluster (3) and the Mn₃Pyr cluster (4), the miniemulsion polymerization led to moderate phase separation and significant phase separation for the VBA-substituted cluster (2), based on transmission electron microscopy (TEM). Given the reduced relaxivity of (2) and (3), and the solubility challenges of all three, these clusters were not explored further.

Fortunately, the unsubstituted cluster, Mn₃(O₂CCH₃)₆(Bpy)₂ (1), termed Mn₃Bpy, was highly soluble in acrylamide and easily formed a uniform polymer nanobead using the inverse-miniemulsion technique.⁴³ To aid in suppressing oil phase nucleation, the hydrophilic azo initiator 2,2′-azobis(2-methylpropionamidine dihydrochloride) or AAPH was chosen, which has a relatively low half-life temperature of 54 °C.⁴⁴ Frequently, ammonium persulfate is used as an initiator for polyacrylamide;⁴⁵ however, low initiator efficiency made the use of AAPH more favorable. The water solubility, low initiation temperature, and lack of persulfates were additional advantages of this initiator. Cyclohexane was chosen as the continuous phase and the miniemulsion droplets were stabilized by the nonionic surfactant sorbitan monooleate (or Span 80). The stability of inverse miniemulsions can typically be enhanced by the addition of a strong lypophobe,⁴⁶ such as NaCl; however, we found that this destabilized the material and therefore it was not used. Although the inverse-miniemulsion approach is less developed than the direct miniemulsion,⁴⁷ by careful adjustment of the sonication time, emulsifier, initiator, and concentration of reactants, we were successful in preparing homogenous nanobeads, described as Mn₃Bpy-PAm.

Characterization of Mn₃Bpy-PAm Nanobeads.

Compared with direct miniemulsion polymerization, the polymer nanobeads from inverse miniemulsion have greater size dispersion. Here, the prepared Mn₃Bpy-PAm nanobeads had size in the range of 40–200 nm with an average diameter of 124 ± 35 nm, confirmed by TEM (histogram given for 650 particles, see Figure 2). Using dynamic light scattering (DLS), there was no evidence of multiple populations of particles, and the hydrodynamic radius was found to be slightly higher than by TEM, with an average diameter of 158 nm (Supporting Information S4). The nanobeads have been imaged by TEM extensively and no evidence for phase separation was observed. The direct miniemulsion polymerization encapsulation of inorganic nanoparticles, typically involves the dispersion of the magnetic nanoparticles, such as Fe₃O₄, in the monomer (e.g., styrene) followed by miniemulsion polymerization. While the nanobead diameters are monodispersed, the metal content is often segregated within populations of high metal nanobeads (because of magnetic aggregation), and pure polymer nanobeads.⁴⁷ Using highly soluble clusters in miniemulsion leads to nanobeads with very homogeneously dispersed metal. We have demonstrated this for both direct and indirect miniemulsion polymerization, to form Mn₈Fe₄-co-PS or Mn₃Bpy-PAm, respectively. Shearing time, amount of surfactant, and polymerization time were explored to optimize the size of the nanobead. The conditions that yielded a stable system with the smallest size and dispersity are reported in the Experimental Section.

Figure 2.

TEM of a single Mn₃Bpy-PAm nanobead (left), TEM of a collection of nanobeads illustrating the dispersity (center), and histogram of 650 particles with an average diameter of 125 ± 35 nm.

In order to isolate the nanobeads for aqueous suspension, they were freeze-dried and resuspended in water. The suspension in water caused significant swelling because of the hydrophilicity of the polymer, causing the hydrodynamic diameter to increase to ~221 nm (Supporting Information S4). The metal content, based on inductively coupled plasma-mass spectroscopy (ICP–MS), was comparable to our Mn₈Fe₄-co-PS nanobeads, with a per mass content of ~2.34%. The cluster integrity inside the polymer matrix was investigated using Fourier-transfer infrared spectroscopy (FTIR); however, the polyacrylamide polymer had strong absorption causing significant background interference. Therefore, for cluster identification, FTIR spectra were background-corrected against a pure polyacrylamide sample, leading to visible peaks that emanate from the cluster. Low-frequency bands between 400 and 700 cm⁻¹ corresponding to the Mn–O stretches were preserved as well as the carboxylate group stretches at around the 1400 and 1600 cm⁻¹ region (Supporting Information S5). Although we do not have additional evidence of the cluster structure after miniemulsion, our electron paramagnetic resonance data support the fact that the divalent oxidation state of manganese is retained (Supporting Information S6). The presence of elemental Mn inside the beads was shown in scanning electron microscopy/energy-dispersive system (SEM/EDS) micrographs (Supporting Information S7). Attempts to perform TEM/EDS line scanning were unsuccessful because of the volatility of the polymer and relatively low metal content.

