Nano assembly and encapsulation; a versatile platform for slowing the rotation of polyanionic Gd³⁺-based MRI contrast agents

Abstract

Encapsulating discrete Gd³⁺ chelates in nano-assembled capsules (NACs) is a simple and effective method of preparing an MRI contrast agent capable of delivering a large payload of high relaxivity imaging agent. The preparation of contrast agent containing NACs had previously focussed preparations incorporating GdDOTP⁵⁻ into the internal aggregate. In this report we demonstrate that other Gd³⁺ chelates bearing overall charges as low as 2- can also be used to prepare NACs. This discovery opens up the possibility of using Gd³⁺ chelates that have inner-sphere water molecules that could further increase the relaxivity enhancement associated with the long τ_R that arises from encapsulation. However, encapsulation of the q = 1 chelate GdDTPA²⁻ did not give rise to a significant increase in relaxivity relative to encapsulation of the outer-sphere chelate GdTTHA³⁻. This leads us to the conclusion that in the NAC interior proton transport is not mediated by movement of whole water molecules and the enhanced relaxivity of Gd³⁺ chelate encapsulated within NACs arises primarily from second sphere effects.

TOC image

Nano-assembled capsules (NACs) can be used to deliver large payloads of high relaxivity Gd³⁺-based contrast agent; a potential solution to the problem of high contrast agent detection limits in MRI. NACs can be formed with a range of polyanionic Gd³⁺ chelates if the overall chelate charge 2- or more negative. Examination of the effect of this encapsulation on mono-hydrated and non-hydrated Gd³⁺ chelates reveals that proton, rather than whole water, transfer is the key mechanism of action in these systems.

Introduction

Clinical MRI as currently practiced often employs small chelates of Gd³⁺ as contrast agents for anatomical imaging. These agents operate by shortening the longitudinal water proton relaxation time constant (T₁) of proximate solvent water; their ability to do this is characterized by their relaxivity (r₁). The relaxivity of current clinical contrast agents is quite poor, typically 3 – 4 mM⁻¹s⁻¹.(1,2) The comparative inefficiency of these clinical contrast agents in generating measurable changes in water proton T₁ leads to high detection limits for these agents. In the emerging paradigm of “personalized medicine” one critical component is the ability to report and map specific biomarkers of pathology; an approach termed “molecular imaging”. The aim is to use this molecular imaging data to tailor therapies specific to each patient. However, the biomarkers of interest for molecular imaging are often present in very low concentrations.(3) The development of molecular imaging probes that can bind to specific receptors, proteins or other biomarkers of disease require the development of contrast media that are many times more effective than the current crop of clinical MRI contrast agents.

Although the limits of detection for the molecular imaging probes of other imaging modalities (e.g. positron emission tomography, PET) are much lower than MRI, MRI offers several advantages. MRI does not necessitate the use of radioactive probes or ionizing radiation; and affords images of exquisite spatial resolution. However, if MRI is to have a legitimate role in molecular imaging, and play a complementary role to nuclear medicine modalities, the question of poor agent detectability must be addressed. The detection limits of MRI contrast agents could be driven lower by enhancing their per-Gd³⁺ relaxivity. However, the relaxivity of Gd³⁺-based contrast agents can be improved by only factor of 10 – 15 at best. Clearly, if the detection limit of a molecular imaging probe for MRI is to be driven as low as those available for PET then it will be necessary not only to develop high relaxivity Gd³⁺-based contrast agents, but also deliver a high payload of high relaxivity agents.(4,5)

These complications with the Gd³⁺-based agents are the reason why many approaches to molecular imaging have focused on the development of super paramagnetic nanoparticle-based contrast media.(4–8) Although this type of contrast agent undoubtedly has the advantage of higher relaxivity on low detection limits these gains come with drawbacks. These super paramagnetic nanomaterials are generally used as susceptibility (T₂*) agents, rather than T₁-shortening agents, and therefore afford less desirable negative image contrast.(9) Furthermore, the relaxivities (r₁, r₂, r₂*) of these super paramagnetic agents are highly dependent upon their proximity to other super paramagnetic particles.(9) This renders it effectively impossible to perform quantitative molecular imaging with these agents because relaxivity and concentration no longer have a direct relationship, as they do for Gd³⁺-based agents.