Relaxivity of the Cluster-Nanocarrier Mn₃Bpy-PAm.

The relaxivities of the Mn₃Bpy-PAm nanobeads were determined using NMR (400 MHz, room temperature, pH 7) of nanobeads suspended in D₂O to avoid both receiver saturation and radiation damping effects. In addition, MR contrast was measured at 7T by imaging phantoms containing nanobeads dispersed in 2% agar at room temperature. The increase in per metal relaxivity for the cluster-nanocarrier combination, Mn₃Bpy-PAm, compared to the cluster alone, was extraordinary, with r₁ = 54.4 ± 1.02 mM⁻¹ s⁻¹ and r₂ = 144 ± 2.88 mM⁻¹ s⁻¹ (data provided in Supporting Information S8). In comparison, clinically used Gd-DTPA is only r₁ = 2.123 ± 0.04 mM⁻¹ s⁻¹, r₂ = 2.626 ± 0.05 mM⁻¹ s⁻¹ under the same conditions (Supporting Information S3). The r₁ value for the cluster nanocarrier is close to that of a Mn chelate coordinated to human serum albumin, with a reported r₁ ≈ 48 mM⁻¹ s⁻¹.⁴⁸ According to the bulk susceptibility effect, a higher magnetic moment contributes to r₂, and incorporation in polymer beads was not anticipated to affect this. Unsurprisingly, only a modest change in the r₂ was seen for the Mn₃Bpy-PAm beads.

The striking r₁ relaxivity value of Mn₃Bpy-PAm was verified by MR imaging of the phantoms, where for the same concentration of the metal, the Mn₃Bpy-PAm nanobeads were significantly brighter (positive contrast) than clinically used Gd-DTPA for T₁-weighted imaging (Figure 3). Additionally, in T₂-weighted images, the Mn₃Bpy-PAm nanobeads produced a darkening effect (negative contrast) that was comparable to commercially available Fe₃O₄ microspheres (Bangs Laboratories, also in Figure 3). We determined the contrast to noise ratio (CNR) by comparing the intensity of the sample in agar with blank agar, as shown in Figure 4. Previous efforts to prepare dual modal MRI contrast agents with strong T₁ and T₂ properties have been hindered, in part by the fact that the magnetic field of the core super-paramagnetic T₂ material adversely affects the relaxation of the paramagnetic shell T₁ agent.⁴⁹ Having the potential to have both types of relaxivity is a significant advantage. Because the artifacts associated with T₁ (e.g., fat, calcification) and T₂ (e.g., hemorrhage, blood clots and air) are different, it is likely that the combination of T₁ and T₂ would amplify the signal and reduce confounding results associated with these artifacts in a clinical setting.⁵⁰

Figure 3.

T₁ (left) and T₂ (right)-weighted phantom MRI images of Mn₃Bpy-PAm, Gd-DTPA and Fe₃O₄ microspheres (Bangs laboratories) in 2% agar, comparing similar manganese, gadolinium and iron metal content, respectively. (Water is at the bottom row for context).

Figure 4.

MRI signal intensity (CNR), for T₁ comparing Mn₃Bpy-PAm-vs-Gd-DTPA, and T₂ comparing Mn₃Bpy-PAm-vs-Fe₃O₄-MS.

Solution Stability and Metal Leaching.

Unlike the Mn₈Fe₄-co-PS cluster-nanocarrier system, the cluster in Mn₃Bpy-PAm is not chemically cross-linked but rather trapped within the network of the polymer. Because it was unclear whether cross-linking was critical for preventing metal leaching, we carefully investigated the stability of the nanobeads toward metal loss. Several methods were used to assess the solution concentration of free Mn²⁺ released from the Mn₃Bpy-PAm nanobeads. First, as a sensitive measure of the metal environment, we quantified the T₁ and T₂ relaxation of nanobeads in D₂O over 17 weeks (Figure 5). The observed stability demonstrates that there is no change in the coordination sphere of the metal that would reflect either metal loss or chemical instability. To further quantify potential metal leaching under more rigorous conditions, the nanobeads were suspended in acidic, neutral, or physiological pH conditions, and subjected to dialysis using a 10 kDa membrane in order to measure free Mn²⁺. The dialysate was tested daily for manganese using ICP–MS over the course of seven days and found to be quite low (see Figure 6). To mimic a physiological environment, 10% human serum [in Dulbecco modified Eagle medium (DMEM)] was used, and the highest leaching observed over the 7 day period was a 0.35% of the starting composition, 10 ppb, well below the lowest level reported to be toxic in the blood and the safe standard for drinking water (0.3 and 0.5 ppm, respectively). We conclude that trapping the cluster within a highly cross-linked polymer nanobead is quite effective at preventing metal leaching. The effort to co-polymerize the ligand of the cluster to the polymer may not further increase the stability or reduce metal leaching.