Clearly, molecular imaging agents that are based on Gd³⁺ but have the detection limits of super paramagnetic nanoparticles, or better, are to be preferred. Nano assembled capsules (NACs) are one possible means of achieving this goal.(10–12) We have previously shown that NACs can be produced (Figure 1) by preparing an aggregate of GdDOTP⁵⁻ (Chart 1) and a cationic polymer which is subsequently encapsulated within a silica shell by addition of silica nanoparticles (SNPs).(13) The molecular tumbling of GdDOTP⁵⁻ is effectively coupled to that of the nanoparticle increasing the per-Gd³⁺ relaxivity of the contrast agent.(1,13,14) In addition each capsule contains many thousands of Gd³⁺ chelates, so the per-NAC relaxivity is extremely high. NAC preparation was found to be robust with respect to the nature of the cationic polymer employed for aggregate formation. Herein, we report the results of our study into the extent to which variations in the polyanionic contrast agent are tolerated in NAC formation. In particular, we were interested to know whether enhanced per-Gd³⁺ relaxivities could be achieved by substituting GdDOTP⁵⁻ for a chelate with a water accessible inner-sphere coordination site.

Figure 1.

A schematic representation of the preparation is SiO₂-NP coated NACs.

Chart 1.

Results and Discussion

As mentioned previously NAC preparation is tolerant of variation in the nature of the cationic polymer.(13) The nature of the polymer was, however, found to have a strong influence on the size of the NACs produced with poly-allylamine hydrochloride (PAH) producing the smallest NACs. NAC size and per-Gd³⁺ relaxivity were also found to be inversely related and NACs formed using PAH as the cationic polymer afforded higher per-Gd³⁺ relaxivities than those formed with any other polymer.(13) In order to reduce the number of variables in this study, only PAH was investigated as a cationic polymer for NAC formation. In addition to the cationic polymer, both the solvent system in which the NACs are prepared and the ratio of charges (R, Eqn. 1) employed influence the average size of the NACs produced.(13) In this study a single solvent system (3:2 v/v acetonitrile/water) was investigated and the charge ratio R varied from 0.3 – 2.0.

$$R=\frac{\left[\mathit{chelate}\right]\times \left|Z-\right|}{\left[\mathit{polymer}\right]\times \left|Z+\right|}$$ (1)

Four discrete low molecular weight Gd³⁺ chelates were investigated: GdDOTA⁻, GdDTPA²⁻, GdTTHA³⁻ and GdDOTP⁵⁻ (Chart 1). In addition to spanning the Z- range from pentaanionic down to monoanionic, this selection of chelates also includes representatives of each primary mechanism of action. GdTTHA³⁻ is an “outer-sphere chelate” – relaxivity arises primarily by means of long range dipole-dipole interactions as the chelate diffuses through the solvent water. All four chelates have an outer-sphere contribution to relaxivity (r₁^os) which contributes about 1.8 mM⁻¹s⁻¹ to the overall discrete chelate’s relaxivity.(1,2,15) But, for GdTTHA³⁻ this contribution makes up almost the entire relaxivity of the chelate.(16) In every case this is expected to disappear once the chelate is encapsulated and no longer diffusing close to solvent water. GdDOTP⁵⁻ is a “second sphere chelate” – relaxivity arises from exchange of water between the bulk and a second hydration sphere of water molecules closely associated with the chelates through hydrogen bonding to the phosphonate groups.(17–19) Phosphonates are effective groups for generating second sphere relaxivity, carboxylates less so.(17,19) Although almost(15) every chelate has some second sphere relaxivity (r₁^ss), it is typically only about 0.8 mM⁻¹s⁻¹ (20 MHz, 25 °C)(15,17) for the carboxylate-based chelates, but rises to about 2.9 mM⁻¹s⁻¹ (20 MHz, 25 °C) for GdDOTP⁵⁻.(1,2) GdDOTA⁻ and GdDTPA²⁻ are “inner-sphere chelates” – relaxivity arises through exchange of water between the inner coordination sphere of Gd³⁺ and the bulk solvent. The inner-sphere contribution to relaxivity (r₁^is) is about 2.0 mM⁻¹s⁻¹.(1) The contribution of each of these effects to the overall relaxivity is given by Eqn. 2.(1,14)