Figure 5.

Stability of the Mn₃Bpy-PAm nanobeads analyzed by T₁ and T₂ relaxation times measured over 17 weeks in solution.

Figure 6.

Plot of free manganese measured in dialysate of suspended Mn₃Bpy-PAm beads in different pH and serum over 7 days.

Surface Agents.

There are a number of chemical approaches to conjugate reporter and targeting groups to the nanoparticle surface. One of the well-explored advantages of nanoparticles for biomedical imaging is the ease of designing multimodal agents for MRI and optical imaging. Fluorescence imaging is one of the most commonly used tools in preclinical studies, and coupled with MRI should greatly enhance our understanding of disease progression. For example, the combination of diffuse optical tomography in the near-IR range concurrently with MRI, allowed for histopathological information of the human breast in vivo.⁵¹

Thus, our first goal was to label the nanobeads with a near-IR dye for optical imaging, which involved synthesizing amine-terminated nanobeads and using this anchor for postsynthetic surface modification. To synthesize amine-terminated nanobeads, the miniemulsion polymerization synthesis was modified by the addition of an amine-terminated comonomer,⁵² N-(3-aminopropyl)methacrylamide (APMA).^53,54 The presence of a propyl chain on APMA was chosen to support interfacial migration during droplet formation. Our initial attempts with 5% APMA with respect to acrylamide, led to fused beads (left, Supporting Information S9); however, well-formed beads were acquired by reducing the content to 2.5% (right, Supporting Information S9) (termed Mn₃Bpy-PAm-co-APMA). The presence of a positive charge on the latex particles from zeta (ζ) potential measurements (6.68 ± 0.05 mV at pH 7), confirmed the successful amine modification. The iso-electric point for the Mn₃Bpy-PAm-co-APMA is shifted to a higher pH compared to Mn₃Bpy-PAm, as expected (see Supporting Information S10). Subsequently, the surface primary amines were conjugated through an aminolysis reaction with Cyanine7 N-hydroxysuccinimide (NHS) ester reagent, to covalently couple the near-IR dye to the bead surface (λ_max = 775 nm, Supporting Information S11).

Our second goal was to conjugate a targeting molecule to the surface. Surface groups on nanoparticles play two important roles in modulating the biological interaction (pharmacokinetics and toxicology).⁵⁵ One is a shielding effect, by either avoiding opsonization or recognition by the MPS, which is often accomplished using polyethylene glycol, or polyphosphoester-coated polymer nanoparticles.^56,57 The other is to enhance targeting and control biodistribution through the choice of ligands that bind to specific receptors.^58,59 One of the targets of interest for small molecules is the folate receptor,^60,61 which is overexpressed on the cell membrane of many tumor cells. In addition to folic acid, the analog methotrexate (MTX) targets the folate receptor, and also exhibits therapeutic effects for a range of cancer types. MTX has been conjugated to nanoparticles with the goal of targeting⁶² and drug delivery.⁶³ The carboxyl groups of MTX were activated by a carbodiimide reaction with the coupling agent (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide) (EDC), followed by NHS and subsequent reaction with the beads. UV emission data revealed successful conjugation with peaks appearing below 300 and at 388 nm (Supporting Information S12) in comparison to unconjugated beads. A similar approach has been demonstrated for conjugating polyethylene glycol to the surface of polyacrylamide nanobeads.⁵⁴ Cell uptake studies comparing cells with or without folate receptor expression would have to be performed to determine the viability, but the feasibility of covalent conjugation has been demonstrated.

Cytotoxicity and Cell Imaging.