$$r_1^{\mathit{obs}}=r_1^{\mathit{os}}+r_1^{\mathit{ss}}+r_1^{\mathit{is}}$$ (Eqn 2)

NAC preparation followed the previously published protocol.(13) Stock solutions of PAH and each Gd³⁺ chelate were prepared in 3:2 v/v MeCN/H₂O at pH 9. For each R value an appropriate volume of PAH stock solution, followed by contrast agent stock solution, was added to vigourously stirred 3:2 v/v MeCN/H₂O at pH 9. Aggregate formation was characterized by the reaction medium becoming immediately turbid. The extent of the reaction turbidity, and presumably aggregate formation, was highly dependent upon the contrast agent charge. GdDOTP⁵⁻ produced the most turbid reaction mixtures whereas GdDTPA²⁻ the least turbid, regardless of the R value employed. No turbidity whatsoever was observed after the addition of GdDOTA⁻ at any R value. The turbid reaction mixture was vigourously stirred for a further 15 seconds, after which a suspension of SNPs (2 wt%) in 9:1 MeCN/H₂O was added. The constitution of the 2 wt% SNP stock suspension was found to be critical to the formation of NACs, if the SNPs were suspended in water alone, or a 3:2 v/v MeCN/H₂O mixture, then NAC formation was not observed. Reactions were stirred for a further 15 seconds and the NACs allowed to age for 10 minutes without stirring. NACs were separated from unreacted starting materials by repeated centrifugation and resuspension in water in 10 kDa MWCO centrifugal filter units.

The extent to which each Gd³⁺ chelate was incorporated into the NAC was determined by ICP-OES. Samples of NACs were digested in 15.7 M nitric acid and the Gd³⁺ concentration in both the NAC digest and the reaction mixture measured. In general >99 % incorporation of Gd³⁺ chelate was achieved for NACs from with GdDTPA²⁻, GdTTHA³⁻ and GdDOTP⁵⁻ (Table 1). In two cases slightly lower chelate incorporation rates were observed ~ 94 %, both for GdDTPA²⁻ NAC preparations. The formation of NACs is readily confirmed by scanning electron microscopy (SEM), samples images of NACs produced with each contrast agent are shown in Figure 2. It is evident that the monoanionic chelate GdDOTA⁻ is incapable of aggregating with the polymer in a manner that allows NAC formation and no NAC formation was detected by SEM for any R value tested. In contrast, all the other Gd³⁺ chelates formed aggregates with PAH that could subsequently be encapsulated to form NACs. In the chosen solvent system the R value had an important effect on whether NACs formed successfully or not. NAC formation could be readily achieved in the range R = 0.3 – 1.25. However, NAC formation was inconsistent when R = 2.0: large quantities of unreacted starting materials being recovered from the reaction.

Table 1.

Summary of the variations in size and relaxivity arising from changing the chelate charge and R value during the preparation of NACs.