We assessed the cellular toxicity of the nanobeads by measuring the cell viability in vitro. For this purpose, PC3 prostate cancer cells, and patient-derived primary pancreatic cancer cells termed PDAC patient 1 (established by the conditional reprogramming approach),^64,65 were assessed for cell viability with the cell proliferation reagent WST-1 (Roche). The percent cell viability was determined 24 and 48 h after exposure of the PDAC patient 1 cells and PC3 cells to different concentrations of nanobeads (see Figure 7). At 24 and 48 h posttreatment, cytotoxicity was low at most concentrations for both cell lines. At concentrations ≥250 μg/mL, the cell viability decreases for both cell lines with pronounced effects after 48 h. Although we find it interesting that the patient-derived PDAC patient 1 cells are more resistant than the commercially available PC3 cell line, the different origin of the cell lines is a limitation preventing conclusions to be drawn. Nevertheless, the high viability of both cell lines verified after 48 h of treatment underline the lack of toxicity of the nanobeads. In addition to these cell studies, blood compatibility is important, so we have used hemolysis experiments to determine the effect of the nanoparticles with the integrity of red blood cells. As shown in Figure 8, the hemolytic activity was quite low compared to other nanoparticles, with a dose of 300 mg/mL below the 5% limit criterion [ASTM E2524-08(2013)].

Figure 7.

Comparative WST-1 cytotoxicity assays of Mn₃Bpy-PAm beads against PDAC Patient 1 cells (green), and PC3 cells (red), for 24 and 48 h.

Figure 8.

Hemolysis studies of Mn₃Bpy-PAm nanobeads.

In order to demonstrate the multimodal imaging capabilities of Mn₃Bpy-PAm beads in bio-labeling applications, the cellular internalization by both cancer cell lines was assessed using confocal laser scanning microscopy. The nanobeads were labeled with Cyanine7 (red), and the cells were stained with DAPI (blue, nucleus), and phalloidin (green, F-actin). The PDAC patient-1 cells showed lower uptake compared with PC3 cells (Supporting Information S13). Using PC3 cells, as shown in Figure 9, the cells exhibited a normal morphology and the beads were dispersed throughout the cytoplasm and could also be seen localized within the perinuclear region. This suggests that the beads could be used for cellular targeting with minimal particle-induced cell death. To confirm that the beads were internalized and not merely bound to the cell surface membrane, optical z stacking was performed to gain an orthogonal view (Figure 9, merged inset). The cell nuclei appeared on the same plane as the beads suggesting they had in fact been taken up by the cell and internalized. The information from the in vivo fluorescence complements the MR imaging, and provides a bridge between studies at the cellular or tissue level to living organisms.

Figure 9.

Confocal laser scanning microcopy images of PC3 cells treated with Mn₃Bpy-PAm-Cy7 beads, by incubating cells in culture medium containing nanoparticles (50 μg/mL) for 24 h. Excitation of the stains and nanoparticles was performed with 405, 488, and 633 nm lasers. The orthogonal z-stacks shown inset for merged single cell.

Potential for Contrast Agent Development.

In order to determine the contrast capability of the nanobeads for in vivo imaging, we performed MRI after injecting the beads both in healthy mice and mice with xenograft tumors (see Figure 10). First, healthy NCI/nude mice were subcutaneously injected into the nuchal area to evaluate the in vivo feasibility of the agent. The brightening was significant in a T₁-weighted image (Figure 10, top, green arrow). Second, PDAC patient 1 cancer cells were xenografted subcutaneously on both flanks of a nude mouse. After sufficient tumor growth, the Mn₃Bpy-PAm nanobeads and commercial Gd-DTPA (Magnetvist) were injected into each, respectively. The difference in contrast is apparent; where the Mn₃Bpy-PAm nanobeads (Figure 10, bottom, red arrow) showed a marked increase in signal compared to Gd-DTPA (Figure 10, bottom, blue arrow). Consistent with the relaxivity measurements, the Mn₃Bpy-PAm nanobeads exhibit a striking brightening effect in T_1-weighted images that is significantly stronger than Gd-DTPA.

Figure 10.

T₁-weighted images of the nuchal area injected with Mn₃Bpy-PAm nanobeads (green arrow), (top). PDAC patient 1 tumor sites injected with Mn₃Bpy-PAm nanobeads (red arrow) and Gd-DTPA (blue arrow), (bottom).

The in vivo behavior of nanomaterials has been the subject of intense study, but several critical factors have emerged notably in size, type of surface coatings, and surface charge.^66,67 Size has been demonstrated to place clear limitations on the blood residence time, and renal excretion. Unfortunately, because the aqueous hydrodynamic radius of Mn₃Bpy-PAm is >200 nm, we anticipate a rapid rate of clearance, and concentration of nanoparticles in the spleen and liver.⁶⁸ The surface chemistry is important because opsonization, the formation of a protein corona around the particle, is typically the first step in the clearance of nanomaterials from the blood by the mononuclear phagocytic (reticulo-endothelial) system, and generally determines the “biological identity” of the nanoparticle.⁶⁹ The next step in the development of the cluster-nanocarrier is to further develop the synthesis of smaller particle diameters. Reducing the particle size will be challenging because of the lower stability of the inverse miniemulsions, but sizes of <50 nm have been reported for direct miniemulsions.

CONCLUSIONS

The enhanced relaxivity of Mn₃Bpy is strong evidence to suggest that clusters may greatly exceed the relaxivity of chelates. The utility of our cluster-nanocarrier approach is further demonstrated by the ultrahigh T₁ contrast of the Mn₃Bpy-PAm, which is an order of magnitude brighter than gadolinium at higher fields. This combined with a pronounced T₂ negative contrast, similar to that seen with iron oxide nanoparticles, opens the possibility to combine T₁-weighted and T₂-weighted data to enhance sensitivity. Our data show that the nanocarrier not only enhances the relaxivity, it efficiently prevents the leaching of manganese metal under a variety of physiological conditions. The amino-modified beads also demonstrated the synthetic ease of coupling, for example, to the near-IR dye, Cyanine7, forming a multimodal imaging agent. Optical imaging, in combination with MRI, can be powerful in preclinical studies. The cellular uptake of the Mn₃Bpy-PAm nanobeads was quite efficient while simultaneously exhibiting low cytotoxicity. The nanobeads showed strong contrast in vivo as well. Future work will focus on size control, greater colloidal stability, and developing the surface chemistry in order to determine the blood concentration curves and biodistribution data.

EXPERIMENTAL SECTION

Materials.

Manganese(II) acetate tetrahydrate (>99%), 2,2′-bipyridine (>99%), acrylamide (>98%), 4-vinylbenzoic acid (97%), methacrylic acid (99%), N,N′-methylenebis(acrylamide) (MBA) (99%), 2,2′-azobis(2-methylpropionamidine)dihydrochloride (AAPH) (>97%), APMA hydrochloride (98%), sorbitan monoleate (Span 80), diethylenetriaminepentaacetic acid gadolinium(III) dihydrogen salt hydrate (97%), sodium bicarbonate (>99.5%), NHS (98%), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (98%), methotrexate hydrate (>98%). Manganese AAS standard (1000 ppm, TraceCERT), bovine serum albumin (96%) and cell proliferation reagent WST-1 were purchased from Sigma-Aldrich and used as received. The iron oxide microspheres were purchased from Bangs Laboratories Inc. (diameter 0.96 μm), DAPI solution was purchased from Thermo Fisher Scientific and Phalloidin-iFluor 488 reagent was acquired from Abcam. Cyanine7 NHS ester was obtained from Lumiprobe. Demineralized ultrapure water was used (18 Ω).

[Mn₃(O₂CCH₃)₆(Bpy)₂] or Mn₃Bpy (1).

This cluster was synthesized as previously reported.³¹ Briefly, manganese(II) acetate tetrahydrate (0.245 g, 1 mmol) was dissolved in 10 mL ethanol under nitrogen. 2,2′-Bipyridine (0.156 g, 1 mmol) in 10 mL ethanol was cannula transferred to the former and the resulting mixture was stirred under nitrogen for 10 min. The solvent was then partially evaporated to yield pale-yellow microcrystals which were isolated via vacuum filtration and copiously washed with cold ethanol. IR (cm⁻¹): 3425 (br), 3000 (w), 1603 (s), 1477 (w), 1425 (s), 1019 (m), 771 (s), 649 (m), 511 (w), 415 (w). Anal. Calcd: C, 46.23; H, 4.12; N, 6.74. Found: C, 46.34; H, 4.10; N, 6.72.

[Mn₃(O₂CC₆H₄CH=CH₂)₆(Bpy)₂] or Mn₃BpyVBA (2).

The ligand exchange was carried out as previously reported with slight variations.⁶² In a Schlenk flask, Mn₃(O₂CCH₃)₆(Bpy)₂ or 1, (0.104 g, 0.13 mmol) was dissolved in 20 mL warm ethanol. Once fully dissolved under stirring, 4-vinylbenzoic acid (0.222 g, 1.5 mmol) was added to the clear yellowish solution. After 20 min a pale-yellow powder was precipitated, which was isolated by vacuum filtration and washed with cold ethanol. Large crystals suitable for single X-ray analysis were obtained by letting the flask slowly cool and allowing to sit for two days. IR (cm⁻¹): 3063 (br), 1598 (s), 1551 (s), 1475 (w), 1441 (w), 1402 (s), 1179 (w), 1106 (m), 1017 (m), 912 (w), 867 (m), 795 (s), 762 (m), 650 (w), 615 (w), 451 (w). Anal. Calcd: C, 65.34; H, 4.30; N, 4.12. Found: C, 65.10; H, 4.29; N, 4.13.