Chelate	R value	% Gd3+ incorp.a	Diameter (nm)b	r1 (mM−1s−1)c
GdDTPA2−
	R = 0.30	99.97	127 ± 58	23.3 ± 0.3
	R = 0.50	94.5	645 ± 267	14.9 ± 0.1
	R = 0.75	93.3	1088 ± 232	8.5 ± 0.05
	R = 1.00	99.0	1542 ± 465	7.8 ± 0.01
	R = 1.25	99.9	1802 ± 581	4.8 ± 0.1
GdTTHA3−
	R = 0.30	99.6	105 ± 60	22.7 ± 0.1
	R = 0.50	99.9	180 ± 110	14.0 ± 0.1
	R = 0.75	99.9	370 ± 260	7.4 ± 0.3
	R = 1.00	99.0	1100 ± 500	7.1 ± 0.3
	R = 1.25	99.9	1651 ± 445	6.5 ± 0.1
GdDOTP5−
	R = 0.30	99.9	320 ± 260	46.4 ± 0.3
	R = 0.50	99.1	620 ± 235	19.8 ± 0.1
	R = 0.75	98.9	1147 ± 368	15.8 ± 0.1
	R = 1.00	98.6	1232 ± 650	7.1 ± 0.1
	R = 1.25	99.5	1598 ± 401	6.3 ± 0.2

^aThe percentage of Gd³⁺ chelate added to the NAC synthesis that was found to be incorporated into the NAC;

^bthe mean average hydrodynamic diameter of 6 NAC preparations ± the 1^st standard deviation of the size distribution, determined by dynamic light scattering;

^cthe mean longitudinal relaxivity of 6 NAC preparations ± the 1^st standard deviation, determined at 20 MHz and 25 °C.

Figure 2.

SEM images of NAC preparations in 3:2 v/v MeCN/H₂O with PAH and: A) GdDOTA⁻ R = 1.0; B) GdDTPA²⁻ R = 0.75; C) GdTTHA³⁻ R = 0.75; D) GdDOTP⁵⁻ R = 0.75.

The distribution of hydrodynamic diameters in each NAC preparation was determined by dynamic light scattering (DLS). The average size of the NACs was found to increase as the charge ratio R increased. No consistent correlation between NAC size and R was observed across the three Gd³⁺ chelates studied. Indeed, the different Gd³⁺ chelates give rise to different relationships between R and NAC size (Figure 3, Table 1). For both GdDTPA²⁻ and GdDOTP⁵⁻ the NAC size increases sub-linearly with increasing R, although the two do not have identical relationships. In contrast, for GdTTHA³⁻ NAC size appears to increase super-linearly with increasing R. On the basis of chelate structure it is hard to rationalize this observation; outwardly GdTTHA³⁻ has more structural similarity to GdDTPA²⁻ than either chelate has with GdDOTP⁵⁻. The only notable difference is the presence of a discrete anionic group, the uncoordinated carboxylate group, in GdTTHA³⁻ which neither GdDTPA²⁻ nor GdDOTP⁵⁻ possess. The overall charge in both GdDTPA²⁻ and GdDOTP⁵⁻ is therefore delocalized over the entire chelate, whereas in GdTTHA³⁻ a portion of the overall charge is localized on a single carboxylate. We speculate that this difference may be the origin of the different relationships between the R value and NAC size.

Figure 3.

The effect of changing the R value on the average hydrodynamic diameter of the NACs produced for GdDTPA²⁻ (purple), GdTTHA³⁻ (red) and GdDOTP⁵⁻ (green). Dashed lines are a guide to the eye only, error bars represent the 1^st standard deviation of the size distribution of the capsules. Sizes and their distributions are the average of 6 NAC preparations.