[Mn₃(O₂CC(CH₃)=CH₂)₆(Bpy)₂]or Mn₃BpyMA (3).

In a Schlenk flask, Mn₃(O₂CCH₃)₆(Bpy)₂ (0.1039 g, 0.125 mmol) was dissolved in 20 mL warm ethanol. Once fully dissolved under stirring, methacrylic acid (0.1272 mL, 1.5 mmol) was added to the clear yellowish solution. The yellow solution remained after 4 h of stirring and it was then layered with 40 mL hexanes and allowed to sit undisturbed in the dark for approximately 14 days. The formed microcrystals were then isolated by vacuum filtration and washed with cold ethanol. 3444 (br), 3106 (w), 2952 (w), 2922 (w), 1644 (m), 1567 (s), 1415 (s), 1099 (w), 1017 (m), 863 (w), 831 (m), 765 (s), 649 (w), 509 (w), 412 (m). Anal. Calcd: C, 53.50; H, 4.69; N, 5.67. Found: C, 53.29; H, 4.68; N, 5.66.

[Mn₃(O₂CCH₃)₆(Pyr)₃] or Mn₃Pyr (4).

This cluster was synthesized as previously reported.³² Briefly, manganese(II) acetate tetrahydrate (2.0 g, 8.15 mmol) was dissolved in a solvent mixture containing 10 mL glacial acetic acid and 20 mL pyridine. Upon dissolution, N-n-Bu₄MnO₄ (0.76 g, 2.10 mmol) was added slowly with stirring. The resulting deep brown/black solution was allowed to stand undisturbed for 48 h. Large reddish brown crystals appeared which were isolated via vacuum filtration and washed with pyridine. IR (cm⁻¹): 3419 (br), 3005 (w), 1725 (w), 1616 (s), 1487 (m), 1413 (s), 1337 (m), 1221 (s), 1075 (m), 755 (m), 685 (m), 648 (m), 616 (m). Anal. Calcd: C, 45.14; H, 4.50; N, 6.58. Found: C, 46.18; H, 3.97; N, 6.62.

Preparation of Mn₃bpy-PAm Nanoparticles.

Composite latex particles with incorporated clusters were prepared using the inverse miniemulsion polymerization process. In a typical synthesis, acrylamide (0.476 g, 6.7 mmol), Mn₃bpy (0.075 g, 0.09 mmol), MBA (0.0238 g, 0.154 mmol), AAPH (0.0274 g, 0.101 mmol) were dissolved in 0.5 g water and stirred under nitrogen to constitute the aqueous phase and was added to the oil phase containing Span 80 (0.167 g, 0.39 mmol, 0.170 mL) and cyclohexane (3.895 g, 5 mL). In the case of amino-functionalized particles, APMA (0.0119 g, 0.067 mmol) was also added to the aqueous phase. After 30 min of preemulsification, the mixture was homogenized in an ice bath using an ultrasound sonicator set at an output of 10 W for 240 s. The polymerization was then performed under nitrogen for 1 h at 65 °C with magnetic stirring set at 300 rpm. The particles were then washed with tetrahydrofuran (THF) and water, resuspended in water, and recovered by the freeze-drying method to yield Mn₃Bpy-PAm solid powder. The % Mn was 2.34%, 124 ± 35 nm diameter based on TEM, 158 nm based on DLS. Select FTIR peaks were 1597 cm⁻¹ (COO⁻ asym), 1417 cm⁻¹ (COO⁻ sym), 400–700 cm⁻¹ (Mn–O stretch). The zeta potential (ζ) was −3.05 ± 0.78 mV at pH = 7.

Conjugation of Cyanine7.

As synthesized amino-functionalized Mn₃bpy-PAm nanoparticles (12.5 mg) were washed with THF and resuspended in 2.5 mL of 0.1 M sodium bicarbonate (pH 8.2) solution. The dye, Cyanine7 NHS ester (0.25 mg, 0.34 × 10⁻³ mmol) was added to dimethyl sulfoxide (DMSO) (100 μL) and added to the amine-functionalized Mn₃Bpy-PAm nanoparticle THF solution, dropwise and stirred for 20 min in the dark. After washing with THF, the particles were suspended in 1 mL H₂O and used for cell uptake studies.

Conjugation of MTX.