The per-Gd³⁺ relaxivity of NACs derived with each chelate are shown in Figure 4 and Table 1. The effect of changing the R value on the per-Gd³⁺ relaxivity for NACs prepared using PAH and GdDOTP⁵⁻ is similar to that previously reported for other NACs prepared using the same polymer and the anion in different solvent systems.(13) For NACs formed using higher R values (1.0 and 1.25) the per-Gd³⁺ relaxivity is comparatively low – only slightly higher than that of the relaxivity of the free chelate. However, as the R value and therefore the size of the NAC decreases the per-Gd³⁺ relaxivity is found to increase dramatically. Notably, the highest per Gd³⁺ relaxivity (20 MHz, 25 °C) reported herein – for NACs R = 0.3 – is significantly higher (46.3 mM⁻¹s⁻¹) than the 24.2 mM⁻¹s⁻¹ (23 MHz, 25 °C) obtained for the slightly smaller NACs produced with PAH and GdDOTP⁵⁻ in 1:1 MeCN/H₂O.(13) This suggests that changes to the aggregate formation step can result in differences in the nature of the aggregated NAC interior that alter the capacity of relaxed protons in the interior to exchange with those of the bulk solvent.

Figure 4.

The effect of changes in the overall Gd³⁺ chelate charge on the relaxivity of the NACs produced for different R values.

This observation is further supported by the results of changing the polyanionic chelate used in aggregate formation. The effect of changing R value on the per-Gd³⁺ relaxivity for NACs prepared using GdDTPA²⁻ and GdTTHA³⁻ is qualitatively similar to that observed for NACs produced with GdDOTP⁵⁻. Low per-Gd³⁺ relaxivity values were obtained for the larger NACs (higher R value) but again relaxivity increases as the NACs become smaller (lower R values). Although the per-Gd³⁺ relaxivity for all larger NACs (R = 1.0 or 1.25) is comparable regardless of the Gd³⁺ chelate employed, the Gd³⁺ chelate has a substantial effect on the per Gd³⁺ relaxivities obtained for smaller NACs. The per-Gd³⁺ relaxivity of small NACs produced with GdDOTP⁵⁻ is substantially higher than that of small NACs produced with the carboxylate containing chelates GdDTPA²⁻ and GdTTHA³⁻. This is presumably because the phosphonate groups of GdDOTP⁵⁻, which are known to promote strong second-sphere interactions are able to pull the cationic amino groups of PAH into the chelate’s second-sphere more tightly than the carboxylate groups of the other two chelates. This may have two effects: the molecular tumbling of the chelate should be most effectively coupled to that of the NAC in this situation and the amino protons of PAH are likely to be held more tightly in the chelate’s second-sphere, increasing τ_M and decreasing r_GdH. Both of these will tend to increase the relaxivity of NACs encapsulating GdDOTP⁵⁻ over those containing GdDTPA²⁻ and GdTTHA³⁻. However, surprisingly the per-Gd³⁺ relaxivity of NACs containing GdDTPA²⁻ and GdTTHA³⁻ are almost the same for each R value studied, this despite the vastly greater potential for the inner-sphere GdDTPA²⁻ to exhibit gains in relaxivity from reduced tumbling than the outer-sphere GdTTHA³⁻.(1,2,14) Both inner- and second-sphere relaxivities can be described by the Solomon-Bloembergen-Morgan (SBM) equations and will increase as the chelate tumbles more slowly (longer τ_R). However, owing to the increased proximity of inner-sphere water to the paramagnetic metal ion, the effect of making τ_R long is known to be more profound for inner-sphere relaxation.(1,2,17) GdDTPA²⁻ and GdTTHA³⁻ have approximately the same second-sphere relaxivity contributions and these would reasonably be expected to increase the same amount as tumbling slowed. Although there are differences in NAC size between GdDTPA²⁻ and GdTTHA³⁻ at some R values, at R = 0.3 the NACs are about the same size (Figure 3). At R = 0.3 the NACs formed with GdDTPA²⁻ and GdTTHA³⁻ have very similar per-Gd³⁺ relaxivities which suggests that inner-sphere water exchange is either extremely slow, or simply does not occur to any meaningful extent in the interior of a NAC. Because the chelate exhibit high per-Gd³⁺ relaxivities, ¹H relaxation of protons proximate to Gd³⁺ must still occur and these protons must then make their way through the NAC interior, across the silica shell, and into the bulk solvent. We speculate that the interior of a NAC is a comparatively water scarce environment and that inner-sphere water exchange is not a significant contributory factor to the overall relaxivity of a chelate once incorporate within the interior aggregate of a NAC. The transfer of relaxation therefore most probably occurs through proton exchange.