The targeting agent, methotrexate MTX (2 mg, 4.4 × 10⁻³ mmol) was added to the coupling agent EDC (3 mg, 15.6 × 10⁻³ mmol) in DMSO (250 μL) to activate the carboxylic groups of MTX. After adding NHS (1 mg, 8.7 × 10⁻³ mmol) in H₂O (50 μL), the solution was added dropwise to a 2.5 mL aliquot of THF-washed nanoparticles in sodium bicarbonate (pH 8.2) and was stirred overnight. The MTX-conjugated nanoparticles were then washed with THF and suspended in DMSO for UV analysis.

Characterization.

High-resolution TEM was performed on a JEOL JEM-2100F FEG operated at 100 kV at the Advanced Imaging and Microscopy Lab of University of Maryland on dispersed samples, drop-casted onto amorphous carbon-coated Cu grids. SEM was performed on the same grids using a Zeiss SUPRA 55-VP operated at 15 kV. UV–vis spectra were obtained by diluting samples in DMSO using a Cary 5000 UV–vis–NIR spectrometer. ICP–MS measurements were carried out on diluted HCl-digested samples on an Agilent 7700 series ICP–MS with Ar plasma using a calibration curve prepared by a manganese standard (TraceCERT). Elemental analysis (C, H, and N) was performed on a PerkinElmer 2400 microanalyzer using acetanilide as a standard. Infrared spectra were collected in the range 400–4000 cm⁻¹ from pressed pellets in KBr on a PerkinElmer FTIR Spectrum 2 system. A Malvern ZetaSizer Nano-ZS was used to determine the hydrodynamic diameter and the zeta potential (10⁻³ M KCl, 25 °C, pH adjusted with standard 0.1 N HCl or NaOH solutions) of dispersed particles.

Single-Crystal X-ray Diffraction.

Single crystals of each compound 2 and 3 were mounted under mineral oil on glass fibers and immediately placed in a cold nitrogen stream at 100(2) K prior to data collection. Data were collected on a Siemens SMART three-circle X-ray diffractometer equipped with an APEX II CCD detector (Bruker-AXS) and an Oxford Cryosystems 700 Cryostream. Full spheres of data were collected (0.3° or 0.5° ω-scans; 2θ max = 56°; monochromatic Mo Kα radiation, λ = 0.7107 Å) and integrated with the Bruker SAINT program. Structure solutions were performed using SHELXS, and nonhydrogen atoms were refined with aniostropic thermal parameters and hydrogen atoms were included in idealized positions (see Supporting Information Table S1).

NMR Relaxivity Measurements.

Relaxation data was collected in a 9.4 T Varian MR 400 MHz spectrometer using standard 5 mm quartz tubes. The proton relaxivities of Mn₃bpy-PAm nanoparticles were measured in 99% D₂O to avoid both receiver saturation and radiation damping effects. Fresh solutions of the nanoparticles were made using freeze-dried powder immediately prior to use and then sonicated for better dispersion. The concentration range was based on the maximal amount of solid powder that could be suspended to form a homogenous solution during the time required for the experiment. The cluster Mn₃Pyz was not soluble without high concentrations of ligand (acetic acid), and was prepared as a 60% acetic acid 40% D₂O solution. To compare with this high acid concentration, Mn₃Bpy was measured both at pH = 7, and pH = 2. The water relaxation times correspond to the residual H₂O in the sample and all measurements were done at pH 7 at room temperature (25 °C), in triplicate. The T₁ and T₂ values were determined by standard inversion recovery and CPMG sequences, respectively, using variable waiting periods (τ). The corresponding r₁ and r₂ values were calculated from the slope of a plot of 1/T₁ or 1/T₂ against concentration, in terms of mM manganese (based on ICP–MS).

Phantom MRI.

MRI was performed in the Georgetown-Lombardi Preclinical Imaging Laboratory. Phantoms with metal concentrations ranging from 0.1 to 0.4 mM metal (manganese for Mn₃Bpy-PAm nanoparticles, iron for BANGS particles, gadolinium for Gd-DTPA) in 2% agar were imaged in a 7T Bruker Biospin (Germany/USA) MR imager using ParaVision 5.1 software. The T₁-weighted images were obtained via a rapid acquisition with relaxation enhancement (RARE) protocol using echo time (TE) 8.7 ms, repetition time (TR) 470.4 ms, image matrix 256 × 256, slice thickness 1.0 mm, and field of view (FOV) of 6.00 cm. For T₂-weighted imaging, the protocol used was a spin-echo fast imaging technique, TurboRARE-T₂ with the same parameters as mentioned above except for TE 140 ms and TR 2000 ms.

Cytotoxicity of Mn₃bpy-PAm Nanoparticles.

The WST-1 assay was performed to assess the viability of human-derived PC3 and PDAC patient 1 cells seeded in 96-well cell culture plates at a density of 4 × 10³ cells/well in 100 μL of culture medium (DMEM supplemented with l-glutamine, 10% fetal bovine serum, 1% penicillin streptomycin, and 1% sodium pyruvate). After incubating for 48 h in a humidified atmosphere containing 5% CO₂ at 37 °C the culture medium was replaced with fresh medium containing varying concentrations of nanoparticles and incubated for 24 and 48 h. Subsequently, the supernatant was aspirated and washed twice with Phosphate-buffered saline (PBS). Culture medium was again added and followed by 10 μL of WST-1 reagent per well. After incubating for 4 h, the plates were shaken and the absorbance was measured at 450 nm using a BioTek ELx800 microplate reader in the Georgetown-Lombardi Genomics and Epigenomics Shared Resource. Untreated cells in medium served as the control and results were calculated as follows against a blank background. (n = 6), [(Abs_treated − Abs_blank)/(Abs _control − Abs_blank) × 100%].

Cellular Uptake and Imaging.

Human-derived PC3 cells were seeded in 6-well cell culture plates on gelatin-coated coverslips at a density of 50 × 10³ cells/well in 1 mL supplemented DMEM culture medium. After incubating for 48 h in a humidified atmosphere containing 5% CO₂ at 37 °C, the supernatant was replaced with culture medium containing Cyanine7-conjugated Mn₃bpy-PAm nanoparticles (50 μg/mL), and was again incubated for 24 h. After incubation, the cells were washed twice with 1× PBS, fixed with 4% formaldehyde, permeabilized with 0.25× Triton and stained with DAPI (nuclear), and Phalloidin-iFluor 488 (F-actin). The stained cells were then imaged in the Georgetown-Lombardi Microscopy Shared Resource using a Leica SP8 confocal microscope. Excitation of the stains and nanoparticles was performed with 405, 488, and 633 nm lasers. Emission and transmission light was passed through AOBS emission filters and was collected by HyD and PMT detectors. The HC PL APO CS2 1.4NM lens was used. Data acquisition of single plane and z-stacks was performed with the Leica LASX control software and images were exported in TIFF formats.

Hemolysis.

Freshly drawn citrated blood was used. PBS was added and centrifuged at 1000g for 10 min (4 °C) to separate red bloods cells (RBCs) from plasma. RBCs were then washed three times and resuspended in PBS; 200 μL was mixed with serial concentrations of Mn₃Bpy-PAm nanoparticles, mixed by gentle vortex and incubated at 37 °C, 5% CO₂ for 4 h. The samples were centrifuged again and the supernatant was transferred to a 96-well plate. Free hemoglobin was measured by the absorbance at 570 nm using a BioTek ELx800 microplate reader. RBCs incubated with 2% Triton X-100TM and PBS were used as the positive and negative controls, respectively. The percent hemolysis was calculated as follows. (n = 6), untreated cells in medium served as the control and results were calculated as follows against a blank background. (n = 6), [(Abs sample − Abs negative control)/(Abs positive control − Abs negative control) × 100%].

In Vivo MRI.

For flank tumor induction, 2.5 × 10⁵ primary pancreatic cancer cells in sterile matrigel and F media (1:1) were injected subcutaneously in 7–9 week old NCI/nude mice using a 27-gauge needle. The xenografts were grown until a maximum tumor volume of 1.0 cm³. Tumor measurements were taken twice weekly and volumes calculated using the prolate ellipsoid formula 4/3Pi (X × Y × Z). After sufficient growth, the tumors were injected with Mn₃Bpy-PAm nanobeads and Gd-DTPA (20 μL of 0.5 mM metal solution) via subcutaneous injection. MRI was performed in the Georgetown-Lombardi Preclinical Imaging Laboratory in a 7-T horizontal Bruker spectrometer run by Paravision 5.1 as previously described.⁷⁰ Anesthetized mice (1.5% isoflurane in a gas mixture of 30% oxygen and 70% nitrous oxide) were placed on our in-house designed and custom-manufactured stereotaxic device (ASI Instruments, Warren, MI) with built-in temperature and cardio-respiratory monitoring engineered to fit a 40 mm Bruker mouse body volume coil. For T₁-weighted imaging, a two-dimensional RARE sequence was used with: TR, 1009.8 msec; TE, 9.4 msec; inversion time, 650 msec; FOV, 3.0 cm; RARE factor, 2; matrix, 256 × 256; averages, 1; slice thickness, 1 mm. For T₂-weighted images, TR was 3000 msec and TE was 240 msec.