Inside the NAC the second “hydration sphere” is probably replaced by protons from the PAH amino groups that hydrogen bond to the chelate, forming the aggregate.(6) The T₁ of these protons is then short, owing to their proximity to the paramagnetic Gd³⁺ ion. Exchange of these protons across sites on polymers, chelates and whatever water molecules are present in the aggregate facilitate proton movement to the shell interior. Exchange onto water and subsequent whole water exchange across the silica shell presumably completes the journey to the bulk solvent. Thus, GdDPTA²⁻ and GdTTHA³⁻ have comparable abilities to generate relaxivity within the NAC but neither is able match GdDOTP⁵⁻ with its greater capacity to from hydrogen bonding networks. With this in mind we might very well envisage that any change to the nature of the encapsulated aggregate (be it a different chelate, R value, polymer, solvent) will result in a different network of interactions across the aggregate interior of the NAC, each with a different ability to transport protons to the exterior which will directly impact relaxivity beyond a simple question of NAC size.

Conclusions

Contrast agent based NACs represent a robust platform by which large payloads of high relaxivity contrast agent may be delivered to sites of interest. Furthermore, this can be achieved without the need for difficult and expensive chemical modifications to the ligand structure.(20,21) Not only the nature of the cationic polymer employed in the preparation, but also the nature and overall charge of the contrast agent employed in NAC formation may be varied substantially. Overall chelate charges as low as 2- are well tolerated in the synthesis of NACs. The nature of the encapsulated contrast agent, its charge, the polymer and solvent used,(13) and the R value employed do affect the size of the NAC produced. Although the general trend observed is clearly that smaller NACs are produced by reducing the R value, precise control of NAC size is highly dependent upon all the factors involved in the aggregate step.

The highest relaxivities are achieved by using GdDOTP⁵⁻ as the basis for aggregate formation at low R values. Employing chelates, such as GdDTPA²⁻, with the potential to provide inner-sphere contributions to relaxivity does not appear to be a route to more effect NAC-based contrast agents. It seems that in the aggregate interior of a NAC inner-sphere exchange does not occur to any meaningful extent. The increased per-Gd³⁺ relaxivity observed for each chelate incorporated into a NAC seems to derive entirely from second-sphere contributions enhanced by slowed molecular tumbling (long τ_R). We therefore speculate that the exchange of relaxed protons from the chelate’s second-sphere across the aggregate interior and silica shell to the bulk solvent occurs primarily through proton exchange.(22) Whole water exchange is probably only involved in the final exchange step across the silica shell.

These results demonstrate the robustness of the NAC formation process to alterations in each constituent part of the aggregate interior. Furthermore, they begin to demonstrate the best approaches to tailoring NAC formation to a given imaging application.

Experimental

General remarks

‘Water’ refers to deionized water with a resistivity of 18.2 MΩ. All solvents and reagents were purchased from commercial sources and used as received unless otherwise stated. H₂GdDTPA, H₆TTHA and PAH were purchased from the Sigma-Aldrich corporation. The molecular weight quoted by Sigma-Aldrich of the PAH used in this study is 56,000, not as previously reported 70,000.(13) H₈DOTP was purchased from Macrocyclics. GdDOTA⁻, GdTTHA³⁻ and GdDOTP⁵⁻ were prepared by previously published methods.(16,23,24) Silica nanoparticles were purchased from Nissan chemicals Snowtex O type as a 20.2 w/w suspension in water (pH 3.5).

General Procedure for the Preparation of Gd³⁺-based NACs

An 8.9 μM stock solution of PAH was prepared by dissolving PAH (0.0498 g, 0.89 μmol) in a 3:2 v/v mixture of acetonitrile and water (10 mL). Stock solution of each contrast agent anion was prepared by diluting a concentrated aqueous solution of the anion into a 3:2 v/v mixture of acetonitrile and water to afford stock solutions of 0.757, 1.26, 1.89, 3.78 mM for GdDOTP⁵⁻, GdTTHA³⁻, GdDTPA²⁻, and GdDOTA⁻, respectively. The pH of the chelate stock solution was adjusted to 9, by addition of 1M NaOH solution. A stock 2% suspension of SNPs was prepared by adding 1 mL of a commercially available 20.2 wt% suspension of SNPs in water to 9 mL of MeCN.

For NACs R = 0.5, stock solutions of the polymer (20 μL) and Gd³⁺ chelate (120 μL) were added to a stirred solution of 3:2 v/v MeCN/H₂O (1 mL). Upon addition of the chelate stock solution the reaction rapidly became turbid (except in the case of GdDOTA⁻), a sign that aggregate formation had occurred. The reaction was vortexed at a medium speed for 15 seconds. Aggregates were then allowed to age for 10 minutes before addition of the 2% SNP stock solution (100 μL). The reaction was vortexed for a further 10 seconds at medium speed and then allowed to age without agitation for 30 minutes. Excess SNPs were removed by filter centrifugation using 10 kDa MWCO filter centrifuge tube at 6,000 rpm. The NACs retained in the filter centrifuged tube were washed with water and filtered by centrifugation for 30 minutes at 14,000 rpm a total of six times. The NACs were then taken up into water (1 mL) and recovered into a sample vial.

Microscopy

Scanning electron microscopy (SEM) was performed on a FEI Sirion FEG electron microscope equipped with an energy dispersive X-ray (EDX) detector. A droplet of NACs suspension was placed on the aluminium stub and dried in air, the sample was then sputter coated with gold for 55 seconds. Secondary electron images were taken at 5kV with a working distance between 5–10mm.

Light Scattering

Dynamic light scattering was performed on a Horiba LB-550 dynamic light scattering instrument. For these measurements freshly syringed filtered samples were dispersed in water and measured at four dilutions to ensure size distributions were independent of concentration effects. Samples were regularly agitated to guard against settling of larger particles.

Relaxometry

Water proton T₁s were measured on a 0.47 T Bruker MiniSpec contrast agent analyzer operating at 19.99 MHz using an inversion recovery pulse sequence. NACs were suspended in water (1 mL) at Gd³⁺ concentrations ranging from 0.22mM to 2.37 mM. Samples were regularly agitated to guard against settling of larger particles. Relaxivity values were determined by linear regression analysis of the experimentally determined R₁ values as a function of Gd³⁺ concentration in Excel.

ICP-OES Gd³⁺ Concentration Determinations

Concentration determination of Gd was performed using a Perkin Elmer Optima 2000 inductively coupled plasma optical emission spectrometer (ICP-OES). Gd³⁺ standards were produced by quantitative serial dilution of a commercial 1000 mgL⁻¹ gadolinium in 2 % nitric acid standard (Fluka Analytical) into a 0.1% nitric acid solution. Samples were prepared by digesting the dried samples with 70% HNO₃ (100 μL) and water (900 μL). The resulting analyte was vortexed thoroughly, then filtered to remove any particulate using a 0.2μm filter to produce the final sample for ICP-OES analysis. Sample analyte concentrations were calculated to fall in the middle of the constructed ICP-OES Gd³⁺ calibration curve to ensure accurate Gd spectral readings. Readings were taken in triplicate and averaged. The highest percent relative standard deviation allowed between these replicates was 1%, to ensure precise Gd³⁺ spectral readings for the samples analyzed in this work. The concentration of Gd³⁺ was determined for each sample and then multiplied by the dilution factor.